Source(s) of the information:
Source(s) of the information:

Lawyers and other representatives

Decision on Counterclaims

TABLE OF ABBREVIATIONS AND DEFINITIONS
2nd SMCCRespondent [Ecuador]'s Second Supplemental Memorial on Counterclaims of 24 April 2012 (corrected as of 7 May 2012)
Andrade ERExpert Report of Fabián Andrade Narváez of 20 February 2013
APIAmerican Petroleum Institute
Arbitration RulesICSID Rules of Procedure for Arbitration Proceedings
AwardDecision on Reconsideration and Award
Bedón ER11st Expert Report of Dr. René Bedón of 27 September 2012
Bedón ER22nd Expert Report of Dr. René Bedón of 2 July 2013
BITBilateral Investment Treaty; specifically "Treaty between the United States and Ecuador concerning the Encouragement and Reciprocal Protection of Investments" of 11 May 1997
BPBritish Petroleum Development Limited
BurlingtonBurlington Resources Inc.
CEPECorporación Estatal Petrolera Ecuatoriana
Civil Code or CCEcuadorian Civil Code
CMCCClaimant [Burlington]'s Counter-Memorial on Counterclaims of 29 September 2012
ConsortiumConsortium formed by Perenco Ecuador Limited and Burlington Resources Oriente Limited
CPFCentral Production Facility
C-PHBClaimant [Burlington]'s Post-Hearing Brief of 3 October 2014
C-PSVBClaimant [Burlington]'s Post-Site Visit Brief of 15 July 2015
CPUFCoca-Payamino Unified Field
Crespo ERExpert Report of Ricardo Crespo Plaza of 29 September 2011
Decision on Liability Decision on JurisdictionDecision on Liability of 14 December 2012 Decision on Jurisdiction of 2 June 2010
EISEnvironmantal Impact Study
EMLEnvironmental Management Law (Ley de Gestión Ambiental)
ERExpert Report
Exh.Exhibit
Exh. C-Claimant [Burlington]'s Exhibits
Exh. CL-Claimant [Burlington]'s Legal Exhibits
Exh. E-Respondent [Ecuador]'s Exhibits
Exh. EL-Respondent [Ecuador]'s Legal Exhibits
GSIGSI Environmental Inc.
GSI ER11st Expert Report of GSI of 20 September 2012
GSI ER22nd Expert Report of GSI of 2 July 2013
HLHydrocarbons Law (Ley de Hidrocarburos)
ICSIDInternational Centre for Settlement of Investment Disputes
ICSID ConventionConvention on the Settlement of Investment Disputes between States and Nationals of other States
IDWInverse Distance Weighting
IEMSIntegrated Environmental Management Services, S.A. de C.V.
IEMS ER11st Expert Report of IEMS of 11 January 2011
IEMS ER22nd Expert Report of IEMS of 29 September 2011
IEMS ER33rd Expert Report of IEMS of 23 April 2012
IEMS ER44th Expert Report of IEMS of 18 February 2013
MABUNESCO Man and the Biosphere Program
NTUNephelometric Turbidity Units
P&TPump-and-Treat
PerencoPerenco Ecuador Limited
PHBPost-Hearing Brief
PO [No.]Procedural Order [number]
PRBPermeable Reactive Barriers
ProtocolJoint Site Visit Protocol signed by both Parties of 20 January 2015
PSCProduction Sharing Contract
PSEPrediction Standard Error
RAOHEEnvironmental Regulation for Hydrocarbon Operations in Ecuador (Reglamento Ambiental para las Operaciones Hidrocarburíferas en el Ecuador)
RECRecognized Environmental Condition
RejoinderClaimant [Burlington]'s Rejoinder on Counterclaims of 8 July 2013
ReplyRespondent [Ecuador]'s Reply on Counterclaims of 18 March 2013
ROHRegulation of Hydrocarbon Operations (Reglamento de Operaciones Hidrocarburíferas)
R-PHBRespondent [Ecuador]'s Post-Hearing Brief of 3 October 2014
RPSRPS Group Plc
RPS ER22nd Expert Report of RPS of February 2013
R-PSVBRespondent [Ecuador]'s Post-Site Visit Brief of 15 July 2015
Saltos WS11st Witness Statement of Wilfrido Saltos of 28 September 2012
Saltos WS22nd Witness Statement of Wilfrido Saltos undated
Site VisitThe inspection of a number of sites within Blocks 7 and 21
SMCCRespondent [Ecuador]'s Supplemental Memorial on Counterclaims of 30 September 2011
SNAPNational System of Protected Areas (Sistema nacional de áreas protegidas)
SPLPSynthetic Precipitation Leaching Procedure
TCLPToxicity Characteristic Leaching Procedure
TPHTotal Petroleum Hydrocarbons
Tr. [(Day)] [(Language)] [page:line]Transcript of the Hearing on Counterclaims of 1-7 June 2014, English or Spanish version, as indicated
Tr. Site Visit [(Day)] (ENG) [page:line]Transcript of the Site Visit of 29 March to 1 April 2015, English version
TULASUnified Text of Secondary Environmental Legislation (Texto Unificado de Legislación Ambiental Secundaria)
UNESCOUnited Nations Educational, Scientific and Cultural Organization
UPLUpper Prediction Limit
USDUnited States Dollar
USEPAUnited States Environmental Protection Agency
VOCVolatile Organic Compounds
WSWitness Statement

 

 

 

I. THE PARTIES

A. The Counter-Claimant

1.
The Counter-Claimant is the Republic of Ecuador ("Ecuador" or the "Counter-Claimant" or the "Respondent").
2.
The Counter-Claimant is represented in these proceedings by Dr. Diego García Carrión, Procurador General del Estado, Dr. Blanca Gómez de la Torre, Directora de Asuntos Internacionales y Arbitraje, Dr. Christel Gaibor, and Dr. Diana Moya at the PROCURADURÍA GENERAL DE ECUADOR; by Professor Eduardo Silva Romero, Mr. José Manuel García Represa, Mr. Philip Dunham, Mr. Alvaro Galindo, Ms. Maria Claudia Procopiak, Ms. Audrey Caminades and Ms. Gabriela González Giráldez of the law firm DECHERT (Paris) LLP; and by Professor Pierre Mayer who left DECHERT (Paris) LLP on 1 June 2015.

B. The Counter-Respondent

3.
The Counter-Respondent is Burlington Resources Inc. ("Burlington" or the "CounterRespondent" or the "Claimant"), a corporation created under the laws of the State of Delaware, United States of America, in 1988 and active in the exploitation of natural resources. On 31 March 2006, Burlington was acquired by ConocoPhillips, a multinational energy company with headquarters in the State of Texas, United States of America.
4.
The Counter-Respondent is represented in these proceedings by Mr. Nigel Blackaby, Ms. Noiana Marigo, Ms. Lauren Friedman, Mr. Leon Skornicki and Ms. Giulia Previti of the law firm FRESHFIELDS BRUCKHAUS DERINGER US LLP; by Ms. Tracie Renfroe, Mr. Wade Corriell, Mr. Esteban Leccese, Ms. Jamie M. Miller, Ms. Anisha Sud and Ms. Sara McBrearty of the law firm KING & SPALDING; and by Mr. Javier Robalino-Orellana of the law firm FERRERE.

II. PROCEDURAL HISTORY

5.
The counterclaims addressed in this decision (the "Decision") were filed as part of ICSID Case No. ARB/08/5, which has already given rise to decisions on jurisdiction and liability and which will end by the notification of an Award shortly following this Decision.

A. Initial Phase

6.
Pursuant to Rule 40(2) of the ICSID Arbitration Rules, Ecuador asserted two counterclaims in its Counter-Memorial on Liability of 17 January 2011. Although Burlington initially stated that it would challenge the Tribunal's jurisdiction, the Parties later executed an agreement dated 26 May 2011 by which Burlington accepted the jurisdiction of the Tribunal over the counterclaims.1

B. Written Phase

7.
After having consulted with the Parties, the Tribunal issued Procedural Order No. 8 on 21 July 2011 (Procedural Orders are referred to herein as "PO" followed by their respective numbers; for example, Procedural Order No. 8 is "PO8"), which referred to and annexed the Parties' jurisdictional agreement and contained the calendar for the written phase of the counterclaims proceedings.
8.
In accordance with this calendar as later amended on the Parties' request, Ecuador filed a Supplemental Memorial on Counter-Claims on 30 September 2011 and a Second Supplemental Memorial on Counter-Claims on 24 April 2012. Burlington then filed its Counter-Memorial on Counter-Claims on 29 September 2012. Ecuador filed its Reply on 18 March 2013 and Burlington its Rejoinder on 8 July 2013.
9.
On 6 July 2012, the Tribunal issued PO9 with regards to Burlington's request for production of documents. On 14 December 2012, the Tribunal issued PO10 with regard to Ecuador's request for production of documents. On 22 May 2013, the Tribunal issued PO13 with regard to Burlington's additional requests for production of documents.
10.
On 8 April 2014, the Tribunal issued PO16 with regard to Ecuador's request to add new evidence to the record. On 16 April 2014, the Tribunal issued PO17 with regard to Burlington's request to add new evidence to the record. On 16 May 2014, the Tribunal issued PO20 with regard to Burlington's request to add new evidence in response to exhibits admitted by the Tribunal in PO16.

C. Hearing on Counterclaims

11.
Pursuant to PO14 dated 16 July 2013, the hearing on counterclaims (the "Hearing") was initially set to take place from 26 August to 31 August 2013. Due to the Proposal for Disqualification of Professor Francisco Orrego Vicuña submitted on 25 July 2013, the proceedings were suspended in accordance with Rule 9(6) of the ICSID Arbitration Rules. The proceeding was resumed on 10 January 2014 following the disqualification of Professor Orrego Vicuña and the appointment of Mr. Stephen Drymer as arbitrator. The Hearing was thus rescheduled and, in accordance with PO19 dated 16 May 2014, took take place from 1 to 7 June 2014.
12.
The Hearing was held at the World Bank offices in Paris. The following persons attended the Hearing:

The Tribunal

Members of the Tribunal

Professor Gabrielle Kaufmann-Kohler, President

Professor Brigitte Stern, Arbitrator Mr. Stephen L. Drymer, Arbitrator Secretary of the Tribunal Mr. Marco Tulio Montañés-Rumayor Assistant to the Tribunal Mr. Magnus Jesko Langer

Ecuador's representatives Counsel

Mr. Pierre Mayer Dechert (Paris) LLP

Mr. Eduardo Silva Romero Dechert (Paris) LLP

Mr. Philip Dunham Dechert (Paris) LLP

Mr. José Manuel García Represa Dechert (Paris) LLP

Mr. Timothy Lindsay Dechert LLP

Ms. Meredith Bloch Dechert (Paris) LLP

Ms. Audrey Caminades Dechert (Paris) LLP

Ms. Gabriela González Giráldez Dechert (Paris) LLP

Support

Ms. Djamila Rabhi Dechert (Paris) LLP

Mr. Jeremy Eichler Dechert (Paris) LLP

Mr. Pedro Arcoverde Dechert (Paris) LLP

Mr. Oswaldo Santos Davalos Dechert (Paris) LLP

Ms. Katerine Marami Dechert (Paris) LLP

Parties

Dr. Diego García Carrión Procuraduría General del Estado

Dr. Blanca Gómez de la Torre Procuraduría General del Estado

Dr. Diana Moya Procuraduría General del Estado

Burlington's representatives

Counsel

Mr. Nigel Blackaby Freshfields Bruckhaus Deringer

Ms. Noiana Marigo Freshfields Bruckhaus Deringer

Ms. Lauren Friedman Freshfields Bruckhaus Deringer

Ms. Giulia Previti Freshfields Bruckhaus Deringer

Mr. Leon Skornicki Freshfields Bruckhaus Deringer

Ms. Sarah Gans Freshfields Bruckhaus Deringer

Mr. Anthony Ogunseye Freshfields Bruckhaus Deringer

Ms. Cassia Cheung Freshfields Bruckhaus Deringer

Ms. Tracie Renfroe King & Spalding

Mr. Wade Coriell King & Spalding

Mr. Esteban Leccesse King & Spalding

Ms. Jamie Miller King & Spalding

Ms. Anisha Sud King & Spalding

Ms. Sara McBrearty King & Spalding

Ms. Veronica Garcia King & Spalding

Ms. Pui Yee (Lisa) Wong King & Spalding

Mr. Thomas Norgaard Debevoise & Plimpton

Ms. Floriane Lavaud Debevoise & Plimpton

Paz Horowitz Robalino Garces

Mr. Javier Robalino

Abogados

Mr. James Haase FTI Consulting

Parties

Ms. Janet Kelly ConocoPhillips Company

Mr. Clyde Lea ConocoPhillips Company

Ms. Laura Robertson ConocoPhillips Company

Ms. Suzana Blades ConocoPhillips Company

Mr. Fernando Avila ConocoPhillips Company

Ms. Ann Morgan ConocoPhillips Company

Mr. Jared L. Richards ConocoPhillips Company

Mr. Rick Greiner ConocoPhillips Company

Ecuador's witnesses and experts

Witnesses

Mr. Saulo Bernabe Carrasco Paredes Agencia de Regulación y Control Hidrocarburífero del Ecuador (ARCH)

Mr. Pablo Alberto Luna Hermosa Petroamazonas

Mr. Diego Fernando Montenegro Munoz Petroamazonas

Mr. Manuel Solís Petroamazonas

Mr. Marco Puente Petroamazonas

Experts

Mr. Ricardo Crespo Universidad San Francisco de Quito

Mr. Fabian Alexander Andrade Narvaez Universidad San Francisco de Quito

Mr. José Rubén Villanueva Peon IEMS (Integrated Environmental Management Services, S.A. de C.V.)

Mr. José Francisco Alfaro Rodriquez IEMS (Integrated Environmental Management Services, S.A. de C.V.)

Mr. Jonathan Green IEMS (Integrated Environmental Management Services, S.A. de C.V.)

Mr. Henry Chaves Kiel IEMS (Integrated Environmental Management Services, S.A. de C.V.)

Ms. Kathleen Kerr RPS

Ms. Martha Pertusa RPS

Mr. Scott Crouch RPS

Burlington's witnesses and experts

Witnesses

Mr. Alex Martinez Burlington Resources Peru Ltd

Mr. Wilfrido Saltos Perenco Ecuador Limited

Mr. Eric d'Argentré Perenco Ecuador Limited

Experts

Mr. John Connor GSI Environmental Inc.

Mr. Gino Bianchi GSI Environmental Inc.

Mr. Danny Bailey GSI Environmental Inc.

Ms. Claudia Sanchez de Lozada GSI Environmental Inc.

Mr. Geoffrey R. Egan Intertek

Mr. Dr. René Bedón Albán Bedón Macías & Associates

Mr. Shahrokh Rouhani NewFields

13.
The Tribunal heard opening and closing statements by counsel as follows:

• For Ecuador: Mr. Eduardo Silva Romero, Mr. Pierre Mayer, Mr. José Manuel García Represa, Mr. Philip Dunham, and Mr. Timothy Lindsay.

• For Burlington: Mr. Nigel Blackaby, Mr. Wade Coriell, and Ms. Tracie Renfroe.

14.
The Tribunal also heard evidence from the fact witnesses and experts listed in paragraph 12.
15.
The Hearing was interpreted to and from English and Spanish. It was also sound-recorded and transcribed verbatim, in real time, in both English and Spanish. Copies of the sound recordings and the transcripts were delivered to the Parties. At the end of the Hearing, the Tribunal and the Parties held a procedural discussion in relation to posthearing matters, including the advisability and feasibility of a site inspection.

D. Post-Hearing Phase

16.
Pursuant to PO21, and after several postponements, the Parties simultaneously filed Post-Hearing Briefs on 3 October 2014.
17.
On 22 December 2014, the Tribunal issued PO24 with regard to Burlington's requests for production of documents, of which one request partially related to the counterclaims proceedings.

E. Site Visit

18.
Having secured the Parties' consent at the Hearing, the Tribunal proceeded with the inspection of a number of sites within Blocks 7 and 21 (the "Site Visit"). The Tribunal informed the Parties by letter of 23 July 2014 that the main objective of the Site Visit was to focus on (i) soil contamination, including issues of topography and groundwater and (ii) land use, including the Sumaco Biosphere Reserve.2 Upon the invitation and with the guidance of the Tribunal, the Parties drew up a Joint Site Visit Protocol (the "Protocol"). For this purpose, the Parties and the Tribunal held two telephone conferences, respectively on 8 October 2014 and 4 November 2014. The Protocol was executed by the Parties on 20 January 2015.
19.
The Protocol was incorporated into PO30 dated 25 March 2015. It set out the scope of the Site Visit, provided for the preparation of joint site information packets ("Joint Site Packets"), addressed pre-Site Visit inspections, detailed the Site Visit itinerary and schedule, provided for a confidentiality regime as well as immunities, and touched on miscellaneous other issues, such as transportation and security, accommodation, food, safety and health.
20.
Up to 12 representatives per Party were permitted to attend at each visited site.3 Counsel for the Parties were invited to make oral presentations at each site and to respond to the Tribunal's questions. The Tribunal also addressed questions to the group of experts and to the fact witness designated by each Party for a given site.4 Follow-up questions by the Parties were not allowed, subject to the Tribunal's permission under certain circumstances (such as the particular relevance of the topic and available time), which was in fact granted on many occasions.5
21.
Under the Protocol, Party presentations were considered to be in the nature of submissions and expert and witness responses to Tribunal questions were deemed to be evidence. Accordingly, the experts and witnesses who participated in the Site Visit made the declaration provided in Rule 35(2) or (3) of the ICSID Arbitration Rules at the outset of the Site Visit.6
22.
The Protocol further specified that any objection regarding the conduct of the Site Visit had to be raised immediately whenever possible and at the latest at dinner time on the day when the incident complained of occurred.7 In fact, informal debriefing meetings were held each day before dinner to discuss the conduct of the Site Visit that day and to seek to agree on possible improvements.
23.
The Joint Site Packets included information relating to the specific sites to be visited (for instance, excerpts from expert reports, pleadings, photographs, maps and/or historical records), their content being limited to evidence already in the record.8 The Parties agreed that no sampling was to be undertaken and that no media would be disturbed (soil, groundwater, sediment, or surface water) before or during the Site Visit.9 Off platform locations could be inspected upon request and disagreements concerning the planned Site Visit were to be resolved by the Tribunal prior to the Site Visit.10
24.
The Site Visit was conducted in English and Spanish with simultaneous translations.11 It was audio- and video-recorded, the audio/video record capturing (i) the presentations by the Parties, (ii) questions from the Tribunal, (iii) answers to such questions, and (iv) any features of the sites requested by the Tribunal.12
25.
The Site Visit took place from 29 March 2015 to 1 April 2015, with 2 April 2015 in reserve. The following persons attended the Site Visit:13

For the Tribunal

Professor Gabrielle Kaufmann-Kohler, President of the Tribunal

Professor Brigitte Stern, Arbitrator

Mr. Stephen Drymer, Arbitrator

Secretary of the Tribunal

Mr. Marco Tullio Montañés-Rumayor

Assistant to the Tribunal

Mr. Magnus Jesko Langer

Ecuador's representatives

Dr. Diego García Carrión, Procuraduría General del Estado Dr. Blanca Gómez de la Torre, Procuraduría General del Estado Dr. Diana Moya, Procuraduría General del Estado

Mr. Eduardo Silva Romero, Dechert (Paris) LLP Mr. José Manuel García Represa, Dechert (Paris) LLP Ms. Audrey Caminades, Dechert (Paris) LLP

Ms. Gabriela González Giráldez, Dechert (Paris) LLP

Burlington's representatives

Ms. Suzana Blades, Burlington/ConocoPhillips Mr. Rick Greiner, Burlington/ConocoPhillips Ms. Laura Robertson, Burlington/ConocoPhillips Mr. John Urby, Burlington/ConocoPhillips14 Mr. Nigel Blackaby, Freshfields Bruckhaus Deringer Ms. Tracie Renfroe, King & Spalding LLP Ms. Jamie Miller, King & Spalding LLP Mr. Thomas Norgaard, Debevoise & Plimpton LLP Mr. Javier Robalino, Paz Horowitz

Ecuador's witnesses and experts Ms. Martha Pertusa, RPS

Mr. Scott Crouch, Di Sorbo Consulting

Mr. Francisco Alfaro, IEMS

Mr. Henry Chaves, IEMS

Mr. Rodrigo Anota, IEMS

Mr. Marcel López, IEMS

Mr. Marco Puente, Petroamazonas

Mr. Pablo Luna, Petroamazonas

Burlington's witnesses and experts Mr. John Connor, GSI Environmental Mr. Gino Bianchi, GSI Environmental

Mr. Danielle Kingham, GSI Environmental Mr. Wilfrido Saltos, Perenco

Technical team

Mr. Daniel Giglio, Interpreter Mr. Charles Roberts, Interpreter Mr. Favio Claure, AV Contractor

Mr. Christian Richard Winter, AV Contractor

26.
The Tribunal and the Parties visited the following sites:

On day 1 (29 March 2015): Yuralpa Pad A;

On day 2 (30 March 2015): Payamino 2/8, Payamino 1/CPF, Coca 8 and Coca CPF;

On day 3 (31 March 2015): Jaguar 1, Jaguar 7/8 and Mono CPF;

On day 4 (1 April 2015): Gacela 2, Gacela CPF, Gacela 1/8 and Coca 15.15

27.
While the Protocol provided that there would be no record of the Site Visit other than the transcript, audio and video recordings mentioned above, it was arranged with the consent of the Parties that an official photograph be taken on day 2 at the Coca 8 site. Such photograph, reproduced below, shows the Tribunal, its Secretary and Assistant, the Party representatives and their Counsel, the experts, the fact witnesses, and the interpreters.

F. Post Site Visit Phase

28.
Pursuant to paragraph 31 of the Protocol and paragraph 2 of PO31 dated 21 April 2015, the ICSID Secretariat made copies of the unedited audio/video recording available to the Parties on 21 April 2015. It also provided the Parties with the "floor" transcripts on 19 May 2015 and circulated the translated transcripts on a rolling basis from 31 May 2015 to 6 June 2015.
29.
Burlington objected to Ecuador's wish to make public use of the audio/video recording. The Tribunal decided that absent the consent of both Parties, under the Protocol and Regulation 22(2) of the ICSID Administrative and Financial Regulations, the audio/video recording and its transcript could not be made available to the public or used outside the present proceedings. However, in light of Ecuador's insistence on its duty of transparency within its domestic legal framework, the Tribunal indicated that if Ecuador wished to use the audio/video recording and/or transcript, it could apply for leave from the Tribunal by (i) identifying with precision the excerpts which it wished to use, (ii) describing the specific purpose for which the identified excerpts would be put to use, and (iii) explaining the reasons why such publication is deemed necessary.16 Ecuador did not seek leave from the Tribunal thereafter.
30.
In accordance with PO31, the Parties filed Post-Site Visit Briefs on 15 July 2015. In PO31, the Tribunal had asked the Parties to jointly file certain depth-integrated maps for the sites inspected during the Site Visit so as to enable the Tribunal to more easily compare the Parties' respective delineations of the alleged contamination and of the areas to be remediated as a result. This request raised difficulties for the parties in terms of the data required to generate these maps and the potential use of data and maps not previously filed in the record. As a result, after a lengthy series of correspondence, the Tribunal advised the Parties on 12 August 2015 that, in light of such unexpected difficulties, it preferred not to receive these maps.
31.
Following the issuance on 14 August 2015 of the Interim Decision on Ecuador's Environmental Counterclaim in ICSID Case No. ARB/08/6, Perenco Ecuador Limited v. the Republic of Ecuador ("Perenco v. Ecuador"), the Tribunal invited the Parties to provide their comments on that decision, which they did on 18 September 2015.
32.
In the course of its deliberations, the Tribunal provided the Parties with status reports on its progress on 13 November 2015, on 25 February 2016, on 27 July 2016 and on 29 November 2016. Upon the invitation of the Tribunal, the Parties filed their cost submissions on 2 May 2016, and their reply submissions on 23 May 2016.
33.
By letter of 18 January 2017, Ecuador requested leave from the Tribunal to submit the audio and video recording of the site visit to Blocks 7 and 21 conducted in Perenco v. Ecuador. By letter of 20 January 2017, Burlington objected to Ecuador's request. After due deliberation, the Tribunal reached the conclusion that the submission of the audio and video recording of the site visit in Perenco v. Ecuador would not assist the resolution of the counterclaims in these proceedings. Accordingly, by letter of 25 January 2017, the Tribunal denied Ecuador's request to file the audio and video recording of the Perenco site visit into the record of this arbitration. Further, by means of the same letter, the Tribunal, declared the proceedings closed as of 25 January 2017 in accordance with ICSID Arbitration Rule 38(1).

III. GENERAL INFORMATION REGARDING BLOCKS 7 AND 21

34.
To the extent relevant for purposes of the counterclaims, the descriptions of fact contained in decisions previously rendered in this arbitration, i.e. the Decision on Jurisdiction and the Decision on Liability, are incorporated into the present Decision. Similarly, terms and abbreviations used there are used here with the same meaning.
35.
This section addresses various characteristics of Blocks 7 and 21 not previously discussed, in particular as regards their geography and the oilfield operations that have been conducted in the Blocks, including by Petroamazonas after July 2009. It is based on allegations of facts that were not disputed or on facts that the Tribunal considers established. Where a fact referred to has been disputed it is noted as such.
36.
Blocks 7 and 21 are located in the provinces of Francisco de Orellana, Napo and Pastaza, in the Amazon region of the Ecuadorian Oriente.17 Situated east of the Andean mountains in North-Eastern Ecuador, the Oriente is located in a slightly undulated plain between 200 and 350 meters above sea-level, characterized by the presence of swamps and numerous water bodies flowing into the Napo River.18

Topographic Map of Blocks 7 and 2119

A. Block 7 (including the Coca-Payamino Unified Field)

37.
Block 7 and the Coca-Payamino Unified Field (the "CPUF"), which extends to the northeast of Block 7, cover an area of nearly 2,000 square kilometers located approximately 160 kilometers east of the capital Quito.
38.
It is also a relatively populated area: the city of Francisco de Orellana (also known as "Coca"), capital of Orellana province, is located on this Block at the confluence of the Napo, Coca and Payamino Rivers.20 Orellana province has a population of 136,000, of which about 73,000 live in the city of Coca.21 Coca is the main hub for oilfield contractors and functions as the platform for numerous oilfield operations throughout the region. The following map shows Block 7:22
39.
A large part of Block 7, including the city of Coca, overlaps with the Sumaco Biosphere Reserve,23 and with the Yasuní Biosphere Reserve, both of which are UNESCO MAB reserves established pursuant to the UNESCO Man and Biosphere Program.24 The following map shows the Sumaco and Yasuní Biosphere Reserves:25
40.
Other than some exploratory wells, such as Zorro and Cóndor, which were drilled by Texaco in the 1970s, Block 7 was first developed by British Petroleum Development Limited ("British Petroleum" or "BP") under a Service Contract concluded in December 1985 ("Service Contract") with the then Ecuadorian State Petroleum Corporation (Corporación Estatal Petrolera Ecuatoriana ; "CEPE").26 At that time, Block 7 did not include the Coca field or the portion of the Payamino field that extended to the north of Block 7 and that were operated by CEPE.27 Following BP's 1986 discovery of oil in the Payamino field, in the northeastern corner of Block 7,28 CEPE's drilling of the Payamino 2 well in 1987 and the drilling by CEPE's upstream subsidiary Petroproducción of the Coca 4 and Coca 7 wells, the conclusion was reached in 1989 that the oil reservoir from the Payamino field extended beyond the borders of Block 7 into the Coca field, thus forming a "common oil deposit".29
41.
As a result, negotiations began for a Unified Exploitation Operating Agreement for the Common Oil Deposits of Basal Tena, Napo "U", and Hollín Superior and Hollín Principal of the CPUF ("Unitization Agreement").30 On 11 October 1990, Oryx (which had acquired 100% of BP's rights under the Service Contract in September 1990) and Petroecuador (which had replaced CEPE in September 1989) executed the Unitization Agreement, providing for alternate operatorship of the CPUF between Oryx and the State-owned enterprise Petroproducción (which exercised Petroecuador's rights under the Unitization Agreement).31 Accordingly, Petroproducción operated the CPUF between February 1991 and February 1994. Oryx operated the CPUF between February 1994 and June 1997, and Petroproducción reassumed operations from June 1997 through February 2000.32
42.
The CPUF is depicted on the map below:33
43.
As shown in the maps at paragraphs 38 and 412 above, Block 7 (including the CPUF) features seven active oil fields: Payamino, Coca, Lobo, Mono, Jaguar, Oso, Gacela, and Cóndor.34 As noted above, Oryx acquired British Petroleum's interest in the Block 7 Service Contract in 1990. In 1993, Santa Fe Minerales del Ecuador, S.A. ("Santa Fe"), Sociedad Internacional Petrolera ("Sipetrol") and Sigdoil, S.A. (later renamed Compañía Latinoamericana Petrolera Numero Dos, S.A. or "Clapsa II") acquired interests in the Service Contract.35 In 1995, Santa Fe transferred its rights to Preussag Energie GmbH ("Preussag").36
44.
In 1999, Oryx became Kerr McGee Ecuador Energy Corporation ("Kerr McGee" or "KmG").37 In March 2000, an amendment to the Service Contract was executed and the contractual framework was migrated into a Production Sharing Contract ("PSC"), which was entered into between Petroecuador on the one hand, and Kerr McGee, Preussag, Sipetrol and Clapsa II on the other hand, for the joint operation of Block 7 and the Coca-Payamino Unified Field.38 As a result of that transaction, Kerr McGee became the joint operator of Block 7 and the Coca-Payamino Unified Field, and the two areas have been jointly administered ever since.39 Under the Participation Contract, the Contractor companies were to invest in the Block at their own risk in exchange for a share of the crude oil produced.40
45.
In the course of 2001, Perenco Ecuador Limited ("Perenco") and Burlington both acquired interests in the Block 7 PSC, Perenco holding 45%41 and Burlington 42.5%.42 Perenco became the operator of the block on 12 December 2002. The Perenco-Burlington Consortium (the "Consortium") then became the 100% operator of Block 7 in September 2005.
46.
By the time the Consortium ceased operating the Blocks in July 2009, Block 7 included six CPFs; multiple platforms – 7 in Gacela, 11 in Mono, 7 in Jaguar, 16 in Coca, 10 in Oso, 17 in Payamino, 4 in Lobo, and a platform in the Cóndor field (a field where no oil was found); fluid lines and pipelines connecting these facilities with each other and beyond; three camps for employees; and a waste management area.43 With Perenco as operator, the production in Block 7 increased from approximately 13,000 barrels a day in February 2003 to approximately 19,000 barrels a day in February 2009.44

B. Block 21

47.
To the south of Block 7, Block 21 covers an area of nearly 1,550 square kilometers in the Eastern-Central Oriente basin about 240 kilometers southeast of Quito.45 Block 21 is shown on the following map:46
48.
Several indigenous communities reside in Block 21, including fourteen Kichwa and three Huaorani communities.47 Block 21 partly overlaps with the Sumaco Biosphere Reserve, the Yasuni Biosphere Reserve, as well as with the Huaorani Indigenous Reserve located in the eastern part of the Block.48
49.
Prior to 1995, Block 21 had been explored by Texaco and Exxon, each drilling a single well in 1972 and 1987 respectively.49 On 20 March 1995 Oryx and Ecuador entered into a Participation Contract for Block 21.50 Burlington acquired a 37.5% interest in the Block in late 2001 and a further 8.75% in 2005 (for a total of 46.25%). Perenco owned the remaining share and was the operator of the Block.
50.
Although exploitation of the Block began on 8 June 2001,51 as late as 2002 Block 21 remained sparsely developed52 with only a small number of wells and no CPF.53 Actual production in the Block began only in 2003 and it was largely developed by the Consortium.54 By July 2009, the Consortium had increased the number of wells from 3 to 3155 and Block 21 contained a total of 6 platforms in the Yuralpa field, 2 platforms in Chonta and Waponi-Ocatoe, two injection wells in Sumino and Nemoca, one camp, one CPF, one waste management area and a pipeline linking Yuralpa and Puerto Napo.56
51.
Burlington highlights the fact that, in contrast to Block 7 which had already been developed by prior operators, Block 21 only accounts for 8.3% of Ecuador's soil remediation claims and 16.7% of the total cost for groundwater remediation.57

IV. SUMMARY OF THE PARTIES' POSITIONS AND REQUESTS FOR RELIEF

A. Ecuador's Position and Request for Relief

52.
Ecuador's case on the counterclaims can essentially be summarized as follows:

i. Under Ecuadorian law, Burlington is strictly liable for any environmental damage found in Blocks 7 and 21. There is environmental damage in Blocks 7 and 21 in the form of significant soil and groundwater pollution. IEMS, Ecuador's environmental experts, discovered that close to 2.5 million cubic meters of soil and all groundwater locations tested (18) are polluted with hydrocarbons, heavy metals or both. Burlington is liable for this damage.

ii. Under the PSCs and Ecuadorian law, Burlington was bound to maintain the infrastructure, and return it to Ecuador, in good working condition in accordance with industry standards. Burlington breached this obligation and is thus liable.

iii. As a result of these breaches, Burlington must pay damages to Ecuador in the total amount of USD 2,797,007,091.42 composed of USD 2,507,107,626 for soil remediation and USD 265,601,700 for groundwater remediation, USD 3,380,000 to complete groundwater studies, USD 3,500,000 for the abandonment of wells in Block 7, and USD 17,417,765.42 for infrastructure damage,58 plus interest and costs.

53.
On this basis, in its Post-Hearing Brief, Ecuador requests that the Tribunal to render an award:

1. Declaring

(i) That Burlington is liable towards Ecuador for the costs of restoring the environment in areas within Blocks 7 and 21 of the Ecuadorian Amazon Region;

(ii) That Burlington is liable towards Ecuador for the costs required to remedy the poor state of the infrastructure of Blocks 7 and 21 left behind by Burlington;

2. Ordering

(i) Burlington to pay damages to allow the State to proceed with full restoration of the environment in areas within Blocks 7 and 21 of the Ecuadorian Amazon Region quantified at USD 2,507,107,626 (or, alternatively, USD 790,465,298) for soil clean up costs and USD 265,601,700 for groundwater remediation costs;

(ii) Burlington to pay the costs for Ecuador to complete groundwater studies in 52 additional locations quantified at USD 3,380,000;

(iii) Burlington to pay the costs of abandonment of wells in Block 7, quantified at USD 3,500,000;

(iv) Burlington to pay damages for its failure to return the Blocks' infrastructure in good condition to Ecuador in an amount quantified at USD 17,417,765.42 with interest at an adequate commercial interest rate from the date of disbursement thereof until the date of the Award;

(v) Burlington to pay all the costs and expenses incurred in this arbitration in connection with Ecuador's counterclaims, including but not limited to Ecuador's legal and expert fees and costs and ICSID's other costs; and

(vi) Claimant to pay interest at an adequate commercial interest rate on all amounts stated in the preceding paragraphs from the date of the Award until the date of full payment; and

3. Awarding

(i) Such other relief as the Tribunal considers appropriate.59

54.
In its Post-Site Visit Brief, Ecuador has expressly maintained these prayers for relief.60

B. Burlington's Position and Request for Relief

55.
In essence, Burlington makes the following submissions:

i. Ecuador's multi-billion environmental claim is no more than tactical retaliation fabricated after Burlington lodged its ICSID claim, with the objective of offsetting the significant damages that are owed to Burlington by Ecuador for its unlawful seizure of the Blocks.61 Similarly, the infrastructure claim is opportunistic and lacks merit.62 It was fabricated to provide additional support for Ecuador's environmental claim. Not satisfied in having cashed in over USD 4 billion in "unanticipated" oil revenues from the seizure of the Blocks, Ecuador now seeks USD 2.6 billion for an alleged "tremendous environmental harm" caused by a "series of small incidents", and an additional USD 17 million to upgrade infrastructure it took from the Consortium.63

ii. Ecuador in fact seeks to impose on Burlington responsibility for environmental conditions preexisting the Consortium's operation of Blocks 7 and 21, and even preexisting any human activity on the Blocks. Moreover, whatever environmental damage may be present in the Blocks today is the result of, or has been severely exacerbated by, Ecuador's own operation and indeed expansion of oilfield operations.

iii. Ecuador's claims must be rejected in toto, since Ecuador relies on invented legal tests that have never been applied in practice and are scientifically unsupported. Alternatively, the Tribunal should evaluate the reasonable costs of remediating "the pockets of exceedances at the two sites that are possibly attributable to the Consortium, as would have occurred in a normal orderly handover, absent Ecuador's unlawful actions".64 These costs amount to USD 1.09 million and include the remediation of Yuralpa Pad A and Jaguar 1 Area 3T.

