Author: Mosaix

The European Union Deforestation Regulation (EUDR) enacted through Regulation (EU) 2023/1115 shifts the compliance basis for agricultural commodities from document-based traceability to geolocation evidence and land history verification. Coffee, alongside cocoa, palm oil, soy, rubber, timber, and cattle, must now be demonstrably free of association with deforestation or forest degradation after the cut-off date of 31 December 2020.

This article presents the results of a Deforestation Risk Assessment (DRA) and forest-to-coffee conversion analysis for the period 2001–2024 across three major coffee-producing provinces in Vietnam’s Central Highlands: Dak Nong, Lam Dong, and Dak Lak. The assessment is not designed to render a legal verdict on Vietnamese coffee. It is a prioritisation tool — a framework that helps supply chain actors identify where evidence needs strengthening, where geolocation verification should be prioritised, and how compliance resources can be deployed proportionally.

Results reveal sharp differences across provinces and districts. Dak Nong recorded the largest conversion: 21,121 ha or 4.97% of its year-2000 forest cover, concentrated in two dominant districts: Dak Song (10,982 ha, 22.58%) and Dak R’Lap (9,272 ha, 7.73%). Lam Dong recorded 9,518 ha with hotspots in Di Linh (5,769 ha) and Bao Lam (3,276 ha). Dak Lak recorded 5,731 ha with high local intensity in Cu M’gar (9.81%) and Krong A Na (7.53%).

This article builds a narrative from the assessment results: reading spatial conversion patterns as a risk-filtering basis, identifying priority districts, and translating analytical findings into operational implications for due diligence, evidence packs, and supplier engagement.

---

EUDR and the Shift to Spatial Evidence

For years, questions about coffee origin were answered by country name, purchasing region, or cooperative network. Under EUDR, that answer is no longer sufficient. Origin must be demonstrable as a production location tied to specific land, spatially verifiable, and unconnected to deforestation after the 31 December 2020 cut-off.

EUDR requires operators and traders to maintain evidence-backed due diligence statements. For plots larger than four hectares, the perimeter must be described by a polygon; for smaller plots, coordinate points may be used. This means compliance is moving toward precision sourcing: every batch must be traceable to a specific, verifiable production location.

Three evidentiary nodes of EUDR must operate simultaneously: geolocation as the entry gate, deforestation risk analysis as the interpretive layer, and the evidence pack as the auditable output. Without all three, traceability becomes merely a transaction record — not a guarantee of deforestation-free sourcing.

Coordinates only answer ‘where’. The next questions — whether the plot was previously forested, whether there was any tree cover loss after the cut-off, and whether the supply chain is free of blending — can only be answered through systematic spatial analysis.

---

Assessment Methodology

This assessment employs two complementary analytical layers: (1) a historical forest-to-coffee conversion analysis for the period 2001–2024, and (2) a predictive model-based Deforestation Risk Assessment. Both are grounded in coffee farm geolocation data and consistent forest cover datasets.

1. Forest-to-Coffee Conversion Analysis

The first layer reads what has historically occurred. Tree cover loss data from Hansen et al. (2013) was combined with coffee plantation layers to identify areas where forest was converted to coffee cultivation over a two-decade period. The analysis produces per-province and per-district metrics: absolute conversion area (ha), year-2000 forest cover as baseline, and percentage of forest replaced as a measure of relative intensity.

The Area of Interest (AOI) encompasses all districts across the three provinces: Dak Nong, Lam Dong, and Dak Lak. The primary unit of analysis is the administrative district, which provides sufficient resolution for operational prioritisation without losing provincial landscape context.

 

2. Deforestation Risk Assessment

The second layer constructs a forward-looking risk probability surface. The DRA workflow begins with geolocation data preparation: invalid polygons are corrected, very small artefacts are removed, and plots are clipped to the AOI. A working coffee mask is generated through rasterisation at 100-metre resolution.

The forest baseline is built from historical forest extent minus tree cover loss through the end of the analysis window. Recent loss is read within a five-year rolling window as the model’s target label. Spatial features used include: proximity to settlements (GHSL), permanent water bodies (JRC), coffee plantations (coffee mask), elevation and slope (SRTM), and administrative boundaries (FAO GAUL).

A Random Forest model is trained on balanced samples of recent loss and stable forest. The output is a relative risk probability surface aggregated to approximately 1 km grids and classified into five priority tiers: Low, Medium Low, Medium, Medium High, and High.

3. Assessment Results: Provincial Overview

At the provincial level, the assessment produces three headline figures that serve as the entry point for risk interpretation. Total indicated forest-to-coffee conversion across the entire AOI reached 36,371 ha — unevenly distributed across three provinces with markedly different profiles.

Table 1. Summary of assessment results by province. Source: Spatial forest-to-coffee conversion analysis, 2001–2024.

Dak Nong dominates with 21,121 ha — more than 58% of total AOI conversion. More significant than the absolute figure is its relative intensity: nearly 5% of all Dak Nong forest cover from the year 2000 has been converted to coffee plantations over two decades. This is an ecologically significant figure and directly relevant to due diligence frameworks.

