Remote sensing for ESG due diligence
Satellite data is increasingly used to verify environmental claims — from deforestation commitments to emissions. What satellite-derived ESG indicators can and cannot reliably measure.
The verification problem in ESG
Environmental, social, and governance reporting has historically relied on self-reported data. Companies disclose emissions, land use, water consumption, and supply chain practices through questionnaires and audit processes that are resource-intensive, geographically limited, and subject to the incentive problems inherent in self-reporting. Satellite-based verification offers an independent, spatially comprehensive, and continuously updated alternative for a subset of environmental claims.
The subset matters. Remote sensing is well-suited to verifying claims about land cover, land-use change, surface water, and some atmospheric properties. It is poorly suited to verifying claims about internal processes, social conditions, or governance structures that have no direct surface expression. Understanding precisely what satellite data can and cannot reliably measure is a prerequisite for using it responsibly in due diligence workflows.
What satellite data can reliably measure
Deforestation and land-use change are among the best-validated satellite applications in ESG due diligence. High-resolution optical imagery combined with change detection algorithms can identify forest clearing events with high accuracy and short latency (days to weeks after clearing in cloud-free conditions). This makes satellite-based deforestation monitoring actionable for supply chain sourcing decisions, particularly for commodities with high deforestation exposure (soy, palm oil, cattle, timber).
Surface water dynamics — reservoir levels, wetland extent, flood inundation — are reliably measurable using synthetic aperture radar (SAR), which penetrates cloud cover and operates independently of solar illumination. Vegetation health indices derived from near-infrared and red-edge bands provide leading indicators of drought stress and ecosystem degradation. Thermal infrared observations can detect industrial heat sources and, with appropriate atmospheric correction, estimate surface temperature anomalies indicative of certain industrial emissions.
What satellite data cannot reliably measure
The limits of satellite-based ESG verification are as important as its capabilities. Emissions from specific industrial sources — methane from livestock operations, N2O from agricultural soils, CO2 from combustion — can be detected at large scales using hyperspectral sensors, but attribution to specific sites at the resolution required for facility-level due diligence remains technically challenging and subject to significant uncertainty. Conflating satellite-derived area estimates (reliably measured) with emissions estimates (much less reliable) is a common source of overconfidence in satellite-based ESG assessments.
Social and governance indicators have almost no direct satellite expression. Labor conditions, community consultation processes, corruption in permitting — these are not observable from orbit. Due diligence workflows that rely heavily on satellite data risk systematically underweighting social and governance factors relative to environmental ones, not because the former are less important, but because they are less measurable with available remote sensing technology.
Integrating satellite data into due diligence workflows
Effective integration of satellite data into ESG due diligence requires a layered approach. Satellite screening identifies geographic areas or suppliers with elevated environmental risk based on observable proxies (high deforestation rates, proximity to protected areas, declining vegetation health). This narrows the scope of more intensive investigation — on-site audits, certification verification, stakeholder engagement — to the cases where it is most needed.
The screening layer needs to be calibrated to the risk tolerance of the due diligence process. High-recall screening that flags all potentially problematic cases generates audit burden; high-precision screening that minimizes false positives risks missing real issues. The right operating point depends on the cost structure of the downstream investigation process and the materiality thresholds of the organization's environmental commitments.
Forward-looking applications
The satellite data landscape is evolving rapidly. Commercial hyperspectral constellations are increasing the feasibility of facility-level emissions attribution. Higher revisit rates (daily or near-daily for some commercial systems) are enabling near-real-time monitoring. SAR interferometry is opening applications in ground subsidence and permafrost degradation monitoring relevant to physical climate risk assessment.
Organizations building satellite-based ESG verification capabilities now will have a significant advantage as these data sources mature. The modeling and validation infrastructure developed for current applications — change detection pipelines, anomaly detection systems, spatiotemporal data stacks — transfers directly to new sensor types and new applications. The investment in rigorous methodology compounds over time in ways that ad hoc applications of new data sources cannot match.
Working on a complex AI problem?
Our research orientation means we're always interested in technically challenging problems. Let's explore what's possible.