56.
Accordingly, in its Post-Hearing Brief, Burlington requests the Tribunal to grant the following relief:

(a) DISMISS with prejudice Ecuador's environmental counterclaims in their entirety;

(b) In the alternative, determine that the reasonable cost of remediating the exceedances possibly attributable to the Consortium, including the costs of closing four pits and abandoning seven wells is no more than USD 1.09 million;

(c) DECLARE that, beyond the exceedances assessed by the Tribunal, Burlington has no further liability for environmental harm in Blocks 7 and 21;

(d) DISMISS with prejudice Ecuador's infrastructure counterclaims in their entirety;

(e) ORDER Ecuador to pay all of the costs and expenses of this arbitration, including Burlington's legal and expert fees, the fees and expenses of any experts appointed by the Tribunal, and ICSID's other costs;

(f) ORDER Ecuador to pay compound interest on the sum awarded in (e) above, until the date of effective and complete payment, at a rate of 4 percent compounded annually, or at such a rate and for such a period of compounding as the Tribunal considers just and appropriate in the circumstances; and

(g) AWARD such further and other relief as the Tribunal considers appropriate.65

57.
The Post-Site Visit Brief did not change these requests for relief.

V. ANALYSIS

58.
Having considered the Parties' extensive submissions as well as the very extensive evidentiary record, including the evidence gathered during the Site Visit, the Tribunal now proceeds with its analysis of the counterclaims. It will first address various preliminary matters (Section A). Thereafter, it will review Ecuador's environmental counterclaims (Section B) and then turn to the infrastructure counterclaims (Section C) before concluding on the counterclaims, including on the avoidance of double recovery in connection with the counterclaims pending against Burlington's consortium partner Perenco (D). Finally, the Tribunal will discuss interest (VI) and costs (VII) before setting out its decision (VIII).

A. Preliminary Matters

59.
The present section deals with (1) the Tribunal's jurisdiction over the counterclaims, (2) the scope of this Decision, (3) the relation between this arbitration and the Perenco arbitration, (4) the law applicable to the merits of Ecuador's counterclaims, and (5) certain evidentiary matters.

1. Jurisdiction

60.
The Tribunal's jurisdiction in respect of Ecuador's counterclaims is not challenged, and rightly so. As noted in PO8, it derives from an agreement entered into by the Parties on 26 May 2011.66 In this agreement, Burlington and Ecuador expressed their agreement and consent that this arbitration is the "appropriate forum for the final resolution of the Counterclaims arising out of the investments made by Burlington Resources and its affiliates in Blocks 7 and 21, so as to ensure maximum judicial economy and consistency".67
61.
Burlington committed not to raise any jurisdictional objections.68 The Parties further agreed that (i) the Tribunal's decision would be final and binding,69 (ii) Ecuador (including its emanations, agencies, instrumentalities, subdivisions and controlled corporations) waived its right to file the counterclaims against Burlington, its subsidiaries or any other corporation in the ConocoPhillips Group before "any jurisdiction whatsoever whether arbitral or judicial, national or international except for this Arbitration".70
62.
Since the agreement just described deals with jurisdiction over counterclaims, one must in addition refer to Article 46 of the ICSID Convention. Article 46 allows for counterclaims "arising directly out of the subject-matter of the dispute provided that they are within the scope of the consent of the parties and are otherwise within the jurisdiction of the Centre".71 These conditions are met here: (i) the counterclaims arise directly out of the subject-matter of the dispute, namely Burlington's investment in Blocks 7 and 21; (ii) they are within the scope of the Parties' consent to ICSID arbitration which is manifested in the agreement just referred to; and (iii) they also fall within the jurisdiction of the Centre as circumscribed by Article 25 of the ICSID Convention (legal dispute arising out of an investment, and nationality requirement).

2. Scope of this Decision

63.
This Decision resolves one part of the dispute which forms the subject matter of this arbitration, namely Ecuador's environmental and infrastructure counterclaims raised in the course of the proceedings over Burlington's claims against Ecuador. Burlington's claims, which form the other part of this dispute, have been dealt with in part in the Decision on Jurisdiction of 31 May 2010 and the Decision on Liability of 14 December 2012. Ecuador's request for reconsideration of the latter decision as well as the quantification of Burlington's claims will be settled in the forthcoming Award.

3. Relation with the Perenco arbitration

64.
In the separate ICSID arbitration initiated by Burlington's Consortium partner, Perenco, Ecuador has raised counterclaims essentially resting upon the same facts and legal arguments as those before this Tribunal. On 11 August 2015, the Perenco tribunal issued an Interim Decision on the Environmental Counterclaim (the "Perenco Decision"), in which it ruled on certain issues of fact and law with a view to narrowing its analysis of the counterclaims.72
66.
With respect to item (i), Ecuador opposed the possibility of a consolidation of the counterclaim proceedings.74 While acknowledging that the two arbitrations were "parallel", Ecuador disagreed with Burlington that they were "nearly identical".75 At the end of the Hearing, the Tribunal enquired whether the Parties would accept some level of coordination between the two tribunals.76 Subject to consulting with Perenco, Burlington was in favor of such coordination if it could reduce the risk of contradictory decisions.77 By contrast, Ecuador maintained its opposition to consolidation and stated that it was "too late" for a coordination between the two tribunals.78 Lacking the consent of the Parties, the Tribunal had no choice but to refrain from any coordination with the Perenco tribunal.
67.
In connection with item (ii) above, the issue arose whether the Parties could rely on the transcripts of the Perenco hearing. Ecuador opposed the introduction into the record of the Perenco transcripts on the grounds that it "would threaten the structural integrity of the [present arbitration, as well as] violate the separate identity of the proceedings and create an unacceptable risk of pre-judgment".79 Following various exchanges, the Parties agreed to admit the Perenco transcripts, but only for impeachment purposes during cross-examination of each other's witnesses.80 Accordingly, except for a few instances where specific passages of the Perenco transcript were used for this purpose at the Hearing, the Perenco transcripts are not part of this record.81
68.
The Parties were asked to comment on the third issue, concerning the impact of the Perenco Decision on this case, after that decision was rendered. In its comments dated 18 September 2015, Ecuador reiterated that the two sets of proceedings are separate and invited the Tribunal to consider only the arguments and evidence before it in this present arbitration.82 For its part, Burlington commented that the Perenco Decision "substantially narrowed the disputed issues" and therefore invited the Tribunal to carefully assess that decision's conclusions.83 Burlington further stated that the Site Visit effectively eliminated the need to appoint an independent expert or to engage in further fact-finding.84
69.
The Tribunal is mindful of the separate nature of the two arbitrations and of its duty to resolve the dispute before it solely on its own record and merits. This said, the Tribunal is also mindful of the risk of double recovery, to which it will revert, and of the potential risk of contradictory decisions. For reasons linked to the value of coherence of the legal system, it considers that contradictory decisions on identical issues should be avoided to the degree possible without sacrificing any party's rights of due process or fairness. While ruling on the basis of the record in this case exclusively, the Tribunal will refer to the Perenco Decision in those instances where, in spite of the desire to avoid contradictions, it reaches a conclusion different from that of the Perenco tribunal.

4. Applicable Law

71.
The law governing the conduct of the procedure on the counterclaims is addressed in the rules of the ICSID Convention, the ICSID Arbitration Rules and the Tribunal's procedural orders, in particular PO8 (Proceedings on Counterclaims), PO14 (Hearing on Counterclaims), PO21 (Post-Hearing Matters), PO30 (Site Visit), and PO31 (Post Site Visit Matters).
72.
As regards the substance of the dispute, it is undisputed that Ecuadorian law applies to both the environmental and the infrastructure counterclaims. This being so, there is no common ground between the Parties with respect to the applicability to the environmental counterclaims of the Block 7 and Block 21 PSCs and the relevance of international law.
73.
With respect to the environmental counterclaims, Ecuador affirmed throughout the proceedings that its action is solely based on Ecuadorian tort law, as opposed to contract law, although it relies on several provisions of the PSCs to support its interpretation of the strict liability regime under Ecuadorian law.89
74.
According to Article 42(1) of the ICSID Convention, the Tribunal shall decide the dispute "in accordance with such rules of law as may be agreed by the parties", absent which "the Tribunal shall apply the law of the [host State] … and such rules of international law as may be applicable".90 Here, Ecuador brings a tort action, and neither Party has argued that the choice of (Ecuadorian) law in the PSCs encompasses torts. Therefore, the Tribunal will apply Ecuadorian tort law, not as the law chosen by the Parties under the first leg of Article 42(1) of the ICSID Convention, but as the law of the host State under the second leg of that provision. The relevance of this distinction is that, under the second leg, international law also "may be applicable". According to prevailing case law,91 it is left to the Tribunal's discretion to apply either municipal or international law depending on the type of issue to be resolved. Subject to any particular matter that may call for the application of international law, which will be discussed if and when it arises in the analysis, the Tribunal will thus apply Ecuadorian law to the environmental counterclaims. The fact that the dispute is governed by tort law does not mean, however, that the Tribunal may not look to the PSCs to inform its understanding of the obligations assumed by the Consortium in terms of environmental stewardship.
75.
With respect to the infrastructure counterclaims, Ecuador argues that Burlington's liability for the poor condition of the infrastructure arises both under the Block 7 and 21 PSCs and under Ecuadorian law.92 In this regard, the PSCs contain a choice of Ecuadorian law and, accordingly, the Tribunal will apply such law (under the first leg of Article 42(1) of the ICSID Convention) as well as any relevant contractual provisions of the PSCs.

5. Assessment of evidence

76.
At the Hearing, in accordance with its powers in evidentiary matters, the Tribunal suggested to the Parties that an inspection of various sites would assist its understanding of several key issues connected to the environmental counterclaims, such as soil and groundwater contamination, the experts' sampling programs, land use, the topography of the Blocks and specific sites, and the extensive modelling and delineation employed by the Parties' respective experts. As discussed above, both Parties having given their consent prior to the close of the Hearing, the Tribunal decided to conduct a Site Visit in accordance with ICSID Arbitration Rule 34(2)(b).93
77.
Further, in light of the technical nature of many of the factual issues at stake in the environmental counterclaims, including issues explored during the Site Visit, the Tribunal put to the Parties whether, in their view, it should seek the assistance of a Tribunal-appointed expert in addition to the experts offered by the Parties. The Tribunal ultimately concluded in the negative, because (i) neither Party requested the appointment of a Tribunal expert (for example, Burlington stated that with the benefit of the Site Visit it considered that the Tribunal was fully equipped to decide the environmental counterclaims "without the need for an independent expert or additional fact finding"),94 (ii) it had before it extensive evidence from qualified experts, (iii) it also had detailed evidence from fact witnesses, including persons with long-time familiarity with the fields, (iv) it had visited relevant sites and had thus acquired a further, firsthand, visual understanding of the physical environment in which the oil exploration and production operations were conducted, of the situation of the platforms and the other installations, and of the terrain and areas modelled and delineated by the experts. In addition, (v) developing and engaging in a process involving a Tribunal-appointed expert would have caused delays and costs which the Tribunal considered would be disproportionate to the possible usefulness of such an exercise in the circumstances, and contrary to the Tribunal's duty of efficiency.
78.
Several issues also arose in the context of the document production exercise, which will be addressed if and to the extent relevant to the Tribunal's analysis of particular claims or submissions discussed later in this Decision.

B. Environmental Counterclaims

79.
In a nutshell, Ecuador claims that the Consortium abandoned Blocks 7 and 21 leaving behind a massive environmental catastrophe, a claim which Burlington strongly denies (save for an admission of liability in an amount of approximately USD 1.09 million). The Parties are in sharp disagreement on nearly all issues of fact and law raised by these counterclaims. To untangle this disagreement, the Tribunal will first set out the Parties' positions on matters of law (Section 1). It will then review the regulatory framework applicable to oilfield operations in Ecuador, in particular as it relates to environmental protection and stewardship (Section 2). This will then allow the Tribunal to assess the legal positions advocated by the Parties (Section 3). In the light of this assessment of the law, the Tribunal will thereafter summarize the Parties' positions on the facts and proceed to analyze such facts on soil contamination (Section 4), mud pits (Section 5), groundwater (Section 6), and well site abandonment (Section 7).

1. Parties' positions

1.1 Ecuador's position

1.1.1 Burlington is liable for all environmental harm in Blocks 7 and 21

80.
According to Ecuador, Burlington is incorrect when alleging, and in any event it fails to establish, that Ecuador's counterclaims are a mere set-off strategy. As Professor Crespo explained at the Hearing, Ecuador's position in and approach to these proceedings are proof that Ecuador is pursuing its constitutional obligation to vindicate any environmental harm caused by the Consortium in Blocks 7 and 21.95 Furthermore, Burlington's novel theory on sustainable development does not withstand scrutiny. Sustainable development cannot amount to a "license to pollute".96 Quite to the contrary, this principle encompasses the polluter-pays principle, and Ecuador only allowed oil operations to be carried out because such operations are highly regulated and any environmental harm must be fully restored under the maxim ubi emolumentum, ibi onus.97 If one were to accept Burlington's permissible limits theory, it would signify that "the Ecuadorian environment and the society at large will be forced to assume part of the costs of Burlington's oil activities […]"; an unacceptable position.98 As stated by Professor Crespo, sustainable development implies that operators internalize environmental costs.99

1.1.2 Burlington is strictly liable for any environmental harm found in Blocks 7 and 21

81.
According to Ecuador, under the 2008 Constitution, oil operators bear strict liability for environmental harm. Strict liability means that oil operators are liable upon a mere showing of environmental harm in the areas where they performed oil operations. The State need not prove that the oil operator is at fault, nor that there is a causal nexus between a breach of the duty of care and the environmental damage.100
82.
Under the strict liability regime, the State must only prove the existence of an economic activity in the relevant area entailing serious risks to the environment and a negative impact on the environment of the type that comes with such perilous economic activity. For Ecuador, Burlington is wrong in seeking to downplay the significance of the impact of its operations on the environment. Indeed, Burlington's "significant admission" of liability to restore soils in the amount to USD 10 million is itself a manifestation of a "significant impact".101

1.1.3 The 2008 Constitution applies to damage discovered after its entry into force

83.
In terms of temporal application, Ecuador submits that the 2008 Constitution applies to any environmental harm discovered after its entry into force, no matter when the harm actually occurred. Burlington did not contradict Ecuador's legal experts, Professors Crespo and Andrade, who stated that the 2008 Constitution applies, at the very least, to all environmental harm that occurred after its entry into force. This follows from Articles 11.3 (direct application)102 and 396 of the 2008 Constitution (strict liability for environmental torts), the latter providing for full reparation of any environmental harm ("todo daño al ambiente").
84.
Professor Bedón, Burlington's expert, also did not deny that the 2008 Constitution could apply to environmental harm preceding the entry into force of the Constitution if a particular norm dictated its application.103 Such norm does exist, so argues Ecuador, as the court confirmed in Nelson Alcívar in reliance on the principle in dubio pro natura (enshrined in Article 395(4) of the 2008 Constitution),104 in the following terms:

"It is important that we highlight that the constitutional principal [sic] that states that in doubt, the rule that most favors environmental protection shall apply, as it is a standard that generates ample protection, the matters contemplated by the current Constitution will apply, as it is a standard that creates a broad protection to the environment, above the matters outlined by the Environmental Management Law or the 1998 Constitution. The provisions of the current Constitution as it relates to environmental issues and its protection is also preferably applied, since, in procedural matters, the rules in force at the time of filing the action apply, and not those that were in force when the legal situation was created".105

1.1.4 Environmental claims are imprescriptible

85.
The Respondent further emphasizes that in Nelson Alcívar the court also held that, under Article 396(4) of the 2008 Constitution, environmental claims are "imprescriptibles", that is subject to no statute of limitation, and that the Constitution applies to all environmental claims filed after its entry into force.106 Accordingly, since Ecuador filed its environmental counterclaims in January 2011, i.e. after the entry into force of the 2008 Constitution, Burlington cannot rely on Article 2235 of the Ecuadorian Civil Code (the "Civil Code" or "CC") to argue that claims relating to environmental harm caused prior to January 2007 are prescribed, i.e. time-barred.
86.
Burlington's argument that Ecuador's claims are time-barred with regard to harm resulting from acts prior to January 2007, i.e. four years before Ecuador first filed its counterclaims, is also misplaced. Burlington ignores that Article 2235 of Ecuador's Civil Code exclusively governs fault-based liability. The present environmental claims are for strict liability and are not subject to a statute of limitation according to Article 396(4) of the 2008 Constitution, as acknowledged by the Claimant's expert Professor Bedón at the Hearing.107 Even if Article 2235 of the Civil Code were to apply, the time period for prescription (the "clock") would only start running upon discovery of the harm, which in this case was after the Consortium abandoned the Blocks in July 2009.108
87.
This understanding of Ecuadorian law is reinforced by Burlington's policy of concealing and failing to report environmental harm,109 which incidentally also does away with Burlington's "utterly misplaced" argument that environmental impact studies approved by Ecuador would somehow allow Burlington to evade liability.110
88.
In any event, by acquiring its interest in the Blocks and contractually assuming the environmental liabilities of prior operators, Burlington waived its right to invoke any statute of limitations.111
89.
In the alternative, so argues Ecuador, Burlington has not denied that norms of public order may be applied with retroactive effect, as was held in Baquerizo.112 The Claimant's expert Professor Bedón acknowledged at the Hearing that Article 14(2) of the 2008 Constitution characterizes the constitutional environmental protection regime as a matter of public order.113 Incidentally, the same approach has been adopted in various other jurisdictions.114
90.
Be that as it may, for the Respondent, even if the 2008 Constitution had no retroactive effect, the result would be the same. Read in conjunction with the Delfina Torres decision, the environmental regime set out in the 1998 Constitution also imposes strict liability (responsibilidad objectiva) for environmental torts.115 As explained by the Respondent's expert Professor Andrade at the Hearing, the only differences between the 1998 and the 2008 regimes concern the burden of proof and the statute of limitations.116 Similarly, the court in Aguinda v. Chevron confirmed the objective nature of the liability, holding that "[l]iability for environmental harm is always strict, regardless of whether a legal provision exists that establishes so".117

1.1.5 Burlington bears the burden to prove the inexistence of harm

91.
As regards the burden of proof, Ecuador notes that Burlington "conveniently remained silent" on Article 397(1) of the 2008 Constitution, which places the burden of proving the inexistence of potential or real environmental damage on the operator. Its expert Professor Andrade explained that Ecuador must only establish the existence of a "negative impact", that is an "alteration of the natural environment" (something far less than "harm") for the burden of proof to shift to Burlington:118

"[W]hat the Constitution did in relation to these two components – harm on the one hand and quantum on the other, the quantum of environmental harm – what it did it was transfer the burden of proof to the person carrying out the activity or the defendant. So, the only modification of the 2008 Constitution in connection with this matter is that what must be shown, the evidence that must be given by the Claimant, is the negative impact, obviously the impact, negative impact upon nature. This is as far as he has to go".119

92.
More specifically, Ecuador argues that it only needs to establish that the Blocks have been "adversely impacted", upon which the burden shifts to Burlington to show that this negative impact is "not significant within the meaning of the Ley de Gestión Ambiental".120 It is sufficient for it to establish that Burlington performed environmentally risky activities in a certain area and that an adverse environmental impact of the kind that may be caused by such activities is present in that area.121 Finally, Professor Andrade confirmed without being contradicted, that pursuant to Article 2217 of the Civil Code, even if the harm to the Blocks were caused by several persons, including prior operators, Burlington remains jointly liable with all third parties for all such harm.122

1.1.6 Under strict liability, causation is presumed

93.
Ecuador argues primarily that under the strict liability regime, causation is presumed and Burlington can only escape liability if it shows that force majeure, Ecuador, or a third party caused the environmental harm.123 Burlington cannot, however, escape liability for the actions or omissions of prior operators since the PSCs require the operator to restore the Blocks so as to "allow the potential return to environmental conditions similar to those encountered at the beginning of the operations".124 In other words, Burlington contractually agreed to repair any environmental harm found in the Blocks that may have been caused by prior operators.125 Nor can Burlington escape liability by showing that it took appropriate measures to prevent damage during its tenure as operator.126
94.
More specifically, the Respondent cites to Article 396 of the 2008 Constitution under which "[l]iability for environmental harm is strict".127 It also refers to Article 397(1) pursuant to which the burden to prove the inexistence of potential or real environmental harm falls on the operator.128 In addition, in its Final Report on Natural Resources and Biodiversity, the Constituent Assembly of Ecuador indicated that it is not for the claimant to prove causation, but for the respondent to disprove its existence.129 This approach was first applied in Delfina Torres130 and again in Aguinda where the National Court of Justice confirmed that strict liability presumes causation.131
95.
Once the existence of harm in the area of oil operations is proven, the oil operator is liable, unless it can show that the harm results from force majeure, the actions or omissions of the victim, or a third party. As Professor Crespo explained, the rationale for such a regime is that in most instances it is very difficult or impossible to establish that the operator's fault caused the environmental damage.132
96.
Ecuador concludes that, in the present case, there is environmental harm in Blocks 7 and 21. Since the Consortium has not demonstrated that the damage is due to force majeure, a third party, or the victim, the Consortium is strictly liable for the harm and must fully restore the ecosystems or pay damages to allow the State to proceed with such restoration.133
97.
Furthermore, Ecuador argues that, contrary to Burlington's assertion, the presumption of causation applies not only to entities operating simultaneously in the Blocks but also to successive operators. Article 2217 of the Civil Code reinforces this view by providing for joint liability "where a single harm has been caused by several tortfeasors".134 Joint liability in environmental law precisely addresses the issue of contamination resulting from years of environmentally risky operations. In any event, according to the Respondent, Burlington failed to establish that Petroamazonas caused any harm in Blocks 7 and 21.135
98.
Finally, Ecuador argues that it has proven that Blocks 7 and 21 have suffered widespread environmental harm and must be fully restored to background values, or, in the alternative, to the sensitive ecosystem standard set out in Reglamento Ambiental para las Operaciones Hidrocarburíferas en el Ecuador ("RAOHE"). It also emphasizes that that Burlington must bear the full costs of restoration.

1.1.7 Burlington must fully restore the damaged environment to background values or, in the alternative, to the sensitive ecosystems standard

a. Environmental harm is distinct from violations of regulatory permissible limits

99.
In reliance on the 2008 Constitution, Ecuador argues that Burlington is under an obligation to repair the harm "found in the Blocks" and to restore the environment to its original condition (restitutio in pristinum).136 Ecuador also points to Burlington's identical obligation in Clause 5.1.20.10 of the Block 7 PSC and Clause 5.1.20 of the Block 21 PSC.137
100.
As Ecuador's expert Professor Crespo explains, the notion of "environmental harm" under the Constitution was intentionally left undefined in order to cover as many types of environmental harm as possible. The notion includes the loss of biodiversity as well as impacts on air quality, health of local populations, cultural heritage, and the local economy.138
101.
According to the Respondent, Burlington's reliance on the regulatory permissible limits is misplaced. As Ecuador puts it, "[b]elow the numerical permissible limits, pollution is not simply dirt that can be swept under the carpet".139 While a harmful activity may be conducted lawfully, it still gives rise to an obligation to repair where harm occurs. The Claimant confuses fault and harm, which are independent from each other.
102.
In reliance on decisions of the Colombian Constitutional Court and on scholarly writings, specifically Betancor Rodríguez and Peña Chacón, Ecuador submits that RAOHE and the Texto Unificado de Legislación Ambiental Secundaria ("TULAS") do not define compensable environmental harm. That definition is found in the Environmental Management Law of Ecuador (Ley de Gestión Ambiental) ("EML"). The EML defines environmental harm as "all loss, reduction, detriment or significant damage to the preexisting conditions in the environment or one of its components. It affects the functioning of the ecosystem or the renewal of its resources".140
103.
Therefore, under the EML, "the significance of the harm depends exclusively on the magnitude of the impact on the environment", whereas RAOHE and TULAS make environmental harm dependent on the economic use of the land affected.141 In other words, the significant harm standard under the EML "is perfectly operational, without the need for numerical standards",142 which is reinforced by the fact that the EML makes no reference to limits.143 In addition, asserts the Respondent, Article 396 of the 2008 Constitution mandates that all environmental damage is to be fully repaired, and Article 397 similarly mentions full restoration. In any event, TULAS also requires that contaminated soils be restored to their "prior condition".144 This "prior condition" is represented by the background values established by IEMS.

b. Burlington must remediate contaminated soils back to background values

104.
What Ecuador refers to as its "main case" is its claim that full restoration under Article 396(2) in conjunction with Article 71 of the 2008 Constitution requires remediation of all environmental harm to the level of the so-called background values, i.e. restoration of the environment to a condition equivalment to its condition prior to the initiation of oil operations. In Ecuador's submission, any "alteration to the natural state"145 qualifies as contamination or harm and must be fully remedied back to so-called background values. There is no "right to freely pollute up to the permissible limits".146
105.
This approach is consistent with case law of the highest court in Ecuador147 and was endorsed by Professor Andrade's oral evidence.148 From a scientific point of view, IEMS also testified that the regulatory criteria are insufficient to fully protect the environment or the health of the population. In IEMS's opinion, there is contamination whenever a contaminant exceeds the background values triggering the duty of the operator to "remediate the affected area so as to return it to its natural state".149
106.
Ecuador argues six grounds for its background values case. First, the Parties' experts agree that the 2008 Constitution requires full restoration. Since the permissible limits set forth in RAOHE and in TULAS do not represent full restoration, the Tribunal should disregard them and apply the background values established by IEMS.
107.
Second, as stated above, RAOHE and TULAS do not define compensable environmental harm. They merely set limits within which hazardous activities may be lawfully conducted in a manner so as to prevent to the extent possible the occurrence of harm.150
108.
Third, the Hearing confirmed that RAOHE and TULAS do not purport to define environmental harm, but that their purpose is to regulate the activities of operators.151 The "hollow theory of tolerable and intolerable impact" put forward by the Claimant's expert Professor Bedón was put to rest when, confronted with Professor Peña Chacón's comprehensive quote, he had to admit that the strict liability regime excluded a requirement of fault. For Ecuador, "[i]f the fault element disappears with strict liability, then infringing regulations such as those contained in RAOHE and TULAS regulations becomes irrelevant to assessing the existence and extent of environmental harm".152
109.
Fourth, RAOHE and TULAS are merely risk management tools. They seek to prevent environmental harm and therefore apply ex ante.153 Compliance with administrative regulations is no shield against environmental harm. For instance, harm can occur as a result of the "cumulative effect of certain emissions […] where no permissible limits have been infringed".154
110.
Fifth, as Professor Crespo confirmed, RAOHE and TULAS govern administrative liability, not civil liability for environmental harm. While the administrative liability regime aims at preventing the occurrence of environmental harm, the civil liability regime provides for full restoration.155
111.
Sixth and last, environmental harm cannot be determined in the abstract by ‘blindly' applying the permissible limits provided in RAOHE and TULAS. It must be assessed on the facts specific to each case.156 Referring to Professor Andrade's evidence, Ecuador submits that the permissible limits in RAOHE and TULAS are irrelevant "because any harm must be assessed on site "157 and because "ex ante empirical data cannot determine, in advance, whether the renewability of a specific ecosystem has been impaired, let alone significantly impaired".158 More importantly, RAOHE and TULAS do not distinguish between "types and characteristics of lands (deserts, forest, rainforest, mountains, islands…)", making them unsuitable for the sensitive ecosystem of the Amazon rainforest.159 Finally, while the limits set in RAOHE and TULAS are bound to vary over time, thus making them unreliable, the notion of environmental harm in the EML will not change over time.160
112.
In sum, it is Ecuador's submission that remediation to background values is no extraordinary requirement. Comparative law corroborates this approach.161 Burlington's argument that it has the right to pollute up to the permissible limits is incompatible with "two of the most fundamental principles of environmental law".162 The first is the polluter-pays principle according to which operators must internalize the environmental costs of their economic activities, as enshrined in the 2008 Constitution, general tort law, and the Rio Declaration. The second fundamental principle is that nature itself is a bearer of rights entitled to protection and reparation in the event of environmental harm under Article 71 of the Constitution.163 According to Ecuador, the 2008 Constitution also establishes that, in case of doubt, "environmental regulations should be interpreted in the sense most favourable to the protection of Nature" (in dubio pro natura).164 Hence, absent proof to the contrary, any damage found must be considered significant and thus compensable.

c. In the alternative, Burlington must remediate contaminated soils back to the sensitive ecosystems standard

113.
Alternatively, if the Tribunal were to discard the background values approach and refer to RAOHE and TULAS as the basis for defining environmental harm, it should apply the permissible limits for sensitive ecosystems in those laws,165 for the two following reasons. First, the future land use, taken into account under RAOHE to determine the applicable limits, must be taken to refer to the time "when oil operations will end", as opposed to the end of a given operator's activities, as was clarified at the Hearing.166 Oil operations are temporary and the oilfields in Blocks 7 and 21 are "destined to be absorbed again by rainforest".167 As a result, the sensitive ecosystem standard must apply, which implies restoration to the state of pristine rainforest.
114.
Second, the oilfields were "carved out" of the sensitive ecosystems of the Amazonian rainforest, an area of "significant ecological importance" vulnerable to change and in need of protection.168 According to Ecuador, Block 7 is situated in a "fragile and diverse ecosystem" with numerous rivers and streams, such as the Napo River crossing the entire Block, the Suno River crossing the Oso field, and the Payamino River in the northern part of Block 7.169 Similarly, Block 21 is an ecologically rich but vulnerable area, characterized by rainforest and drained by several rivers, of which the most important is the Napo River.170 Burlington's argument that Blocks 7 and 21 do not qualify as sensitive ecosystems under RAOHE and TULAS, because they are not so-called "designated protected areas" under Ecuadorian law, is ill-founded. According to Ecuador, by devising a distinction in RAOHE between industrial use, agricultural use and sensitive ecosystem, "the Ecuadorian legislator obviously intended for the Amazonian rainforest, which does not fulfill any industrial or agricultural use, to be included in the latter category".171
115.
In this respect, Ecuador stresses that the map showing the ‘ Sistema nacional de áreas protegidas ' ("SNAP"), filed by Burlington – disproves the latter's position that would imply that all non-colored areas of the map would be industrial or agricultural land.172 Even GSI, the Claimant's environmental expert, acknowledged at the Hearing, says Ecuador, that some sites within Blocks 7 and 21 were "sensitive ecosystems".173 More significantly, Burlington's position is contradicted by the wording of RAOHE, which defines sensitive ecosystems in a non-exhaustive way by referring to areas "such as the National Heritage of Natural Areas and others identified in the corresponding Environmental Study (tales como Patrimonio Nacional de Areas Naturales y otros identificados en el correspondiente Estudio Ambiental)".174
116.
Ecuador further emphasizes that Blocks 7 and 21 are located in the Napo River basin, which is "one of the most bio-diverse regions in the world".175 Similarly, it notes that the "greater part of the Blocks is entirely within the Sumaco and Yasuni Biosphere Reserves", while a significant portion of Block 21 lies additionally in the Huaorani indigenous reserve.176
117.
For Ecuador, the "mega-diversity" of the area is further supported by environmental impact studies carried out in 2010 as well as by IEMS, which note the high diversity of the area and the fact that some species are endangered.177 More importantly, in various environmental impact studies, the Consortium itself concluded "that the characteristics of the Blocks 7 and 21 area were those of a sensitive ecosystem".178 In addition, Burlington's witness Mr. Wilfrido Saltos also recognized the dynamic nature of the ecosystem around the platforms.179

1.2 Burlington's position

118.
In brief, Burlington's submission is that Ecuador's USD 2.6 billion environmental claim is a "work of fiction", unsupported by the facts and ill-founded as a matter of law.180 Essentially, to succeed with its claim, the Respondent must show that (i) the Consortium acted in breach of its duty of care; (ii) Ecuador suffered harm; and (iii) there is a causal nexus between the act and the harm. Ecuador has not established any of these elements.
119.
In support of its defense, Burlington contends that (i) the Consortium was a responsible operator; (ii) Ecuador misstates the applicable legal standards; (iii) it uses the wrong definition of harm; (iv) it erroneously relies on background values; (v) it applies the wrong standards when it alternatively refers to regulatory criteria; (vi) it does not prove causation; and (vii) it raises a claim that is time-barred.

