Robust ESG data for biodiversity: towards a spatially-sensitive approach to Sustainable Finance
Lead Research Organisation:
UK CENTRE FOR ECOLOGY & HYDROLOGY
Department Name: Biodiversity (Wallingford)
Abstract
This project seeks to construct an interdisciplinary expert and stakeholder community to uncover and pilot using 'discoverable' data to construct a usable and scientifically rigorous biodiversity dataset. To achieve that, the project will be led by the Oxford Sustainable Finance Group, in partnership with WCMC and Catapult, in four key stages:
Stage One: Network building
Stage Two: construction of a strawman for integrating biodiversity in ESG metrics
Stage Three: Testing and revising the strawman
Stage Four: White Paper, Next Steps and Capacity Building
The project will result in:
(1) a community of committed stakeholders;
(2) a concept note of use cases of biodiversity metrics, existing data and capacity availability and gaps;
(3) a tested blueprint of what a biodiversity finance database would look like;
(4) an agreed-upon plan of next steps forward.
Stage One: Network building
Stage Two: construction of a strawman for integrating biodiversity in ESG metrics
Stage Three: Testing and revising the strawman
Stage Four: White Paper, Next Steps and Capacity Building
The project will result in:
(1) a community of committed stakeholders;
(2) a concept note of use cases of biodiversity metrics, existing data and capacity availability and gaps;
(3) a tested blueprint of what a biodiversity finance database would look like;
(4) an agreed-upon plan of next steps forward.
People |
ORCID iD |
| Richard Pywell (Principal Investigator) |
| Description | Our two use-cases covered a broad perspective within sustainable finance with the Brazilian cattle farming use-case, an example of green finance - using the power of investment to drive the transition of a sector with a poor historic track record to more sustainable operations and the deployment of nature-based solutions in UK water utility use case, an example of financing green - to drive investment to nature-positive outcomes. The two examples also covered different sectors, geographies, financial assets and AI modelling techniques, providing an overview on how AI could be applied to different challenges relating to nature's integration into finance. Our asset level modelling approach, with the help of AI, exploits a wide range of detailed local heterogenous data that can attribute biodiversity loss and gain to each actor in the value-chain. This allows investors, and others, to see the specific impacts of investments, mitigates greenwashing and, with Generative AI, could provide a means for accurate, frequent and automated disclosure. Consequently, we were able to propose E-scores for each organisation in the value-chain. Uncertainty is modelled explicitly and this both informs and directs further efforts for data gathering and model refinement. Although the approach described above represents an overall improvement in biodiversity risk measurement robustness, there are also limitations. Practically, AI techniques typically require much more computing power than more traditional methods. Furthermore, our approach requires asset-level data to achieve the necessary granularity, which has proven challenging as companies either do not have data to the required granularity or they are unwilling to share it voluntarily. To conduct scenario analysis, we would have to specifically build scenarios, which might require a new network although some scenarios can be run on the same network by tweaking initial conditions. Finally, our modelling approach is currently quite bespoke and rather involved, requiring intensive engagement and time investment on behalf of the financial institutions, supply chain partners and ecological experts, which could be challenging if not a company priority. However, by exploiting the Bayesian approach, we can present our model in an intuitive and accessible manner with diagrams of actor relationships, which provides structure to allow meaningful engagement with experts with different areas of expertise. The proposal envisioned a strong component of network building across the research, business, governmental, and investment communities. It was envisioned that this network will collaboratively engage candidly in interdisciplinary conversations on existing data use, data limitations, and future needs in nature-positive investment in the UK, Brazil, and Southeast Asia. However, we experienced difficulties in engaging with the investment community, particularly the banking sector since the beginning of the project. The challenge of identifying AI-ready projects relating to biodiversity and finance should not be underestimated. Due to the interdisciplinarity of the topic, a viable use case requires people in finance, supply and ecology all interested and willing to engage. We believe that this will happen in the near future. However, there is currently little risk incentive for banks to commit sufficient resources. Many financial institutions we engage with were in the early stages of exploring their nature impacts and dependencies. The current landscape of nature discourse in financial institutions involves the definition of scope, metrics, and methodology of stocktaking and strategising the management of various nature risks. Currently, the quality of data infrastructure and level of understanding of key data analytical demands require refinement to effectively co-create specific formulae required for AI integration. Therefore, to establish these use cases, we had to delve much deeper into the ecological and financial technicalities than expected to outline their relationships to be able to