Satellite Deforestation Checks for EUDR: How They Work

Uploading GPS coordinates to a compliance platform and getting a "deforestation-free" result back feels like a black box. Where does that result come from? What data is it based on? And what exactly does it mean for your Due Diligence Statement? This article explains how satellite-based EUDR deforestation checks actually work, so you understand what you're relying on before you submit a DDS.

Why GPS coordinates alone don't prove deforestation-free status

A set of GPS coordinates tells you where a farm is. It doesn't tell you what was there before. A farmer could send you a perfectly formatted polygon in WGS84 decimal degrees for a plot that was covered in closed-canopy forest two years ago and converted to cocoa in early 2021. The coordinates are accurate. The deforestation already happened. The EUDR doesn't care whether you knew.

The regulation's Article 10 requires you to assess whether the plot of land was deforested after December 31, 2020. That requires comparing the current land cover at those coordinates against historical imagery, and that comparison has to be based on something more than the supplier's word. Coordinates are the input to the deforestation check. They're not the check itself.

This is also why "I got coordinates from my supplier" doesn't complete your due diligence. Getting the coordinates is Step 1. Running an actual satellite check against them is Step 2. Your DDS needs evidence for both.

How satellite deforestation detection works (without the jargon)

Satellites capture imagery of the Earth's surface at regular intervals. When you stack those images over time, you can see where forest cover existed at one point and was gone at another. That change signal is what deforestation detection algorithms look for.

The detection process compares canopy cover in a baseline period (pre-2021) against canopy cover in more recent imagery. Where the signal drops significantly — where what was dense green canopy is now bare ground or low vegetation — that's a deforestation event. The system logs the approximate date of the change and the coordinates. For EUDR purposes, any event that falls after December 31, 2020 is within scope.

In practice, the analysis runs a set of coordinates against a precomputed change detection layer. Rather than downloading raw satellite imagery and running classification from scratch for each plot, most platforms (including TracePlot) use change detection datasets that have already processed the raw imagery. Your result comes back in seconds because the computation already happened.

The data sources: Sentinel-2, Hansen Global Forest Watch, EU forest maps

Three data sources form the foundation of most EUDR satellite verification workflows.

Sentinel-2 is the European Space Agency's optical imaging satellite constellation, operated as part of the EU's Copernicus program. It captures imagery at 10-metre resolution and revisits any given point on Earth roughly every 5 days under cloud-free conditions. That combination of resolution and revisit frequency makes it the most practically useful open satellite dataset for deforestation monitoring. A 10-metre pixel means you can detect clearing events at the scale relevant to smallholder plots; a 5-day revisit means the temporal resolution is tight enough to date events with reasonable precision.

Hansen Global Forest Watch is a canopy cover change dataset produced by the University of Maryland and published through the World Resources Institute's Global Forest Watch platform. It's derived primarily from Landsat satellite data (30-metre resolution) and covers global forest cover loss annually from 2000 onwards. The dataset is widely used for EUDR verification because it has consistent global coverage, it's openly licensed, and it provides a time-series that extends well before the 2020 cut-off date.

The EU's own forest reference maps, produced through the Copernicus Land Monitoring Service, provide a complementary layer, particularly for verifying baseline forest cover within EU-adjacent origins and for cross-checking results from the other sources.

No single dataset is perfect. Using multiple sources together, where one dataset flags a potential deforestation signal and another confirms or contradicts it, produces a more reliable result than any source in isolation.

The deforestation cut-off: why December 31, 2020 is the date

The EUDR doesn't prohibit importing commodities grown on land that was ever deforested. It prohibits importing commodities grown on land that was deforested after December 31, 2020. That specific date is written into Article 2 of Regulation (EU) 2023/1115, and it hasn't changed despite the enforcement deadline being pushed back to December 2026.

The cut-off anchors the regulation to a specific policy moment. Forest loss before that date, while environmentally significant, falls outside EUDR scope. A farm cleared in 2019 is out of scope. A farm cleared in 2021 is within scope. For your deforestation check, the only period that matters is January 1, 2021 onwards.

This means when you look at a change detection result, you need the event date, not just a binary "deforestation detected" flag. An algorithm that detects a change but dates it to mid-2019 is an out-of-scope event. One dated to March 2022 is an in-scope deforestation indicator that must be addressed in your Article 10 risk assessment before you can submit a DDS.

What satellite verification can and can't tell you

Satellite-based deforestation checks are genuinely useful for EUDR purposes, but there are real limits to what they can confirm.

What they can tell you: whether a canopy cover change event consistent with deforestation has been detected within or adjacent to a plot boundary after the cut-off date, with what confidence level, and from which data source. A clean result across multiple datasets for a plot in a low-cloud-cover region is meaningful evidence. It won't be airtight, but it's the kind of documented, methodology-backed assessment the regulation requires.

What they can't tell you: whether a change event was actually deforestation (as opposed to a natural disturbance like fire or storm damage), whether cleared land was actually converted to agricultural use, or whether a missing signal means no deforestation occurred versus that cloud cover obscured the imagery. West Africa is where cloud cover uncertainty is most significant in practice. Persistent cloud cover over cocoa-growing regions in Ghana and Ivory Coast means imagery gaps can extend for months. A clean result for a plot in those regions carries more uncertainty than a clean result for a plot in Colombia.

Free tools like the Global Forest Watch map viewer let you manually inspect any coordinates in a browser and are a reasonable first-pass check. They don't produce a downloadable audit record, don't integrate with DDS workflow, and don't give you a documented methodology that you can attach to your compliance file. That gap matters when an authority asks to see how you conducted your risk assessment.

How TracePlot uses free public satellite data for the Article 10 risk check

TracePlot runs the deforestation check automatically once you upload GPS coordinates. The check runs against the Hansen Global Forest Watch change detection dataset and Sentinel-2 derived change layers, covering the period from January 1, 2021 through the date of your check. Each plot gets an individual result: clean, flagged, or uncertain (where cloud cover or imagery gaps reduce confidence in the result).

The underlying data is the same public satellite data available to anyone. The difference is that TracePlot logs the specific dataset version, the date the check was run, and the result for each individual plot, giving you a documented methodology record rather than a screenshot from a web viewer. That record is stored against your DDS and is retrievable if you're audited.

For flagged plots, TracePlot doesn't automatically block the plot from your DDS. A flag means a potential deforestation signal was detected and you need to decide how to handle it: whether that's requesting supplementary documentation from your supplier, commissioning a ground-truth check, or deciding the plot represents too much risk to include. The risk decision stays with you as the operator; the satellite check gives you the evidence base to make it.

If you're preparing your EUDR compliance workflow, the satellite check is one step in a longer process that includes supplier data collection, DDS preparation, and TRACES NT submission. TracePlot's methodology page covers how those steps fit together.

TracePlot runs the deforestation check automatically for every plot you upload, logs a dated methodology record against your DDS, and flags any in-scope signals that need your attention before submission. Plans start from EUR 59/month, with a EUR 49 deposit to get started. Start your EUDR compliance workflow today.