1.2.1 The Consortium was a responsible and environmentally conscientious operator

120.
Ecuador's attempt to depict the Consortium as a careless operator is contradicted by Ecuador's own data. Indeed, there were no material oil spills during the Consortium's operations and there are only limited pockets of regulatory exceedances in the Blocks.181 In reality, the Consortium "was a responsible operator and good corporate citizen that prioritized environmental conditions, complied with government regulations, promptly reported and remediated spills and invested in the neighboring communities".182 In particular, the environmental audits attested to the Consortium's "overall regulatory compliance and sound stewardship of the Blocks".183 In any event, Burlington asserts that there can be no liability so long as the Consortium acted as a bonus pater familias or a reasonably prudent operator.184
121.
As a responsible member of the Consortium, Burlington is willing to accept responsibility for exceedances above regulatory criteria for two sites "at which it cannot definitively link harm to non-Consortium activities (Yuralpa Pad A and Jaguar 1 Area 3T)" and to bear the costs to close four open pits and abandon seven wells.185

1.2.2 Ecuador misstates the applicable legal standards

122.
First, fault-based and not strict liability governs Ecuador's environmental claim. Ecuadorian law imposed strict liability in October 2008. Under Article 7 of the Civil Code, "the law provides only for the future; it has no retroactive effect [...]" (Translation by the Tribunal). Hence, the strict liability regime of the 2008 Constitution does not govern Ecuador's claim in respect of harm that predates the entry into force of the Constitution in October 2008.186 Accordingly, Burlington cannot be held liable unless it is proven that the Consortium breached its duty of care.187 No such proof has been addressed.
123.
Burlington further notes that, as Professor Bedón explains, the relevant date to determine the applicable legal regime is the date of the occurrence of the act. This date is consistent with the statute of limitation for torts in the Civil Code, which runs from the date on which the harmful act occurred.188 As a result, it rejects Ecuador's argument that strict liability governs because that regime was only in place at the time of IEMS's inspection in 2011. Similarly, it discards Ecuador's view that the Ecuadorian Supreme Court imposed strict liability as early as 2002. Indeed, in Delfina Torres, the Supreme Court merely discussed strict liability in dicta, and in fact applied fault-based liability with a reversal of the burden of proof.189 In any event, Burlington reiterates that whatever changes were made in 2008, the new Constitution "cannot have any retroactive effect",190 since the applicable legal regime is the one in force at the time of the allegedly harmful act.191 Accordingly, the 2008 regime only applies to acts that postdate the Constitution's entry into force.192
124.
The Claimant identifies a second misstatement of the law argued by Ecuador, relating to the burden of proof. For Burlington, Ecuador has the burden of proving environmental harm.193 Proof of harm is an essential element of tortious liability. Ecuador's position that any effect on the environment is an environmental harm and that it only needs to make a showing of a "negative impact" defies common sense. An effect on the environment in compliance with environmental regulations cannot constitute recoverable "harm". Otherwise, there would be no purpose in regulations and government authorizations.194 Astonishingly, Professor Crespo erroneously believes that Ecuador need only "allege" or "point out the probability of harm".195 If that were true, Ecuador's theory would imply that "any participant in human activity in Ecuador is presumptively liable for an environmental damage figure chosen by any plaintiff", since any human activity causes some impact to the environment.196
125.
Third, contrary to Ecuador's contention, Ecuadorian law requires causation even for strict liability. The Supreme Court of Ecuador held so and Article 396 of the 2008 Ecuadorian Constitution expressly provides that persons are strictly liabile for damages which they have "caused". Thus, Ecuador is wrong to argue that it "need only establish that environmental harm exists".197 In addition, it must prove that Burlington and Perenco caused such harm.198

1.2.3 Ecuador's definition of environmental harm is baseless, since a significant impact requires an exceedance of permissible limits

126.
According to Burlington, Ecuador's allegation of widespread environmental harm in the Blocks is a "fantasy"199 based on novel and unsupported theories about the definition of environmental harm and on flawed reports from its expert IEMS. Among other things, IEMS applied incorrect legal standards, analyzed the soil in mud pits as though it were regular soil, overestimated the volume of contaminated soil through unreliable modelling, and failed to properly filter groundwater samples.
127.
Burlington notes that it is common ground that the EML defines harm under Ecuadorian law and that the 2008 Constitution left this definition untouched. Ecuador is wrong to deny that the notion of harm is further specified in RAOHE and TULAS, and to make artificial distinctions between the notion of harm and the idea of setting limits to prevent harm. Ecuador's first argument in its Second Supplemental Memorial that its own regulations were "arbitrary" is just as erroneous as its second argument that RAOHE and TULAS do not define harm but merely seek to set limits on dangerous activities.200
128.
In conformity with the EML, "harm" requires a "significant negative impact". As the Respondent's experts Professors Andrade and Crespo acknowledged, this wording implies that some negative impacts are significant, while others are not.201 If all types of negative impacts were prohibited, there could be no development at all.202 While the EML defines "harm", it does not set specific permissible limits for each contaminant. Burlington therefore considers it necessary to resort to Ecuador's implementing regulations, specifically RAOHE and TULAS, where the notion of "harm" is further specified.203
129.
Burlington also asserts that reliance on the limits set in RAOHE and TULAS conforms with the "constitutional right of legal certainty".204 In other words, "there must be a clear dividing line determining the point at which environmental impacts become significant – a dividing line between ‘insignificant impacts' and ‘harm'".205
130.
The Claimant further points to Ecuador's practice after 2008, which shows that it refers to permissible limits, not to background values, when it defines "harm". For instance, the Ministerial Accord 169 provides that environmental harm must be restored to the conditions "determined by the environmental authority", and the Ministerial Resolution No. 1 refers to "harm" by reference to permissible limits.206 Ecuadorian case law and the practice of Ecuador's regulatory agencies equally demonstrate that harm is always defined by reference to permissible limits. With respect to case law, Professor Andrade, the Respondent's expert, conceded at the Hearing that no Ecuadorian court had ever applied background values.207 The Los Vencedores and the Municipality of Orellana decisions support the proposition that there is no harm if there is no exceedance of permissible limits.208 In respect of Aguinda, Professor Andrade admitted that the court speaks of "full reparation", but does not once make reference to – let alone apply – background values.209
131.
In connection with regulatory practice, Ecuador's agencies have never applied background values either.210 In support, Burlington cites the following elements: (i) Ecuador's Ministry of the Environment applies Ministerial Resolution No. 1, i.e. permissible limits, when a spill occurs; (ii) Petroamazonas remediates spills by reference to permissible limits; (iii) Professor Andrade acknowledged that if a spill occurs, agencies apply permissible limits; (iv) an agency of the Ministry of the Environment called PRAS issued operational policies and norms explicitly based on RAOHE; (v) IEMS applied regulatory criteria in their first expert report before being instructed by Ecuador to discard them; and (vi) Professor Bedón testified that he had never seen background values applied.211 In conclusion, Ecuador has provided no evidence of even a single instance where background values have been applied, whereas Burlington has shown that in fact Ecuador applies permissible limits in determining the existence of environmental harm.212

1.2.4 Ecuador uses background values to falsely portray and inflate alleged environmental harm

132.
Burlington opposes Ecuador's use of background values, including IEMS's position that only background values are sufficiently protective of the public health. Burlington highlights that IEMS was instructed by Ecuador to employ background values and never actually analyzed whether the ecosystem was in fact affected.213 IEMS's first expert report actually referred to permissible limits as applicable standard.214 "Astonishingly – argues Burlington – IEMS admitted that if the Consortium were operating today […], it would be subject to RAOHE and TULAS regulations, not ‘background values'".215 Thus, had Burlington not been expropriated, it would be subject to the regulatory standards. Having been expropriated, it is somehow subject to background values "resulting in billions of dollars of bogus damages".216
133.
In support of its argumentation, the Claimant also relies on the purpose of RAOHE, which in the regulation's words, is to set "[p]ermissible limits for the identification and remediation of contaminated soils in all phases of the hydrocarbon industry".217 Should background values govern, as Ecuador claims, RAOHE would serve no apparent purpose. Further, IEMS itself acknowledged, in its first report and in the City Oriente v. Ecuador arbitration, that RAOHE establishes the criteria to define the scope of environmental harm.218 If background values were relevant, then active oil fields would have to be restored to a pristine state of nature even as the oil fields continue to be in operation. This is absurd, especially considering Ecuador's aggressive expansion of oil production.
134.
In refutation of Ecuador's reliance on RAOHE Table 6, Burlington remarks that Table 6 states that background values may be used to increase permissible limits, not to decrease them. Thus, if the natural concentration of a given element in the soil is higher than its regulatory limit, the permissible limit may be increased to that naturally occurring level – the reverse is not true.219
135.
Burlington also rebuts Ecuador's further argument according to which background values alone comport with the 2008 Ecuadorian Constitution. First, the 2008 Constitution cannot apply retroactively and thus would cover at most a 10-month period (between its coming into force in October 2008 and Ecuador's takeover of the Blocks in July 2009). Second, Article 396 of the 2008 Constitution only comes into play in the event of "harm", which is defined in environmental regulations and in particular RAOHE.220
136.
Burlington further disputes IEMS's assertion that the Ecuadorian regulatory criteria "might not be enough to protect the health of the local population [...]".221 On the contrary, says Burlington, "Ecuador's regulatory limits are remarkably stringent".222 Ecuador is more protective than other oil producing regions such as Venezuela, Texas and Louisiana. Its regulations are also more stringent than the Risk-Based Corrective Action ("RBCA") methodology widely adopted in European countries, the United States, and other parts of the world. In a nutshell, RAOHE criteria are highly protective of human health and no public health rationale warrants the application of a more stringent standard.223
137.
Burlington moreover criticizes IEMS's methodology in calculating the background values that it employs, which resulted in labeling clean samples as contaminated ones. When aggregated across contaminants, Ecuador's background values result in 84% of clean samples being labeled as contaminated.
138.
Finally, according to Burlington, Ecuador's reliance on background values, as opposed to RAOHE criteria, has enormous consequences for its counterclaims. An amount of damages of approximately USD 1.3 billion of Ecuador's claim depends on the application of background values. If the background values are discarded, Ecuador's claims shrink from USD 2.2 billion to USD 895 million.224

1.2.5 IEMS's "regulatory case" based on sensitive ecosystems is likewise incorrect

139.
Burlington also challenges Ecuador's alternative submission based on regulatory criteria because it exclusively relies on the sensitive ecosystem standard, which is inapposite in most cases. Burlington disagrees with Ecuador's assertion that the entirety of the Blocks is situated in a "sensitive" region or that they are located in a biosphere reserve, noting that only a very limited number of areas are designated by Ecuador's Ministry of the Environment as "sensitive ecosystem".225
140.
Burlington further argues that, according to RAOHE Table 6, the permissible levels of contamination depend on the "use" of the land, which is classified by RAOHE as industrial, agricultural or sensitive ecosystem, the first being the most permissive and the last the least permissive. IEMS's calculation of contamination levels based on the sensitive ecosystem threshold is unwarranted as only a small number of sites in the Blocks intersect with designated "sensitive ecosystems" areas.226 Most of the areas where the Consortium's active operations take place are a "quintessential example" of land use for industrial purposes, with the surrounding areas being used mostly for agriculture.227
141.
For Burlington, Ecuador's argument that the Blocks are to be considered as sensitive ecosystem in their entirety is inconsistent with RAOHE and with Ecuador's own application of RAOHE other than in this arbitration.228 Under RAOHE, the "posterior use that will be given to the remediated soil" determines the land use. Posterior use is "immediate subsequent use to the seizure", and not use when the oil operations are completed as advocated by Ecuador.229 The areas within the oilfields, such as platforms and drilling equipment, are used to conduct hydrocarbon operations – an industrial use. This is the sole purpose, present and foreseeable, of these areas. The lands surrounding the oilfields, in turn, are used mostly for pastures, crops, or wood gathering, that is, for agricultural purposes. According to GSI, "95% of the areas surrounding the inspected platforms constitute agricultural areas or secondary forests".230
142.
Burlington also considers that Ecuador's exclusive reliance on the sensitive ecosystem standard is contrary to its own practice. According to Burlington, that practice confirms that most of the land in the Blocks is destined for either "industrial" or "agricultural" use. On multiple occasions, Ecuador approved, or did not object to, plans where the land in the Blocks was classified as either "industrial" or "agricultural".231 For instance, IEMS conceded at the Hearing that the ex-post studies on the environmental condition of the Blocks after Ecuador's takeover applied the agricultural use standard.232 In application of RAOHE, the "posterior use" of any soil remediated will continue to be either "industrial" or "agricultural". In short, the "sensitive ecosystems" classification bears no relationship to the actual use of the land and must be rejected.233
143.
IEMS also incorrectly relies on Table 2 of Annex 2, Book VI of TULAS. For the Claimant, this table only establishes pre-estimated background values for substances not included in RAOHE Table 6 (such as barium and vanadium). It does not express remediation criteria, which are found in Table 3 of Annex 2, Book VI of TULAS, entitled "Criteria for Remediation and Restoration of Soils". TULAS expressly states that the standards for remediation or restoration "are included in Table 3",234 a fact that IEMS itself has acknowledged in the City Oriente arbitration. Hence, Burlington views TULAS Table 3 as the appropriate table to determine soil remediation standards.235
144.
As a final point, Burlington stresses that the land use designations in RAOHE are sufficiently protective of the human health and the environment, since Ecuador's standards are amongst the most stringent worldwide.236
145.
As noted above, for Burlington, applying the proper regulatory criteria has an enormous impact on Ecuador's damages claim. In particular, GSI found that 74% of the soil samples analyzed by IEMS that show exceedances of the "sensitive ecosystems" limits are in reality compliant with Ecuadorian regulations for industrial and agricultural land.237 Applying correct regulatory criteria, Ecuador's alleged damages are further reduced by USD 885 million to a total of approximately USD 10 million, for which Burlington accepts liability in the amount of USD 1.09 million.238

1.2.6 Ecuador entirely ignores its duty to prove causation

146.
Burlington opposes Ecuador's view that Ecuadorian law imposes a rebuttable presumption of causation. For Burlington, the 2008 Constitution imposes the burden to prove causation on Ecuador and Burlington is only liable for harm caused during its tenure of the Blocks, not for the one caused by Ecuador itself or prior operators.

a. Ecuador bears the burden to prove causation

147.
Burlington's submission is that causation is not presumed, and that Ecuador must prove causation. Professor Bedón gave clear evidence to this effect at the Hearing and even Professor Crespo acknowledged that environmental harm must "be caused by the operator's activity".239 Ecuadorian courts consistently require plaintiffs to prove causation, including in strict liability cases.240 For instance, in Medardo Luna, the Ecuadorian Supreme Court held that the injured party must prove (i) the fact, (ii) the damage, and (iii) the causal nexus. The same conclusion was reached by the re-named National Court of Justice in Aguinda, where it held that in the post-2008 strict liability regime the plaintiff must only "corroborate the risky activity and the relation of causality".241

b. Burlington is only liable for harm caused during its tenure of the Blocks

148.
The Claimant further submits that it is only liable for damage caused while the Consortium operated the Blocks. Causation must be proven because each operator is responsible for the damage which it has caused, not for harm caused by someone else.242 Since Professor Crespo testified that Ecuador must prove that the damage was caused "during the time of the Consortium's operations", it follows that Petroamazonas, who is presently operating the Blocks, is presumptively liable for any harm that occurred after July 2009. In this context, Professor Andrade's explanations that Petroamazonas somehow plays by different rules must be rejected.243
149.
Consequently, Burlington also disputes that it is jointly liable for harm caused by prior operators.244 Joint and several liability can only result from "the same tort or wrongful act, not successive ones".245 Hence, Burlington is only liable for the conduct of Perenco, as operator for the Consortium, and not for the conduct of previous or subsequent operators.246

c. Contamination in Coca-Payamino was caused by Ecuador itself

150.
Burlington submits that the environmental harm in Coca-Payamino, which constitutes more than half of Ecuador's claim, was "almost certainly" caused by CEPE and Petroproducción when they operated the fields in the 1980s and during the latter's alternate operatorship with Oryx of the CPUF in the 1990s.247
151.
Indeed, there is abundant evidence of oil spills, problems with mud pits, lack of remediation, and continuous discharges into the environment during CEPE and Petroproducción's tenures of the Blocks. For instance, an environmental audit in 1999 found Petroproducción's operating practices to be "reactive and not proactive".248 Furthermore, Burlington notes that, notwithstanding the express order of the Tribunal, Ecuador failed to produce any spill report or remediation program from that period.249 In support of its allegations, Burlington cites the following facts: (i) in Coca CPF, Petroproducción "permanently discharged production water" into the environment prior to 1994;250 (ii) in Payamino 2 and 8, the crude oil and heavy metals found in the Jungal swamp originate from a pre-1992 rupture of the mud and production test pits built by CEPE in 1987;251 (iii) in Coca 6, a serious spill was recorded by Oryx in 1999 and the pit in the northern part of the platform was built by Petroproducción in 1989, while the 2007 spill was fully remediated;252 (iv) in Payamino 4, CEPE built pits in 1988 without using liners and filled them with oil-based drilling muds;253 and (v) in Coca 4, drilling was undertaken by CEPE in 1989 which explains the presence of barium in that area.254

d. Ecuador fails to account for environmental harm caused by Petroamazonas after July 2009

152.
Burlington calls attention to the fact that Ecuador's claim ignores that the latter took over the Blocks in July 2009 and aggressively expanded operations. In this context, Ecuador's insistence on the sensitive ecosystem standard is "absurd" when one considers that Ecuador itself has actively increased the industrial use of the Blocks.255 According to Burlington, Ecuador's expansion of the fields has obviated the need for any remediation, as is shown by three illustrations.
153.
First, the Coca 13 site has been expanded by over 40,000 square meters, 11 new wells, four new mud pits and a new CPF. Doing so, Ecuador has subsumed and "wiped out" the alleged zone of contamination for which it claims that Burlington is liable. Yet, it claims USD 26.5 million in damages under its background values case and USD 7.4 million under its sensitive ecosystem case for this site.256
154.
Second, Oso A has also been dramatically expanded, with Ecuador drilling 14 new wells. As with Coca 13, "Petroamazonas's extensive development of Oso A since July 2009 has obviated the need for any remediation and rendered IEMS's conclusions obsolete – if contamination even existed in the first place".257
155.
Third, Burlington argues that Ecuador failed to mention or account for the occurrence of a spill in Mono CPF that occurred in 2010 or 2011. In this context, Burlington highlights that IEMS's first report identified this spill as a recognized environmental condition ("REC"), but "curiously" did not include it in its third report "despite the fact that Mono CPF represents one of the single largest claims by dollar amount in all of Ecuador's counterclaims".258 Therefore, Ecuador's claims for over USD 100 million must fail because it cannot show that the alleged contamination is attributable to the Consortium.259

e. The PSCs exonerate the Consortium for harm caused by others

156.
Burlington notes that Ecuador relies on the PSCs to absolve it from having to prove causation, the Consortium being allegedly liable for environmental harm caused by prior operators. Not only is Ecuador's reference to the PSCs barred by the tortious nature of its counterclaim, it is also wrong in and of itself. Indeed, the PSCs "expressly exonerate the Consortium from responsibility for harm caused by others".260
157.
In this context, Burlington further observes that the limited exceedances above regulatory criteria found at 17 sites were not caused by the Consortium, but by prior operators, including State-owned operators. For instance, there is ample proof that the contamination of the Jungal swamp in Payamino 2 and 8 was caused by CEPE between 1988 and 1992. In fact, Burlington produced evidence linking the contamination found at all but two of the 17 sites to the activities of parties other than the Consortium. As a responsible member of the Consortium, Burlington is "willing to accept liability for the two sites at which it cannot definitely link harm to non-Consortium activities" for a total cost of USD 1.09 million.261

1.2.7 Ecuador's claims are time-barred under Ecuadorian law

158.
Burlington's last defense is that the counterclaim is time-barred. Article 2235 of the Civil Code enshrines a four-year limitation period. Ecuador's reliance on the so-called ‘discovery rule', which it borrows from other jurisdictions, is ill-founded. Indeed, Ecuador fails to cite a single case in support of this rule. Even if such a discovery rule were to exist, the period would start to run only when Ecuador knew or should have known ("constructive knowledge") about the alleged environmental harm.262 In light of the fact that the Consortium was strictly supervised by Ecuadorian regulatory authorities, that Ecuador consistently approved the Consortium's environmental impact studies, that the Consortium regularly filed reports and audits, and that regular inspections were carried out, Ecuador must be deemed to have had constructive knowledge of any alleged harm throughout the Consortium's operations.263 Thus, even if Ecuador had legitimate claims, they are now time-barred under Article 2235 of the Civil Code.

2. Applicable legal framework

159.
Since the 1970s, Ecuador has incorporated environmental norms into its legal order, including with respect to hydrocarbons activities, at the constitutional, legislative and regulatory levels, as discussed below.264 After setting out the relevant provisions of the Hydrocarbons Law (2.1), the Tribunal will review the evolution of Ecuador's environmental framework for oilfield operations, including the Law on the Prevention and Control of Environmental Contamination (2.2), the 1978 and 1998 Constitutions (2.3), the Law on Environmental Management (2.4), the Substitute Environmental Regulation for Hydrocarbon Operations in Ecuador (2.5), the Unified Text of Secondary Environmental Legislation (2.6), and the 2008 Constitution (2.7). Finally, it will set out relevant provisions of the PSCs (2.8).

2.1 The Hydrocarbons Law and regulation of hydrocarbons industry

160.
The Hydrocarbons Law (Ley de Hidrocarburos ("HL")) was enacted in 1971.265 It contained the first expression of the State's general duty to ensure that the oil industry "does not cause harm to people, property or the environment", obligating the State to conduct periodical socio-environmental audits.266 The Ecuadorian legislature progressively imposed environmental obligations on oilfield operators, such as the 1982 amendment of the HL providing for the obligation of operators to comply with applicable environmental laws and regulations267 and to conduct environmental impact studies and prepare environmental management plans.268 In 2002, the Regulation of Hydrocarbon Operations (Reglamento de Operaciones Hidrocarburíferas ("ROH")) was enacted.269 It contains rules applicable to all phases of the hydrocarbon operations, exploration, drilling and exploitation, as well as an annex with definitions of terms.270 Article 7 prescribes that oilfield operators must comply with applicable laws and regulations relating to the protection of the environment.271

2.2 The Law on the Prevention and Control of Environmental Contamination

161.
In 1976, Ecuador adopted the Law on the Prevention and Control of Environmental Contamination, establishing for the first time rules to prevent soil, water and air contamination, without however specifying any parameters.272
162.
As regards water, for instance, Article 16 prohibited the dumping of waste water containing contaminants harmful to human health, fauna or property "into the sewage network, or into streams, irrigation channels, rivers, natural or artificial lakes, or in the ocean, or to let it infiltrate onto land".273 In connection with soils, Article 20 prohibited the discharge of "any type of contaminants that could alter the quality of the soil and affect human health, flora, fauna, natural resources and other goods" without complying with applicable technical and regulatory standards.274

2.3 The 1978 and 1998 Constitutions

163.
Protection of the environment was enshrined in the constitution for the first time in the 1978 Constitution.275 Article 19 of such Constitution enshrined the fundamental right to live in an environment free of contamination, specifying that environmental protection would require laws restricting the "exercise of certain rights or liberties".276 In 1996, a constitutional reform instituted the collective right to a healthy environment and declared the following to be of public interest: (i) the protection of the environment and the conservation of ecosystems and biodiversity, (ii) the principles of prevention and sustainable exploitation of natural resources, and (iii) the establishment of a system of protected areas.277
164.
The 1998 Constitution further elaborated the environmental regime by enshrining the principles of sustainable development, prevention, precaution, rehabilitation, and participation.278 Furthermore, Article 91 provided for the State's liability, including that of its "delegates and concessionaires", for environmental harm.279

2.4 The 1999 Environmental Management Law (EML)

165.
In order to implement the environmental provisions of the 1998 Constitution, Ecuador adopted in 1999 the EML.280 The EML established the principles and guidelines of Ecuador's environmental policy, determining the obligations, responsibilities and levels of participation of the public and private sectors in environmental management and indicating "the permissible limits, controls and sanctions" in such matters.281 It set out the scope and principles of environmental management; the institutional regime underlying environmental management; the various instruments of environmental management; financial mechanisms; as well as provisions on information and participation, and the protection of environmental rights.
166.
With respect to the instruments of environmental management, the EML provided for the assessment of environmental impacts and environmental controls. For instance, Article 21 EML provides that licenses for economic activities may only be granted if environmental management systems are drawn up, which must include "base line studies; an environmental impact assessment; risk assessment; management plans; risk management plans; monitoring systems; contingency and mitigation plans; environmental hearings and plans for withdrawal".282 The EML also defines many relevant terms, such as contamination, environmental harm, and environmental impact.283

2.5 The Environmental Regulation for Hydrocarbon Operations in Ecuador (RAOHE)

167.
In 1992, Ecuador enacted RAOHE,284 which was amended in 1995285 and replaced in 2001 by an expanded and updated version.286 RAOHE contains various chapters and annexes that apply to all phases of oilfield operations in Ecuador,287 including prospection, exploration, exploitation, industrialization, storage and transportation, and commercialization of crude oil, its derivatives, and natural gas susceptible of causing environmental impacts.
168.
RAOHE requires operators to submit environmental programs and audits to the relevant Ministry on a regular basis,288 to undertake regular internal monitoring of environmental conditions in relation to atmospheric emissions, solid and liquid discharges, as well as remediation of contaminated soils or pits.289 It also requires operators to present environmental impact studies prior to any new phase in oilfield operations,290 including by providing environmental management and monitoring plans to mitigate and control adverse impacts.291
169.
Article 42 of RAOHE further provides that operators must conduct biennial environmental audits,292 in order to determine whether their oilfield operations comply with applicable environmental standards, including the management and monitoring plans.293 Article 42 reads as follows:

ART. 42.– Environmental audit.– The Undersecretariat of Environmental Protection, through the National Environmental Protection Directorate, shall audit, at least every two years, or whenever the Undersecretariat of Environmental Protection so orders upon detecting non-compliance with the Environmental Management Plan, the environmental aspects of the various hydrocarbons activities conducted by the ones subjected to control.

The Undersecretariat of Environmental Protection, through the National Environmental Protection Directorate (DINAPA), shall determine the type and scope of the Environmental Audit for the operations of those subjected to control based on compliance with the Environmental Management Plan.

At least every two years, those subjected to control shall conduct an Environmental Audit of their activities, following approval of the corresponding Terms of Reference by the Undersecretariat of Environmental Protection, and they shall submit the respective audit report to the Undersecretariat of Environmental Protection.

Additionally, the parties, upon the termination of hydrocarbons exploration and exploitation, or in the event of a change of operator, shall conduct the audit referenced in Art. 11 of the Regulation to Law 44, amending the Hydrocarbons Law.

To the effect of the aforementioned audits, the ones subjected to control shall select an environmental auditor qualified by the Undersecretariat of Environmental Protection to carry out the monitoring and verification of compliance with the Environmental Management Plan, in accordance with the Terms of Reference previously approved by the Undersecretariat of Environmental Protection, in which the documentary framework is determined against which the audit shall be conducted (Translation by the Tribunal).

170.
Article 16 of RAOHE stipulates that the operator must present for approval to the Undersecretariat of Environmental Protection (Subsecretaría de Protección Ambiental) a remediation plan in cases of remediation of pits, contaminated soils, and "major accidents with spills of more than five barrels of oil, combustibles or other product".294 Within 15 days after completion of the remediation works, the operator must submit a technical evaluation report to the Undersecretariat for Environmental Protection through the National Direction of Environmental Protection (Dirección Nacional de Protección Ambiental).295
171.
RAOHE also requires that operators comply with permissible limits, i.e. with maximum benchmark values set out in the annexes. Thus, with respect to atmospheric emissions, liquid discharges and solid wastes, Article 86 of RAOHE requires that operators comply with the permissible limits set out in Annexes Nos. 1, 2 and 3, which constitute the "minimum program for the internal environmental monitoring" of the operator.296 If a permissible limit set in the annexes is exceeded, the operator must immediately notify the Undersecretariat of Environmental Protection and explain what corrective measures have been applied.297
172.
Annex 2 contains six tables detailing "parameters, maximum benchmark values and permissible limits" for routine environmental monitoring and control relating to atmospheric emissions (Table 3), liquid water and effluent discharges (Table 4), black and grey water discharges (Table 5), soil remediation (Table 6), mud pit leachates (Table 7), and waste disposal (Table 8). Finally, Annex 3 contains additional parameters, benchmarks and limits for more in-depth environmental monitoring and control.298
173.
RAOHE Table 4 establishes permissible limits for the permanent environmental monitoring of water and effluent discharges during all phases of hydrocarbons operations. It requires oilfield operators to routinely monitor discharges in accordance with the limits for discharge points (Table 4a) and for points at the receptor bodies (Table 4b).299
174.
RAOHE Table 6 sets out permissible limits for the identification and remediation of contaminated soils "in all phases of the hydrocarbons industry".300 It specifies that "the permissible limits to be applied to a determined project depend on the subsequent use (uso posterior) to be given to the remediated soil".301 As already mentioned, Table 6 distinguishes between three types of land use: industrial, agricultural and sensitive ecosystems.302 Industrial land use criteria apply to "sites of industrial use (constructions, etc.)"; agricultural land use criteria focus on the "protection of soils and crops"; and sensitive ecosystem land use criteria are employed for the "protection of sensitive ecosystems such as the National Heritage of Natural Areas and others identified in the corresponding environmental study".303
175.
The permissible limits set in RAOHE Table 6, Annex 2 are the following:304

RAOHE Table 6, Annex 2
ParameterExpressed inUnitAgricultural UseIndustrial UseSensitive Ecosystems
Total hydrocarbons TPH mg/kg<2500<4000<1000
Polycyclic Aromatic Hydrocarbons (PAHs) C mg/kg<2<5<1
Cadmium Cd mg/kg<2<10<1
Nickel Ni mg/kg<50<100<40
Lead Pb mg/kg<100<500<80

176.
Table 6 also contemplates the possibility of increasing the maximum allowable limits where uncontaminated soils show natural concentrations, in other words background values, above the limits of Table 6. In such a case, RAOHE stipulates that these background values must prevail.
177.
According to Table 6, contaminated soils must be monitored at least every six months with at least one sample and a final characterization once remediation works are finalized.
178.
RAOHE Table 7 spells out the permissible limits for leachates during the "final disposal at surface-level of mud drillings and cuttings".305 The limits vary depending on whether a pit is lined or not (has an "impermeabilization at its base or not").306 Table 7a applies to unlined pits and Table 7b to lined pits. Table 7a contains the following permissible limits:

RAOHE Table 7a, Annex 2
a) WITHOUT impermeabilizaron at the base
ParameterExpressed inUnitPermissible limit value
Hydrogen potential pH 6
Electrical conductivity CE pS/cm 4,000
Total hydrocarbons TPH mg/l<1
Polycyclic Aromatic Hydrocarbons (PAHs) C mg/l<0,003
Cadmium Cd mg/l<0.05
Total chromium Cr mg/l<1.0
Vanadium V mg/l<0.2
Barium Ba mg/l<5

179.
And Table 7b contains the following limits:

RAOHE Table 7b, Annex 2
b) WITH impermeabilizaron at the base
ParameterExpressed inUnitPermissible limit value
Hydrogen potential pH 4
Electrical conductivity CE pS/cm 8,000
Total hydrocarbons TPH mg/l<50
Polycyclic Aromatic Hydrocarbons (PAHs) C mg/l<0,005
Cadmium Cd mg/l<0.5
Total chromium Cr mg/l<10.0
Vanadium V mg/l<2
Barium Ba mg/l<10

180.
RAOHE Table 7 further indicates that, in addition to the initial analysis of the mud drillings and cuttings for final disposal, the operator must undertake periodic sampling and analysis of the pit (i) seven days after the disposal, (ii) three months after the disposal, and (iii) six months after the disposal.307

2.6 Unified Text of Secondary Environmental Legislation (TULAS)

181.
TULAS was promulgated on 31 March 2003 in furtherance of the principles established in the EML.308 It deals inter alia with the environmental authorities, environmental management, forest regime, biodiversity, coastal resources, environmental quality, the special regime of the Galápagos, the Institute for the eco-development of the Amazon region or ECORAE, and a special tax regime for the use of resources under the authority of the Ministry of the Environment.
182.
The provisions relevant to the present case are found in Book VI of TULAS, which is composed of a number of sections and annexes. TULAS establishes a Single Environmental Management System (the "SUMA") in order to ensure inter alia the State's oversight of compliance with environmental management plans of regulated entities.309 One of the main objectives of TULAS, in particular in Title IV, is to provide the necessary regulations in pursuance of the Environmental Management Law for the Prevention and Control of Environmental Contamination.310 Article 42 of Title IV provides that TULAS seeks to "determine, at the national level, the permissible limits for discharges into water bodies or sewer systems; emissions in the air, including noise, vibrations and other forms of energy; wastes, application and disposal of liquids, solids or a combination, into the soil", and to "establish the quality criteria of a resource and criteria or objectives of remediation of an impacted resource".311
183.
Article 45 of Title IV sets forth general principles of environmental stewardship of regulated entities, such as sustainability, equity, prior informed consent, valid representation, coordination, precaution, prevention, mitigation and remediation of negative impacts, solidarity, co-responsibility, co-operation, recycling and reuse of wastes, conservation of resources in general, minimization of wastes, use of clean technologies, alternative and environmentally sound technologies, and respect for traditional cultures and practices as well as ancestral possessions.312
184.
Chapter IV of Title IV addresses environmental control techniques, such as environmental impact studies, environmental management plans, environmental audits, and inspections.313 In this context, Article 70 specifies that the approval of environmental management plans or other environmental studies cannot exonerate operators from liability for environmental contamination.314 Chapter V sets out the rights and obligations of regulated entities, such as the duties to submit annual reports on environmental monitoring,315 to obtain approval of environmental management plans and environmental audits,316 to immediately notify emergency situations,317 and to obtain permits for discharges or emissions.318 Finally, Chapter VIII addresses environmental norms, including the various stages for the elaboration of standards in pursuance of Article 4 of the EML.
185.
According to Ecuador, TULAS sets out general rules for the evaluation of environmental impacts of water and soil resources "from a technical perspective".319 TULAS is therefore complementary to RAOHE, since it supplies chemical parameters not present in RAOHE and also provides for the application of background values.320
186.
By contrast, according to Burlington, RAOHE is the primary source of environmental obligations of oilfield operators in Ecuador, because, unlike the more general TULAS, it expressly and specifically applies to oilfield operations, and TULAS merely supplements RAOHE. Burlington submits that the rules in RAOHE trump the more general regulations contained in TULAS, except where RAOHE is silent. This is so, for instance, with respect to some soil contaminants that are not listed in RAOHE Table 6, Annex 2.321
187.
The Annexes to Book VI of TULAS establish the limits for different environmental media.322 For present purposes, Annex 1 relating to the quality of water resources and effluent discharge standards and Annex 2 relating to soil quality standards and remediation criteria are of particular relevance.
188.
Turning first to soil quality, Annex 2, entitled "Environmental Quality Standard for Soil Resources and Remediation Criteria for Contaminated Soils", sets out regulatory standards in furtherance of the EML and the Regulation of the EML for the Prevention and Control of Environmental Contamination. Its application is mandatory across the entire Ecuadorian territory.323 Its main objective is to prevent and control soil contamination, so as to preserve the integrity of humans, ecosystems and their interrelationships, as well as the environment at large.324 To this end, it establishes (i) norms of general application for soils with distinct uses, (ii) soil quality standards, (iii) remediation criteria for contaminated soils, and (iv) technical norms for the evaluation of the agrological capacity of soils.325
189.
Article 4.1.2 deals with activities that are degrading the soil quality.326 It provides that public or private entities engaged inter alia in hydrocarbons exploration and exploitation must take the necessary measures to avoid harm to soils.327 Article 4.1.3, relating to contaminated soils, states that contamination due to spills or leaks must be remediated taking into account "the remediation criteria for contaminated soils found in [TULAS]".328 It prescribes that the competent regulatory authority must request that remediation and monitoring be undertaken until the "remediation objectives or remediation values" established in TULAS are achieved.329 It further states that "in case of the inapplicability for the specific case of any parameter established in [TULAS] or in case of the absence in the regulation of a relevant parameter for the soil under study", the regulated entity must establish the "background or reference value of the parameter of interest present in the soil" in the affected area, in order to compare the actual values with these background values.330 This provision further indicates that a concentration exceeding three times the background value for the soil constitutes contamination that requires "immediate attention" on the part of the environmental control authority.331 In such cases, the regulated entity must remediate the contaminated soil until the concentration is lower or equal to 1.5 times the background value.332
190.
Article 4.2 establishes soil quality standards and remediation standards. It defines soil quality standards as "approximate background values or analytical detection limits for a contaminant in the soil".333 Background values refer to "representative environmental levels for a contaminant in the soil", which can reflect natural geological variations of undeveloped areas or areas free of the influence of generalized industrial or urban activities.334 Table 2 sets out soil quality standards, comprising 3 general parameters (electric conductivity, pH and sodium adsorption ratio), 19 inorganic parameters (including arsenic, barium, cadmium, total chromium, mercury, nickel, lead, vanadium and zinc), and 14 organic parameters (including polycyclic aromatic hydrocarbons).335 The table below sets out the relevant parameters:

Excerpts of TULAS, Book VI, Annex 2, Table 2
SubstanceUnits (Concentration in Dry Weight)Soil
General Parameters
Conductivity mmhos/cm 2
pH 6 to 8
Inorganic Parameters
Arsenic mg/kg 5
Barium mg/kg 200
Cadmium mg/kg 0.5
Total Chromium mg/kg 20
Nickel mg/kg 20
Lead mg/kg 25
Vanadium mg/kg 25
Organic Parameters
Polycyclic Aromatic Hydrocarbons mg/kg 0.1

191.
In this context, Article 2.38 of Annex 2 defines background values as the "prevailing environmental conditions, prior to any disturbance", that is, the "conditions that would have predominated in the absence of anthropogenic activities, with only natural processes being active".336
192.
Article 4.2.2 defines soil remediation or restoration standards in accordance with the use of the soil, distinguishing agricultural, commercial, residential and industrial uses. The standards set out in Table 3 are meant to determine the "maximum contaminant concentration levels for soil in the process of remediation or restoration".337 It contains the following relevant parameters:

Excerpts of TULAS, Book VI, Annex 2, Table 3
SubstanceUnitsLand Use
Agricultural Residential Commercial Industrial
General Parameters
Electrical Conductivity mmhos/cm 2 2 4 4
pH 6 to 8 6 to 8 6 to 8 6 to 8
Inorganic Parameters
Arsenic mg/kg 12 15 15 15
Barium mg/kg 750 500 2000 2000
Cadmium mg/kg 2 5 10 10
Total Chromium mg/kg 65 65 90 90
Nickel mg/kg 50 100 100 100
Lead mg/kg 100 100 150 150
Vanadium mg/kg 130 130 130 130
Organic Parameters
Polycyclic Aromatic Hydrocarbons mg/kg< 2< 5< 1

193.
Coming now to groundwater, Annex 1 of Book VI of TULAS determines (a) permissible limits for discharges into water bodies or sewer systems, (b) water quality standards according to their various uses, and (c) methods and procedures to determine the presence of contaminants in water.338 Its objective is to prevent and control contamination of water resources, to preserve the integrity of humans, ecosystems and their interrelationships, as well as the environment more generally.339
194.
Article 4.1.3 addresses quality standards for groundwater.340 It stipulates that any proven alteration of the quality of groundwater triggers the duty to remediate the contaminated groundwater and the affected soil.341 Table 5 sets out the quality standards for groundwater "considering a soil with clay content between (0-25.0)% and organic material content between (0-10.0)%".342 It contains the following relevant parameters:

