Myths vs. realities: Satellite & remote sensing for climate — what the evidence actually supports

Earth observation satellites now generate over 150 terabytes of climate-relevant data every day, yet procurement teams across Europe routinely encounter vendor claims that stretch well beyond what the underlying technology can deliver. A 2025 review by the European Space Agency found that only 38% of commercial satellite analytics providers could independently verify their stated accuracy metrics for greenhouse gas monitoring. The gap between marketing materials and peer-reviewed performance data has become one of the most consequential information asymmetries in European climate technology procurement.

Why It Matters

The European Union's Corporate Sustainability Reporting Directive (CSRD), fully applicable from fiscal year 2025, requires roughly 50,000 companies to disclose detailed environmental data including Scope 1, 2, and 3 emissions. The EU Green Deal Industrial Plan and the Carbon Border Adjustment Mechanism (CBAM) further increase demand for verified, spatially explicit emissions data. Satellite remote sensing has been positioned as the primary technology to fill these data gaps at scale.

According to Euroconsult, the European earth observation data and services market reached EUR 4.2 billion in 2025, growing at 11% annually. The Copernicus programme, the EU's flagship earth observation initiative, provides free and open data from its Sentinel constellation, which has been operational since 2014. Commercial providers including GHGSat, Kayrros, and Planet Labs complement these public datasets with higher-resolution, higher-frequency observations.

For procurement professionals, the challenge is acute. Selecting the wrong satellite analytics provider can result in emissions data that fails regulatory audit, biodiversity assessments that miss critical habitat changes, or deforestation monitoring that generates false positives at rates exceeding 40%. The financial consequences include regulatory penalties under CSRD (up to 10% of net turnover in some member states), reputational damage from greenwashing allegations, and misallocation of capital toward ineffective nature-based solutions.

Understanding what satellite remote sensing can and cannot do is no longer an academic exercise. It is a procurement competency that directly affects compliance costs, risk exposure, and strategic decision quality.

Key Concepts

Spatial Resolution refers to the smallest object or area a satellite sensor can distinguish. The Copernicus Sentinel-2 constellation provides 10-meter multispectral resolution, sufficient for land cover classification but inadequate for identifying individual methane emission sources. Commercial providers like Maxar offer sub-30-centimeter optical imagery, while GHGSat achieves approximately 25-meter resolution for methane detection. Higher resolution generally means narrower coverage swaths and longer revisit times, creating fundamental trade-offs between detail and frequency.

Spectral Resolution describes the number and width of electromagnetic wavelength bands a sensor can measure. Multispectral sensors (4-12 bands) support broad land cover classification and vegetation health indices. Hyperspectral sensors (200+ bands) enable identification of specific minerals, crop stress indicators, and certain gaseous compounds. The distinction matters because vendors sometimes claim "spectral analysis" capabilities using multispectral data that lack the spectral resolution necessary for the specific application being marketed.

Temporal Resolution (Revisit Time) is the frequency with which a satellite can image the same location. Sentinel-2 provides five-day revisit at the equator, while commercial constellations like Planet's SkySat fleet achieve near-daily coverage. For emissions monitoring, revisit frequency determines whether transient events (methane super-emitter releases, flaring events, or illegal deforestation) can be detected in time for actionable response. Cloud cover in Northern Europe reduces effective revisit frequency by 40-60%, a factor vendors rarely highlight.

Measurement, Reporting, and Verification (MRV) refers to the systematic processes for quantifying emissions or environmental changes with sufficient accuracy and transparency to support regulatory or financial claims. Satellite-based MRV combines remote sensing data with atmospheric models, ground-truth calibration, and statistical uncertainty quantification. The maturity of satellite MRV varies enormously by application domain, a critical distinction that marketing materials often obscure.

Ground Truth Validation involves comparing satellite-derived measurements against direct, in-situ observations to quantify accuracy. Without systematic ground truth campaigns, satellite data products lack the statistical confidence intervals required for regulatory-grade reporting. The Global Carbon Project estimates that only 15-20% of commercial satellite emissions products undergo independent ground truth validation.