identify where AI could be applied. As a result we found ourselves generating financial models that capture the transitional risks associated with investments, including reputational and financial risks. Although, without investor validated models, it is difficult to determine exactly how AI would be useful in each use-case. With that, our models should be taken with a "pinch of salt" and, instead, represent a "flavour" of potential AI applications in this space. Through developing the models ourselves, we were able to engage with all partners in a coherent way on the illustrated problem. With a proposed model and, particularly with the network approach, we were able to engage with financial institutions and demonstrate the detail required for a successful partnership. Furthermore, we found that the model anchored the conversation and that we were able to apply each organisation's expertise in a meaningful and specific manner without requiring much cross-disciplinary debate. It was particularly challenging yet important to leverage each type of required expertise, be it ecological or financial, as we were unable to identify experts with a deep interdisciplinary understanding of nature-finance relationships. Objectives Yes Expand Further Details Taken Forward To successfully implement AI solutions, we needed engagement with actors across the problem space, all contributing their portion at the required detail. In our engagements, financial institutions could not articulate what information was needed but currently lacking to integrate nature into financial decision-making. Without a clear understanding of the gaps in information required, it is not possible to identify existing data and models that might be able to fulfil their requirements and it is inappropriate to spend a lot of effort to employ potential AI solutions at this moment. For AI to be an appropriate solution, the problem and information required for decision-making, including appropriate levels of granularity and uncertainty, must be clearly defined and articulated. In the case of finance and nature, the relationship between ecological and financial systems as well as the specific information requirements of financial institutions, must be better understood and articulated. As these pre-conditions have not been met, this project will not continue into Phase 2 of the Integrating Finance and Biodiversity project. Oxford Sustainable Finance Group is creating a team focussed on finance for biodiversity and we are encouraging a modelling component to this which will exploit the results from the REDB project. |
| Exploitation Route | To successfully implement AI solutions, we needed engagement with actors across the problem space, all contributing their portion at the required detail. In our engagements, financial institutions could not articulate what information was needed but currently lacking to integrate nature into financial decision-making. Without a clear understanding of the gaps in information required, it is not possible to identify existing data and models that might be able to fulfil their requirements and it is inappropriate to spend a lot of effort to employ potential AI solutions at this moment. For AI to be an appropriate solution, the problem and information required for decision-making, including appropriate levels of granularity and uncertainty, must be clearly defined and articulated. In the case of finance and nature, the relationship between ecological and financial systems as well as the specific information requirements of financial institutions, must be better understood and articulated. As these pre-conditions have not been met, this project will not continue into Phase 2 of the Integrating Finance and Biodiversity project. Oxford Sustainable Finance Group is creating a team focussed on finance for biodiversity and we are encouraging a modelling component to this which will exploit the results from the REDB project. |
| Sectors | Agriculture Food and Drink Digital/Communication/Information Technologies (including Software) Environment Financial Services and Management Consultancy |
| Description | To successfully implement AI solutions, we needed engagement with actors across the problem space, all contributing their portion at the required detail. In our engagements, financial institutions could not articulate what information was needed but currently lacking to integrate nature into financial decision-making. Without a clear understanding of the gaps in information required, it is not possible to identify existing data and models that might be able to fulfil their requirements and it is inappropriate to spend a lot of effort to employ potential AI solutions at this moment. For AI to be an appropriate solution, the problem and information required for decision-making, including appropriate levels of granularity and uncertainty, must be clearly defined and articulated. In the case of finance and nature, the relationship between ecological and financial systems as well as the specific information requirements of financial institutions, must be better understood and articulated. As these pre-conditions have not been met, this project will not continue into Phase 2 of the Integrating Finance and Biodiversity project. Oxford Sustainable Finance Group is creating a team focussed on finance for biodiversity and we are encouraging a modelling component to this which will exploit the results from the REDB project. |
| First Year Of Impact | 2024 |
| Sector | Agriculture, Food and Drink,Digital/Communication/Information Technologies (including Software),Environment,Financial Services, and Management Consultancy |
| Impact Types | Economic |