Excerpts of TULAS, Book VI, Annex 1, Table 5

ParametersExpressed AsUnitMaximum Permissible Limit
Arsenic (total) As Mg/l 35
Barium Ba Mg/l 338
Cadmium Cd Mg/l 3,2
Cobalt Co Mg/l 60
Copper Cu Mg/l 45
Total Chromium Cr Mg/l 16
Molybdenum Mo Mg/l 153
Mercury (total) Hg Mg/l 0,18
Nickel Ni Mg/l 45
Lead Pb Mg/l 45
Zinc Zn Mg/l 433
Total Petroleum Hydrocarbons Mg/l 325

2.7 The 2008 Constitution

195.
Ecuador entered a new stage of environmental protection with the adoption of the 2008 Constitution. Most notably, the new Constitution bestows rights to nature (called Pacha Mama using the term employed in traditional Andean cultures) and codifies the fundamental principles of environmental stewardship. It declares that the elimination of poverty and the promotion of sustainable development, as well as the equitable redistribution of resources and wealth are fundamental duties of the State.343 It also specifies that constitutional principles and rights, including those relating to environmental matters, are "unalienable, obligatory, indivisible, interdependent and of equal importance".344
196.
Addressing first the rights of nature, the Tribunal notes Article 10 providing that "[n]ature shall be the subject of those rights that the Constitution recognizes for it". These rights are then specified in Articles 71 to 74.
197.
Article 71 declares that "Nature or Pacha Mama, where life is reproduced and occurs, has the right to integral respect for its existence and for the maintenance and regeneration of its life cycles, structure, functions and evolutionary processes".
198.
Article 72 further provides that "Nature has the right to be restored" and that the duty to restore nature runs in parallel to the duty to compensate individuals and communities that depend on affected natural systems.345 The exploitation of non-renewable natural resources leading to "severe or permanent environmental impact" must be subject to "the most effective mechanisms to achieve the restoration", including through "adequate measures to eliminate or mitigate harmful environmental consequences".346
199.
Article 73 requires the State to apply "preventive and restrictive measures on activities that might lead to the extinction of species, the destruction of ecosystems and the permanent alteration of natural cycles".347
200.
Article 74 then qualifies the rights of nature by stating that persons, communities, peoples and nations have the "right to benefit from the environment and the natural wealth" so as to enable them "to enjoy the good way of living", while specifying that environmental services are not subject to appropriation.348
201.
Turning then to environmental protection in general, the Constitution declares as matters of "public interest", inter alia, environmental conservation, the protection of ecosystems, biodiversity, the prevention of environmental damage and the recovery of degraded natural spaces.349 The Constitution further recognizes the "right of the population to live in a healthy and ecologically balanced environment that guarantees sustainability and the good way of living (sumak kawsay)".350 In this context, the State is called upon to promote the use of environmentally clean technologies, nonpolluting and low-impact alternative sources of energy, while certain noxious substances, such as highly toxic persistent organic pollutants or toxic wastes, are forbidden.351
202.
The Constitution also highlights the special value of the Amazon ecosystem, providing that, in addition to the regular territorial subdivisions, special systems may be established for reasons of environmental conservation.352 With respect to the Amazonian region, Article 250 insists on the conservation of its ecosystem as part of the planet's environmental equilibrium:

"The territory of the Amazon provinces is part of an ecosystem that is necessary for the planet's environmental balance of the planet [sic]. This territory shall constitute a special territorial district, for which there will be integrated planning embodied in a law including social, economic, environmental and cultural aspects, with land use development and planning that ensures the conservation and protection of its ecosystems and the principle of sumak kawsay (the good way of living)".353

203.
To implement these objectives, Article 259 requires the State to adopt sustainable development policies:

"With the aim of safeguarding the biodiversity of the Amazon ecosystem, the central State and decentralized autonomous governments shall adopt sustainable development policies which shall also offset disparities in their development and consolidate sovereignty".354

204.
In connection with the economic aspects of environmental protection, the Constitution reserves the State's powers to administer, regulate, monitor and manage strategic sectors,355 such as the hydrocarbons industry,356 in accordance with "the principles of environmental sustainability, precaution, prevention and efficiency".357 Only on an exceptional basis may the State delegate to private enterprise the management of strategic resources, including the sustainable use of natural resources.358
205.
Non-renewable natural resources are declared part of the "unalienable heritage of the State and are not subject to a statute of limitations".359 Their management must take into account the "responsibility between generations" and "the conservation of nature" as well as the objective to "minimize the negative impacts of an environmental, cultural, social and economic nature".360 Water, as a "vital element for nature and human existence", forms part of the "country's strategic heritage for public use" which is "unalienable" and imprescriptible property of the State.361
206.
Of particular relevance for present purposes is Chapter II of Title VII which deals with biodiversity and natural resources management as part of the so-called Good Way of Living System. It addresses fundamental environmental principles,362 biodiversity conservation,363 natural assets and ecosystems,364 natural resources,365 soils,366 water,367 and the biosphere, urban ecology and alternative sources of energy.368
207.
Within Chapter II, Article 395 sets forth fundamental environmental principles, such as sustainable development, based on a model that is "environmentally balanced and respectful of cultural diversity, conserves biodiversity and the natural regeneration capacity of ecosystems, and ensures meeting the needs of present and future generations",369 integration,370 and public participation.371 Article 395(4) embodies the principle in dubio pro natura in the following terms:

"In the event of doubt about the scope of legal provisions for environmental issues, it is the most favorable interpretation of their effective force for the protection of nature that shall prevail".

208.
Article 396 then addresses the risks and consequences of human-induced environmental impacts. Having articulated the principles of prevention and precaution,372 it provides in paragraph 2 that strict liability governs environmental harm and that full restoration of ecosystems must be achieved in addition to compensating affected persons and communities:

"Responsibility for environmental damage is objective. All damage to the environment, in addition to the respective penalties, shall also entail the obligation to integrally restoring the ecosystems and compensating the affected persons and communities".373

209.
The third paragraph of Article 396 then imposes on all economic actors the duty to prevent environmental impacts, and to mitigate and repair the "damages caused" by them to the environment:

"Each one of the players in the processes of production, distribution, marketing and use of goods or services shall accept direct responsibility for preventing any environmental impact, for mitigating and repairing the damages caused, and for maintaining an ongoing environmental monitoring system".374

210.
Finally, the fourth paragraph of Article 396 provides for the imprescriptibility of environmental claims in the following terms:

"The legal proceedings to prosecute and punish those responsible for environmental damages shall not be subject to any statute of limitations".375

211.
In case of environmental harm, Article 397 provides that the "State shall act immediately and with a subsidiary approach to guarantee the health and restoration of ecosystems". That provision further specifies that, in addition to sanctions, "the operator of the activity that produced the damage" must procure "integral reparation, under the conditions and on the basis of the procedures provided for by law".
212.
In this context, the (individual and collective) right to live in a healthy and ecologically balanced environment is secured by a guarantee of access to justice, ascertaining "effective custody in environmental matters".376 This entails among other things that the operator of the impugned activity bears the burden of proving the "absence of potential or real danger".377
213.
In this section of Chapter II dealing with fundamental principles, the State further pledges to "ensure the intangibility of protected natural areas, so as to guarantee the conservation of biodiversity and the maintenance of the ecological functions of the ecosystems".378 Finally, Article 399 vests a decentralized national environmental management system with overall stewardship over the environment and nature.379
214.
Among the other provisions of Chapter II which may be relevant for the resolution of this dispute, the Tribunal notes in Section III: Article 404 under which ecosystems380 must be managed "in accordance with land use planning and ecological zoning, in compliance with the law"; Article 405 referring to the "national system of protected areas" which serves to guarantee the conservation of biodiversity and the "maintenance of ecological functions"; Article 406 mandating that the State regulate "fragile and threatened ecosystems", such as "high Andean moorlands, wetlands, cloud forests, dry and wet tropical forests and mangroves, marine ecosystems and seashore ecosystems"; and Article 407 prohibiting extractive industries of non-renewable natural resources in protected areas and in "areas declared intangible assets", subject to limited exceptions.
215.
In the remainder of Chapter II, attention is called to norms of the following content: Section IV which deals with non-renewable natural resources as part of the "unalienable property of the State, immune from seizure and not subject to a statute of limitations",381 the production of which must be "in strict compliance with the environmental principles set forth in the Constitution";382 soil conservation as a matter of "public interest and national priority";383 the State must regulate water resources and the equilibrium of ecosystems,384 mitigate climate change, limit greenhouse gas emissions, deforestation and air pollution, as well as conserve "forests and vegetation".385
216.
This overview demonstrates that environmental protection is one of the 2008 Constitution's fundamental pillars and that environmental stewardship has taken on a new dimension in Ecuadorian society, where nature (or Pacha Mama) is itself bearer of constitutional rights.

2.8 The PSCs

217.
Although its claim is not contractual in nature, Ecuador has relied on the PSCs to strengthen its contention that Burlington is liable for the environmental conditions found in Blocks 7 and 21.
218.
Under the Block 7 and Block 21 PSCs, the Contractor agreed to comply with "all laws, regulations and other provisions" of Ecuador that are applicable to the contracts.386 The Consortium also gave its consent to a series of reporting requirements, such as keeping Petroecuador informed of the activities carried out under the contracts, including by submitting environmental management plans;387 carrying out environmental impact studies in accordance with RAOHE;388 providing the Ministry of Energy and Mines and Petroecuador with copies of environmental information related to the Consortium's activities;389 and conducting periodical socio-environmental audits "in order to prevent, as much as possible, any negative impact of the Contractor's activities on human settlements and on the environment".390
219.
Under Clause 5.1.20 of the Block 7 PSC, the Contractor agreed to "preserve the existing ecological equilibrium in the Contract Area" in accordance with all pertinent standards and the environmental impact studies.391 The Contractor also accepted to "[t]hroughout the term of this Contract, take all necessary measures to conserve and safeguard life and property and to protect the environment".392 In order to achieve these goals, the Contractor further undertook to employ "qualified personnel, equipment, machinery, materials, operational procedures and, in general, technology which complies with environmental protection standards and practices used in the international petroleum industry, subject to compliance with existing standards in Ecuador".393
220.
The Block 7 and Block 21 PSCs also addressed the issue of remediation and postoperations cleanup, in particular by providing that the Contractor would not be liable for preexisting environmental conditions (prior to the Services Contract for Block 7, and prior to the Participation Contract for Block 21), or for environmental conditions resulting from operations that are subsequent to the expiry of the PSCs. Clause 5.1.20.10 of Block 7 PSC reads as follows:

"The Contractor will take responsibility for cleanup and reforestation of the area with species similar to those originally found at the site, in order to, with time, allow the site to return to its original potential, with environmental conditions similar to those found at the beginning of the operations. The Contractor will also be responsible for abandonment of the wells and installations for which the Contractor has been responsible under the terms of this Contract. The cleanup, reforestation and return to similar conditions and abandonment activities will comply with the Environmental Regulations for Hydrocarbon Operations and with the provisions of the Environmental Impact Study. The Contractor will not be liable for preexisting environmental conditions at the beginning of the Service Contract activities. In situations where government authorities order remediation of the environment in the Contract Area due to preexisting conditions, the Contractor will not be liable for the costs and services. Moreover, the Contractor will not be liable for environmental conditions resulting from operations conducted by PETROECUADOR or third parties after the Contractor has returned the Contract Area".

221.
Similarly, Clause 5.1.20 of the Block 21 PSC reads as follows:

"The Contractor shall be responsible for the clean-up, reforestation, and abandonment of non-productive wells and facilities as a result of this Contract. These activities shall be carried out in accordance with the legislation in effect in Ecuador at the time that such clean-up, reforestation, or abandonment is carried out and as contemplated in the Environmental Impact Study.
The Contractor shall not be responsible for pre-existing environmental conditions at the start of the operations under the Contract.
In such cases where the competent authorities order mitigation of the environment in the Contract Area due to pre-existing conditions, any costs incurred in connection with such activities shall be assumed by the Ecuadorian State.
Nor shall the Contractor be responsible for environmental conditions resulting from operations by PETROECUADOR or third parties after the Contract Area is returned by the Contractor".394

222.
Finally, in the context of the Contracts' termination, the PSCs specify that the Consortium was to undertake a "comprehensive environmental audit" two years prior to the termination of the PSC. Clause 5.1.20.7 of the Block 7 PSC, for instance, reads as follows:

"Two (2) years prior to the termination date of this Contract, the Contractor will contract a comprehensive environmental audit of the Contract Area, based on the Environmental Impact Studies, and this audit must be completed no later than six (6) months prior to the termination date of this Contract. The final results of the audit will be binding for the Parties. The cost of performing the audit will be paid by the Parties in equal proportions (50/50). This audit will be performed by a company selected by mutual agreement of the Parties in a written agreement and the audit will comply with applicable Ecuadorian Law. The parties who perform these studies and audits must be previously qualified by the Ministry of Environment through the Undersecretary of the Environment".395

3. Conditions of liability

223.
The Tribunal will start by reviewing the liability regime applicable to hazardous activities, such as oilfield operations, in Ecuador (3.1). It will then address the statute of limitation (3.2) and the liability for successive operators (3.3).

3.1 The liability regime for hydrocarbons operations in Ecuador

224.
The Tribunal is called to rule on Burlington's possible liability in tort, as opposed to contractual liability under the PSCs. It will thus review the tort liability regime for environmental harm as it arises from the 2008 Constitution (3.1.1) and as it existed before (3.1.2), to then draw the necessary conclusions for the case at hand.

3.1.1 The strict liability regime under the 2008 Constitution

225.
The Parties agree, and rightly so, that the 2008 Constitution establishes strict liability for environmental harm.396 The Constitution also sets the following rules of the strict liability regime: (i) the burden of proof of the inexistence of harm falls on the operator;397 (ii) the operator is only responsible for the harm caused by him;398 and environmental claims are imprescriptible.399
226.
The constitutional regime has been clarified in case law. Specifically, the burden of proof of environmental harm was addressed in Los Vencedores, a decision dealing with a claim against Perenco for soil contamination in Block 7 that occurred after the entry into force of the 2008 Constitution.400 The Court held that the burden of proving harm had been reversed by Article 397(1) of the 2008 Constitution, which provides that "[t]he burden of proof regarding the inexistence of potential or real harm shall lie with the manager of the activity or the defendant". The Tribunal's understanding is that the plaintiff still has the burden to make a showing of harm plausibly connected to the defendant's activities and the defendant then has the burden of proving its absence.
227.
Applied to the present case, the rule contained in Article 397(1) means that once Ecuador has made a showing of the existence of environmental harm reasonably related to the Consortium's risky activities, for example by way of the IEMS sampling exercise, Burlington then carries the burden of demonstrating that there is no harm or, if there is harm, what its limits are.
228.
Ecuadorian courts have also held that fault is not a requirement of strict liability. The National Court of Justice explained that strict civil liability is the equivalent of "risk liability" or "fault-less liability" that is triggered "regardless of whether there was willful misconduct or negligence from the person".401 The Court added that the duty to repair environmental harm is based on the fact that the harm caused "may be materially attributed to a certain activity, not on the existence of fault".402
229.
Applied to this case, the absence of a requirement of fault implies that Burlington may not avoid liability by raising that it acted diligently.
230.
Another element of strict liability addressed by the courts is the requirement of causation. Under the strict liability regime, causation is presumed. In Aguinda, the National Court of Justice referred to the Delfina Torres decision, where it was held that under strict liability the operator could only escape liability by showing that the harm was caused by force majeure (hecho fortuito), a third party or the victim.403
231.
The reliance on a presumption of causation under the constitutional regime is further substantiated by the Final Report of Committee 5 on Natural Resources and Biodiversity of the Constituent Assembly that prepared the 2008 Constitution. This report explains that the reversal of the burden of proving harm is due to the need to establish a "presumption of innocence" in favor of the environment,404 which, in turn, entails a reversal of the burden of proving the causal nexus between the act and the damage:

"The presumption of innocence should be established in favour of the environment. From this perspective, the burden of proof should be reversed to be against the defendant […], in other words, counter to the former principle which assumes innocence until guilt is proven, strict liability would be the defense in environmental matters. The plaintiff will not be required to prove causal nexus, and the burden of proof would correspond to the defendant. This principle eliminates one of the barriers impeding access to environmental justice, which is the cost of producing evidence and the technical requirement to prove environmental damages".405

232.
For present purposes, this holding means that Burlington can be exempted from liability if it proves that the harm was caused by force majeure, by a third party, or by Ecuador and in particular by Petroamazonas after the takeover of the Blocks.406 Burlington accepts that the strict liability regime established under the 2008 Constitution involves "some burden shifting within the concept of causation". It insists, however, that Ecuador needed first to prove that the harm "occurred on the Consortium's watch", after which Burlington could show that the harm was caused by force majeure, the victim or a third party.407 While the Tribunal will revert to the issue of successive tort liability, it disagrees with Burlington's position that Ecuador must prove that the harm was caused during the time of the Consortium's operations. Indeed, proof of causation is not required. Causation is presumed, with the result that liability ensues from the mere exercise of a risky activity and the occurrence of harm that is plausibly connected to such activity as far as the type and location of the harm is concerned.
233.
The constitutional regime just described has no retroactive effect. Indeed, Article 7 of the Civil Code restates the general principle that laws have no retroactive effect.408 It is true that the decision in Baquerizo held that rules of public order do apply retroactively.409 However, environmental protection is not a matter of public order under the 2008 Constitution. It is merely a matter of public interest. Public interest, which is dealt with in Article 14 of the Constitution cannot be equated with public order, the latter notion being more restrictive than the former. As a result, there is no room to apply a rule on environmental protection such as the imprescriptibility of claims under Article 396(4) retroactively.

3.1.2 The liability regime for hydrocarbons operations prior to the 2008 Constitution

234.
The strict liability regime of the 2008 Constitution not applying retroactively, the Tribunal must assess the rules governing liability for hydrocarbons operations prior to the 2008 Constitution.
235.
The Parties diverge on whether these rules provided for fault-based or strict liability. Burlington argues in favor of the former, with the result that Ecuador must prove the existence of environmental harm and that the harm was caused by the Consortium's lack of diligence, that is, its fault. In other words, for Burlington, Ecuador must prove harm, fault and causation. For its part, Ecuador argues that the Consortium's activities were always subject to strict liability, or at the latest since the 2002 Delfina Torres decision of the Ecuadorian Supreme Court. Accordingly, says Ecuador, Burlington can only be exonerated if it establishes that the harm was caused by force majeure, a third party or the victim.410 A showing of diligence, or lack of fault, would not suffice.
236.
Prior to the entry into force of the 2008 Constitution, tort liability was governed by the Civil Code. While liability was fault-based, the Civil Code also provided for an objective liability standard applicable to particular situations.411 Moreover, the Ecuadorian courts, especially the Supreme Court in the Delfina Torres case, had extended objective liability to hazardous activities such as oilfield operations based on the theory of risk.412 In Delfina Torres, the Supreme Court of Justice provided the legal foundations for risk based liability in the following language:

"Today's world and the coming world, with its extraordinary and progressive accumulation of risks, requires greater defense of the values of man, created by a technology that, while it enables everything, it also threatens everything. The variety of real contingencies of dangers and risks that are actually drawn as a result of dissatisfaction and aside from any idea of damages, led to a slow evolution of elements and knowledge that enabled the most advanced legal systems to enter into socialization regimes of risk leading to the proposition that the victim is not left defenseless. This gave rise to the theory of risk, according to which, he who uses and takes advantage of any type of means that endow it with benefits, generates social risks through such means, and, as a result of that circumstance, must assume liability for harm caused by them […]. No fault or wrongfulness is required to recognize tort liability. It is enough that the harm is a direct consequence of the event that has caused it. It is purely strict liability".413

237.
The Supreme Court of Justice then identified the elements of strict liability showing that it is for the alleged tort feasor to prove that he or she did not cause the harm:

"Hence, the need arises to promulgate a new type of liability for this type of harm, eliminating the criteria of fault through strict liability or the establishment of an absolute presumption of such. The owner of an exploitation or industry should respond directly for harm caused by the referenced industry or exploitation, and he may only exempt himself from liability if he shows that the harm did not arise from the exploitation, and arose instead from a foreign act (force majeure, fault of a third party or the fault of the victim itself)".414

238.
Accordingly, the following elements are characteristic of the strict liability regime for environmental harm under Ecuadorian law as established in Delfina Torres :

(i) the plaintiff must prove harm connected to the defendant's activities;

(ii) fault is not required; and

(iii) causation is presumed, the defendant being exonerated if he or she proves that the harm was caused by force majeure, the victim or a third party.

239.
The Supreme Court of Justice also held in Medardo Luna that fault is not required for strict liability.415 In the context of the Code of Aeronautics, the Medardo Luna court held that it was irrelevant whether the company followed applicable regulations and whether the pilot acted with care.416
240.
The Tribunal notes that the Parties' experts disagree on whether fault is not a requirement of strict liability at all or whether it remains a requirement but is presumed. Differently worded, they disagree on whether there is a rebuttable presumption of fault (Burlington) or an irrebuttable one (Ecuador). The Respondent's expert Professor Andrade speaks of an irrebuttable presumption ("presunción de culpabilidad absoluta"),417 while the Claimant's expert Professor Bedón considers the presumption to be iuris tantum, that is rebuttable.418 The analysis of the Delfina Torres and Medardo Luna decisions just referred to show that the presumption of fault cannot be rebutted by a showing that all precautions have been followed. Only a showing of force majeure, an act of a third party or of the victim can exonerate the defendant.419
241.
Even if the Tribunal were to adopt the position that the pre-2008 strict liability regime implies a rebuttable presumption of fault, any exceedance of RAOHE or TULAS values would defeat a due diligence argument. In addition, the non-compliances identified by Ecuador's expert RPS in connection with the biennial audits for instance,420 as well as evidence of Mr. Saltos, one of Burlington's witnesses, on various practices of the Consortium (discharges, mud pits, spills), would produce the same result.
242.
More specifically, RPS reviewed the compliance record of the Consortium in environmental matters and identified the following non-compliances during the Consortium's operatorship: (i) failure to conduct biennial environmental audits in 2004 in breach of Article 42 of RAOHE and Article 61 of Book VI of TULAS; (ii) lack of key environmental documents, such as environmental management plans and environmental licenses; (iii) irregularities in the management of drill cuttings and mud pits, such as exceedances in maximum permissible limits of leachates in unlined pits; (iv) exceedances in maximum permissible limits for wastewater, as well as soil quality; (v) additional non-compliances, such as irregular waste and chemical management, or failure to respect a "highly sensitive ecological area" in Block 21; and (vi) environmental liabilities associated with the Payamino field operations resulting in offsite contamination and affecting third parties.421
243.
These instances refute Burlington's contention that the Consortium complied with all regulatory requirements.422 It is further preoccupying that the Consortium was operating the Coca-Payamino, Jaguar, Mono and Gacela fields without any proper environmental license423 and was relying on an outdated environmental impact study for the Coca-Payamino Field.424 And while the Consortium sought to downplay the fact that it produced no biennial environmental audit in 2004, the Subsecretary of Environmental Protection held this omission to be a serious breach of RAOHE.425 In this overall context, the Tribunal further observes that Burlington has not put into the record any baseline studies or internal due diligence studies generated at the outset of operations, allowing to assess the environmental condition of the Blocks at the start of operations as well as preexisting environmental liabilities.426
244.
The testimony of Mr. Saltos, also revealed various instances of Consortium malpractices. For instance, Mr. Saltos mentioned that production water was discharged on the sand bank of the Napo River at the Oso 4 site, so as to filter these discharges before they reached the water body.427 Mr. Saltos further testified that the Consortium was using auxiliary mud pits, such as the well workover pit in Area 2MT in Payamino 1/CPF, which is located beneath the CPF close to the Añango River.428 Mr. Saltos also conceded that, during well workovers, the Consortium was moving drilling muds from one site to another site with open pits, eventually even reopening sealed pits to introduce newly contaminated soils.429 In addition, as the example of auxiliary pits shows, the Consortium also made use of pits that were not reported to the authorities, thus forcing Ecuador's experts to discover unreported mud pits as a result of their sampling campaigns (see, for instance, Payamino 1/CPF; Payamino 16).430 Mr. Saltos also confirmed that the collapse of the pit at Cóndor Norte due to a landslide in 2006 was not remediated until the Consortium left the Blocks in 2009.431
245.
With respect to spills, the evidence provided by Mr. Manuel Solís of Petroamazonas shows that Mr. Saltos' list of spills during the Consortium operatorship is incomplete, further showing that the Consortium did not report all spills to the authorities. Mr. Saltos confirmed that at Jaguar 1, for instance, the 2005 spill in Area 3T was left unreported and, worse, was left unremediated with a 30cm layer of clean soil having been put on top of the affected area, which is not more than 8 meters away from a water body in close proximity to human dwellings and a school.432 Mr. Saltos also testified at Coca CPF during the Site Visit that, while the malpractice of prior operators in discharging produced water into the drainage leading to the Chalá swamp was discontinued and remediated, oil seeping continued during the Consortium's operatorship without any proper remediation being undertaken by the Consortium.433
246.
While the Tribunal accepts Mr. Saltos' representation that the risk of spills is always present in oilfield operations "no matter how cautious the operator is", the Tribunal has much more difficulty in accepting his statement that the Consortium had a "robust spill prevention system" in place.434 The Tribunal is even less inclined to accept Mr. Saltos' attempts to minimize the importance of certain spills, which he qualified as "casi derrames" or "liqueos",435 since the accumulation over time even of minute amounts of certain contaminants can in certain circumstances lead to significant environmental harm if left unremediated. Finally, while Mr. Saltos is right to argue that discharging effluents into a river was not prohibited prior to the entry into force of TULAS on 31 March 2003,436 this practice does not support Burlington's contention that the Consortium was a responsible operator.
247.
To conclude, the Tribunal reaches the conclusion that in Ecuadorian law strict liability governed instances of environmental harm at the latest since the Delfina Torres decision was handed down in 2002.437
248.
The Tribunal is mindful that the Perenco tribunal held that fault-based liability governed Ecuador's claims for environmental harm that predates the 2008 Constitution. While the Tribunal agrees with the finding in the Perenco Decision that "an operator can in general be held only to the legal standards that applied to its conduct at the time",438 it has difficulty following the Perenco tribunal's view that decisions of the Ecuadorian courts have merely "strengthened the presumptions in favour of a finding of liability in the case of damage caused through hazardous activities".439 It finds indeed that these courts have established a strict liability regime for hazardous activities, in particular oilfield operations.
249.
In addition, it is not disputed that Burlington may be held liable for the harm caused by its Consortium partner, Perenco, which operated the Blocks during the Consortium's tenure.440 It is a different issue whether Burlington may be held liable for harm caused by prior operators as Ecuador claims and Burlington opposes, to which the Tribunal will revert after having addressed the limitation period.441

3.2 Statute of limitations

250.
Burlington submits that most of Ecuador's claims are time-barred under the four-year statute of limitations of Article 2235 of the Civil Code,442 which runs from the date when the allegedly harmful act occurred.443 For Burlington, Ecuador is wrong in seeking to import the so-called "discovery rule" into the Ecuadorian legal system or to argue that the imprescriptibility rule of the 2008 Constitution could apply retroactively.444 Accordingly, Burlington submits that claims arising out of incidents that occurred before 18 January 2007 are time-barred and inadmissible.445
251.
Ecuador opposes Burlington's argument mainly on the ground that Article 2235 of the Civil Code only governs liability based on fault.446 Claims for strict liability are subject to the limitation set forth in Article 396(4) of the 2008 Constitution, as the Claimant's expert Professor Bedón acknowledged at the Hearing.447 Even if Article 2235 of the Civil Code were to apply, says Ecuador, the period would start to run when the harm is discovered, which is after the Consortium abandoned the Blocks in July 2009.448 Ecuador adds that its argumentation is reinforced by Burlington's concealment policy and lack of reporting,449 and that by acquiring its interest in the Blocks, Burlington contractually took over the environmental liabilities of prior operators.450
252.
It is common ground that environmental claims for harm caused after the entry into force of the 2008 Constitution are imprescriptible. The Parties disagree, however, on the prescriptibility of claims for harm caused prior to the Constitution's entry into force. They diverge on whether such claims are subject to the four-year limitation of Article 2235 of the Civil Code or whether they are imprescriptible.
253.
The Tribunal is not convinced by Ecuador's argument that questions related to the statute of limitations are procedural rather than substantive. The approach in Nelson Alcívar in this regard is not convincing, nor is its criterion of the date of the filing of the claim.451 Article 7(20) of the Civil Code is of no help either here, since it leaves unaffected terms that have begun to run prior to a change in law.452 More generally, the Tribunal is of the view that in comparative law and private international law the trend is to consider the statute of limitation as a substantive rather than a procedural issue.453
254.
In the Tribunal's view, it is correct that in principle the four-year statute of limitation period provided in Article 2235 CC applies to harm caused prior to the entry into force of the 2008 Constitution, subject to the following specifications.454 The date of the discovery must apply for environmental harm, not the date of the act, since strict liability is focused on the result, not on the act causing the harm. This is so because it is difficult to identify the precise time when environmental harm is generated455 and because, in oilfield operations in Ecuador, the identification of environmental harm is essentially dependent on the operator's self-reporting.456 In the same vein, reliance by Burlington on environmental audits produced by the Consortium as a means to establish constructive knowledge is misplaced here, since (i) environmental audits have no preclusive effect as a general matter under Ecuadorian law,457 and (ii) the Consortium did not generate the required 2004 biennial audits.458 Accordingly, the limitation period only starts to run from the discovery of the harm. The same applies to harm that has been insufficiently remediated, since Ecuador could only assume on the basis of remediation plans and reports submitted by the Consortium that identified harm had been properly remediated. In this context, the notion of "pasivo ambiental" in Article 1 of the Ministerial Agreement No. 169 (2012), which refers to insufficiently remediated harm, further reinforces the Tribunal's findings above.459
255.
In sum, (i) environmental harm caused and/or discovered after the entry into force of the October 2008 Constitution is actionable (Articles 396(4) and 11(3) Cst.), (ii) environmental harm caused between 17 January 2007 and the entry into force of the 2008 Constitution is actionable, and (iii) environmental harm caused prior to 17 January 2007 is actionable to the extent that it was discovered thereafter (Article 2235 CC read in light of strict liability). This extends to harm that has been insufficiently remediated, since any remaining exceedances were only discovered by IEMS in 2011 or thereafter.

3.3 Successive liability of operators

256.
Burlington rejects Ecuador's attempts to hold it liable for harm caused by prior operators as a matter of tort law (successive tort liability) or pursuant to Clause 5.1.20.10 of the Block 7 PSC and Clause 5.1.20 of the Block 21 PSC (successive contractual liability).460 After addressing the issue of Burlington's liability for the acts of prior operators, the Tribunal will also address the matter of liability for acts of Petroamazonas, the operator that succeeded the Consortium and is presently operating the Blocks.
257.
With respect to successive tort liability for prior operators, the Tribunal has already held above that causality is a component of strict liability (both under the liability regime of the 2008 Constitution and under the anterior regime elaborated by Ecuadorian courts), although it is presumed. Accordingly, to be exonerated, Burlington must prove that any alleged harm was caused by force majeure, the victim, or a third party. In other words, under the strict liability regime, Burlington cannot be held liable for harm which it can prove to have been caused by prior operators.
258.
The situation is different under the PSCs. As was seen above, Clause 5.1.20.10 of the Block 7 PSC excludes the Consortium's liability for environmental harm caused prior to the first service contract, i.e., prior to January 1986 and after return of the contract area, i.e. in July 2009.461 A similar provision is found in the Block 21 PSC, with the difference that the starting point for the Consortium's liability is set at the conclusion of the first participation contract in 1995.462 Within these time spans, the Consortium is under a contractual obligation to remediate any environmental harm caused by prior operators.
259.
The question is thus whether the Tribunal should apply this contractual obligation for harm caused by prior operators in the present context. Ecuador's position has somewhat fluctuated on this issue.463 At the Hearing, it sought to clarify matters as follows: "Now, our case, Members of the Tribunal – and we don't want you to have any confusion about it – our case is not based upon any contractual liability, but rather of a tort liability".464 In spite of this clear statement, Ecuador sought relief in reliance on the Constitution and the PSCs465 and one part of its oral presentation focused on the construction of the PSC provisions on successive operators.466 On a question from the Tribunal, Ecuador further explained that "the basis for Ecuador's claim is principally the 2008 Constitution tort liability principles; and […] tentatively the Contract would apply, if you don't find that a [sic] tort liability principles apply here. Second point, is that, even if you have, as the main case of Ecuador, these tort liability principles encompassed in the 2008 Constitution, it is our position […] that there are some contractual provisions that have a bearing on the issue of causation, which you understood especially in this scenario of succession of operators […]".467 In its Post-Hearing Brief, Ecuador then invoked strict liability and stated that, in addition, the obligation to fully restore the Blocks also arises from the PSCs468 and the transfers authorized by the Ministry,469 which obligation extends to the CPUF.470
260.
On this basis, the Tribunal understands that (i) Ecuador's environmental counterclaim is a strict liability claim in tort and that (ii) Ecuador refers to the PSCs first, to provide principles of liability in the event that such principles are lacking in the legal framework governing strict liability and, second to regulate causation in respect of other operators in the field.
261.
For its part, Burlington regards this counterclaim as one based exclusively on tort law.471 It also considers that the outcome under the PSCs would be no different, as "the PSCs exonerate Burlington from liability for harm caused by others".472 Before proceeding further, the Tribunal notes that the outcome would, however, be different, as the contractual liability starts to accrue from the first contracts for the Blocks and not from the beginning of the Consortium's operations, as is shown by Clause 2.1 in conjunction with Clause 5.1.20.10 of the Block 7 PSC and by Clause 5.1.20 of the Block 21 PSC.
262.
The Tribunal does not consider that resort to the PSCs is well-founded here. The 2008 Constitution and Ecuadorian case law provide the relevant tort liability principles, as established above, and there is no basis to import liability principles from the PSCs. As regards the contractual rules on successive liability of operators more specifically, Ecuador has not convincingly explained why the Tribunal would need to apply such contract rules to a tort claim in derogation from the tort liability rules contained in the 2008 Constitution and case law. The same reasoning applies to Ecuador's argument in respect to the successive transfers of the prior contractor's obligation to restore the Blocks which were authorized by the Ministry of Mines and Energy.473 While Ecuador characterizes these transfers as a "parallel source of Burlington's obligation to fully restore the environment in Blocks 7 and 21",474 it has not sufficiently substantiated that this obligation must take precedence over the general tort liability regime discussed above.
263.
Ecuador has also relied on Article 2217 of the Civil Code to argue that, where harm is attributable to several authors, each one is liable for the full amount.475 There is no need to dwell on the rather obvious point that this rule concerns a situation in which several actors create the same harm; it does not govern a situation in which an operator which has not caused the harm may, by operation of law or contract, be liable for the harm caused by another.
264.
Accordingly, Burlington can only be held liable in these proceedings for harm caused by the Consortium. In other words, Burlington can avoid liability if it shows that the alleged harm was caused by prior operators. This conclusion is nuanced to take account of the possibility that pre-existing environmental harm may have continued to expand or disperse as a result of the Consortium's breach of its duty to monitor the environmental conditions in the Blocks and remediate any exceedances, as well as for the possibility that the Consortium increased the environmental liabilities of an already contaminated site. In either case, such harm could properly be said to have been at least partially caused by the Consortium.
265.
In application of the principles of causation discussed above, the Consortium cannot either be held liable for harm caused by Petroamazonas, the operator that took over the oilfields in 2009 and expanded operations since then. The Tribunal, however, rejects Burlington's argument that the Consortium bears no liability for the environmental conditions of the Blocks simply by virtue of the fact that Petroamazonas took over the Consortium's rights and obligations by substituting itself to the Consortium.
266.
Having set out the key components of the strict liability regimes, the Tribunal will now address the Parties' positions and set out the Tribunal's approach regarding soil contamination (Section 4), mud pits (Section 5) and groundwater contamination (Section 6). Finally, the Tribunal will address the issue of well site abandonment (Section 7).