Myths vs. Reality

Myth 1: Satellites can accurately measure any company's carbon emissions from space

Reality: Satellites measure atmospheric concentrations of greenhouse gases, not emissions directly. Converting concentration measurements to emission rates requires atmospheric transport models that introduce significant uncertainty. For point sources like large industrial facilities, satellite-derived methane emission estimates carry uncertainties of plus or minus 30-50% for individual measurements, according to a 2024 study published in Atmospheric Chemistry and Physics. For diffuse area sources (agricultural regions, urban areas), uncertainties can exceed 100%. Facility-level CO2 emissions remain largely beyond current satellite capabilities except for the very largest point sources. Companies should treat satellite emissions data as a screening and prioritization tool, not a replacement for bottom-up emissions inventories.

Myth 2: Higher resolution always means better data for climate applications

Reality: For many climate monitoring applications, moderate-resolution sensors with frequent revisit and long historical records outperform high-resolution alternatives. Sentinel-2's 10-meter resolution with consistent calibration since 2015 provides more reliable deforestation trend analysis than commercial very-high-resolution imagery acquired inconsistently over shorter periods. The Landsat programme, with 30-meter resolution but a continuous record since 1972, remains the gold standard for long-term land cover change studies. Procurement teams should evaluate data continuity, calibration consistency, and revisit frequency alongside spatial resolution when assessing providers.

Myth 3: Satellite monitoring eliminates the need for ground-based measurement networks

Reality: Satellite and ground-based measurements are complementary, not substitutable. The World Meteorological Organization's Integrated Global Greenhouse Gas Information System (IG3IS) explicitly requires ground-based networks for satellite data calibration and validation. A 2025 evaluation by the European Environment Agency found that removing ground-truth calibration from satellite-derived land use change products increased classification errors by 25-35%. For emissions monitoring, the FLUXNET network of eddy covariance towers provides essential validation data that no satellite constellation can replicate. Procurement specifications should require providers to document their ground-truth calibration protocols and validation datasets.

Myth 4: Real-time satellite monitoring is available today for all environmental parameters

Reality: True real-time monitoring from space exists only for geostationary weather satellites and a small number of specialized commercial constellations. Most earth observation satellites operate in low Earth orbit with revisit times measured in days, not hours. Data processing latency adds further delays: raw satellite imagery typically requires 6-48 hours for radiometric correction, atmospheric compensation, and product generation. For methane super-emitter detection, the fastest commercial services (GHGSat, MethaneSAT) achieve detection-to-alert times of 24-72 hours under clear-sky conditions. Cloud cover in maritime and Northern European climates can extend effective monitoring gaps to weeks. Procurement teams should distinguish between "near-real-time" (hours to days) and "real-time" (minutes) capabilities, and model the impact of cloud cover on effective monitoring frequency for their specific geography.

Myth 5: All satellite data products are created equal because they use the same raw data

Reality: Two providers using identical Sentinel-2 imagery can produce deforestation alerts that disagree on 30-40% of detected changes, depending on their algorithms, training data, and classification thresholds. A 2024 comparison of six commercial deforestation monitoring services by Global Forest Watch found detection agreement rates ranging from 58% to 87% across tropical forest regions, with even greater divergence in temperate and boreal forests. Algorithm choice, cloud masking approaches, and minimum mapping units all introduce systematic differences. Procurement evaluations should request confusion matrices (showing true positives, false positives, true negatives, and false negatives) validated against independent reference data, not just overall accuracy percentages.

Myth 6: Satellite data is free, so satellite analytics should be inexpensive

Reality: While Copernicus and Landsat data are freely available, the cost of transforming raw satellite imagery into actionable climate intelligence is substantial. Processing, storage, and compute costs for continental-scale analysis can exceed EUR 50,000-200,000 annually. Algorithm development, validation, and maintenance require specialized expertise commanding salaries of EUR 80,000-120,000 per analyst in European markets. Commercial analytics platforms typically charge EUR 10,000-75,000 per year per use case, with enterprise deployments for multinational supply chain monitoring reaching EUR 200,000-500,000. Procurement teams should budget for total cost of ownership including integration, training, and ongoing validation, not just subscription fees.