| Description | Integrating Finance and Biodiversity for a Nature Positive Future |
| Amount | £3,021,258 (GBP) |
| Funding ID | NE/Z503368/1 |
| Organisation | Natural Environment Research Council |
| Sector | Public |
| Country | United Kingdom |
| Start | 01/2024 |
| End | 04/2026 |
| Title | Bayesian Finance Impact Model |
| Description | The project report "Robust ESG data for biodiversity: towards a spatially-sensitive approach to Sustainable Finance" describes two models that integrate data and capture complex down stream relationships for a greening finance use-case (beef supply in Brazil) and a financing green use-case (Domestic water supply in the UK). Our first model adopts a Bayesian approach to analyze relationships within a complex system involving Brazilian banks, meat processors like JBS, abattoirs, and farms. The model is used to assess risks related to environmental impact, regulatory compliance, and reputational damage for investors in the supply chain. Key points include: (1) Bayesian Approach: Utilizes Bayesian theory to encode probabilistic relationships between parameters, allowing for the expression of weak information and adaptability to new data; (2) Supply Chain Dynamics: Describes the intricate relationships in the supply chain, highlighting how Brazilian banks provide debt to meat processors, sourced from abattoirs and farms, with potential environmental implications like deforestation and human rights abuses; (3) Investor Risks: Identifies reputation loss and unrealized financial gain as chief concerns for financiers, proposing probabilistic models to assess the impact of investments on nature and regulatory compliance; (4) Data Challenges and Solutions: Addresses data gaps related to Forest Code compliance and the Nature Risk Profile, proposing AI-driven solutions such as satellite imagery analysis and remote sensing to validate reports and monitor ecosystem impact; (5) Modelling Investor Risks: Develops probabilistic frameworks to model asset relationships, assess reputational and financial risks, and calculate E-scores reflecting investments' impact on nature; (6) Scaling the Model: Explores scaling the model regionally or nationally to capture broader supply chain dynamics and potential scenario analyses considering legislative changes or divestment strategies; (7) Combining Corporate Controversy Reports: Demonstrates how text-based reports can be integrated into the Bayesian model to reduce uncertainty in reputational risk assessment. Overall, the text outlines a comprehensive approach to assessing the environmental and reputational risks associated with investments in the Brazilian meat supply chain, leveraging Bayesian modeling and AI-driven data analysis. Our second proposed model aims to address an optimization problem that balances the expenses related to fines and operational costs against payments for ecosystem services (PES). This approach incorporates a time series framework to allow for strategic adjustments at specific intervals, aligning with bond (and loan) refinancing schedules. In a particular scenario, pension funds have tasked an asset manager with investing their capital for appropriate returns, often seeking low-risk, long-term investments. The asset manager invested in bonds issued by Thames Water, a UK water utility regulated by the government for various performance metrics, including pollution and leakage targets, with fines imposed for non-compliance. To mitigate risks associated with environmental performance, pension funds, represented by RLAM, are pressuring Thames Water to reduce pollution and negative biodiversity impact through a PES scheme, which involves compensating upstream farmers to implement solutions that reduce water pollution. Two nature-based solutions (NbS) options are explored, focusing on buffer strips to reduce nitrate and pesticide runoff into water sources. E-planner, developed by UKCEH, assesses the potential applicability of NbS using environmental indicators. Proposed AI-driven extensions to E-planner include combining crop planting history with environmental maps and predicting spatial suitability of NbS. The NbS options involve establishing buffer strips either on cultivated land or improved permanent grassland, with specific requirements and monitoring mechanisms. Farmers must provide various data and milestone reports, which can be analyzed using AI tools. The model aims to assist Thames Water in investing in NbS across farms in the Thames catchment area, selecting cost-effective options to mitigate nitrate and pesticide runoff. It involves monitoring farms, discontinuing ineffective NbS strategies, and potentially switching to optimal strategies during the bond's existence. Additionally, the model proposes predicting river pollution levels based on NbS efficacy, estimating chemical cleanup costs, optimizing NbS allocation to fields, and introducing an E-score to assess environmental outcomes. Overall, the model integrates financial considerations with environmental goals to optimize investment in nature-based solutions while addressing regulatory requirements and mitigating risks related to water pollution and ecosystem health. |
| Type Of Material | Computer model/algorithm |
| Year Produced | 2024 |
| Provided To Others? | No |
| Impact | None so far. The REDB Report will be published end of March 2024 and we will use it to seek further feedback and engagement. |
| Description | Biodiversity & Water Quality Uses Cases |
| Organisation | Severn Trent Water |
| Country | United Kingdom |
| Sector | Private |