4. Soil contamination

267.
With respect to soil contamination, the Tribunal will start by addressing the notion of environmental harm (4.1), followed by general considerations on permissible limits (4.2), land use criteria (4.3), as well as guidelines to calculate impacted areas and volumes of impacted soils (4.4) and the remediation costs (4.5). This will then allow the Tribunal to outline its approach to assessing the environmental conditions in the Blocks and engage in a site-specific analysis (4.6).

4.1 The notion of environmental harm

268.
The Parties disagree on the definition of environmental harm. The opposition hinges on whether environmental harm is defined by reference to "permissible limits" set out in applicable regulations (as argued by Burlington) or to "background values" that reflect environmental conditions as they were prior to any human interference (as argued by Ecuador);476 and, if the permissible limits approach is adopted, whether the relevant limits are those applicable to industrial or agricultural land use (as argued by Burlington) or to sensitive ecosystems (as argued by Ecuador in the alternative).
269.
The consequences of adopting one or the other of these positions are significant in terms of the monetary relief for soil remediation claimed by Ecuador. Indeed, Ecuador's primary case is for soil remediation of any environmental harm in the Blocks back to background values, for which it claims USD 2,507,107,626 (the "background values case"). In the alternative, Ecuador claims soil remediation back to the sensitive ecosystems criteria, i.e. the most protective land use standards under RAOHE Table 6, in the amount of USD 790,465,298 (the "regulatory criteria case"). The Tribunal will therefore first assess the merits of Ecuador's argument that background values should apply as a general matter, and it will only address Ecuador's alternative case if its background values case is ill-founded.

4.1.1 Parties' positions

270.
In support of the background values case, Ecuador essentially argues that the notion of environmental harm under the 2008 Constitution and the EML is distinct from the notion of regulatory permissible limits enshrined in RAOHE and TULAS.477 For Ecuador, the Tribunal must determine environmental harm by examining whether there is a "negative impact" of any sort on the environment, that is, whether the alleged alteration affects "the functions of the environment or the renewability of the resources" in accordance with the EML.478 The Consortium, Ecuador states, had no right to pollute up to regulatory permissible limits and it must repair all harm found in the Blocks so as to restore the environment "to its original ‘background' condition".479 This must be so, according to Ecuador, because (i) Burlington's permissible limits theory is contrary to the Constitution's aim of full restoration,480 (ii) the notion of environmental harm is not defined in RAOHE and TULAS,481 which are mere instruments (iii) aimed at regulating the activities of oil operators,482 (iv) and at operating as risk management tools,483 which (v) govern administrative liability, not civil liability for harm.484 In the end, Ecuador argues that (vi) the significance of a "negative impact" can only be assessed on the specific facts of each case and environmental harm "cannot per se be defined through quality standards or permissible limits",485 in particular because RAOHE and TULAS do not differentiate the "types and characteristics of lands (deserts, forest, rainforest, mountains, islands…)".486
271.
For Burlington, the background values case has no technical or scientific basis, and only serves to inflate Ecuador's claims in the arbitration.487 In particular, Burlington highlights that IEMS, acting independently, initially applied the permissible limits in RAOHE and TULAS to determine the existence of harm in the Blocks, but was subsequently instructed by Ecuador to apply background values without ever evaluating whether the ecosystem had in fact been impacted.488 IEMS further confirmed that if the Consortium were still operating today it would be subject to RAOHE and TULAS, not background values,489 and RPS agreed that it is appropriate to consider local regulations when determining volumes of contaminated soils.490 According to Burlington, meeting permissible limits avoids environmental harm and it is these limits that determine whether remediation is warranted.491 While Ecuador failed to present a single example where background values had been applied in practice,492 the record is replete with examples where permissible limits were applied, including instances of remediation by Petroamazonas pursuant to RAOHE and TULAS. The 2010 Ex-Post Studies also evaluated the environmental condition of the Blocks under regulatory criteria.493 In addition, Burlington argues that IEMS applied a flawed methodology for calculating background values of metals resulting in 84 percent of IEMS's clean samples being deemed contaminated.494 In the end, Burlington submits that

"[r]emediation must be limited to those soils that exceed Ecuador's protective regulatory limits", and therefore Ecuador's background case should be dismissed in its entirety.495

4.1.2 Discussion

272.
The Tribunal will first ascertain the meaning of environmental harm within the Ecuadorian regulatory framework (Section a) and then address Ecuador's argument that background values provide the relevant threshold for determining the existence of environmental harm (Section b).

a. Environmental harm under Ecuadorian law

273.
The Tribunal starts its analysis by noting that neither the 2008 Constitution, nor RAOHE or TULAS contain a definition of environmental harm. Paragraph 2 of Article 396 of the Constitution provides that any environmental harm triggers the duty to fully restore the affected ecosystem. It reads in relevant part:

"All damage to the environment, in addition to the respective penalties, shall also entail the obligation of integrally restoring the ecosystems and compensating the affected persons and communities".496

274.
The Tribunal further observes that Article 71 of the Constitution states that nature has the "right to integral respect for its existence and for the maintenance and regeneration of its life cycles, structure, functions and evolutionary processes", and Article 72 provides for nature's "right to be restored". This latter provision further stipulates that in cases of "severe or permanent environmental impact" caused inter alia by the exploitation of nonrenewable resources, "the State shall establish the most effective mechanisms to achieve the restoration and shall adopt adequate measures to eliminate or mitigate harmful environmental consequences".497
275.
Article 397 further states that, in case of environmental harm, the State is to act immediately to ensure the restoration of affected ecosystems, and that, in addition to sanctions, "the State shall file against the operator of the activity that produced the damage proceedings for the obligations entailing integral reparation, under the conditions and on the basis of the procedures provided for by law".498 Thus, under the 2008 Constitution, any environmental harm triggers the duty of full restoration but under the conditions and on the basis provided for by law. The concept of integral or full restoration has been addressed in various instances. The travaux préparatoires of the 2008 Constitution indicate, for example, that full restoration means returning things to their original state ("volver las cosas a su estado original o a su lugar").499 And in Aguinda, the National Court of Justice held that integral reparation served to achieve the ultimate finality of "global restoration of environmental goods".500 In sum, the Constitution does not define environmental harm, but states that such harm must be fully restored according to the conditions set out in the law. The Tribunal therefore finds no support for Ecuador's argument that the notion of full reparation or full restoration under the 2008 Constitution mandates a return to pre-human conditions or, as Ecuador puts it, the return to a state of the environment "prior to any form of contamination".501 In the end, Ecuador's reliance on the notion of full restoration sheds no light on the notion of environmental harm. The question remains whether relying on permissible limits to define environmental harm would be contrary to the 2008 Constitution as Ecuador contends.502
276.
The notion of environmental harm is defined in the glossary of the EML, which also contains definitions of the terms contamination and environmental impact.503 Environmental harm is defined as:

"any significant loss, decrease, detriment or impairment to the preexisting conditions in the environment or one of its components. It affects the functioning of the ecosystem or the renewability of its resources".504

277.
Accordingly, environmental harm is more than a mere "negative impact" (as argued by Ecuador), it implies a significant loss or impairment. Ecuador's definition also fails to capture de minimis or socially and environmentally tolerable impacts, that is, as IEMS puts it, impacts which do not "significantly [affect] human health or the environment".505 The second sentence of the EML definition provides some criteria for determining when an environmental impact is "significant" and thus constitutes environmental harm, namely when it "affects the functioning of the ecosystem or the renewability of its resources".
278.
Neither Party has put much emphasis on the second sentence in the definition. Burlington has tendered evidence showing that the permissible limits in RAOHE and TULAS pose no risk to human health,506 but has not otherwise sought to demonstrate that compliance with these standards is sufficiently protective to ensure the ecosystem's functioning and the renewability of its resources. For its part, Ecuador has offered a preliminary analysis of certain macro-invertebrates,507 but has otherwise not provided an ecological risk assessment or any other analysis of possible impacts of the Consortium's operations on the ecosystem of the allegedly contaminated sites. The Tribunal will thus form its opinion on the background values case on the basis of other elements in the record.
279.
In this context, the Tribunal notes that the EML not only provides a definition of environmental harm, it also refers to permissible limits. Article 1 states that the EML establishes principles and guidelines of environmental policy, determines obligations and responsibilities, and "indicates the permissible limits, controls and sanctions in this matter". Moreover, Article 33 stipulates that "environmental quality parameters" shall be established as "instruments for the application of environmental norms", which shall be "governed by the respective regulation".508 This provision tends to support the proposition that RAOHE and TULAS refine and implement the general provisions of the EML.
280.
The Tribunal further notes that Ecuador's adoption in 1999 of the definition of environmental harm in the EML, was followed shortly thereafter by revisions to the permissible limits in RAOHE in 2001 and by the adoption of limits in TULAS in 2003. It is also particularly noteworthy that the definition of environmental harm under the EML, as well as the permissible limits under RAOHE and TULAS, remained unchanged following the adoption of the 2008 Constitution.
281.
It is also relevant that RAOHE Table 6 and TULAS Table 3 apply specifically to soil contamination and remediation. RAOHE Table 6 is entitled "Permissible limits for the identification and remediation of contaminated soils in all phases of the hydrocarbons industry". In this context, it is noteworthy that the glossary in RAOHE Annex 6 defines permissible limits as the "maximum value of concentration of element(s) or substance(s) in the different components of the environment, determined through standardized methods, and regulated through legal instrument".509 Article 4.2.2 of TULAS which comprises TULAS Table 3 is similarly entitled "Soil Remediation or Restoration Criteria", further specifying that the values in Table 3 aim at "establish[ing] the maximum contaminant concentration levels for soil in the process of remediation or restoration".510
282.
The Tribunal agrees with Ecuador that the constitutional duty of full or integral restoration is not in all cases necessarily exhausted by referring to permissible limits under RAOHE and TULAS, but it cannot agree with Ecuador's position that remediating back to permissible limits for the relevant parameters set out in these regulations would be unconstitutional.
283.
Harm to the environment could indeed be caused in a myriad of different ways, but in the Tribunal's view, when it comes to allegations of soil contamination by one of the parameters set out in RAOHE Table 6, then the dividing line between impact and harm (or between impacto ambiental and daño ambiental) is encapsulated in the very notion of permissible limits.

b. The subsidiary nature of background values

284.
It is only when there is soil contamination by an element not contemplated in RAOHE Table 6, such as for instance barium or arsenic, that recourse must be had to supplementary means such as TULAS, which contains permissible limits for additional parameters, and otherwise enshrines rules prescribing recourse to background values as a subsidiary means to determine the proper remediation standard. Of particular relevance in cases of soil contamination is Article 4.1.3.3 of TULAS, which specifically contemplates having recourse to background values only in the absence of a parameter in TULAS or in case of inapplicability of a certain parameter:

"In case of the inapplicability for the specific case of any parameter established in the present regulation, or in case of the absence in the regulation of a relevant parameter for the soil under study, the Environmental Control Entity shall adopt the following evaluation criteria:

The regulated entity must establish the background or reference value of the parameter of interest present in the soil. The regulated entity shall determine the present or current concentration of the parameter under study in the affected area. It shall then proceed to compare the obtained results for the concentration present in the soil against the background values. In general, it is considered that a present concentration greater than three times the background value for the soil denotes contamination that requires immediate attention on the part of the Environmental Control Entity".511

285.
Accordingly, recourse to background values is subsidiary and only permitted in two cases, namely when a specific parameter is inapplicable or the absence of a relevant parameter, an issue to which the Tribunal will revert further below. In contrast to the regime established in RAOHE, TULAS further differentiates between screening criteria and remediation criteria by stating that any concentrations exceeding three times the mean background value (screening criteria or action level) must be remediated back to 1.5 times the background value (remediation criteria).

"If the concentration is found to be three times higher than the background value, the Environmental Control Entity shall give immediate attention to this situation and must oblígate the regulated entity to remediate the soil until the present concentration is less than or equal to 1.5 times the background value".512

286.
This provision not only demonstrates the subsidiary or residual nature of recourse to background values to determine remediation criteria, but also puts to rest Ecuador's thesis that Burlington must remediate any alleged harm back to background values, since any remediation would ex hypotesi be limited to reducing any contamination present in soils to 1.5 times the mean background value of the relevant parameter. In sum, in addition to the fact that Ecuador has not pointed to a specific provision in its legal order which would mandate having recourse to background values as remediation criteria, the existing references to background values in Ecuadorian law have a much more limited scope of application than suggested by Ecuador.513
287.
Two further considerations are of relevance here, namely the actual practice in Ecuador when it comes to the remediation of environmental harm and the methodology initially adopted by Ecuador's experts from IEMS. With respect to actual practice, the Tribunal notes that Ecuador has not provided a single example where background values have been applied in practice. All remediation reports in the record show that instances of contamination were remediated back to permissible limits as specified in RAOHE and TULAS, be that before or after the promulgation of the 2008 Constitution.514 The Los Vencedores case is particularly apposite here since it deals with a 12-barrel crude spill that occurred in Block 7 on 11 June 2009 when the Consortium was still operating the Blocks. The court in that case repeatedly referred to permissible limits in RAOHE Table 6 and ultimately found that Perenco had properly remediated the spill within those permissible limits.515 The same can be said of the environmental audits in the record, whether commissioned by the Consortium or by Petroamazonas, which identify environmental conditions by reference to permissible limits.516 Interestingly, the court in Los Vencedores also referred to numeral 11 of Article 1 of the preliminary title of TULAS (which was not added to the record by Ecuador), which expressly refers to permissible limits and indicates that the duty to repair and restore the environment arises once these limits are exceeded.517
288.
It is also noteworthy that the 32 Environmental Impact Statements and Environmental Management Plans issued since 2001 for both Blocks (that is, after Decree 1215 was enacted) and reviewed by GSI show that RAOHE and TULAS were consistently referred to as the applicable regulations.518 GSI further pointed to 20 remediation projects undertaken in other Blocks in the Oriente region, all of which applied regulatory criteria under RAOHE Table 6 and none requiring remediation back to background levels.519
289.
Of further relevance is the fact that Ecuador's own expert IEMS initially sought to determine the environmental conditions of the Blocks by reference to the permissible limits in RAOHE and TULAS,520 before being instructed by Ecuador to apply background values on the grounds that permissible limits were not sufficiently protective of the environment.521 Ecuador admitted at the Hearing that it instructed IEMS to apply background values522 and explained that IEMS employed permissible limits in RAOHE and TULAS during its first sampling campaign "for the sake of information only", since IEMS did not yet have the reference background values at its disposal.523 IEMS moreover opined that Ecuadorian benchmarks are "arbitrary" and "not sufficiently protective of the health of the people who live in the area".524
290.
Whatever the merits of these explanations, it remains that IEMS chose to use regulatory limits in its first report and then shifted to background values under the pretext that, in its expert opinion, the regulations were not protective enough, but in reality because it was instructed to do so.525 As to Ecuador's argument that its own regulatory limits are not sufficiently protective of the environment or human health, the Tribunal observes that Ecuador set the permissible limits in RAOHE and TULAS further to the adoption of the EML, that it did not amend these limits following the entry into force of the 2008 Constitution, and that it is not for the Tribunal to modify Ecuador's duly enacted regulatory framework in place of the competent authorities. The same applies to Ecuador's argument that its regulations are deficient since they do not differentiate between various ecosystems in Ecuador, and that the Amazon region should find particular protection. As already noted above, the 2008 Constitution makes special reference to the Amazon region and this element will be of some relevance when it comes to determining the applicable land use, but here again it falls within the remit of the Ecuadorian authorities to set different criteria for the Amazon region, if they so wish; it is not for this Tribunal to do so.
291.
In light of the foregoing, it is the Tribunal's view that environmental harm is defined by reference to regulatory criteria. In the case of oilfield operations, these regulatory criteria are to be found primarily in RAOHE, and subsidiarily in TULAS. In other words, an oilfield operator could not be considered to have caused environmental harm if permissible limits were observed, since precisely these permissible limits allow determining when a negative impact crosses the threshold of harm.
292.
Consequently, the Tribunal will resort to RAOHE and TULAS to define harm under the EML. Hence, the limits set out in these regulations establish when impacts become significant, and thereby become harm, thus allowing the inference that these limits determine when the functioning of the ecosystem or the renewability of its resources are affected.526 In other words, any exceedance of applicable limits triggers extracontractual civil liability and the ensuing obligation of full restoration back to these limits, independently from parallel administrative liability and possible sanctions.
293.
Having rejected Ecuador's background values case, the Tribunal will now assess the regulatory criteria cases presented by the Parties respectively, starting with the applicable permissible limits for soil remediation and thereafter setting out the applicable land use criteria. Following its analysis for soil remediation, the Tribunal will address the applicable regulatory framework for mud pits and for the alleged groundwater contamination.

4.2 Permissible limits for soil remediation

294.
Turning to Ecuador's alternative case for soil remediation based on regulatory criteria, the Tribunal must now determine which standards apply to the allegedly impacted areas in the Blocks. The Parties agree that under the regulatory criteria scenario, RAOHE Table 6 applies for soil remediation, but they disagree whether the values in that table must be adjusted to higher natural values and, if so, what the proper values should be (4.2.2). They further disagree whether TULAS Table 2 (as Ecuador contends) or Table 3 (as Burlington contends) applies for parameters not contemplated in RAOHE Table 6 (4.2.1).527

4.2.1 Applicable table under TULAS for soil remediation

295.
TULAS Table 2, entitled "soil quality criteria", is part of Article 4.2.1 and sets out soil quality criteria for different "parameters" or substances. Article 4.2.1 defines soil quality criteria as "approximate background values or analytical detection limits for a contaminant in the soil". It further provides that background values "refer to the representative environmental levels for a contaminant in the soil" and that these values "reflect the natural geological variations of undeveloped areas or areas free of the influence of generalized industrial or urban activities". In other words, Table 2 only provides for approximate background values reflecting natural variations in undeveloped areas. By contrast, Table 3 is entitled "Soil Remediation or Restoration Criteria" and sets forth the "maximum contaminant concentration levels for soil in the process of remediation or restoration". In light of the subject matter of the two tables, Table 3 is the proper source for soil contamination parameters not contemplated in RAOHE Table 6.

4.2.2 Adjustment to higher natural values

296.
The Tribunal now turns to the need to adjust the permissible limits to account for higher natural values. After setting out the adjusted values adopted by the Parties as well as the justifications provided by the Parties (Section a), the Tribunal will address the methodologies adopted by the Parties to calculate their adjusted values and then establish the adjusted values applicable to the present case (Section b).

a. Parties' positions

297.
The Parties adjusted the permissible limits of various compounds to take into account higher natural concentrations in the soils, but reached different conclusions in great part because they applied different methodologies. The Tribunal will first set out the adjusted values adopted by the Parties ((i) below) and then the justifications provided by the Parties ((ii) below).

(i) The adjusted values adopted by the Parties

298.
Ecuador's expert IEMS collected 192 soil samples in areas not impacted by oilfield operations to calculate the natural concentrations of (i) cadmium, nickel and lead, which are regulated under RAOHE Table 6;528 and (ii) barium and vanadium, which are regulated under TULAS Table 3.529 In addition to making its own calculations for the values of these compounds, Burlington's expert GSI also calculated the natural value for chromium in Block 7 but not Block 21.530 As a result, each Party adopted the higher values which it had reached as remediation criteria for barium, cadmium, lead, nickel and vanadium.531 The relevant values adopted by the Parties are the following:532

Chemical IEMS Soil Criteria (mg/Kg) GSI Soil Criteria (mg/kg)
IEMS "Background" Criteria IEMS "Regulatory" Criteria Applicable Rule Industrial Agricultural Sensitive Ecosystem Residential
\CPUF/Block 7
Barium 213.1 639.3 TULAS 2000 750 - 706
Cadmium 2 1 9967 RAOH 10 3.8 3.8 -
Chromium - 60 TULAS 90 65 - 65
Lead 1841 80 RAOH 500 100 80 -
Nickel 14.41 40 RAOH 100 50 40 -
Vanadium 106.71 320.75 TULAS 311 311 - 311
TPH - 1000 RAOH 4000 2500 1000 -
Block 21
Barium 113.93 600 TULAS 2000 750 - 500
Cadmium 1,775 1,775 RAOH 10 2 2 -
Chromium - 60 TULAS 90 65 - 65
Lead 1841 80 RAOH 500 100 30 -
Nickel 34.13 40 RAOH 100 50 47 -
Vanadium 62.31 186.93 TULAS 130 130 - 130
TPH - 1000 RAOH 4000 2500 1000 -

299.
Accordingly, with respect to Block 7/CPUF, the Parties agree that RAOHE Table 6 be adjusted for cadmium, not for lead and nickel, but disagree on the adjusted value and the relevant land use. Ecuador argues that the cadmium value be adjusted to 2 mg/kg for sensitive ecosystems, whereas Burlington argues that the cadmium value be adjusted to 3.8 mg/kg for sensitive ecosystem as well as agricultural use. And with respect to compounds regulated under TULAS, the Parties agree that TULAS Table 3 be adjusted for barium and vanadium, and not for chromium. They disagree, however, on the adjusted values. Ecuador argues that the barium value for sensitive ecosystem land use should be increased from 500 mg/kg to 639 mg/kg, whereas Burlington argues for an increase to 706 mg/kg. With respect to vanadium, Ecuador's adjusted value of 320.75 mg/kg is higher than Burlington's adjusted value of 311 mg/kg.
300.
With respect to Block 21, the Parties disagree on the adjusted value for cadmium for sensitive ecosystems, Ecuador arguing for an increase to 1.77 mg/kg and Burlington to 2 mg/kg. They further disagree on whether the value for nickel for sensitive ecosystems should be adjusted, Burlington arguing that the value for nickel should be increased from 40 mg/kg to 47 mg/kg. In addition, Ecuador calculated the background value for barium at 133 mg/kg and the remediation criteria at 600 mg/kg, while Burlington provided no adjusted value for barium. Finally, Ecuador proposed an adjusted value of 186 mg/kg for vanadium, while Burlington did not adjust the vanadium limit.

(ii) Justifications

301.
In general terms, the disagreement between the Parties on adjusting regulatory criteria to reflect higher natural concentrations in the soils essentially turns on when a specific TULAS parameter is inapplicable and, when it is, whether the adjusted value should be calculated using the mean value of all samples (as Ecuador and its experts contend) or the 99% upper prediction limit ("UPL") (as Burlington and its experts contend).
302.
The approach adopted by Ecuador's experts can be summarized as follows. IEMS gathered 192 samples from the Blocks to calculate the background values.533 For IEMS, background values under TULAS are determined by calculating the population mean, i.e. the mean of the complete set of samples. IEMS thus calculated the population mean by taking "all particles of soil from the unaffected areas in each block, to average them, and to calculate an average concentration of all the particles".534 Since IEMS did not sample the entire Blocks, it constructed a confidence interval ("CI") around "a range that contains the population mean with a specified probability".535 IEMS estimated the 99% CI of the mean concentrations, such that there is, in its opinion, a "99% probability that the true but unknown background mean concentration is below the value presented by IEMS and, conversely, only a 1% probability that it be above".536 This approach, says IEMS, is "consistent with the provision of TULAS, although slightly more conservative", since using a 95% CI would result in lower background concentrations thus increasing the size of the contaminated area.537 Thus, for barium in Block 7 for instance, IEMS calculated the population mean to be 183.65 mg/kg, with a 99% CI value set at 213.10 mg/kg whereas a 95% confidence interval would have resulted in a value of 203.21 mg/kg. In other words, if the 99% CI method is chosen, as advocated by IEMS, there is a 99% probability that the background value for barium is below 213.10 mg/kg and a 1% probability that it is above that value.
303.
IEMS then compared the values of the 99% CI with the values in TULAS Table 2 and then estimated the adjusted regulatory criteria by multiplying by three the site-specific background value "or 3 times the Table 2 criteria (general background values), whichever was higher".538 IEMS stressed that, by contrast, GSI's "selective use of the criteria" by estimating background values as the 99% UPL is "erroneous and unreliable", since it is "inconsistent with regulatory requirements".539 For IEMS, no provision states that the background values must be estimated by comparing the 99% UPL to any table in TULAS, let alone to Table 3 as GSI did. Instead of using the 99% UPL, GSI should have calculated the "true population mean" of all 192 samples to estimate the 99% CI.540 IEMS further argued that GSI engaged in "biased manipulation", since the "99th percentile is the value that is above 99% of the dataset", and thus artificially increased the "tolerance to contamination".541 In other words, had GSI applied a 100% UPL, the background value would have equalled the highest measured value in clean soil samples. Similarly, after GSI eliminated outliers, its 99th percentile approach resulted in a background value that is above 99% of the dataset, and therefore biased and excessively conservative.
304.
Ecuador's experts from RPS concur with IEMS that GSI's use of the 99% UPL is inconsistent with the methodology specified in TULAS. RPS explains that remediation criteria for a particular parameter are not applicable if "the remediation criterion for a parameter is less than the background concentration".542 In such cases, the "background threshold values" ("BTVs") is equal to three times the mean of the background value. In other words, if concentrations at an oilfield site exceed the BTV, "then remediation is required and the actual cleanup level (remediation criteria) is set equal to 1.5 times the mean of the background data set".543 While GSI's block-specific remediation criteria equal to 3 times the mean background value are correct, RPS argues that GSI should also have shown "the actual remediation criteria as 1.5 times the mean background concentration".544 Instead, GSI's approach underestimates the volume of impacted soils to be remediated. In addition, RPS highlights GSI's use of "incorrect values for the background soil concentrations", pointing to discrepancies between the data used by GSI in the ProUCL software with the background soil concentration included in GSI's Table D.18 Analytical Results for Background Metal Concentrations.545
305.
GSI adopts a different approach to estimate background values. While GSI accepts that IEMS's tests may be representative of background soil conditions, GSI criticizes the "fundamental conceptual error" committed by IEMS when calculating their background metals concentrations.546 By calculating the 99% upper confidence limit on the mean concentration, GSI considers that IEMS adopted an adjusted mean concentration that is only slightly above the average "but far below the upper range of concentrations observed in clean samples", thus resulting in "approximately 80% of the clean soils at all sites" being declared by IEMS as being contaminated.547
306.
According to GSI, background values are in practice "commonly established at or near the upper range of observed concentrations (e.g., 99th percentile) of a compound in unaltered "clean soils".548 According to GSI, the use of the UPL is "a well-recognized statistic" that "can be employed as a conservative measure of the upper range of normal metals concentrations", where UPL is defined "as the value below which a given percentage of the background sample population falls" and corresponds to the "upper end of the normal range of measured background concentration".549 To that end, GSI conducted additional background soil sampling, collecting 91 additional samples, including 12 duplicate samples in the CPUF and Block 7.
307.
GSI's approach can be summarized as follows. As a first step, GSI compiled all background samples in a Microsoft Access database and compared the population distributions of metal concentrations in both Blocks. Using the Wilcoxon-Mann-Whitney statistical test, GSI then determined that the mean background concentrations of all metals, except for lead and nickel in Block 7, were different from those in Block 21, thus justifying evaluating the background soil samples from CPUF/Block7 and Block 21 as "two distinct background populations, representing site-specific conditions".550 Since the comparison of IEMS and GSI samples collected in CPUF/Block 7 presented similar background populations, GSI combined its analytical results with those of IEMS "to create an encompassing background soil dataset (CPUF/Block 7) comprised of 243 unaltered soil samples".551
308.
As a next step, GSI excluded data outliers from the background populations in the calculation of the UPL by using Rosner's Outlier Test in the ProUCL software.552 GSI also used half of the detection value for non-detect concentrations. After deriving the 99% UPL for each data set,553 GSI evaluated the 99% UPL value against the population of metals measured in background soil samples.554 It finally proceeded to adjust the soil cleanup criteria based on the applicable regulations in RAOHE and TULAS.555 For GSI, the fact that outliers were excluded and more than 1% of clean soil samples exceeded the 99% UPL for several metals (barium, cadmium and nickel), confirms that its approach is conservative.556 GSI rebuts IEMS's criticism that outliers should only be removed after careful review and that Rosner's Outlier Test should only be used when data follows a normal distribution, by stating that the inclusion of the outliers would have skewed the 99% UPL towards less conservative estimates.557
309.
With respect to RAOHE Table 6, GSI considers that the 99% UPL background level is the appropriate cleanup criteria for chemicals for which the relevant 99% UPL background level exceeds the regulatory criteria.558 With respect to TULAS, GSI opines that IEMS not only misinterpreted TULAS but also misrepresented GSI's approach,559 which it explains in the following terms: "for chemicals for which the relevant 99% UPL background level exceeded the regulatory cleanup level provided in TULAS Table 3 (meaning that the cleanup level logically cannot apply), an action level equal to three times the average background concentration measured for that chemical [is to be used] as the appropriate site-specific cleanup level".560 Accordingly, this approach not only recognizes that remediation below natural background concentrations is "infeasible", but also that "an additional ‘significance' factor" is applied "to more accurately indicate abnormal conditions that may warrant action".561
310.
In other words, since TULAS does not define when a relevant remediation criterion is inapplicable, GSI opined that the "only reasonable interpretation" is that a criterion is inapplicable "when it is less than the observable range of clean background concentrations".562 Thus, if the 99% UPL of a given parameter exceeded the remediation criterion, GSI calculated a substitute criterion as mandated by TULAS by (i) establishing the mean background levels, (ii) calculating a substitute criterion by multiplying the mean concentration of the clean background soil samples by three, and (iii) comparing concentrations in the affected area to the substitute criterion to determine if remediation is required.563 By contrast, the effect of IEMS's approach is "absurd" according to GSI, since 84% of IEMS's actual clean samples require cleanup under its approach, which constitutes a "gross overestimate of the cleanup requirements".564 By ignoring that TULAS specifically requires that soil measurements must be compared to a substituted criterion that is three times the average background concentration, IEMS erroneously concluded that 84% of clean soils are contaminated.565
311.
For GSI, RPS also misinterpreted TULAS by claiming that a relevant criterion is inapplicable when the mean value of clean background samples is greater than the relevant criterion.566 This approach is "irrational" according to GSI, since this would mean that "approximately half of clean samples used to represent naturally occurring background conditions would require remediation under TULAS regulations".567 RPS further misapplied TULAS by claiming that the background threshold value should be calculated using three times the mean background concentration. According to GSI, TULAS instructs that three times the mean value should be used as a replacement regulatory criterion "only when the regulatory criterion provided by TULAS is ‘inapplicable'".568 RPS also misrepresented GSI's actions when claiming that GSI should have calculated the actual remediation criteria as 1.5 times the mean value of the background concentration.569 GSI explained that the volume of impacted soils should be determined by using three times the mean value of the background dataset, and TULAS only requires that this volume of soil should be remediated to 1.5 times the mean background samples.570 Accordingly, RPS's approach for calculating soil remediation criteria violates TULAS, since RPS "replaced the Ecuador criteria with a value equal to 1.5 times the mean background concentration".571

b. Discussion

(i) Preliminary remarks

312.
At the outset, the Tribunal notes that, with slight divergences, both Parties accepted that their background values apply to the entire block, even though the actual background concentrations may vary significantly from one site to another within the same block. The Tribunal further notes that neither Party has pointed to any practice where regulatory criteria have been adjusted to take into account higher natural concentrations in the Blocks.

(ii) Methodologies to adjust values under RAOHE and TULAS

313.
RAOHE Table 6 and Article 4.1.3.3 TULAS allow adapting permissible limits to take into account that natural concentrations of certain chemicals in the soil exceed permissible limits for those substances. However, these two regulations adopt different approaches. RAOHE Table 6 provides for adjusting the Table 6 limits upwards – "se pueden incrementar" – where they are exceeded by natural concentrations.572 The use of the word "incrementar" defeats Ecuador's contention that these limits can also be adjusted downwards. On the other hand, RAOHE Table 6 does not indicate how to assess the natural concentrations or how to adjust the permissible limits to account for these.
314.
Article 4.1.3.3 of TULAS adopts a different approach to the adjustment of the remediation criteria:

"In case of the inapplicability for the specific case of any parameter established in the present regulation, or in case of the absence in the regulation of a relevant parameter for the soil under study, the Environmental Control Entity shall adopt the following evaluation criteria:

The regulated entity must establish the background or reference value of the parameter of interest present in the soil. The regulated entity shall determine the present or current concentration of the parameter under study in the affected area. It shall then proceed to compare the obtained results for the concentration present in the soil against the background values. In general, it is considered that a present concentration greater than three times the background value for the soil denotes contamination that requires immediate attention on the part of the Environmental Control Entity. […] The procedure described shall be coordinated and supervised by the environmental control entity".573

315.
This provision goes on to state that the most reliable way to calculate natural concentrations is to collect samples in areas "immediately outside the area under study" with no local contamination. Where there is a "total absence" of such values in the immediate area surrounding the area under study, the background values may be obtained in regional or national areas. More specifically, it provides some indications on the sampling:

"In order to determine the background or reference value, at least 5 samples must be taken, if between 5 and 20 samples are taken, the mean value must be selected as the background value".574

316.
In sum, with different approaches, both RAOHE and TULAS allow adjusting relevant remediation criteria to take into account higher natural concentrations in soils. On the one hand, RAOHE Table 6 merely states that permissible limits can be adjusted upwards when natural concentrations exceed the regulatory criteria without specifying how this is done. On the other hand, TULAS provides that contamination exists when concentration in the soil is greater than three times the mean background value in case a given parameter is absent or inapplicable. In addition, TULAS provides that background values are to be determined by taking at least 5 samples and that "if between 5 and 20 samples are taken, the mean value must be selected as the background value".575 This formulation is confusing, as it might be read to suggest using another methodology above 20 samples. Since the regulation provides for no alternative methodology, the Tribunal rejects this suggestion. It therefore finds that, where IEMS collected 192 samples, the mean value of these samples must in any event be estimated so as to comply with TULAS.
317.
The Parties diverge significantly on the proper methodology for adjusting permissible limits. Ecuador's experts from IEMS adopted the 99% CI to determine the mean background concentration, whereas Burlington's experts from GSI held that the proper methodology is to estimate the 99% UPL for each metal in clean background soil samples. In this context, the Tribunal notes that neither RAOHE nor TULAS refer to either methodology, TULAS only referring, as seen above, to the "mean" background value (valor promedio).576 The Tribunal further notes that GSI, while invoking "accepted technical guidelines", in particular of the Environmental Protection Agency of the United States but none from Ecuador, indicated that resort to the 99% UPL "can be employed as a conservative measure of the upper range of normal metals concentrations",577 and that it was "our understanding […] that sound technical judgment should be applied in this determination".578 Absent any express stipulation in the regulations, the Tribunal is disinclined to follow GSI's approach and thus relies on IEMS's approach of estimating the 99% CI even if it too is not strictly compliant with TULAS. This is so, because the 99% CI is a slightly more conservative method than the mean value, there being only a 1% probability that the background mean concentration is above the confidence interval calculated by IEMS. Accordingly, the Tribunal will estimate the background values on the basis of the mean background concentration as adjusted by the 99% CI.