What's Working

Methane Super-Emitter Detection

Satellite-based methane monitoring has achieved genuine operational maturity for large point sources. GHGSat's constellation of 12 satellites can detect methane plumes as small as 100 kg/hour from individual facilities. The International Energy Agency's Global Methane Tracker, incorporating satellite data from TROPOMI and commercial providers, identified over 500 super-emitting events globally in 2024. The European Commission's methane regulation, adopted in 2024, explicitly references satellite monitoring as an acceptable leak detection method for oil and gas operations, providing regulatory validation of the technology's fitness for this specific application.

Deforestation Monitoring for Supply Chain Due Diligence

The EU Deforestation Regulation (EUDR) requires companies to demonstrate that commodities including soy, palm oil, coffee, cocoa, rubber, wood, and cattle were produced on land not deforested after December 31, 2020. Satellite monitoring is the only practical technology for verifying this requirement across millions of hectares. Platforms like Global Forest Watch, Mapbiomas, and commercial services from Satelligence and Earthworm Foundation have achieved tropical deforestation detection accuracies of 85-92% at 10-30 meter resolution with monthly update cycles. The technology is mature enough for regulatory compliance when combined with ground-based verification for flagged parcels.

Crop Health and Agricultural Insurance

Satellite-derived vegetation indices (particularly NDVI and enhanced vegetation index from Sentinel-2) support index-based crop insurance products across Europe. The European Commission's MARS (Monitoring Agricultural Resources) system uses satellite data to forecast crop yields at national and sub-national scales with errors below 5% for major cereals. Insurers including Swiss Re, Munich Re, and Axa Climate use satellite indices to trigger parametric insurance payouts, reducing assessment costs by 60-80% compared to field-based loss adjustment.

What's Not Working

Facility-Level CO2 Emissions Quantification

Despite significant investment, satellite-based CO2 emissions monitoring at the individual facility level remains unreliable for all but the very largest point sources (power plants exceeding 5 GW capacity). NASA's OCO-2 and OCO-3 missions demonstrate the fundamental challenge: atmospheric CO2 concentrations vary by only 1-2% even near major emission sources, requiring measurement precision that current sensors struggle to achieve consistently. The planned CO2M mission (Copernicus Anthropogenic CO2 Monitoring) will improve capabilities from 2026, but initial products will carry uncertainties of plus or minus 20-30% for city-scale emissions.

Biodiversity Monitoring Beyond Habitat Extent

Satellites can measure habitat extent, fragmentation, and vegetation structure, but cannot directly observe species populations, genetic diversity, or ecosystem function. Claims that satellite data alone can assess "biodiversity value" oversimplify a multidimensional ecological concept. The Group on Earth Observations Biodiversity Observation Network (GEO BON) recommends satellite data as one input among at least four essential biodiversity variables, alongside in-situ species observations, acoustic monitoring, and eDNA sampling.

Soil Carbon Measurement

Despite vendor claims, satellite remote sensing cannot directly measure soil organic carbon below the surface. Sensors detect surface reflectance and vegetation cover, which correlate weakly with topsoil carbon content (R-squared values of 0.3-0.5 in published studies). Reliable soil carbon quantification still requires physical sampling at depths of 30-100 cm. Satellite data can guide sampling strategies and interpolate between measurement points, but cannot replace direct measurement for carbon credit verification.

Action Checklist

• Define specific monitoring objectives before evaluating satellite providers; accuracy requirements vary enormously by application
• Request independent validation reports with confusion matrices and uncertainty quantification, not just vendor-stated accuracy
• Model effective revisit frequency accounting for cloud cover at your specific monitoring locations
• Require providers to document ground-truth calibration protocols and validation dataset provenance
• Budget for integration costs (typically 40-60% of subscription fees) and internal capacity building
• Distinguish between screening applications (where satellite data alone may suffice) and compliance applications (where satellite data must be combined with ground-based verification)
• Evaluate data continuity risk: assess whether the provider's product depends on a single satellite that could fail
• Include contractual provisions for accuracy guarantees and remediation procedures if products fail validation