| PI Contribution | Showcase of environmental data and modelling approaches to support Green Finance requirements. Identification of ways to use data to meet specific requirements of case study partners. Development of workflows to translate data into required outputs and metrics (e.g. collation of natural capital baseline data, development of environmental management scenarios and modelling of environmental impacts). Co-development of refinements to workflow and presentation of outputs. Examples only. |
| Collaborator Contribution | Identification of key requirements for Green Finance case study. Outreach to investors / identification of case study partner. |
| Impact | Detailed question and response documentation shared between UKCEH and RLAM outlining status quo of water companies reporting on biodiversity and potential requirements / gaps. Future collaboration with Severn Trent (meetings arranged) to gather baseline data > develop metrics > case study > final outputs. |
| Start Year | 2023 |
| Description | Activity Title Interviews and Feedback from Commercial, Finance and Academic Organisations |
| Form Of Engagement Activity | A talk or presentation |
| Part Of Official Scheme? | No |
| Geographic Reach | International |
| Primary Audience | Professional Practitioners |
| Results and Impact | Result Description The project report "Robust ESG data for biodiversity: towards a spatially-sensitive approach to Sustainable Finance" presents a comprehensive overview of engagement efforts and insights derived from stakeholders involved in two distinct projects: the Brazil Beef Supply Use-Case (greening finance) and the Thames Water Use-Case (financing green). In the Brazil Beef Supply Use-Case, the engagement included interviews with Brazilian financial institutions and discussions with biodiversity risk tool providers like Trase and BVRio. The focus was on understanding how these institutions consider biodiversity and nature risks in agricultural sector investments. Key findings indicate that while biodiversity loss is recognized as a growing risk, there's a lack of clarity regarding its integration into financial decision-making. Deforestation emerges as a significant concern, with increasing litigation risks identified. Data availability is acknowledged, but challenges exist in centralizing and making data actionable for decision-makers within banks. Specific suggestions include developing metrics connecting nature risk to legal risk and verifying the accuracy of reported data, particularly Cadastro Ambiental Rural (CAR) data. Notable organizations involved in discussions include JGP, Santander Asset Management, BVRio, BNP Paribas, and Quant Foundry. The Thames Water Use-Case involves engagement with RLAM and the Environmental Change Institute at Oxford University. Discussions revolved around storm overflow damage, water quality improvement costs, and challenges in identifying buffer zones and field margins. RLAM emphasized the need for innovative approaches, especially in addressing operational constraints and regulatory requirements. The Environmental Change Institute provided insights into technology and data sources for monitoring buffers, highlighting the complexity of identifying these zones from satellite imagery. Additionally, discussions with organizations like Quant Foundry and TNFD emphasized the importance of expressing model outputs in relevant frameworks and metrics, as well as the need for collaboration and further research. General feedback from stakeholders highlighted support for the proposed models and methodologies, with suggestions for refinement and collaboration. Overall, the engagements underscore the importance of integrating biodiversity and nature risks into financial decision-making processes and developing actionable metrics to address emerging environmental challenges in the agricultural and water sectors. |
| Year(s) Of Engagement Activity | 2023,2024 |
| Description | NERC Joint Integrating Finance and Biodiversity and Economics of Biodiversity Conference |
| Form Of Engagement Activity | Participation in an activity, workshop or similar |
| Part Of Official Scheme? | No |
| Geographic Reach | International |
| Primary Audience | Study participants or study members |
| Results and Impact | Purpose: presentation and discussion of findings from phase I of the IFB programme (UKCEH Project presentation); discussion of the aims of the IFB programme overall including Phase II (Themes and Flagships) and fit with Innovate UK funding (Nick Wells), planning of collaborative activities for Phase II of the IFB programme; planning of collaborative activities with the aligned Economics f Biodiversity (EoB) programme. Impact: Greater collaboration across the IFB and EoB programmes. |
| Year(s) Of Engagement Activity | 2024 |
| Description | NERC Nature Positive Business Roundtable |
| Form Of Engagement Activity | Participation in an activity, workshop or similar |
| Part Of Official Scheme? | No |
| Geographic Reach | National |
| Primary Audience | Industry/Business |
| Results and Impact | The PI for Science-based Markets for Nature Recovery attended a NERC-convened roundtable of industry and environmental science leaders, to be held in Caxton House, London, on Thursday 11th January 2024. The workshop brought together key leaders in this area to help understand the challenges facing businesses in the construction and infrastructure sectors as they seek to conserve and enhance biodiversity through their activities. Invitees included Network Rail, HS2, National Trust, Task Farce for Nature Disclosures, and Food Matters International. The workshop will help NERC identify opportunities where new investment is required into pre-competitive, business-informed R&D that could make a step change in enabling businesses to achieve their ambitions for clean growth and environmental sustainability. |
| Year(s) Of Engagement Activity | 2024 |