(iii) Adjusted values determined by the Tribunal

318.
With respect to RAOHE, the Tribunal notes that the only divergence between the Parties lies in the cadmium values for both Blocks and the nickel value for Block 21. IEMS accepts that the sensitive ecosystem land use criterion for cadmium be adjusted to 2 mg/kg for Block 7, which is the same criterion for agricultural land use, and to 1.77 mg/kg for Block 21. For its part, GSI estimated the background value for cadmium in Block 7 at 3.8 mg/kg, thus requiring in its view an adjustment of both the sensitive ecosystem and agricultural land use criteria. For nickel, GSI estimated the background value in Block 21 to be 47 mg/kg instead of 40 mg/kg. Consistent with its decision to discard GSI's use of the 99% UPL to estimate background values, the Tribunal will adjust the criteria for cadmium as follows: 2 mg/kg for Block 7 and 1.77 mg/kg for Block 21. Similarly, the Tribunal did not adjust the nickel value for Block 21. This approach is consistent with the few instances of practice that the Tribunal discerned in the record. For instance, the Tribunal observes that in two remediation reports for spills in Oso 2 in 2007 and in Mono CPF in 2008, the applicable cadmium value for agricultural land use was held to be 2 mg/kg and not the 3.8 mg/kg advocated presently by Burlington.579
319.
With respect to TULAS, there is a first significant disagreement between the Parties on when a specified criterion is inapplicable under TULAS, thus justifying adjusting that criterion to reflect background conditions. For IEMS, a given criterion is inapplicable if the mean background concentrations exceed the criterion provided in TULAS Table 2.580 RPS also stated that a remediation criterion is not applicable, for instance, if the remediation criterion for a parameter is less than the background concentration.581 For its part, GSI indicated that a remediation criterion is to be considered inapplicable under TULAS "when it is less than the observable range of clean background concentrations", meaning that only if the 99% UPL exceeds the established criterion in Table 3, in which case an adjusted criterion should be established by multiplying by three the mean background concentrations.582
320.
TULAS contemplates adjusting remediation criteria in two scenarios, namely in case (i) of the "inapplicability for the specific case of any parameter established in the present regulation", or (ii) of the "absence in the regulation of a relevant parameter for the soil under study".583 At the outset, the Tribunal stresses that, while TULAS speaks of the "inapplicability of any parameter" relevant for the specific case, it should properly be interpreted for present purposes as meaning the inapplicability of any criterion for a given relevant parameter (since it would be contradictory to establish new criteria for a parameter which is held to be inapplicable). The Tribunal notes that neither Party argued that in the present case a relevant parameter is absent in TULAS, but that they advocated differing interpretations on when a criterion of a given parameter is inapplicable. In this context, the Tribunal recalls that TULAS Table 3 contemplates four land uses, namely industrial, commercial, agricultural and residential, but does not provide specific criteria for sensitive ecosystem land use. Thus, it may reasonably be considered that the remediation criteria established in TULAS are inapplicable for sensitive ecosystem land use and that new criteria should be established for this hypothesis by following the procedure provided for in Article 4.1.3.3. This also avoids having to choose either agricultural or residential criteria as a proxy for sensitive ecosystem land use.
321.
In conformity with its decision not to use GSI's 99% UPL, the Tribunal also rejects GSI's interpretation that a given criterion becomes inapplicable only if the 99% UPL value exceeds the relevant criterion set in Table 3. For the Tribunal, the correct approach to establish new or adjusted permissible limits for sensitive ecosystem land use is to multiply the mean background value by three, and to compare the result with the relevant criterion set out in TULAS Table 3. This approach does away with GSI's criticism that relying on mean background values would imply that clean samples are considered contaminated, an argument that would hold true for Ecuador's background value case (which was rejected above), but not for Ecuador's regulatory criteria case. Thus, for instance, with respect to barium, IEMS calculated a mean background value (adjusted to the 99% CI) of barium in Block 7/CPUF of 213.10 mg/kg, meaning that the permissible limit for barium in that block is 639.30 mg/kg. This value appears reasonable when compared to the permissible limit of 750 mg/kg for agricultural land use and 500 mg/kg for residential land use.584
322.
The issue is somewhat different with respect to vanadium in Block 7, the only other parameter where the Parties disagree on the adjusted value. They concur, however, that the permissible limits for all relevant land uses should be adjusted upwards. This is thus truly a case where the regulatory remediation criteria need adjustment. For the Tribunal, a given criterion becomes inapplicable if (i) the mean background value exceeds the soil quality criterion established in Table 2 and (ii) the newly adjusted criterion calculated as three times the mean background value exceeds any given permissible limit. Thus, for vanadium, the relevant criteria for the contemplated land uses (i.e., industrial and agricultural) must be adjusted since the mean background value calculated by IEMS of 106.71 mg/kg exceeds the 25 mg/kg criterion established in Table 2. Here too, the Tribunal will rely on IEMS's mean background value of 106.71 mg/kg, thus adjusting the applicable permissible limit to 320.75 mg/kg.585 The Tribunal also adopts this value for sensitive ecosystem land use. Although this value is somewhat higher than the one calculated by GSI (i.e., 311 mg/kg), the Tribunal notes that it does not make any difference in terms of actual volumes of soils to be remediated since there are no samples showing vanadium values in between those proposed by either Party. For the same reasons, the Tribunal adopts Ecuador's adjusted value of 186 mg/kg for vanadium in Block 21 for agricultural and industrial land uses, as well as for sensitive ecosystem land use.
323.
Finally, the Tribunal rejects RPS's approach of calculating the remediation criteria as 1.5 times the mean background value concentrations, since the volume of contaminated soils is determined on the basis of the adjusted permissible limit (i.e., a multiple of three of the mean background value), not the adjusted remediation criteria set in TULAS (i.e., a multiple of 1.5 of the mean background value). As GSI correctly stressed, "TULAS never states that soil samples with concentrations exceeding 1.5 times the mean of background samples must be remediated", since this would lead to an overestimation of the volume of soil requiring remediation.586
324.
On this basis, the Tribunal has determined that the applicable remediation criteria are as follows (the adjusted criteria are in bold):

• Under RAOHE for Block 7
ParameterEcuadorian RegulationRegulatory Criteria SE/Agr/IndEcuador's adjusted valueBurlington's adjusted valueTribunal's adjusted valueAdjusted Criteria SE/Agr/Ind
Cadmium RAOHE Table 6 1/2/10 2 3.8 22 /2/10
Lead RAOHE Table 6 80/100/500 80 80 80 80/100/500
Nickel RAOHE Table 6 40/50/100 40 40 40 40/50/100
• Under TULAS for Block 7
ParameterEcuadorian RegulationRegulatory Criteria Res/Agr/IndEcuador's adjusted valueBurlington's adjusted valueTribunal's adjusted valueAdjusted Criteria Res/SE/Agr/Ind
Barium TULAS Table 3 500/750/2000 639 706 639 500/ 639 /750/2000
Chromium TULAS Table 3 65/65/90 60 65 65 65/65/65/90
Vanadium TULAS Table 3 130/130/130 320.75 311 320.75 130/ 320 / 320 / 320

• Under RAOHE for Block 21
ParameterEcuadorian RegulationRegulatory Criteria SE/Agr/IndEcuador's adjusted valueBurlington's adjusted valueTribunal's adjusted valueAdjusted Criteria SE/Agr/Ind
Cadmium RAOHE Table 6 1/2/10 1.77 2 21.77 /2/10
Lead RAOHE Table 6 80/100/500 80 80 80 80/100/500
Nickel RAOHE Table 6 40/50/100 40 47 40 40/50/100
• Under TULAS for Block 21
ParameterEcuadorian RegulationRegulatory Criteria Res/Agr/IndEcuador's adjusted valueBurlington's adjusted valueTribunal's adjusted valueAdjusted Criteria Res/SE/Agr/Ind
Barium TULAS Table 3 500/750/2000 600 500 600 500/ 600 /750/2000
Chromium TULAS Table 3 65/65/90 60 65 65 65/65/65/90
Vanadium TULAS Table 3 130/130/130 186 130 186 130/ 186 / 186 / 186
4.3Land use criteria

325.
Having determined that RAOHE Table 6 and TULAS Table 3 apply to soil remediation, the Tribunal will now address the issue of land use classification. As seen above, both RAOHE Table 6 and TULAS Table 3 set different limits depending on the subsequent use (uso posterior) of the land.
326.
RAOHE Table 6 distinguishes between industrial, agricultural and sensitive ecosystem land uses, the first being more permissive than the last. TULAS Table 3, for its part, distinguishes between industrial, commercial, agricultural and residential land uses. While the first is also here the most permissive, agricultural or residential use is the least permissive depending on the particular parameter. Since TULAS Table 3 does not contemplate land use for sensitive ecosystems, the Tribunal used the adjusted background values calculated by IEMS as the applicable values for sensitive ecosystems for the relevant parameters (i.e., for barium and vanadium).
327.
The Parties diverge on the meaning of uso posterior, and on the question whether the entire Blocks should be classified as sensitive ecosystem (as Ecuador contends) or agricultural land (as Burlington contends). They further disagree whether the notion of uso posterior in RAOHE Table 6 refers to use immediately after remediation (as Burlington argues) or to eventual use at the end of hydrocarbons operations (as Ecuador argues). After setting out the Parties' positions (4.3.1), the Tribunal will first clarify the meaning of uso posterior, before addressing the time at which uso posterior must be assessed and then setting out its own approach to determine the land use at each site (4.3.2).

4.3.1 Parties' positions

328.
For Ecuador, uso posterior refers to future land use. It does not refer to land use after the Consortium leaves the Blocks but to the time "when oil operations will end",587 especially since oil operations are "per se temporary" and that the oilfields are destined "to be absorbed again by rainforest".588 Accordingly, the current use of the Blocks, whether Petroamazonas' expansion or the use made by farmers of the land around active fields "is obviously of no relevance to future land use".589 The Tribunal should therefore characterize all facilities and surrounding areas in the Blocks as sensitive ecosystems,590 especially taking into account IEMS's ecological studies demonstrating that the "areas in question display characteristics similar to those of a National Natural Area".591 Such conclusion is further reinforced by the fact, Ecuador argues, that the Blocks overlap with the Sumaco and Yasuní biosphere reserves, and that a "significant portion of Block 21 lies within the Huaorani indigenous reserve".592
329.
For its part, Burlington rejects Ecuador's "broad and extraneous characterizations of the Amazon rainforest as generally sensitive".593 It points to various instances of practice where Ecuadorian regulators applied the "immediate subsequent use" to classify land use for remediation purposes,594 further highlighting that industrial facilities have consistently been classified as falling under the industrial land use criteria, and that otherwise agricultural land use was applied as a default criteria, except for the few instances where sites have been expressly designated as protected areas under the National System of Protected Areas ("SNAP").595 In this context, Burlington argues that sensitive ecosystem land areas are areas which are either "(1) certified under the National System of Natural Areas, or ‘SNAP' program, or (2) expressly designated as sensitive ecosystem in an Environmental Impact Study".596 Accordingly, Ecuador's "colloquial description" of the Amazon as sensitive ecosystem in its entirety for regulatory purposes should be rejected.597 Burlington further argues that the certificates of intersection issued by the Ministry of Environment under the SNAP program show that "the majority of the Blocks do not intersect with the SNAP program".598 Burlington also indicates that the Environmental Impact Studies, including the 2010 Ex-Post Studies, consistently apply agricultural and industrial land use standards.599 Finally, Burlington calls Ecuador's reliance on the Sumaco Biosphere Reserve a "red herring", and irrelevant, since that reserve has not been designated as sensitive ecosystem in the SNAP program or in any Environmental Impact Study which are "the only two ways to designate an area as sensitive ecosystem under RAOH".600

4.3.2 Discussion

a. The meaning of uso posterior

330.
First, the Tribunal must determine the meaning of uso posterior and at what point in time such use must be assessed. RAOHE Table 6 speaks of "subsequent use to be given to the remediated soil" (uso posterior a darse al suelo remediado).601 It adds that such use must be indicated in the remediation program to be approved by the Subsecretariat of Environmental Protection, which tends to show that the determination is both context-driven and immediately posterior to any remediation. This is reinforced by the use of the expression uso posterior (subsequent use), as opposed to the more indeterminate uso futuro (future use). It thus appears that the uso posterior of land is determined on a case by case basis according to its use immediately after remediation, and that this determination is subject to approval in each instance by the Subsecretariat.602 This reading consequently does not support the argument of a single standard to be applied across the board. Accordingly, the Tribunal cannot accept Burlington's position that the agricultural standard applies by default to all off platform areas not specifically designated as sensitive ecosystems.
331.
For the same reason, the Tribunal cannot accept Ecuador's position either according to which the entire Blocks must be deemed to be sensitive ecosystem or that their future use, as the expression is used in Ecuadorian law, is necessarily and in all cases to revert to rainforest.603 Ecuador's argument that the Consortium must remediate the entire Blocks back to sensitive ecosystem criteria is incorrect for practical reasons. It makes no sense that an operator should be required to restore to sensitive ecosystem values the soil of an operating platform that will continue in operation. This is especially so in a situation where Ecuador itself concedes that Petroamazonas may well continue to operate the Blocks for another 30 years.
332.
Furthermore, the evidence on record does not support Ecuador's view. To the contrary, the record demonstrates that the uso posterior of platforms and other industrial facilities has been consistently classified as industrial.604 This practice has continued after Petroamazonas took over the Blocks in July 2009.605

b. The time at which uso posterior is assessed

333.
This raises the question of the time at which the subsequent land use must be assessed. When the operator alleged to have caused the harm is still operating the facilities, the subsequent use is normally assessed at the time of remediation. The position is different here as the Consortium stopped operations in July 2009 and the harm was allegedly caused before, but only brought forward in 2011. In the circumstances, the Tribunal considers that the most appropriate solution is to assess the subsequent use at the time of the expropriation in August 2009. Indeed, the Tribunal must determine the existing liabilities of the Consortium at the time it was expropriated.606

c. The Tribunal's approach to determining subsequent land use

334.
The Tribunal is of the view that functioning platforms should be classified as ‘industrial'. This applies also to areas that were not industrial in August 2009 but became so under Petroamazonas, with the exception of rights of way which are cleared strips of land and which the Tribunal, in part based on its observations during the Site Visit, classifies as agricultural land.607 The question also arises as to the classification of platforms no longer in operation in August 2009 and which should have been revegetated by the Consortium in accordance with RAOHE.608 Because they were no longer in use in August 2009 (and were not put into use by Petroamazonas), the Tribunal is of the view that these platforms do not qualify as industrial areas and will therefore classify them under the remaining land use categories on a case by case basis according to the circumstances of the relevant site.
335.
The classification of land is more complex with respect to off platform areas. Ecuador argues that all these areas must be classified as sensitive ecosystem. Burlington concedes that sensitive ecosystem criteria apply to a limited number of designated protected areas in Block 7 (Payamino 1/CPF, Payamino 2/8 and Payamino 18) located in the Protected Forests Scientific Station Napo Payamino Cuerpos 1 and 2 Cerro Sumaco, Cuenca Alta of the River Suno Ampliación,609 which form part of the National System of Protected Areas, Protective Forests, and Forestry Assets of the State (SNAP).610 Burlington also agrees that Nemoca and Waponi-Ocatoe in Block 21 fall within a designated area qualifying as sensitive ecosystem.611 At these locations, GSI applied sensitive ecosystem criteria although the actual land use in Payamino 2/8 and Payamino 1/CPF was arguably different. The Tribunal also notes that ConocoPhillips' own environmental assessment of the Blocks in 2006 recognized that the Puerto Napo – Yuralpa pipeline in Block 21 traverses three privately managed protected areas, namely the Venecia Protected Forest, the Jatun Sacha Foundation and the Selve Viva Foundation.612 Beyond this, Burlington takes the position that the subsequent use of all off platform, non-industrial land should be considered agricultural.
336.

RAOHE Table 6 provides that the sensitive ecosystem values apply to "sensitive ecosystems such as Natural Areas of the National Patrimony or others identified in the corresponding environmental study" ("valores límites permisibles para la protección de ecosistemas sensibles tales como Patrimonio Nacional de Areas Naturales y otros identificados en el correspondientie Estudio Ambiental") (emphasis added by the Tribunal). The use of the words "such as" – "tales como" – shows that the reference to formally designated protected areas is illustrative as opposed to exhaustive.613 Hence, the Tribunal cannot accept Burlington's restrictive understanding of sensitive ecosystems. Neither does it agree with Ecuador that all areas within the Sumaco or Yasuní biosphere reserves automatically qualify as sensitive ecosystems. Indeed, the biosphere reserves under the UNESCO Man and the Biosphere ("MAB") Program are not in fact strictly protected areas in which no human economic activity is tolerated.614

337.
UNESCO launched the MAB program in 1971 and the biosphere reserve network in 1976. In 1995, the Seville Strategy was adopted as well as the Statutory Framework of the World Network of Biosphere Reserves, which sets out the key components of MAB reserves, such as designation, support and promotion of biosphere reserves.615 MAB biospheres should "strive to be sites of excellence to explore and demonstrate approaches to conservation and sustainable development on a regional scale". Such sites must fulfill three functions: (i) the conservation of landscapes, ecosystems, species and genetic diversity; (ii) the achievement of human and economic development from a sociocultural and ecological point of view; and (iii) the support of local, regional, national and global projects of education, training, research and monitoring of the environment and sustainable development.616 Biosphere reserves comprise a "mosaic of ecological systems" that are of significance for "biological diversity conservation".617
338.
Land use in these biosphere reserves depends on zoning, each reserve being divided into three zones: (i) a nature reserve or national park as the core area where no economic activity is tolerated (except for traditional uses), (ii) the buffer zone where only activities compatible with the conservation of the ecosystems is allowed, and (iii) the transition zone, which is the external ring of the biosphere and where sustainable development and urbanization can take place.618 The Yasuní reserve was designated by Ecuador as a MAB biosphere in 1989, and the Sumaco reserve in 2000.619
339.
On this basis, the Tribunal notes that neither of the Blocks intersect with the core area or buffer zone of these reserves; the sites under review only overlap with the transition zones.620 Further, the Yasuní and Sumaco biosphere reserves under the MAB program have not been designated as part of the protected areas known as SNAP, with the exception of their core areas (i.e., the Parque Nacional Sumaco and the Parque Nacional Yasuní).621 The partial overlap of the Blocks with the Sumaco and Yasuní reserves does not per se impose a classification as sensitive ecosystems. Finally, the Tribunal notes that the Huaorani reserve partly overlaps with Block 21, but not with the Yuralpa field where most of the environmental harm in that block is alleged to have occurred. It is therefore not material to determine the land use in that field.622
340.
This said, the Tribunal is mindful that the Amazon region in Ecuador is considered to be one of the most diverse and sensitive ecological areas in the world,623 and that it finds special mention in the 2008 Constitution as was seen above.624 The Tribunal also recalls IEMS's evidence that secondary forests surrounding platforms can function as biological corridors connecting preserved areas and modified primary forests, thus fostering biodiversity conservation and protection.625 However, IEMS's biological resource assessment in no way justifies qualifying the entirety of the Blocks as sensitive ecosystems for present purposes.626 Indeed, in light of the existing human dwellings in proximity of many platforms in both Blocks, especially in the northern part of Block 7, the Tribunal considers IEMS's opinion concerning the "regeneration capacity of the rainforest"627 too general and vague to support the adoption of a single, all-encompadsing standard across the Blocks and thus disregard the actual situation on the ground. In line with the Tribunal's earlier conclusion, subsequent land use must be assessed case by case, by reference to the particular circumstances of each site.
341.
The Site Visit provided the Tribunal an important, first-hand view of the actual use of land surrounding the platforms. By way of general observations, the Tribunal noted that the vegetation cover was much denser in Block 21 than in Block 7, and that the CPUF is located in the most developed part of the Blocks. This does not mean that all disputed areas in Block 21 are sensitive ecosystems, nor that all land surrounding platforms in the CPUF is agricultural. Again, a site-specific assessment is needed. Yet, the general characterization just mentioned may inform such assessments.
342.
On the basis of the foregoing considerations, the Tribunal has engaged in a case-bycase assessment of each site to determine subsequent land use. Its assessment is primarily based on evidence gathered during the Site Visit, the explanations provided by the experts,628 and the satellite imagery and aerial photogaphs of the sites closest in time to the date of expropriation. Where the record contains no images or photographs from that period, the Tribunal decided on the basis of the best available evidence.
343.
Consequently, the Tribunal developed the following guidelines to assess land use at any given site:

i. Platforms in operation on or after the date of expropriation have been classified as industrial, as explained above.

ii. Formally designated protected forests are classified as sensitive ecosystems (except for platforms), regardless of other current uses.629

iii. Where platforms are completely surrounded by primary or secondary forest, the Tribunal classified the forested areas as sensitive ecosystem.

iv. Where areas surrounding the platform are largely cleared, the Tribunal classified the land use of those areas as agricultural.

v. Where multiple land uses may apply to a single site,630 the Tribunal looked more specifically at the sampling locations to determine the land use in that specific area.631

vi. Where land use appeared mainly agricultural, the Tribunal nonetheless looked more closely at particular features, such as swamps, creeks and rivers, that may justify stricter criteria.

vii. Where human dwellings, including school or other community buildings are adjacent to a water system and in close proximity to a contaminated site, the Tribunal applied sensitive ecosystem criteria for RAOHE parameters and residential criteria under TULAS Table 3 (for parameters not contemplated in RAOHE Table 6).632

viii. For platforms that have been abandoned, the Tribunal did not apply industrial criteria to the platform, but the relevant land use of the surrounding area.633

ix. In cases of doubt, the Tribunal adopted the most protective standard in conformity with the principles of precaution and in dubio pro natura.

344.
In applying these guidelines, the Tribunal also considered the administrative practice of land classification, as it arises from approved remediation reports. Most remediation reports address on platform exceedances and classify the land as industrial. As regards off platform exceedances, the few remediation reports referred to by Burlington mostly refer to agricultural use.634
345.
In addition to remediation reports, the Tribunal also reviewed the ex-post studies prepared by Entrix (at the request of Petroamazonas) as well as the post-expropriation practice of Petroamazonas. The ex-post reports very generally apply agricultural criteria to the Coca, Payamino, Lobo and Mono fields,635 which appears too wide an approach for the Tribunal to follow without proceeding to its own site-specific assessment. As for Petroamazonas, it applied agricultural criteria to Coca 6 and Yuralpa Pad E. In this context, the Tribunal also observes that IEMS indicated that agriculture partly surrounds the platforms at Coca 6, Coca 8, Lobo 3, Lobo 1, Oso 9, Mono CPF and Payamino CPF.636
346.
This being so, the Tribunal stresses that the administrative practice of Ecuadorian authorities was certainly helpful. At the same time, reliance on such practice cannot excuse the Tribunal from the task of making its own assessment on the basis of the entire record to ensure that it applies the regulatory criteria correctly.
347.
The result of the Tribunal's land use assessment following the parameters just set out is described in the site review (Section 4.6).

4.4 Guidelines for calculating impacted areas and volumes of impacted soils

348.
The Parties advocated divergent approaches for calculating the extent of soil contamination in the Blocks. Whereas Ecuador argued that, due to the impossibility of sampling the entire Blocks, the extent of soil contamination can only be established by extensive modelling (4.4.1), Burlington favored delineation through a process described as "hand-contouring" (4.4.2). After reviewing the Parties' methodologies, the Tribunal will set out its own approach (4.4.3).

4.4.1 IEMS's methodology and Ecuador's position

349.
For Ecuador, GSI's approaches "to sampling, data collection and analysis do not provide a sound or reasonable basis" to quantify the extent of harm in the Blocks, since they are "inherently subjective and arbitrary", and thus should be disregarded.637 By contrast to GSI's "contrived and unreliable sui generis delineation theory", Ecuador submits that IEMS reliably and conservatively quantified the extent of harm through modelling and by using the "widely accepted" Inverse Distance Weighted ("IDW") interpolation technique.638
350.
Ecuador explains that IEMS used the computer software ArcGIS to calculate the total volume of contaminated soil.639 Using ArcGIS, IEMS developed a Geographical Information System for Quantifying Environmental Impact Soil ("SIGAAS"), which mapped every production facility within Block 7/CPUF and Block 21.640 For the areas of contaminated soil, IEMS mapped each sample it collected, delineated a 10-meter radius around it (called a "buffer"), and defined the boundaries of the contaminated area as the outer limits of all buffers (a group of buffers is referred to as "cluster") marked by a rectangle called "bounds" (see figure below).641

IEMS's schematic representation of macro to micro scales642

351.
To calculate the volumes of contaminated soil, IEMS modeled six different layers of depths (from 0 to 1 meter, 1-2 meters, 2-3 meters, 3-4 meters, 4-5 meters, and more than 5 meters deep).643 For each layer, IEMS measured the level of pollution for each contaminant. IEMS explained that its analytic model required at least three polluted samples per layer and that layers with less than 3 polluted samples were disregarded.644 Once the bounds and the layers were established, IEMS (i) divided the area within the bounds into "cells" of one square meter; (ii) categorized each cell where samples were taken as non-contaminated, contaminated in excess of background values but below regulatory levels, or contaminated in excess of regulatory thresholds; and (iii) categorized each surrounding cell applying the IDW method of interpolation with ArcGIS.645 With that method, the value attributed to each surrounding cell is a function of the value of the sample data cells and the distance between the sample cells and the surrounding cells. In other words, the closer a sample data cell is to a surrounding cell, the more impact it has on the surrounding cell's value. In consonance with IEMS's conservative model, argues Ecuador, total volumes of contaminated soil were calculated only within the bounds.646
352.
The IDW interpolation method was necessary, according to Ecuador, due to the sheer scale of the contaminated areas and because hand contouring would require disproportionate resources.647 IDW was chosen among various other models on the basis of the characteristics on the ground, and the model was duly adjusted to account for the particular characteristics of each site. In addition, Ecuador argues that literature confirms that IDW is a recognized and well-understood form of linear interpolation. Contrary to the assertion of Burlington's expert Professor Rouhani, the "deterministic linear interpolator IDW is not conditional on mathematical semivariance (‘ spatial correlation ')", since semivariance is only required with the geostatistical linear interpolation method known as "kriging".648 In this context, Ecuador explains that IDW is a deterministic, not a geostatistical interpolator. The corollary of the foregoing is that IDW assumes spatial correlation. Accordingly, spatial correlation was not tested by IEMS because IDW is not mathematically dependent on semivariance.649
353.
Ecuador further contends that Professor Rouhani's "desktop review" employed a wholly inapplicable analysis.650 For instance, variograms are not used to evaluate IDW on a post hoc basis. Indeed, Professor Rouhani conceded that his approach was "sub-optimal".651 Moreover, IEMS confirmed that the contamination in the Blocks was in any event sufficiently correlated at each particular site, and Professor Rouhani conceded that practical experience trumps statistical analysis.652 Finally, Burlington's reference to literature suggesting that IDW should not be used for decision-making is misplaced, since (i) the literature relied upon by Burlington is published by the software developer of ArcGIS Geostatistical Analyst, (ii) that manual does not exclude the use of IDW but merely cautions users about the reliability of the IDW interpolations, and (iii) IEMS applied IDW conservatively and with all due caution, for instance, by applying control systems and prediction capacity.653
354.
With regard to the last point, Ecuador insists that IEMS's comprehensive model relies on robust sampling taking into account topography and site features. Furthermore, modeling was undertaken within small-bounded and well-sampled areas. These bounds circumscribed areas of known contamination. Also, contrary to Professor Rouhani's assertion, IEMS included clean samples in its modeling. In addition, IDW was only used in those locations with a minimum of three nearby data points, and IEMS weighted the IDW interpolations conservatively by using a power of 3, instead of the standard power of 2, the consequence being that the importance given to each sample decreased rapidly.654 Finally, IEMS did not interpolate between points separated by more than 300 meters.655
355.
Ecuador further asserts that Burlington errs when stating that the IDW method is "conceptually flawed" or "scientifically inappropriate".656 In fact, Professor Rouhani confirmed at the Hearing that IEMS's IDW approach is a "systematic space-wide analysis".657 Indeed, the experts on both sides employed some form of linear interpolation. Yet, as Professor Rouhani noted at the Hearing, GSI used the "rudimentary" hand contouring technique, while IEMS used the more powerful and conservative IDW.658 Ecuador further explains that GSI also used mathematical forms to interpolate the extent of negative impacts. In fact, "all experts agree that a mathematical approach is the scientifically correct way to estimate environmental contamination".659 Moreover, GSI's criticism of using computer software is unfounded, since GSI also used a computer for its "hand-drawn" contour lines.660 Finally, GSI's claim that IEMS extrapolated from a given point outwards is plainly wrong, according to Ecuador. Interpolation was done within well-defined bounds, circumscribing "heavily sampled areas which were known to contain contamination".661
356.
Regarding GSI's quantification of contamination, Ecuador argues that GSI's approach is unreliable for various reasons: (i) GSI did not undertake a site-wide analysis; (ii) GSI's self-made and arbitrary topographical exclusion rules are scientifically flawed; (iii) GSI failed to "step out" to take additional samples until clean soil was reached; (iv) GSI's manual delineation approach is arbitrary and (v) produces massive errors; and (vi) GSI's contingency factor cannot iron out the wide-scale methodological shortcomings of GSI's delineation approach.662
357.
First, Ecuador stresses that GSI limited itself to IEMS's sampling without itself engaging in a site-wide analysis, thus undermining its claim of a comprehensive assessment.663 For Ecuador, proper delineation would require an assessment on a meter-by-meter grid basis. Since it is Burlington's burden to prove the absence of significant negative impacts, says Ecuador, one would have expected GSI to undertake a systematic assessment: "[T]he easiest way for GSI to disprove IEMS' quantification of the negative impact was to sample, meter-by-meter, across the sites and perform a true delineation".664 Critically, Burlington not only decided not to undertake site-wide sampling, it also decided not to test seriously, "whether or not IEMS' sampling program was representative of contamination across the Blocks".665 Ecuador further argues that Burlington's and GSI's punctual criticisms fail to address the more general question "whether, as a whole, IEMS' sample design produced a representative sample set upon which to found their interpolations".666
358.
Second, GSI's "sui generis delineation" rests on a set of misplaced arbitrary and selfmade topographical exclusion rules,667 which GSI nonetheless does not follow, according to Ecuador. As a starting point, Ecuador states that it bears recalling that both RPS and IEMS confirmed that once contamination reaches the subsoil "any surface topographic features will no longer have any impact on how the contamination migrates through the subsoil".668 Importantly, GSI applied its own topographical rules in an arbitrary manner. In doing so, Professor Rouhani admitted that GSI failed to take further delineation samples around exceedance points, such as, for instance, in Coca 8, where GSI did not take any delineation samples to the northeast, east, southeast, south or southwest of exceedance point # 5 (CO08-4M-03).669 Ecuador highlighted that Professor Rouhani accepted that "it would have been much easier to solve this problem by just having one more delineation sample".670
359.
Third, Ecuador criticizes that GSI did not "step outwards" to take additional samples until clean soil was reached.671 Ecuador explained that "[i]n every case GSI's delineations are founded on only two samples – one being an ‘exceedance’ and one being a ‘ non-exceedance '".672 Yet, as IEMS stated, the most obvious step to achieve a true delineation would have been to continue sampling until clean soil was uncovered. To compound matters, Ecuador stresses that GSI willfully put dirty samples back into the ground, discarding them totally.673 This happened not only in a few instances as claimed by GSI, but in at least 6 sites (of the 17 where GSI took delineation samples).674 When asked to comment, Ecuador states, GSI "cynically" explained: "And the reason for that is because we stepped out because we're looking for the clean margin".675 For Ecuador, "[t]rue delineation required continued sampling in all directions until all contamination was captured".676
360.
Fourth, Ecuador states that GSI's delineation approach is in any event wholly subjective and arbitrary.677 GSI cites no literature in support of its sui generis approach. The hand drawn isoconcentration contour lines are a "very rudimentary form" of interpolation, as Professor Rouhani conceded.678 In addition, this technique provides no certainty as it is not possible to assess the errors associated with the tick marks or the hand drawn contour lines.
361.
Fifth, GSI's approach produced massive errors according to Ecuador, notably because of its "sample dilution" (both through the use of composite samples and by putting dirty samples back into the ground).679 Finally, GSI's "all purpose 20% ‘ contingency '" cannot account for the errors in GSI's delineations and is a "completely inadequate margin for the massive limitations to, and errors inherent in, GSI's delineations".680 Ecuador nonetheless points out that both Parties' experts agree that a contingency factor ranging from 20%-30% should be included.681

4.4.2 GSI's methodology and Burlington's position

362.
For Burlington, Ecuador's reliance on IEMS's computer model to estimate volumes of allegedly contaminated soil is "[o]ne of the most consequential flaws in Ecuador's environmental claim".682 The result is to artificially inflate the estimates of soil contamination by relying on modelling instead of delineating actual contamination in the fields.683 Burlington also criticizes IEMS for committing serious errors in constructing its model,684 and argues that had IEMS engaged in validation testing, it would have realized that the model is "highly unreliable".685
363.
According to Burlington, the modelling methodology is incorrect and unreliable,686 in particular because specific circumstances which may justify using computer predictions as a substitute for real data are not present here.687 Relying on Professor Rouhani, Burlington argues that the only reliable method for the Blocks is "site-specific field delineation, relying on real data and observations".688 Therefore, GSI employed the methodology known as "hand-contouring" which involves "sampling known or suspected areas of contamination and then ‘delineating' these areas by taking additional samples until clean soil is reached".689 These clean samples together with topographical features then provide the "contours of the contaminated area".690
364.
With respect to Ecuador's modelling exercise, Burlington argues that IEMS used ArcGIS to model the extent of contamination, although that software is "not designed to extrapolate outwards from a particular data point" but serves only to "interpolate between plotted data values".691 As Professor Rouhani indicated, IDW is "simply not an appropriate method for modeling soil contamination in the Blocks, as it cannot make predictions within any acceptable range of reliability".692
365.
Burlington explains that IEMS should have interpolated whereby "samples are taken moving outward from a contaminated point until compliant samples are found".693 Instead, IEMS used a "fill" function in ArcGIS to connect "non-clean" samples, thus ignoring topography (as the examples of Punino and Coca 1 demonstrate)694 and other physical features (such as low-lying drainages, roads or hills),695 resulting in a modelling exercise where the "vast majority of areas" claimed to be contaminated are in fact areas where IEMS's own samples comply with Ecuadorian regulatory criteria "or areas where no samples were taken at all".696 In addition, IEMS's models extend "far beyond the vicinity of any of IEMS's samples" thus imputing alleged contamination to areas where no sampling was undertaken.697 Burlington also criticizes IEMS's "inconsistent treatment of samples taken from pits", in particular by using those samples to interpolate in areas extending beyond the boundaries of the pits,698 as well as IEMS's substitution of "near-detection-limit values" for non-detect samples.699
366.
According to Burlington, IEMS provided no explanation for choosing the IDW method over other methods.700 IDW only calculates an "average, weighted by distance", based on the assumption that all points are "correlated", irrespective of "how far away one travels from those samples, how unrelated the area being interpolated is to the one from which the samples were taken, or whether there are topographic or other features that would naturally isolate one area from the other".701 In other words, Burlington submits that IEMS's model is "blind to the absence of correlation between the existing data and the point that it has been instructed to predict".702
367.
Ecuador's argument that the assessment of actual soil conditions would have required 1.12 million additional samples is unavailing according to Burlington, especially since (i) sampling was not required in a majority of the Blocks where IEMS did not even identify any exceedances or recognized environmental conditions ("RECs"), (ii) 88% of IEMS's sampling locations comply with regulatory criteria, and (iii) IEMS failed to consider all available data, ignoring "more than 500 soil samples, including all of the data collected by GSI (390 soil samples) as well as IEMS's own samples from its most recent field campaign (157 soil samples)".703 Burlington argues that the difference between IEMS's and GSI's approaches is best illustrated when comparing the results for barium contamination in Mono CPF, where IEMS's model predicts contamination "that has no relation to the topography of the site or the existence of clean samples".704 The following image is used by Burlington to show how GSI used delineation samples and topography to hand-contour the contamination:705
368.
By contrast, Burlington explains that the red area in IEMS's model below706 shows Ecuador's regulatory case and demonstrates how IEMS predicted contamination that has no relation to topography or clean samples:707
369.
According to Burlington, employing appropriate delineation methods would reduce IEMS's estimations "to virtually nothing".708 Burlington further explained that IEMS compounded matters by engaging in a series of "extraordinary errors and omissions" in constructing and applying its method, thus producing distorted results.709 In particular, Burlington highlights the fact that IEMS converted non-detect samples into positive values (so-called false positives) with the result that the ArcGIS software erroneously used "positive data values to generate maps of allegedly "contaminated" areas".710 This resulted, Burlington says, in a manipulation of 79% of IEMS's data points.711 Ecuador's explanation regarding the power function is to no avail since the power function simply instructs the IDW model "how heavily to weight one sample compared to another based on their proximity", but will "always calculate a value for every cell in the modeled area that is between the minimum and maximum sample value, no matter what power is used".712 Instead of using a fixed search radius to avoid overreaching, IEMS used a variable search radius resulting in the model's IDW calculations continuing to run "until it reaches the model's bounds",713 as GSI demonstrated when modifying the bounds in the model.714
370.
Finally, Burlington argues that IEMS's failure to engage in meaningful validation testing proves the unreliability of its model, Professor Rouhani having confirmed that these kinds of models require validation testing.715 As a basic reality check, IEMS should at the very least have compared its model to GSI's 390 soil samples or its own additional 157 samples. Instead, IEMS only plotted its new sample locations on its site maps resulting in clean samples appearing in allegedly contaminated areas such as Coca 8, Coca 9, Coca 18-19, Payamino 4, Mono CPF, Mono 10-12 and Cóndor Norte.716 In this context, IEMS's cross-validation at Coca 8 taking "a single analyte in a single layer in a single site" is unreliable, since cross-validation "only works in points for which the model already contains data" and not at locations distant from known samples.717 In any event, IEMS's cross-validation shows that "the average difference between known values and predicted values" is ten times greater than the acceptable error level calculated by Professor Rouhani.718
371.
Moreover, argues Burlington, IEMS should have engaged in further tests by measuring the correlation of the data and mapping the reliability of the model's predictions through a prediction standard error ("PSE") map.719 With respect to data correlation, Professor Rouhani's variograms showed that "more than half of the model layers" are not correlated with the result that linear interpolation is unreliable.720 IEMS should also have tested the reliability of its interpolation by generating a PSE map showing "zones of reliability" and indicating the range of error within each zone.721 For Burlington, Professor Rouhani's testing of IEMS's model through cross-validation and the use of PSE maps confirmed that the model is not reliable.722