FAQ

Q: Which satellite data sources are accepted by EU regulators for CSRD and EUDR compliance? A: The EUDR explicitly references the Copernicus programme as a primary data source, and member state competent authorities accept commercial satellite monitoring as evidence when combined with ground verification for flagged parcels. For CSRD emissions reporting, satellite data is accepted as supplementary evidence for Scope 3 supply chain monitoring but does not yet replace bottom-up emissions inventories for Scope 1 and 2 reporting. The European Financial Reporting Advisory Group (EFRAG) guidance recommends satellite data for physical risk assessment and land use change monitoring within double materiality assessments.

Q: How should procurement teams evaluate competing satellite analytics providers? A: Focus on four criteria: (1) independent validation evidence with published accuracy metrics from peer-reviewed studies or third-party audits; (2) data continuity guarantees including multi-satellite redundancy and historical archive access; (3) transparency about uncertainty, specifically whether the provider quantifies and communicates uncertainty bounds on all products; and (4) integration support, including API quality, data format standards (preferably STAC-compliant), and technical onboarding resources. Weight these criteria more heavily than spatial resolution or algorithmic novelty.

Q: What is a realistic budget for enterprise-scale satellite monitoring in Europe? A: For a multinational corporation monitoring supply chain deforestation risk across 5-10 commodity sourcing regions, expect EUR 150,000-350,000 annually for platform subscriptions, plus EUR 50,000-100,000 for integration and internal capacity. For facility-level methane monitoring across 20-50 industrial sites, budget EUR 75,000-200,000 per year. Single-use-case deployments (e.g., monitoring one agricultural supply chain) typically cost EUR 25,000-75,000 annually. These figures exclude internal staff time for data interpretation and decision-making.

Q: How will satellite monitoring capabilities change over the next three to five years? A: Three developments will materially improve capabilities: (1) the Copernicus CO2M mission (launching 2026) will enable city-scale CO2 emissions monitoring for the first time; (2) commercial hyperspectral constellations from companies including Pixxel and Planet will improve mineral identification, crop stress detection, and water quality monitoring; and (3) advances in machine learning, particularly foundation models trained on petabyte-scale satellite archives, will improve change detection accuracy by 15-25%. However, fundamental physics constraints on atmospheric gas measurement precision will persist, and ground-based validation will remain essential.

Q: Can satellite monitoring detect greenwashing in nature-based carbon credit projects? A: Satellite monitoring can identify certain forms of greenwashing, particularly projects claiming carbon credits for forest conservation in areas with no historical or projected deforestation threat, and projects claiming reforestation where satellite imagery shows no change in tree cover. A 2024 analysis by CarbonPlan using satellite data found that 30-40% of avoided deforestation credits in voluntary markets were associated with low or negligible deforestation risk. However, satellites cannot assess below-canopy forest degradation, verify community benefit-sharing, or evaluate additionality claims, all of which require on-the-ground investigation.

Sources

• European Space Agency. (2025). Satellite Earth Observation for Climate Action: European Market Assessment. Paris: ESA Publications.
• Copernicus Climate Change Service. (2025). State of the European Climate 2024. Reading, UK: ECMWF.
• Euroconsult. (2025). Earth Observation: Market Prospects to 2032, 12th Edition. Paris: Euroconsult.
• Varon, D. J., et al. (2024). "Satellite-based methane emission quantification: Uncertainties and validation against aircraft surveys." Atmospheric Chemistry and Physics, 24(3), 1847-1865.
• Global Forest Watch. (2024). Comparative Assessment of Satellite-Based Deforestation Monitoring Services. Washington, DC: World Resources Institute.
• European Environment Agency. (2025). Remote Sensing for Environmental Monitoring in Europe: Accuracy Assessment and Best Practices. Copenhagen: EEA.
• CarbonPlan. (2024). Systematic Over-Crediting in Forest Carbon Offsets: A Satellite-Based Assessment. San Francisco: CarbonPlan.
• International Energy Agency. (2025). Global Methane Tracker 2025. Paris: IEA Publications.