Agrifood Value Chains

The pilot explores the strengths and limitations of using EO data to assess nature-related dependencies and impacts linked to investee supply chains. It uses EO data and open-source datasets from initiatives such as Universal Mill List, as well as proprietary databases, to locate operations and suppliers and assess metrics relating to land conversion, carbon stock, fire incidence and biodiversity.

In addition to using raw satellite imagery, the pilot draws on high-quality, analysis-ready products derived from ESA Sentinel 1 and Sentinel 2 data. These include dedicated oil palm classification datasets generated using machine learning, which significantly improve the accuracy of deforestation analysis. This is particularly important in tropical regions, where planted oil palm and natural forests can appear similar in satellite imagery. The pilot also uses above-ground biomass datasets developed under the ESA Climate Change Initiative (CCI), which provide long-term, well-validated estimates of biomass and carbon stocks.

The analysis in this pilot is conducted at the level of individual processing mills, providing an assessment of land conversion and other nature-related metrics linked to specific supply chain nodes. This directly addresses previous survey findings which highlighted limited examples of combining satellite data with company-disclosed supplier lists, as well as the need for precise geolocation data.

The pilot also includes an interactive dashboard that allows early adopters to link company disclosures with EO analysis. This supports a more accessible interpretation of EO-derived metrics and facilitates their integration into risk assessment, client engagement and decision-making processes.

This pilot also assesses how EO data-derived metrics align with those of recognised standards and frameworks:

• Organisations following the recommendations and guidance of the Taskforce on Nature-related Financial Disclosures (TNFD), are required to disclose their status on a number of land use change indicators, including total spatial footprint, extent of land use change by type of ecosystem and type of business activity.
  
• The TNFD’s additional sector guidance for food and agriculture also recommends that organisations:
  
  • Disclose the proportion of supply chain area with native vegetation; the proportion of production volume from land controlled, managed or sourced from that is determined to be deforestation and conversion-free; and land use efficiency (ha land/kg of product).
  • Set targets on natural vegetation/habitat in food production. For example, placing 10%/20%/25% (per square kilometre) of agricultural land under natural and diverse vegetation by 2030
  • Set a deforestation-free target which includes no deforestation for primary deforestation-linked commodities and no conversion of natural ecosystems target (the latter in line with Science Based Targets Network methods).
• The CDP Forest Questionnaire asks companies to track progress on eliminating conversion and degradation of all land-based natural ecosystems, including forests.

The Sustainability Accounting Standards Board (SASB) Agricultural Products Standard requires companies to disclose production by principal crop, number of processing facilities and total land area under active production alongside a description of strategies to manage land use and ecological impacts.

The LEON Requirements Baseline document provides insights into the needs, barriers and expectations of organisations leveraging EO data in the context of nature finance. 15 out of 27 respondents who completed the survey answered the agrifood section of the survey.

The survey found most respondents are at an early stage of assessing nature-related dependencies and impacts linked to their investments, and lack the precise data on clients and their supply chains that is required to make meaningful assessments. 11 out of 15 respondents collect information on land conversion policies of clients/investee companies and seven collect information on sourcing locations of deforestation risk commodities.

Many respondents reported challenges identifying sourcing locations for nature risk commodities. Some respondents had experience using satellite data to monitor clients and their suppliers, but 11 out of 15 said that they need exact geolocation data on suppliers in order to make meaningful assessments. Most respondents described challenges with using company-disclosed supplier lists for nature-related risk assessments. Respondents consider EO data to be useful for overcoming these challenges in identifying suppliers and monitoring nature metrics.

The pilot will be presented as an example for other financial institutions to follow. Guidance and documentation will be produced for use by financial institutions that are maturing on their nature-related journey and will support capacity building to carry out location-specific assessment and monitoring of prioritised sectors, clients or assets.