4.4.3 Discussion

372.
After closely reviewing the evidence and the Parties' arguments, the Tribunal comes to the conclusion that, as between the methods presented to it, the delineation of discrete pockets of contamination by relying on actual sampling is best suited to estimate the impacted areas and to calculate the volumes of contaminated soil requiring remediation. It therefore discards IEMS's modelling, which it considers unreliable, or at least less reliable, in the circumstances. For various reasons, however, the Tribunal does not rely completely on the results of GSI's delineation exercise and has therefore independently calculated impacted areas and volumes of soil. The Tribunal first provides its reasons for discarding IEMS's modelling ((a) below), then addresses the problems with certain of GSI's delineations ((b) below), before explaining its own approach ((c) below).

a. IEMS's modelling

373.
The Tribunal discarded IEMS's modelling for various reasons. In far too many instances there appears to be no reasonable relationship between the model's predictions and the actual data, thus excessively inflating the size of contaminated areas. Among other things, the model does not properly consider clean samples, extends to areas where no sampling was undertaken, and insufficiently takes account of topography and other physical features. In addition, it appears that IEMS did not test the reliability of its conclusions although the ArcGIS user instructions require cross-validation.723 These failings of IEMS's modelling are discussed below.
374.
First, the Hearing and the Site Visit showed that IEMS's model extends to demonstrably uncontaminated areas. It is common ground among the experts that soil sampling data determine the existence of contamination.724 Burlington rightly pointed to Gacela 4 and Jaguar 3 as examples where clean samples collected by IEMS during its fourth sampling campaign are located in areas that were modelled as contaminated after the first three sampling campaigns, thus disproving the model's predictions.725 IEMS's remodelling of Gacela 4 after the fourth campaign still predicted contamination in areas where new clean samples had been found.726
375.
Similarly, during the Site Visit, the Tribunal observed the absence of correlation of the data included in IEMS's models, in particular at Gacela 2 and Mono CPF. The image of Gacela 2 below shows how IEMS's model predicts contamination extending up and across most of the platform to the north (and beyond) to areas where not a single soil sample had been collected.727 In fact, only one sample taken on the platform – indicated by the Tribunal with an arrow – appears to drive the model's predictions (all other samples are to the south or southwest of the platform and lie in lower altitudes than the platform):
376.
In a comparable manner, the image of Mono CPF below depicts low-lying off platform samples to the east of the platform in the area designated by GSI as Area 2M – again indicated by the Tribunal with an arrow – projecting contamination up and across more than a third of the platform although not a single sample was taken on the platform.728
377.
The pictures above also illustrate that the bounds on IEMS's modelling are determined by clusters of samples and/or discrete samples without considering that large areas inside the bounds have not been sampled at all. Much of IEMS's modelling is thus rather speculative. It also involves a degree of arbitrariness as the models commonly predict that contamination abruptly stops at straight lines.
378.
As was already mentioned, IEMS's model also fails to sufficiently factor in topography and other physical features of the sites, which is exemplified in the following figure showing Jaguar 7-8. IEMS's model predicts, without supporting sampling data, that contamination spreads east across a creek for approximately 10 meters and uphill for several meters – indicated by the Tribunal with a yellow arrow –,729 which is contrary to IEMS's own explanation that contamination could conceivably spread uphill some 50 centimetres when the water level rises during the rainy season.730 In addition, IEMS's model ignores GSI's samples on the other side of the creek which show no exceedances.731
379.
IEMS's modelling at Yuralpa Pad A provides another example of the lack of relation to actual site data. On the basis of a single sample taken next to the oil water separator in the southern corner of the platform (i.e., in the area designated by GSI as Area 2M) – indicated by the Tribunal with an arrow – the model predicts that contamination not only spreads downhill to the southeast, but also towards the southwest without any substantiated explanation for this pathway.732 Indeed, there are no IEMS samples showing contamination in that particular area. To the contrary, GSI samples show that the contamination flowed some distance in the southeast direction,733 but that in fact it did not extend to the southwest of the oil water separator.734 This is illustrated by IEMS as follows:735
380.
IEMS's predictions for contamination at Coca 15 also demonstrate the flaws of its model. At that site, a single sample with a TPH value purportedly exceeding permissible limits – signalled by the Tribunal with an arrow – drives the entire model and predicts contamination in excess of regulatory limits across approximately one-fourth of the platform.736 This is illustrated by IEMS as follows:737
381.
During the Site Visit of Coca 15, the Parties' experts disagreed whether the model extended north of the platform where the terrain shows a steep slope upwards.738 While IEMS opined that its model perfectly aligned with topography,739 the Tribunal observes that the red area extends approximately 5 meters beyond the platform (the scale of the map is 1:1000) and thus predicts contamination going up the slope.740
382.
In addition, the Tribunal notes that IEMS did not perform any analysis to determine whether its data was spatially correlated.741 It also conceded at the Hearing that it did not conduct a quantitative review to verify the reliability of its model.742 IEMS actually stated that it could not assess the level of precision or imprecision of its model for lack of quantitative analysis743 and admitted being unable to opine whether a 50 percent margin of error would be acceptable.744
383.
Finally, the Tribunal observes that Ecuador's other expert, RPS, did not peer review IEMS's model predictions or otherwise test the validity of IEMS's work.745 More importantly, it appears to the Tribunal that, although RPS criticised GSI's sampling methodology (such as using composite samples or sampling at different depths than IEMS), it did not express any principled criticism of GSI's contour delineation.746
384.
For these reasons, the Tribunal will not rely on IEMS's modelling. It does not, however, discard the extensive sampling data collected by IEMS. The Tribunal now turns to GSI's delineation approach.

b. GSI's delineation

385.
Before addressing in more detail GSI's approach to delineation (iii), two preliminary issues need to be addressed, namely GSI's indicator parameter theory (i) and GSI's recourse to composite sampling for delineation purposes (ii).

(i) Indicator parameters

386.
GSI only delineated around samples showing TPH or barium exceedances and discarded other heavy metal exceedances as "background metals issues", on the ground that such exceedances are unrelated to oilfield operations.747 Ecuador strongly objected to this practice,748 in particular by stressing that Mr. Saltos confirmed that numerous chemicals are used during various phases of the oilfield operations and that Burlington failed to provide a list of the chemicals the Consortium used.749 The Tribunal rejects Burlington's indicator parameter theory for the simple reason that under RAOHE and TULAS any exceedances of applicable remediation criteria (adjusted as required to account for higher natural concentrations) in the area of the oilfield operations, warrant remediation by the Consortium, unless Burlington can rely on one of the three exculpatory causes referred to in paragraph 238(iii) above.

(ii) Composite sampling

387.
Another issue with GSI's approach lies in the use of composite delineation samples. The Tribunal notes that GSI collected discrete risk characterization and soil confirmation samples and that GSI initially started out by collecting discrete delineation samples, only later collecting composite samples since IEMS was projecting contamination in 1-meter layers.750
388.
In addition to various other criticisms raised by Ecuador with respect to composite samples,751 Ecuador notes the risks of dilution and volatilization.752 Ecuador stressed that GSI conceded that 1-meter composite samples are biased towards the lower value since they only show an average and not the highest concentration.753 For these reasons, IEMS did not rely on GSI's delineation samples to compile its models.754
389.
The Tribunal need not enter with great detail into the Parties' discussion on composite samples.755 It generally agrees that discrete samples allow for greater accuracy in determining the highest concentration of a contaminant in the soil. There is thus a certain risk of dilution inherent in composite sampling which mixes larger intervals of soil instead of testing thinner intervals. On the other hand, IEMS's 30 centimeter interval samples also involve a degree of composition, merely on a smaller portion. Thus, the Tribunal notes that a certain level of vertical compositing is inevitable for any soil sample.
390.
More importantly, the Tribunal notes that Ecuadorian regulations do not mandate a particular sampling method, and, in particular, do not impose discrete sampling. Rather, RAOHE provides that composite sampling is acceptable.756 RAOHE Annex 5 determines various analytical methods and provides that the method for soil sampling is to collect a "composite and representative sample" ("Muestra compuesta y representativa").757
391.
Finally, the Tribunal accepts GSI's composite samples in light of the fact that IEMS itself calculated the allegedly impacted areas on the basis of 1-meter layers, the values of which were determined based on samples that ranged from only 10 to 30 centimeters in length.758
392.
As to the risk of volatilization referred to above, the Tribunal notes that IEMS did not test its samples for Volatile Aromatic Hydrocarbons,759 and that RPS confirmed that GSI screened its samples with a photo ionization detector ("PID") before sending them to the laboratory.760 In any event, the Tribunal is satisfied with GSI's explanation that by the time the sampling took place, these compounds had volatilized to a large extent, since the alleged contamination dated back several years and a key characteristic of these compounds is their high mobility.761

(iii) Linear interpolation

393.
With these considerations in mind, the Tribunal now turns to GSI's approach to delineation through linear interpolation. While the Tribunal generally accepts Burlington's delineation methodology, it cannot simply rely on GSI's calculations for several reasons. First, GSI only investigated 22 sites to confirm and delineate the extent of contamination at 74 facilities (of which GSI inspected 58) based on the erroneous assumption that most off platform areas must be evaluated against agricultural land use criteria.762 Consequently, for the contaminated sites for which GSI provided no delineation, the Tribunal will have to undertake its own assessment.
394.
Second, as discussed above (see paragraphs 318-322) for the 17 sites where GSI identified exceedances, GSI's adjusted remediation criteria for certain metals (i.e., barium, cadmium and nickel) have now been found to be incorrect. As GSI conceded under cross-examination, it would have had to delineate around more IEMS samples if the regulatory criteria had been lower than those which it applied.763 Moreover, GSI did not propose alternative delineations considering different regulatory criteria. In other words, GSI's delineation data is insufficient for purposes of the Tribunal's establishement of soil contamination.
395.
Third, various aspects of GSI's sampling procedure have led the Tribunal to adjust GSI's estimates. In this respect, it may be useful to start by outlining GSI's approach of physical delineation and hand-contouring. GSI first reviewed the laboratory results of soil samples to identify and locate impacted soils.764 It then surrounded these exceedance points with additional soil samples "until a clean perimeter is defined as needed to delineate the extent of the impacted soils".765 This "clean perimeter" is also referred to as the "clean margin".766 It then drew hand contour lines "by interpolating among the sample points" (this process is described below), to finally delineate the impacted soil area and its total volume.767 The following image shows the four steps employed by GSI:768
396.
The contour lines above (so-called isoconcentration lines) circumscribe the area that fails to comply with the applicable regulatory criteria for one of the indicator parameters tested by GSI, namely TPH or barium, for a given layer of soil depth.769 These contour lines were determined through a combination of one or more of (i) linear interpolation, (ii) physical site features and boundaries (such as edges of pits, platforms, oil-water separators, roads), and (iii) topographic features (such as slopes, streams and catchment areas) and field observations.770
397.
GSI relied on linear interpolation using an equation "to mathematically locate the isoconcentration contour line between a sampling location that exceeded the specified criterion to a corresponding sampling location that was below the criterion".771 The interpolated points were indicated on a map using "tick marks" and then connected to create an isoconcentration contour line at the specified regulatory criterion for a certain parameter.772 GSI considered site features "particularly in situations where linear interpolation was not possible".773 It further explained that "[f]or selected portions of potentially impacted areas at some sites", it determined a "set distance beyond points of exceedance" based on local topography and field observations where "linear interpolation was not applicable and/or site features were not available to delineate potentially impacted areas".774 At the Hearing, GSI acknowledged that the methodology involved a degree of uncertainty in some areas, but that by using professional judgment, providing for additional sampling, and allowing for a contingency factor – all of which it did – that uncertainty could be significantly reduced.775
398.
The Tribunal's concerns with GSI's approach can be well illustrated using Coca 8 as an example. Indeed, this could be called the "poster child" of Ecuador's critique of GSI's methodology in view of the prominence that it was given at the Hearing. The figure below shows GSI's delineation for barium in the 1-2 meter layer in the area south of the mud pits (located south of the Coca 8 platform).776
399.
As a first concern, although it does not share Ecuador's criticism that GSI failed to engage in a site-wide assessment,777 especially since the burden to make a showing of harm rests on Ecuador, the Tribunal notes that, as this figure graphically depicts, GSI failed at times to comprehensively collect delineation samples around certain known exceedance points. In the example above, this is the case for instance at: (i) sample location CO08-4M-07 in Area 4M, where no delineation samples were taken to the northwest, west, southwest or south; (ii) sample location CO08-4M-03, where no delineation samples were taken to the northeast, east, southeast, south and southwest; (iii) sample location CO08-4M-04, where no sample was taken to the south; and (iv) sample location CO08-4M-06, where no sample was taken to the west.
400.
The Tribunal is aware of the fact that the barium value for sample location CO08-4M-07 (770 mg/kg) is close to permissible limit for agricultural use (i.e., 750 mg/kg). The Tribunal nonetheless considers that, as a general matter, and in the light of its own description of its methodology, GSI should have taken additional samples, in particular because GSI purportedly "stepped out" of impacted areas to find clean soil.778 This is compounded by the fact that here the applicable criterion is the adjusted sensitive ecosystem limit of 639 mg/kg (see below paragraph 472).
401.
The Tribunal further notes GSI's explanation for sample location CO08-4M-03, where the "potentially impacted area" was extended by 11 meters southeast since that sample location "did not have a bounding delineation sampling location or site feature to the east".779 GSI explained that the 11 meter perimeter was "a conservative estimate based on the relatively flat localized terrain; therefore, the potentially impacted area was assumed to extend radially an equivalent distance from the GSI exceedance point".780 However, Professor Rouhani conceded that "it would have been much easier to solve this problem by just having one more delineation sample".781 Admittedly, he also concurred with GSI's explanation that practical limitations may militate for using "professional judgment" in cases where the results of delineation samples come back from the laboratory with higher than expected values, thus compelling engineers to apply practical approaches instead of relying on purely statistical methods.782
402.
A second and potentially more problematic difficulty is GSI's resort to linear interpolation combined with its practice, during its stepping-out from exceedance points to locate the clean margin, of putting visually "dirty" samples back into the soil without actually testing them in the laboratory. At the Hearing, GSI justified this practice by saying that it was inconsequential to its procedure of "stepping out" to find the "clean margin", that is, a clean sample that could be interpolated with an exceedance point:

"Q. Now, leaving aside the conceptually [sic] caveat you put to your answer, in practice, were there instances in which you made a hole in the ground, you looked at the contents of that soil boring – you took a sample – and in which you did not submit that sample for laboratory testing? Yes or no.

A. I believe so, yes, if I understood correctly your question. So, we saw evidence of contamination, and we did not submit that sample. Did I understand that correctly?

Q. Right.

A. And the reason for that is because we stepped out because we're looking for the clean margin. So, if we're here and we take the sample – the exceedance is that monitor there in the middle, and we come here because we believe we're going to be in a clean zone. We look at the sample and like we're not in a clean zone, so that's what I described in my presentation, I think, that we step out. We go further out so that we make sure we include this point, if appropriate – and that's a key, if appropriate – within that area".783

403.
GSI further indicated at the Hearing that this practice was infrequent, pointing to Coca 8 as one instance.784 But Ecuador reviewed GSI's boring logs and pointed to certain other instances785 at Coca 1,786 Coca 6,787 Jaguar 1,788 Mono CPF789 and Payamino 23.790 Generally speaking, the Tribunal has no issue with and accepts using linear interpolation between an exceedance point and a non-exceedance point in order to determine the impacted area based on regulatory criteria. Stepping out looking for clean samples seems appropriate, especially if the clean sample point is taken as the outer limit of the impacted area. However, it also appears to the Tribunal that, as with any methodology, linear interpolation that is based on incomplete data – in this instance where "dirty" borings were not considered – can distort the results and end up underestimating or otherwise misrepresenting the contaminated area. When linear interpolation between an exceedance point A and a non-exceedance point B is applied, but a "dirty" sample C in between is not considered, there is a risk that the contour line is too close to A or otherwise does not reflect the true contours of the contamined area.
404.
This underestimation became clear during Professor's Rouhani's examination, when he was confronted with the following demonstrative:791
405.
In connection with this demonstrative, the following question was put to Professor Rouhani:

"Q. […] Now, what I sought to do is I sought to put A and C sort of equidistant from the hand contour line, and you will recall that what we discussed earlier was that, if we had an exceedance that was the same level in excess of the regulatory criteria and a non-exceedance that was the same distance below, or the same volume below, we would arrive roughly in the middle. Now, B, B is our hypothetical exceedance. If B has the same exceedance level as A, and we apply the linear interpolation formula that we have been discussing this evening, and C, the nonexceedance, remains the same level, our tick mark is going to end up further to the east; right? It's going to end up outside the hand contour?

A. Yes.

Q. Would you agree with that?

A. Yes".792

406.
In light of the fact that GSI had recourse to this practice in no less than 6 out of 17 sites where GSI delineated impacted areas, the Tribunal cannot but agree with Ecuador that this practice was more systematically used than GSI is willing to concede.793 As discussed above, the Tribunal considers that at certain sites the result of the practice may be to underestimate the contaminated area, in particular where there are no clean samples that define the clean margin in the relevant part of the contour line.
407.
The Tribunal's third concern is that GSI has only delineated areas impacted by its so-called indicator parameters, i.e., barium and TPH.794 As the figure at paragraph 398 above for Coca 8 shows (see also paragraph 471 below), GSI did not delineate areas impacted by what it qualifies as "background metals issues" or areas where metal concentrations exceed regulatory limits but no indicator parameter was found. Sample point CO08-4M-05B demonstrates this flaw with respect to its "background metals" and indicator parameter approach.
408.
GSI indicates that sample CO08-4M-05B in the figure at paragraph 398 above, presents a background metals issue in the layer between 1 to 2 meters. Since the nickel value for that sample at that depth is 53 mg/kg, there is an exceedance at that point both under the agricultural and the sensitive ecosystems limits,795 and GSI should have delineated around that exceedance point, which it has not done. Further, this sample location also serves to disprove GSI's indicator parameter theory. Sample CO08-4M-05B in the layer between 1 to 2 meters has a barium value of 310 mg/kg, which is below regulatory limits. However, the barium value in the layer between 0 and 1 meters at that same location is above regulatory criteria (i.e., 830 mg/kg).796

c. Tribunal's approach to determining the extent of soil contamination

409.
Having determined that it cannot use IEMS's models and that GSI's methodology is affected by certain deficiencies, the Tribunal will now outline its own approach to assess the extent of impacted areas and volumes of contaminated soils. Doing so, the Tribunal relies on all of the elements in the record, including evidence and observations gathered during the Site Visit, which it considers can usefully inform its analysis, in the exercise of its discretion in matters of evidence and quantification of damages.
410.
In application of the regulatory and land use criteria as set forth in paragraphs 324 and 343 above, the Tribunal starts by determining all exceedance points for a particular site as they arise from the sample results provided by the Parties.797 It then assesses whether the contamination can be attributed (at least partly) to the Consortium, it being recalled that (i) fault is not required, (ii) causation is presumed, the Consortium being exonerated if it proves that the harm was caused by force majeure, the victim or a third party (including prior and subsequent operators), and (iii) a finding of fault can lead to the partial attribution of pre-Consortium harm to the Consortium (for instance, with respect to the Consortium's mismanagement of pits built by preceding operators). Where harm is attributed to the Consortium, it then engages in delineating the impacted areas.
411.
For sites delineated by GSI, the Tribunal adjusts as necessary GSI's impacted areas obtained through linear interpolation by extending them to the next "clean" delineation sample (i.e., a sample showing no exceedance under applicable regulatory criteria), where an appropriate "clean" sample exists. Where the Tribunal considers that insufficient delineation samples were taken (such as for example in the area in proximity to Coca 8 sample CO08-4M-03 in the figure reproduced at paragraph 398 above), the Tribunal adopts GSI's set distance (for instance, 8 meters at Payamino 23 and 11 meters at Coca 8), if that distance appears sufficiently protective of the environment. Otherwise, the Tribunal extends that distance as it considers appropriate in the light of all of the circumstances.
412.
For sites not delineated by GSI, the Tribunal seeks to take into consideration all information in the record. Where a number of samples are clustered in a given area, the Tribunal calculates the impacted area on the basis of the scale of the maps, while also considering site features and local topography.
413.
For sites not delineated by GSI but showing discrete exceedance points, the Tribunal also considers site features and local topography, but mostly applies a set radius around the exceedance point to determine a circular impacted area. The length of the radius depends on various factors, such as the type of contaminant, the level of contamination, the location of the exceedance point, the land use, local topography and proximity to human settlements. Thus, the radius is for instance larger for heavy metals than TPH exceedances due to the increased mobility of the former. The radius is also larger for higher than for lower levels of contamination, or if the exceedance is located in a sensitive ecosystem as opposed to on a platform.
414.
In general, the Tribunal applies a 5-meter radius for TPH exceedances (amounting to an impacted area of approximately 80 m2) and an 8-meter radius for heavy metal exceedances, including barium (amounting to an impacted area of approximately 200 m2). It considers that these radiuses are adequately protective of the environment. It, however, also takes into consideration that GSI itself sometimes applied slightly larger distances (see, for instance, above paragraph 401; 11 meters at Coca 8, and 8 meters at Payamino 23). As stated above, the radiuses are thus increased as appropriate at certain sites.
415.
On that basis, the Tribunal establishes the impacted area for each site. To calculate the volume of contaminated soil, it then multiplies the impacted area by the depth of the contamination, rounding-up the depth to the next meter. Doing so, it seeks to take account of differences in contamination per layer where appropriate. For instance, if the impacted area in the layer between 0 and 1 meter is 200 m2 and the impacted area in the layer between 1 and 2 meters is 80 m2, the total volume of contaminated soil would amount to 280 m3. Where feasible and practicable, the Tribunal calculates total volumes of contaminated soil on the basis of layer-by layer values in an effort to avoid overestimating such volumes.798 However, the Tribunal does not always enter into this level of detail. For instance, where the impacted area of the top layer is smaller than the impacted area in a lower layer (and so, perforce, must be removed and remediated along with the larger contaminated area that lies below), or where the difference of impacted areas between layers is not significant, the Tribunal simply multiplies the largest impacted area by the maximum depth of contamination.

4.5 Remediation costs for contaminated soils

4.5.1 Parties' positions

416.
Ecuador calculates the remediation costs by (i) increasing the volume of the compacted clayey soils by a 30% volumetric expansion factor, (ii) applying different costs per cubic meter according to three remediation technologies depending on whether the soil is polluted with TPH only, heavy metals only, or a mixture of both, (iii) adding other costs for studies, remedial design, logistics, transportation, replenishment of excavated areas, safety measures, management and consulting costs, reforestation and environmental monitoring, and (iv) adding a contingency factor.799
417.
Relying on international price comparisons and in particular costs for soil remediation charged by companies in the United States of America, Ecuador envisages ex situ bioremediation for soils mostly contaminated with hydrocarbons at a unit cost of USD 280/m3;800ex situ controlled confinement for soils only contaminated with heavy metals at a unit cost of USD 320/m3;801 and ex situ soil washing for soils contaminated both with hydrocarbons and heavy metals at a unit cost of USD 240/m3.802 It rejects Burlington's proposal to treat and dispose of contaminated soils on site as contrary to the principle of full restoration enshrined in the 2008 Constitution.803
418.
On that basis, Ecuador estimates the total volume of contaminated soil at 654,687 m3 and claims total costs of USD 790,465,298, amounting to an average unit cost of USD 1,220 per cubic meter.804
419.
Burlington contends that Ecuador's remediation costs have no factual basis, and are based on wholly unnecessary remediation technologies and cost elements when equally effective and less costly methods are available, in Ecuador, from Ecuadorian companies.805 Burlington calculates remediation costs by (i) identifying sites exceeding regulatory criteria, (ii) measuring the areas requiring remediation, (iii) evaluating available remediation technologies, (iv) selecting the appropriate remediation technology, (v) creating a site-specific preliminary engineering design, and (vi) calculating the total costs based on local prices.806
420.
With respect to remediation technologies, GSI observes that technologies used in the Oriente region to remediate oil spills include (i) on site bioremediation, such as landfarming, composting or biopiles, (ii) off site treatment and disposal, and (iii) a combination of both.807 Among these, it retained off site treatment and disposal.808 For soils impacted by heavy metals, GSI considered (i) off site treatment and disposal, (ii) on site confinement and (iii) capping in place,809 and contemplated off site treatment and disposal as well as on site confinement.810 Nonetheless, Burlington indicated its preference that soils contaminated with heavy metals should be excavated and confined in pits on site.811
421.
In addition to inflating the cost for soil treatment and disposal,812 Burlington contends that Ecuador includes unnecessary transportation costs to a transfer station,813 inflates backfilling costs,814 and applies excessive mark ups and contingency.815 Relying on the "upper-range" costs of local contractors, Burlington calculates unit costs of USD 80/m3 for off site treatment and disposal (rounded up to USD 100/m3);816 USD 0.40/m3/km for transportation; USD 3/m3 for backfill material; USD 2.50/m3 for excavator rental; and USD 800/day for front loader and bulldozer rental.817 Burlington also includes predesign assessment costs (including for additional soil sampling and testing, as well as travel costs);818 a base cost of USD 20,000 per site for permitting and reporting; and a contingency ranging from 20% (for sites with no potential delineation or access issues) to 30% (for sites with both delineation and access issues).819
422.
In total, Burlington calculates an average all-in unit cost for off site landfill disposal of USD 260/m3, which is comparable to and mid-way between the average cost in Europe of USD 303/m3 and in the United States of America of USD 218/m3.820 On that basis, Burlington arrives at a total cost of USD 10,513,000 for the remediation of the 17 sites where it identified exceedances, including closing four open pits and abandoning seven wells821 (although Burlington insists that it is only responsible for remediation at Yuralpa Pad A for USD 110,000 and Jaguar 1 for USD 213,000, in addition to the four pit closures and the seven well abandonments for a total of USD 1,091,000).822

4.5.2 Discussion

423.
First, the Tribunal must decide whether to apply local or international prices. Contrary to Burlington's computations, Ecuador's remediation costs are mostly based on international prices charged by international contractors. Ecuador's expert, IEMS, rejects the use of Ecuadorian contractors because it claims that they are unable to "provide all of the services required for remediation in compliance with standard[s] which truly protect the environment"823 and because of the extraordinary amount of remediation under Ecuador's background value case:

"the reason we didn't choose Ecuadorian contractors, is because the job that we envisaged, that we costed, is a job that has not been done in Ecuador yet […] It's a job at a level that really has not been done very much in the world. No one has worked those standards before".824

424.
In fact, the volumes of soil requiring remediation fall significantly below Ecuador's estimates (be it under its background value or its regulatory criteria case). Moreover, the type of remediation activities contemplated here are routinely undertaken in the Oriente region.825 Therefore, the Tribunal sees no reason not to use local prices.
425.
Second, the Tribunal must assess whether the remediation must be processed on or off site. Off site remediation appears the more environmentally protective technique. As GSI confirmed, it is indeed used in the Oriente region in about 50% of cases.826 On the basis of the Site Visit and the information on local contractors in the record,827 the Tribunal is satisfied that qualified remediation contractors are available in El Coca and its surroundings which can handle the volumes of contaminated soils off site.828 For these reasons, the Tribunal decides in favor of off site remediation by local contractors.
426.
Third, as to remediation technology, RAOHE Annex 8, which relates to the treatment of wastes, indicates in its sub-section on non-hazardous wastes, that soils and drillings/cuttings contaminated with crude must be remediated through bioremediation or landfarming.829 The regulations are silent about treatment of heavy metals contamination, but Burlington agrees with Ecuador's view that controlled confinement is the proper technology for soils contaminated with heavy metals (although Burlington favors on site confinement, which the Tribunal rejects for the reasons just mentioned) and that soils contaminated with crude be treated off site, for instance, through bioremediation (whether landfarming, composting or biopiles).830 As to soils contaminated both with TPH and heavy metals, the Tribunal accepts Ecuador's suggestion of off site soil washing, especially in light of the fact that this kind of technology has also been applied by other operators in the Oriente region.831
427.
Accordingly, the following remediation technologies appear the most appropriate in the present circumstances: ex situ controlled confinement of soils contaminated with heavy metals; ex situ bioremediation of soils contaminated with crude; and ex situ soil washing for soils contaminated both with crude and heavy metals.
428.
On that basis and taking into account all relevant facts and evidence in the exercise of its discretion in matters of damage quantification, the Tribunal reaches the conclusion that an average all-in unit cost of USD 300/m3 is appropriate in the present circumstances. This unit cost includes all cost items, including pre-design, transportation, additional sampling, backfilling, mark-ups etc, but for a contingency which the Tribunal applies in a range of 20 to 30%. The percentage of the contingency factor depends on whether there are particular remediation challenges in terms of accessibility or exposure issues, such as difficulties in the terrain, larger distances between exceedance points and platform or access road, or proximity to streams and human settlements. Thus, the Tribunal typically applies a 20% contingency to exceedances on platforms and a 30% contingency to exceedance points off platform, unless the latter present no challenges in terms of accessibility or exposure. The all-in unit price used by the Tribunal is applied to volumes of soil as extracted. The Tribunal did not add additional volumes of soil to take account of a possible volumetric expansion, since the all-in unit price already accounts for such expansion.

4.6 Site review

4.6.1 Preliminary remarks

429.
In application of the standards established above, the Tribunal will now review Ecuador's soil contamination claim site by site. Although Ecuador included in its claim for soil contamination the removal of the mud pits, the Tribunal will address that claim separately.
430.
Ecuador investigated a total of 74 sites, sometimes modelling two platforms together.832 IEMS collected a total of 2,950 soil samples to assess the environmental condition in the Blocks, 2,769 in Block 7 and 181 in Block 21.833 Under its regulatory criteria case, Ecuador submits that the impacted area in Block 7 and 21 totals 414,506.14 m2 and that a total of 851,093.75 m3 of soil must be remediated.834 Ecuador's total claim for soil contamination for the regulatory criteria case amounts to USD 790,465,298.835
431.
The Tribunal notes that Ecuador does not claim any soil remediation at Payamino 9, Payamino 13, Payamino 18, Zorro, Yuralpa Pad B, Puerto Napo and Waponi-Ocatoe.836 Neither does it seek soil remedition if regulatory criteria apply at Coca 7, Payamino 6, Payamino 19, Gacela 3, Lobo 2, Lobo 4, Oso 2, Oso 3, Oso A and Yuralpa CPF/Pad F.837 In sum, Ecuador claims for soil remediation at 56 sites.
432.
In addition to relying on IEMS samples, Burlington's experts from GSI collected an additional 535 soil samples, 482 in Block 7 and 53 in Block 21.838 Burlington disputes any liability for Ecuador's soil remediation claims, although it recognizes that 37,555 m3 of non-pit soil may require remediation for a total cost of USD 9,748,000 at 17 sites.839 Burlington identifies six principal sites, namely Payamino 2/8, Mono 1/CPF, Payamino 1, Gacela 1/8, Coca 8 and Jaguar 1, where more than 1,000 m3 of soil may potentially require remediation.840 Burlington also identifies eleven additional sites that may require remediation of minor quantities of contaminated soil, namely Coca 1, Coca 4, Coca 6, Payamino 23, Coca CPF, Gacela 2, Jaguar 7/8, Mono Sur, Payamino 4 and Yuralpa Pad A.841
433.
Burlington acknowledges its liability for the "reasonable costs of remediating the pockets of exceedances at the two sites that are possibly attributable to the Consortium", namely USD 323,000 to remediate Yuralpa Pad A and Jaguar Area 3T, since it could not "definitively link harm to non-Consortium activities" at those sites.842 In addition, Burlington accepts to pay USD 68,000 to close four open and unused pits (Coca 4, Coca 8 and Payamino 15). Moreover, Burlington accepts to pay USD 700,000 for costs linked to the abandonment of seven platforms (Cóndor Norte, Gacela 2, Gacela 3, Jaguar 2, Jaguar 7/8, Jaguar 9 and Lobo 4).843 Accordingly, Burlington accepts to pay a total of USD 1,091,000.
434.
The Tribunal will now review the circumstances of the 56 sites for which Ecuador claims soil remediation, starting with the sites in Block 7/CPUF (4.6.2) before addressing those in Block 21 (4.6.3). Doing so, it will first address claims for non-pit soil and turn to pit soil in the following section. In this context, the Tribunal notes that GSI states that a number of IEMS samples were in reality taken from pits and/or misplaced by IEMS because of inaccuracies in the assignation of coordinates or in the labelling.844 Taking into account the Parties' and their experts' explanations,845 the Tribunal will decide these matters on a case by case basis where relevant.

4.6.2 Block 7/CPUF

a. Coca 1

435.
Ecuador submits that the impacted area in Coca 1 covers an area of 12,960 m2 and it claims for remediation of 16,848 m3 of soil, for a total cost of USD 15,232,093.846 Although Burlington disputes that the Consortium caused any exceedances at Coca 1, it accepts that there is contamination at that site covering 520 m2 in Area 1MT and 70 m2 in Area 2T (see figure in paragraph 437).847
436.
The following figures depict first IEMS's848 and second GSI's849 soil sampling locations at Coca 1:
437.
The potentially impacted areas calculated by GSI, which – as determined in section 4.4.3.c and paragraph 411 – the Tribunal will use as basis for its own delineation, are shown in the following figure:850
438.
Applying the guidelines set out above (section 4.3.2.c) and thus considering that the "subsequent use" of the land in this area is agricultural, the Tribunal identified a TPH exceedance in Area 2T,851 and TPH, barium and cadmium exceedances in Area 1MT.852
439.
Coca 1 was drilled in 1970 by Texaco, but was on stand-by until 1986 when it started operating.853 There are no records of historical spills854 and the Consortium conducted a well workover in 2005,855 although Burlington generally argues that exceedances are linked to drilling activities pre-dating the Consortium‘s time,856 as evidenced in part by the fact that there is no pit at Coca 1 implying that drilling muds were discharged into the environment by prior opertators.857
440.
The Tribunal accepts Burlington's evidence and submissions with respect to Area 1MT. Considering the depth of the exceedances (up to 2.6 meters) and the fact that there is no pit at Coca 1, it finds that contamination in Area 1MT is associated to drilling activities and was thus caused by prior operators. Burlington has thus succeeded in rebutting the presumption of causation with respect to the contamination found in that area. By contrast, Burlington failed to convincingly rebut the presumption that it caused the TPH exceedance in Area 2T.
441.
GSI applied linear interpolation to calculate the impacted area,858 placing certain dirty borings back into the soil during its stepping out procedure859 – although the Tribunal notes that GSI's delineation samples bear relatively low TPH levels.860
442.
On that basis and in accordance with the guidelines established above (sections 4.4.3.c and 4.5.2 above), the Tribunal determines that 200 m2 in Area 2T are impacted. Since remediation is required up to 2 meters depth, the total volume of contaminated soil amounts to 400 m3, for a cost of USD 120,000. Adding a 20% contingency (see paragraph 428 above), the total cost for which Burlington is liable amounts to USD 144,000 for Coca 1.

b. Coca 2 and CPF

443.
Ecuador submits that the impacted area in Coca 2/CPF extends 54,795 m2,861 and claims the remediation of 95,946.50 m3 of soil,862 for a total cost of USD 83,465,389.863 Although Burlington points to prior operators and disputes that the Consortium caused any exceedances at Coca 2/CPF,864 it recognizes that there is contamination in Area 2T extending 630 m2 for a total cost of USD 234,000.865
444.
The following figures depict IEMS's866 and GSI's867 sampling locations at Coca 2/CPF:
445.
The following image shows Areas 1P, 2T, 3P and 4T designated by GSI:868
446.
There are various land uses around Coca 2/CPF.869 With reference to the guidelines set out above (section 4.3.2.c), the Tribunal generally applied agricultural land use criteria to the areas identified by GSI as Areas 4T and 2T (Areas 1P and 3P are pits), especially considering th at the right-of-way in the so-called Chalá swamp in Area 2T was recently created for a Petroamazonas pipeline. It applied sensitive ecosystem criteria in the vicinity of Coca 2,870 near the API separator,871 along the creek,872 and on the other side of the culvert in Area 2T.873
447.
On that basis, the Tribunal identified no exceedances at Coca 2.874 It observed, however, the following exceedances at Coca CPF: one cadmium exceedance to the southwest of the pit area and south of Area 4T used for the treatment of formation water875 and another cadmium exceedance north of the API separator.876 In addition, the Tribunal identified various TPH exceedances in Area 2T.877
448.
Coca CPF is one of the sites where there is a long record of spills caused by prior operators.878 There is also evidence of a 1.5 barrel crude spill on 13 June 2005 at the Coca 9 flow line reaching the CPF manifold and an unreported spill at the washing tank that occurred on 31 August 2008.879 These appear unrelated to the exceedances identified above because of their location. It is nonetheless noteworthy that the 2006 biennial audit identified a non-conformity with respect to wastewater discharges.880 It is further of particular relevance that, between 1989 and 1996, Petroproducción permanently discharged produced water, without skimming the oil, into the environment from the pits to the south of the CPF into the creek flowing southeast towards the Chalá swamp where Area 2T is located.881
449.
While Mr. Saltos indicated that Oryx discontinued this practice and the area was remediated in 1997,882 he also stated that oil seeping continued thereafter, and that this state of affairs persisted during the Consortium's operatorship, and that the Consortium did not reach an agreement with Mr. Chalá because of the takeover in 2009.883 Accordingly, the Tribunal holds that the Consortium is partially responsible for the environmental condition in the Chalá swamp. In the exercise of its discretion in matters of the quantification of damages, it determines that Burlington shall pay 50% of the remediation costs.
450.
In addition, the Tribunal sees no elements in the record to rebut the presumption that Burlington caused the contamination at the other exceedance points mentioned in paragraph 447. Therefore, Burlington is liable to pay the full costs of remediation of these other areas.
451.
With respect to Area 2T in the Chalá swamp, two additional difficulties arise. In this context, it is useful to refer to GSI's impacted area designation in the Chalá swamp:884
452.
The first difficulty in connection with the Chalá swamp relates to the Parties' disagreement on the location of IEMS's sample points in the swamp. Ecuador indicates that IEMS sampled south of the culvert,885 whereas Burlington argues that IEMS sampled north of the culvert.886 The issue was discussed during the Site Visit. While IEMS reiterated that it sampled south of the culvert,887 GSI explained that IEMS did not use a GPS but a tape measure to locate its samples and that one point was missed causing a discrepancy of 100 meters.888 In addition, GSI stated that it found no indication of sampling to the south of the culvert, that GSI itself took verification samples there and found no exceedances, and that TPH exceedances in fact exist to the north of the culvert where GSI sampled.889 On that basis, GSI argues that "both the geographic and the soil data tell us that we were in the right location".890
453.
Because IEMS did not use state of the art (and in fact commonly used) technology to locate its samples, the Tribunal is inclined to rely on GSI's location. Further, the Tribunal notes that GSI places IEMS's samples where the creek flowing from the API separator reaches the swamp area.891 Accordingly, the Tribunal will work on the basis of GSI's locations for IEMS samples.
454.
The second difficulty relates to the fact that Petroamazonas built a pipeline in Area 2T after IEMS collected its samples in 2012.892 Because the right of way built by Petroamaonas partly passes through Area 2T, the Tribunal discarded from its analysis various sample locations, the remediation of which was superseded by the construction of the pipeline.893
455.
On that basis and applying the guidelines set out earlier (section 4.4.3.c and in particular paragraph 414), the Tribunal determined the following impacted areas: 480 m2 in Area 2T,894 200 m2 for the cadmium exceedance southwest of the pit area and 200 m2 for the cadmium exceedance north of the API separator, for a total impacted area of 880 m2. The total soil volume is 1,480 m3: 480 m3 in Area 2T, 400 m3 for the cadmium exceedance southwest of the pit area and 600 m3 for the cadmium exceedance north of the API separator.
456.
In application of the Tribunal's approach explained above (section 4.5.2) and adding a 30% contingency (see paragraph 428 above), the remediation cost for Area 2T is USD 187,200, of which Burlington must bear USD 93,600. Adding a 20% contingency to the cadmium exceedances, the remediation cost for these locations is USD 360,000. Accordingly, Burlington is liable to pay USD 453,600 for soil remediation at Coca CPF.

c. Coca 4

457.
Ecuador submits that the impacted area in Coca 4 extends over 2,732 m2 and claims the remediation of 3,551.60 m3 of soil, for a total cost of USD 3,617,276.895 Although Burlington disputes that the Consortium caused any exceedances at Coca 4, it accepts that there is contamination in Area 1M extending 20 m2 for a total remediation cost of USD 72,000, but argues that this was caused by pre-Consortium drilling activities.896
458.
The following figures depict IEMS's897 and GSI's898 sampling locations:
459.
By reference to the approach set out above (section 4.3.2.c) and applying sensitive ecosystem use for Area 1M,899 the Tribunal identified barium, cadmium and nickel exceedances in Area 1M.900
460.
The potentially impacted area calculated by GSI which the Tribunal will use as a basis for its delineation (see section 4.4.3.c and especially paragraph 411 above) is illustrated below:901
461.
Coca 4 was drilled by CEPE in 1989-1990902 and Burlington links the exceedances to drilling activities that pre-date the Consortium.903 There is no specific record of historical spills.904 Ecuador's witness Mr. Solís testified that Perenco did not report a spill at the production pipeline on 10 September 2006 affecting approximately 4 m3.905 Accordingly, the Tribunal considers that Burlington failed to rebut the presumption that the Consortium caused the exceedances, which appear linked to an overflow of the oil/water separator rather than to drilling mud discharges, in particular since there are eight closed pits at that site and all exceedances are at a relatively shallow depth (<0.7m). Accordingly, Burlington is liable to remedy all exceedances at Coca 4.
462.
On that basis and in application of the approach discussed above (see sections 4.4.3.c and 4.5.2), the Tribunal increased GSI's delineation around sample 7-COC-04-244-MS-E-0,5 from 20 m2 to 100 m2.906 One must then add 80 m2 for the cadmium exceedances between the oil/water separator and the mud pit, and 400 m2 around the nickel exceedances,907 for a total impacted area of 580 m2 and total volume of soil of 580 m3. Adding a 30% contingency (see paragraph 428 above), the total remediation cost for which Burlingtom is liable amounts to USD 226,200 for Coca 4.

d. Coca 6

463.
Ecuador submits that the impacted area in Coca 6 extends over 8,053 m2 and that 10,468 m3 of soil must be remediated for a total cost of USD 9,997,490.908 Although Burlington argues that the exceedances at Coca 6 were caused by a spill in 1999,909 it accepts that there is contamination in Area 1MT extending 780 m2 for a total remediation cost of USD 242,000.910
464.
The following figures depict IEMS's911 and GSI's sampling locations:912
465.
The following image depicts GSI's delineation of the potentially impacted area, from which the Tribunal will start, as explained earlier (see section 4.4.3.c and in particular paragraph 411):913
466.
In reliance on the approach explained above (see section 4.3.2.c) and thus accepting sensitive ecosystem use for Coca 6,914 the Tribunal identified TPH, barium and cadmium exceedances in Area 1MT.915
467.
Coca 6 was drilled by CEPE in 1989, but the well was shut in by the Consortium.916 Burlington links all exceedances to a spill that occurred in 1999.917 However, Ecuador demonstrated that the Consortium conducted a well workover at that site involving drilling and the use of chemicals, which could also be the cause of heavy metal exceedances.918 The Tribunal further notes that there was a spill in 2006 and another one of 10 barrels in 2007.919 In addition, GSI stated that a remediation program was in place when the Consortium took over the operations, but that it was never completed due to unsuccessful negotiations with local communities to gain access to the stream area.920 In light of the fact that the Consortium assumed the remediation program for the 1999 spill, but that it failed to conduct the remediation, and considering the 2002 workover and the 2006 and 2007 spills, the Tribunal holds the Consortium fully responsible for the environmental condition in Area 1MT.
468.
In application of the guidelines set earlier (sections 4.4.3.c and 4.5.2), the Tribunal reaches the following conclusion. Since GSI's delineation does not include all exceedance points in Area 1MT,921 and in light of the high TPH values reaching the creek to the southeast of the swale, the Tribunal determined the impacted area to extend to 1,100 m2 and the total volume of soil to be remediated to amount to 1,100 m3. Considering a contingency of 30% (see paragraph 428 above), the remediation cost amounts to USD 429,000 for Coca 6.

e. Coca 8

469.
Ecuador submits that an area of 12,457 m2 is impacted at Coca 8 and that 36,571.60 m3 of soil must be remediated, for a total cost of USD 34,628,076.922 Burlington argues that the contamination found in Coca 8 is attributable to drilling activities of prior operators,923 that the potentially impacted area is in any event only 2,280 m2,924 and that remediation is warranted for 3,470 m3 of soil, for a total cost of USD 449,000.925
470.
The following images show IEMS's926 and GSI's927 sampling locations at Coca 8:
471.
GSI's delineation of impacted areas and depth in Areas 1M and 4M which will serve as basis for the Tribunal's delineation (see section 4.4.3.c above and in particular paragraph 411) is shown below, first for the 0-1 meter layer928 and then for the 1-2 meter layer:929
472.
There are various land uses around Coca 8. With reference to the Tribunal's approach explained above (section 4.3.2.c), the platform falls within the industrial land use criteria. The area northwest of the platform, along the entry road is used for agricultural purposes. All other areas surrounding the platform are properly classified as sensitive ecosystem.930 Applying these criteria, there are no exceedances on the platform and no exceedances to the north of the platform.931
473.
Further, applying the sensitive ecosystem criteria to the other areas, the Tribunal identified numerous barium, cadmium and nickel exceedances in the area visited by the Tribunal south of the three mud pits bordering the southern edge of the platform, which includes the areas designated by GSI as Area 4M, immediately adjacent to the mud pits, and Area 1M further downhill.932 The highest barium level reaches 23,368.3 mg/kg at the sample location 7-COC-08-TE-104P.933 In addition, there are various barium, cadmium and nickel exceedances to the west of the platform and the pit area.934
474.
The Coca 8 well was drilled by CEPE in 1991 and the Consortium operated the platform from 2002 until the takeover in 2009.935 There is evidence of a 150 barrel crude spill on 4 March 2000 under the operatorship of Kerr McGee due to equipment failure, which affected an area of 600 m2, although 145 barrels were recovered.936 On the other hand, there is no evidence of spills during the Consortium's time or of workovers undertaken by the Consortium.937 As a result, the Tribunal considers that the exceedances must be deemed due to the activities of prior operators.
475.
As the related analysis will show (see paragraph 820 below), there is evidence that the mud pits have been leaking into the surrounding environment, in particular in Areas 1M and 4M. The Consortium was made aware of this situation at the latest in March 2007 when Mr. Noteño, the President of the Corazón de Oriente community, filed a complaint.938 The record does not show that the Consortium took action to mitigate the harm and prevent it from spreading further.939 As a result, the Tribunal finds that Burlington shares a responsibility for this damage and must accordingly contribute 50% to the cost of remediating it. This applies to the area south and west of the pits. By contrast, the two barium exceedances to the west of the platform are unrelated to the condition of the pits at issue and are, therefore, not attributable to Burlington.940
476.
Since GSI applied agricultural criteria, its delineation for Areas 1M and 4M does not encompass all exceedance points.941 With reference to the approach set out above (sections 4.4.3.c and 4.5.2), the Tribunal has thus adopted its own approach and finds that the impacted area south of the pits amounts to 3,270 m2 and the total volume of contaminated soil 6,220 m3.942 Considering a 30% contingency (see paragraph 428), the total cost to remediate Areas 1M and 4M is USD 2,425,800.
477.
The exceedance points to the west of the pits make up five discrete impacted areas of 1,000 m2 in the aggregate (200 m2 each) and a total volume of 1,000 m3 of soil to be remediated,943 for a total cost of USD 390,000 with a contingency of 30% (see paragraph 428).
478.
The exceedance next to the oil/water separator impacted an area of 80 m2 and requires remediation of 160 m3, for a cost of USD 57,600 that includes a 20% contingency (see paragraph 428).944
479.
To conclude, the total impacted area in Coca 8 is 4,350 m2; the total volume of soil to be remediated is 7,380 m3; and the total cost of remediation of regular soil at Coca 8 amounts to USD 2,873,400, of which Burlington shall bear 50%, totalling USD 1,436,700.

f. Coca 9

480.
Ecuador submits that an area of 7,511 m2 is impacted at Coca 9 requiring the remediation of 16,789.50 m3 of soil, for a total cost of USD 16,772,147.945 For its part, Burlington argues that no remediation is called for at this site, since there are no exceedances of any indicator parameters and the vanadium exceedance on the northwestern corner of the platform is a "background metals issue".946
481.
The following image depicts IEMS's sampling locations,947 being specified that GSI collected no samples at this site:948
482.
Coca 9 was drilled by Oryx in 1992 and 1993,949 and it remains in operation under Petroamazonas. There are no records of historical spills at that site.950 In its preliminary assessment, IEMS identified discoloration of the floor in the northwestern corner of the platform and signs of contamination in a marshy area located at the southeast corner of the platform.951 The 2008 audit mentioned the lack of labelling of fuel tanks and chemical products, the lack of signage concerning industrial safety or the handling of chemicals, the lack of maintenance of the perimeter ditch, the unmarked oil well, and the accumulation of loose materials on the platform.952 While these facts are not evidence of environmental harm, they suggest faulty conduct, which reinforces a finding of a causation. In light of these facts and of Burlington's argument described above, Burlington cannot be held to have rebutted the presumption that the Consortium caused the harm found at that site.
483.
Applying industrial use criteria to the platform pursuant to its general approach (see section 4.3.2.c), the Tribunal finds an exceedance with a relatively high level of vanadium (385 mg/kg) on the northwestern corner of the platform. In reliance on its methodology explanined above (sections 4.4.3.c and 4.5.2), the Tribunal calculated an impacted area of 200 m2 and a volume of soil to be remediated of 400 m3, for a total cost of USD 144,000 which includes a 20% contingency (see paragraph 428).953
484.
Considering sensitive ecosystem land use criteria for off platform locations,954 there is one nickel exceedance to the southeast of the platform, amounting to an impacted area of 200 m2, a volume of contaminated soil of 200 m3, and a remediation cost of USD 78,000 that includes a 30% contingency (see paragraph 428).955
485.
Accordingly, Burlington is liable to pay USD 222,000 to remediate the 600 m3 of contaminated soil found covering an impacted area of 400 m2 at Coca 9.

g. Coca 10/16

486.
Ecuador claims USD 252,891 to remediate 175 m3 of soil covering an area of 135 m2 at Coca 10/16.956 Burlington objects that no remediation is required at this site, although it recognizes that a TPH and barium exceedance exists at sample location 7-COC-10-175-MS-E-1,5, if agricultural land use criteria apply.957
487.
The following images depict IEMS's958 and GSI's959 sampling locations at Coca 10/16:
488.
Applying its approach on land use (section 4.3.2.c), the Tribunal observes that there are no exceedances on the platform. By contrast, using sensitive ecosystem criteria to off platform locations,960 the Tribunal identified a TPH exceedance and various barium and cadmium exceedances to the north of the northeastern corner of the platform.961
489.
There are no historical records of spills962 and there is no allegation that the Consortium conducted any well workovers at that site.963 The 2008 audit apparently did not assess Coca 10/16, although Annex B contains some pictures of the site and commentaries, for instance, that the perimeter drain is improperly maintained or that the grease trap is insufficiently cleaned,964 which increases the risk that contaminants flow over the platform into the surrounding environment.965 Because the exceedance points are in the vicinity of the grease trap and because any overflows from that oil/water separator flow directly into the swale identified by GSI as Area 1MT,966 the Tribunal finds that the Consortium is liable for damage to this site.
490.
Ecuador has not increased its claim for Coca 10/16 after GSI's site investigation, although two of GSI's samples show additional exceedances in Area 1MT. Therefore, the Tribunal grants Ecuador's request that 175 m3 of soil be remediated. With a 30% contingency (see paragraph 428), the total remediation cost for this site is USD 68,250 for Coca 10/16.

h. Coca 11

491.
Ecuador claims USD 1,833,746 to remediate 1,765 m3 of soil covering an impacted area of 844 m2.967 Burlington rejects the claim. IEMS only collected three non-pit samples, one on the platform and two outside of it.968 There is no exceedance at that site under any land use criteria.969 Accordingly, Ecuador's claim for Coca 11 is dismissed.

i. Coca 12

492.
Ecuador claims USD 935,761 to remediate 924.30 m3 of soil coming from an impacted area of 283 m2.970 Here again, there are no exceedances at this site under any regulatory criteria,971 with the result that the claim for Coca 12 must be dismissed.

j. Coca 13

493.
Ecuador claims USD 8,176,102 to remediate 8,126.30 m3 of soil from an impacted area of 6,251 m2.972 Here too, the Tribunal dismisses the claim for Coca 13 on the ground that there are no exceedances under any regulatory criteria,973 to which it adds that Petroamazonas has significantly expanded the dimensions of the platform.974

k. Coca 15

494.
Ecuador claims USD 10,987,069 for the remediation of 12,560.60 m3 to be recovered from an impacted area of 7,585 m2.975 The claim for Coca 15 cannot succeed because there are no exceedances under any regulatory criteria.976 Indeed, sample point MS-COC-C2-15.3-0.5, which is within the confines of the platform and which the Parties and the Tribunal had an opportunity to inspect, has a TPH value of 3,571 mg/kg that falls below the 4,000 mg/kg limit for industrial soil.977

l. Coca 18/19

495.
Ecuador claims USD 36,483,570 for the remediation of 37,329.50 m3 of soil from an area of 11,485 m2.978 Burlington initially classified Coca 18/19 as one of five priority sites for potential soil remediation with a possibly impacted area of 160 m2.979 After reviewing the pit closure report for Coca 19, GSI concluded that the exceedances in the area designated as Area 1M were in fact drilling muds disposed in two off platform auxiliary pits.980 On this basis, Burlington opposes this claim.
496.
The following figures show IEMS's981 and GSI's982 sampling locations at Coca 18/19:
497.
Coca 18 was drilled by Kerr McGee in July 2001 and Coca 19 by Perenco in December 2003.983 As will be discussed in the analysis of mud pits, when the Consortium drilled the Coca 19 well in 2003, it deposited drilling wastes in four pits adjacent to that well and in two auxiliary pits located just beyond the southwestern corner of the platform, the latter two corresponding to Area 1M.984 Therefore, the Tribunal will address the content of the auxiliary pits as part of its analysis of mud pits.
498.
In addition to alleged exceedances in Area 1M, the Tribunal identified one barium exceedance on the platform next to the Coca 18 well.985 While that well was drilled by Kerr McGee in 2001, the Tribunal notes that the Consortium conducted some workovers at that well.986 It, therefore, holds the Consortium liable to remediate that exceedance. In application of its methodology (sections 4.4.3.c and 4.5.2), it admits an impacted area of 80 m2987 and a volume of 80 m3, for a total cost of USD 28,800 for Coca 18/19, taking into account a 20% contingency (see paragraph 428).
499.
Finally, applying agricultural land criteria to off platform locations, there are no additional exceedances in areas surrounding the platform.988 Therefore, besides the on platform exceedance mentioned above and subject to the Tribunal's analysis of the mud pits, no additional remediation is required for Coca 18/19.

m. Payamino 1 and CPF

500.
Ecuador claims USD 44,642,039 to remediate Payamino 1/CPF, where it contends that an area of 36,613 m2 is impacted and that 50,659.70 m3 of soil need to be remediated.989
501.
The following figures show IEMS's990 and GSI's991 sampling locations at Payamino 1/CPF:
502.
Further, GSI's delination of Areas 1P and 2MT, which will serve as basis for the Tribunal's determination of the extent of contamination (see section 4.4.3.c above and in particular paragraph 411), is illustrated as follows:992
503.
Burlington classifies Payamino 1 as a priority site, i.e. one where it accepts the existence of contamination potentially requiring remediation for a total cost not exceeding USD 307,000.993 Burlington submits that the source of contamination in Area 1P is a concrete water basin removed in 2001, which thus predates the Consortium's operatorship. According to Burlington, remediating the 800 m3 of potentially contaminated soil would cost no more than USD 243,000.994 Burlington also accepts that 210 m3 of soil potentially require remediation in Area 2MT for a total cost of USD 64,000.995 Finally, Burlington rejects all other claims for remediation in Payamino 1/CPF, in particular by arguing that IEMS samples in Area 3MT are in reality misplaced and were taken from Payamino 2/8.996 With respect to the sample collected close to the exhaust stack of a diesel-powered compressor at Payamino CPF,997 GSI opined that less than 20 liters (0.02 m3) of soot were collected "as a thin film on the ground surface" and that the soils beneath were clean.998 This does not, in Burlington's eyes, justify a claim of over USD 5 million to remediate 5,373 m3 of soil.
504.
After Coca 1, Payamino 1 is the second oldest operating well in Block 7/CPUF. It was drilled by BP in 1986 and evidence suggests that 13 workovers were conducted at that well between 1991 and 2005, including by the Consortium.999 The Payamino 1 well is located north and approximately 30 meters downhill of the CPF, which was put in operation between 1991 and 1992.
505.
Before addressing the off platform exceedances in the areas designated by GSI as Areas 1P, 2MT and 3MT, the Tribunal will address the on platform exceedance of TPH (5,138 mg/kg) adjacent to the diesel-powered compressor at the CPF.1000 Since Petroamazonas operated the compressor after the takeover without addressing that situation, the Tribunal deems it correct to apportion the liability between the Consortium and its successor and, in its discretion in assessing the evidence and quantifying damages, finds that Burlington shall bear 50% of the remediation costs. Because the impacted area appears to be localized, the Tribunal determines that remediation of 30 m3, for a total cost of USD 10,800 including a 20% contingency (see paragraph 428), is adequate and reasonable. Burlington shall thus pay 50% of that cost, i.e. USD 5,400.
506.
Turning now to off platform locations, it is undisputed that sensitive ecosystem land use criteria apply to areas surrounding Payamino 1/CPF, notably because this site is part of a designated protected area under the SNAP system (see section 4.3.2.c above).1001 On that basis, the Tribunal identified two TPH exceedances in Area 1P1002 and various TPH and barium exceedances in Area 2MT.1003 In light of the fact that IEMS confirmed that it mislabeled samples 07-PAY28-SDB3-R and 07-PAY28-SDB4-R and that they were indeed collected in Area 3MT at Payamino 1/CPF,1004 the Tribunal finds that TPH and barium contamination has been shown also in Area 3MT.
507.
In light of the fact that Area 1P hosted a former concrete produced water basin that was remediated in 1999 and then removed in 2001,1005 that this basin was above ground, that a concrete layer now covers the TPH exceedance points, the Tribunal is of the view that the exceedances were not caused by the Consortium but are liabilities of prior operators. Accordingly, Burlington is not liable for the condition of Area 1P.
508.
The situation is the opposite as concerns the catchment area labelled Area 2MT, which appears to be a workover pit located some 120 meters southwest of the Payamino 1 platform in the proximity of the River Añango. This pit was unreported and thus served as a concealed location to dump industrial wastes in the middle of a protected area adjacent to a water body. GSI indicated that 13 workovers had been conducted at Payamino 1 until 2005, including by the Consortium,1006 which Mr. Saltos confirmed.1007 GSI opined that it was unable to determine the cause of the exceedances in Area 2MT.1008 Neither Burlington nor its key witness, Mr. Saltos, appeared to dispute that the Consortium used that workover pit. Accordingly, the Tribunal holds Burlington liable to remediate Area 2MT. By reference to the methodology set out above (sections 4.4.3.c and 4.5.2), the impacted area is 500 m2,1009 the total volume of soil 500 m3, for a total cost of USD 195,000, which includes a 30% contingency (see paragraph 428).
509.
Finally, Burlington argues that the samples showing TPH exceedances in Area 3MT were mislocated, but does not provide an alternative explanation as to the possible cause of these exceedances. Further, the spill record for Payamino CPF, including the area formerly known as Payamino 22, both of which overlook Area 3MT, reveals numerous spills during the Consortium's operations, including an oil runoff from Payamino 22 in 2002, a 0.1 barrel crude spill in 2004,1010 a 2.4 barrel crude oil spill on 2 January 2007,1011 an unreported spill of contaminated fluids on 1 October 2007 affecting the API separator, perimeter drains and the Añango River,1012 a 6.28 barrel crude spill on 6 October 2007 affecting the River Añango,1013 and a 2009 crude spill some 10 meters away from the CPF along the Payamino 19 flowline.1014 These facts support a determination that the exceedances were caused during the Consortium's operatorship, which Burlington has not rebutted. This view is reinforced if one looks to the definition of "affected areas" in the settlement agreement of 11 May 2010 between Mr. Jungal, Ms. Cárdenas Hernández and the Consortium, which includes "landslides on the slope of well Payamino 22" and "erosion and sliding of the slope at Payamino Station".1015
510.
In light of the high levels of TPH present and by reference to its approach explained above (sections 4.4.3.c and 4.5.2), the Tribunal determines a total impacted area of 400 m2, a total volume of soil of 1,000 m3,1016 and a remediation cost of USD 390,000, including a 30% contingency (see paragraph 428).
511.
In conclusion, the Tribunal determines that Burlington must bear the remediation cost for Areas 2MT and 3MT and half of the cost for the TPH exceedance at Payamino CPF, for a total amount of USD 590,400 for Payamino 1/CPF.

n. Payamino 2/8

512.
Ecuador claims USD 31,899,459 for the remediation of 36,643 m3 of soil from an impacted area of 21,600 m2.1017 Burlington includes Payamino 2/8 as one of its principal sites for potential soil remediation, accepting that an area of 15,850 m2 and a volume of 18,345 m3 of soil are affected in an area of the Jungal swamp located to the north and east of the Payamino 2/8 platform and the Payamino 8 mud pit.1018 However, Burlington denies that the Consortium caused the harm, which it attributes to drilling mud discharges and oil spills from the Payamino 2 test pit prior to 1992.1019
513.
The following figures depict IEMS's1020 and GSI's1021 sampling locations at Payamino 2/8:
514.
The potentially impacted areas calculated by GSI, to which the Tribunal will revert, are illustrated below:1022
515.
Payamino 2/8 is one of the most contentious sites and the Parties have extensively debated the potential causes for the environmental condition encountered in the swampy area known as the Jungal swamp. The Tribunal will address Ecuador's submissions on the pits in its analysis of mud pits below. Here it will only deal with the off platform sampling locations to the west, north and east of the Payamino 2/8 platform and adjoining pits.1023 In its review, the Tribunal also considers the samples collected for the Walsh report as GSI relies on these samples (shown as red boxes in the preceding figure).1024
516.
It is undisputed that the land to the east of the Payamino 2/8 platform is located in a designated protected area under the SNAP system that qualifies as sensitive ecosystem.1025 Burlington disputes, however, that the area to the west of the designated area falls within that same category and calls for the application of agricultural land use criteria. In accordance with its methodology (see section 4.3.2.c above), as the areas in question are immediately adjacent to a designated protected area, the Tribunal will apply the sensitive ecosystem criterion to all areas surrounding the platform (except for the access road, where no samples were in any event taken).
517.
On that basis, the Tribunal identified various TPH, barium and cadmium exceedances in the area of the Jungal swamp identified by GSI as Area 2MT,1026 with the highest TPH level reaching 10,661 mg/kg.1027 The Walsh samples identify numerous TPH and barium exceedances, with TPH levels reaching 64'616 mg/kg and the highest barium level at 3,584 mg/kg,1028 when the permissible limits are at 1,000 mg/kg for TPH and 639 mg/kg for barium.
518.
It is not seriously disputed that a major environmental incident occurred in connection with the drilling of the Payamino 2 and Payamino 8 wells between 1987 and 1992. It can be left open for present purposes whether it was caused by the discharge of test crude into the Jungal swamp or by the rupture of the Payamino 2 pit.1029 In any event, it occurred prior to the Consortium's operatorship. While it is true that successive operators did not indicate the presence of contamination at that site,1030 thus suggesting that the area was remediated when Oryx so requested in 1994, the Tribunal notes that Ecuador has provided no evidence of any remediation.1031 Therefore, the Tribunal cannot rule out that the environmental harm identified above was caused by prior operators and continues to be present in the swamp until today.
519.
Nonetheless, there are several reasons for the Tribunal to hold the Consortium partially liable to remediate the exceedances in the Jungal swamp. First, there is contradictory expert evidence generated on the Consortium's behalf regarding the age of the crude found in the swamp. On the one hand, a report by Grüntech dated 2 June 2010 estimated that age to be 4 to 7 years, which suggests that the contamination occurred during the Consortium's time, although the report specified that exact dating was impossible without a patron sample and that other factors influenced the dating procedure.1032 On the other hand, a report by Mr. Douglas from NewFields assessed Grüntech's conclusion and opined that the rate of degradation of petroleum is "highly variable" depending on numerous factors not considered by Grüntech, and finally concluded that any attempt to date the crude oil would be a guess and that it was "equally as possible that the crude oil was released more than 20 years ago".1033
520.
Second, the Consortium was aware of claims of environmental pollution at the latest since early 2007,1034 when Mr. Eduardo Greta Cerda, Political Lieutnant of the Parrish San Luis de Armenia, lodged a complaint with the Provincial Council of Orellana. That council ordered an inspection which took place on 4 April 2007 and discovered "[o]n the #02-08 platform (point 4) and to the side of it, a large area of approximately 20,000 m2 […] containing crude oil residues in large quantities".1035 Burlington's silence about that inspection in these proceedings and its conclusion that crude oil was found on the platform and to its eastern side are disturbing.1036 The same is true of the fact that the Consortium apparently took no immediate action and that no samples were taken for the 2008 audit.1037
521.
Third, an internal memorandum prepared by Perenco in May 2010 "on the characterization of the environmental issues in Payamino 2-8" following a renewed complaint by Mr. Jungal,1038 states that "we [the Consortium] know the true extent of the problem" just as Mr. Jungal. It continues saying that Petroamazonas' health, safety and environment department knows the existence, but not the location and extent of the contamination, and that the local ombudsman knows the location but not the extent. The memorandum then goes on setting out three "possible solutions", namely (i) to remedy the site, (ii) to confine the problem by buying the land, or (iii) to dismiss any liability.1039 About this last option, the memorandum is quite explicit:

"- very probably, Mr. Jungal would sue us before the [criminal] courts for pollution (this crime has already been defined) and will seek multimillion dollar compensation

- the State (Ministry of the Environment and/or Petroamazonas (sic)) will force us to remedy the site under their conditions

- the cost will reach amounts very difficult to estimate now

- the reputational cost to Perenco is also going to be very high

- probably the State will assume that we are hiding many more damages and will scrutinize the operations area in search of more damages and it will probably find them.1040

522.
This memorandum shows that the Consortium was aware of and concerned about the existing liabilities in the Jungal swamp.1041 In 2010, it eventually reached a settlement agreement with Mr. Jungal for an amount of USD 110,000, which agreement contained a confidentiality clause.1042 It is only thereafter that, on 11 June 2010, the Consortium notified the Ecuadorian authorities of the existence of contamination in the Jungal swamp.1043
523.
These facts lead the Tribunal to conclude that the Consortium indeed played some role in causing the environmental condition, or at least in not mitigating the harm when it became aware of it in 2005 and 2007, a role which is much more important than what Burlington would have the Tribunal believe. As a result, on the basis of the record and in the exercise of its powers in the assessment of the evidence and in the quantification of the damage, the Tribunal comes to the conclusion that Burlington must bear 60% of the remediation cost for that site.
524.
As mentioned above, IEMS estimates the impacted area at 21,600 m2, while GSI accepts 15,850 m2. More specifically, GSI resorted to linear interpolation to calculate the impacted area for barium in the layer 0-1 meter (7,460 m2),1044 the impacted area for TPH in the layer 0-1 meter (13,090 m2),1045 and the impacted area for TPH in the layer 1-1.5 meters (4,990 m2).1046 These areas partly overlap and yield a maximum impacted area of 15,850 m2.
525.
Considering that only one exceedance contains TPH to a depth of 1.5 meters, the Tribunal accepts GSI's calculation of the impacted area for TPH in the layer 1-1.5 meters, namely 4,990 m2,1047 amounting to a total volume of 4,990 m3 of soil to be remediated in the 1-2 meter layer. Considering the TPH and barium levels of clean samples, the Tribunal is also in agreement with GSI's delineation of the impacted area of the 0-1 meter layer, considering that GSI extended the potentially impacted areas by 20 meters further to the north of the Walsh samples, which in turn are further north of IEMS's samples (see figure in paragraph 514).1048 In accordance with its general approach set out earlier (section 4.4.3.c), the Tribunal nevertheless extends GSI's maximum impacted area of 15,850 m2 to 17,000 m2 to take into account the relatively high cadmium exceedance at sample point 07-PAY28-SDB8-R.
526.
In light of the foregoing and of its methodology on assessing remediation costs (section 4.5.2), the total volume of contaminated soil is 21,990 m3 and the total cost of remediation, including a 30% contingency (see paragraph 428), is USD 8,576,100, of which Burlington shall pay 60%, i.e. USD 5,145,660 for Payamino 2/8.

o. Payamino 3

527.
Ecuador claims USD 2,207,021 to remediate 2,453.10 m3 of soil from an impacted area estimated at 1,258 m2.1049 Burlington disputes any liability, since the alleged barium exceedance is located in a pit and any other exceedances are related to preConsortium drilling discharges in 1987.1050
528.
Leaving aside pit soil, there are no off platform exceedances under any regulatory criteria.1051 On the platform, there is a high TPH exceedance (10,364 mg/kg) in the south corner "in an area near a soil stockpile".1052 There is no indication that the exceedance is due to pre-Consortium drilling activities.1053 Ecuador also pointed to the fact that no samples were collected at that site for the 2008 audit,1054 although the audit noted that the platform was lacking anti-pollution devices such as perimeter ditches and grease traps.1055
529.
On this basis and by reference to its methodology explained above (sections 4.4.3.c and 4.5.2), the Tribunal accepts that Burlington is liable to remediate this area, which it sets at 80 m2 and 2 meters in depth for a total volume of soil of 160 m3. With a 20% contingency (see paragraph 428), the total cost amounts to USD 57,600 for Payamino 3.

p. Payamino 4

530.
Ecuador claims USD 20,502,145 for the remediation of 20,529.60 m3 of soil covering an impacted area of 10,807 m2.1056 Burlington accepts that 510 m2 are impacted corresponding to 730 m3 of soil requiring remediation for an estimated USD 261,000.1057 It argues, however, that Ecuador "ignores the operational history of the site" and points to the discharges of oil-based drilling muds by CEPE in the unlined pit to the west of the platform.1058 More specifically, Burlington submits that two areas make up the impacted 510 m2, namely 230 m2 in Area 1MT (PA04) corresponding to 230 m3 of soil and a cost of USD 82,000, and 280 m2 in Area 1MT (PA14) corresponding to 500 m3 and