Lam Dong recorded 9,518 ha at a provincial rate of 1.33%. This lower percentage requires contextualisation with district-level data — Lam Dong’s very large forest base (over 716,000 ha) dilutes the provincial intensity, even as several of its districts contain substantial conversion volumes.

Dak Lak recorded 5,731 ha or 1.05%. As Vietnam’s most iconic coffee province, the comparatively smaller total does not imply uniformly low risk. As shown in the district analysis, several areas of Dak Lak exhibit local intensities strong enough to trigger enhanced due diligence.

---

Assessment Results: District Analysis and Hotspots

District-level assessment is the most operationally actionable layer. This is where risk differences become sharp enough to support differentiated procurement decisions. Of the 30 districts within the AOI, the majority of conversion is concentrated in just six.

1.  Conversion Distribution by District

Table 2. Ten districts with the highest indicated forest-to-coffee conversion

Two districts in Dak Nong — Dak Song and Dak R’Lap — dominate strikingly. Dak Song recorded 10,982 ha of conversion at 22.58% intensity: more than one-fifth of its local forest has changed function. This places Dak Song as the highest-risk district across the entire AOI in both absolute scale and relative intensity.

In Lam Dong, Di Linh recorded 5,769 ha — the third largest in the AOI — though its intensity is more moderate at 4.77%. Cu M’gar in Dak Lak shows 9.81% local intensity, significant because of its smaller forest base (34,492 ha).

2.  Reading Two Risk Dimensions: Area vs. Intensity

This assessment employs two dimensions that do not always align: conversion area (ha) as a measure of absolute risk scale, and percentage of forest replaced as a measure of relative intensity against the local forest base. Both must be read together to avoid oversimplification.

Table 3. Risk profile matrix for the six primary hotspot districts.

Dak Song is the only district that stands out in both dimensions: the highest area and the highest intensity. This makes it the unambiguous top-priority district, free from interpretive ambiguity. Dak R’Lap follows with a large scale despite more moderate intensity — which still places it firmly in the high-risk tier.

Cu M’gar is notable for its inverse pattern: smaller area than Di Linh, but local intensity (9.81%) is substantially higher. This means that within the district’s context, conversion pressure on local forest is deeply felt. A similar situation exists in Krong A Na (7.53%) — an intensity that cannot be overlooked despite a smaller absolute volume.

---

Risk Profiles of the Three Provinces

Once district figures are read, the assessment builds distinct risk profiles for each province. These profiles are not merely statistical summaries — they are the basis for proportional due diligence strategies.

Dak Nong is the risk epicentre of this assessment. With 21,121 ha of conversion (58% of total AOI) and a provincial intensity of 4.97%, Dak Nong already stands out at the aggregate level. But the most critical reading is at the sub-level: Dak Song and Dak R’Lap together account for more than 95% of all provincial conversion. Risk in Dak Nong is not evenly distributed — it is spatially concentrated. The due diligence implication is concrete: companies sourcing from Dak Nong must demonstrate that supply does not originate from these two districts — or, if it does, that the geolocation of specific plots does not intersect with historical conversion areas and that no tree cover loss occurred after the 2020 cut-off.

Lam Dong’s provincial percentage (1.33%) can easily be misread as a low-risk signal. The assessment shows that reading is misleading if it stops there. Lam Dong’s very large forest base (716,909 ha) dilutes the provincial percentage, but does not erase the fact that Di Linh (5,769 ha) and Bao Lam (3,276 ha) have recorded real conversion volumes. Di Linh is Vietnam’s most important arabica coffee production centre. Procurement volumes from this district tend to be large in many companies’ portfolios. The combination of substantial historical conversion and Di Linh’s strategic role in premium coffee supply chains makes it a verification priority that cannot be ignored on the strength of a low provincial percentage alone.

Dak Lak is Vietnam’s most famous coffee province — and that reputation can itself become a hidden risk under EUDR. A region’s reputation does not substitute for plot-level proof. The assessment shows that while Dak Lak’s total conversion is smaller (5,731 ha, 1.05%), Cu M’gar (9.81%) and Krong A Na (7.53%) have local intensities that are sufficiently strong. Buon Ma Thuot City, as the largest coffee collection hub, recorded 590 ha of conversion at 5.50% intensity — a figure relevant because of the very large supply volumes passing through and the higher potential for origin blending across districts.

---

From Assessment to Action: The Due Diligence Framework

The assessment produces signals — not final decisions. Its value lies in its capacity to transform complex data into concrete, proportional action priorities.

1. The Agriplot Due Diligence System

The Agriplot Due Diligence System is a purpose-built, web-based platform designed to meet the EUDR’s plot-level compliance demands. The platform integrates supply chain data, multi-temporal satellite imagery, and AI-powered geospatial analytics to generate the plot-level visibility required for due diligence. In the context of Vietnamese coffee, Agriplot functions as the bridge between farm geolocation data in the field and the forest loss database — enabling users to verify whether a given plot has a history of deforestation after the 31 December 2020 cut-off.

Research by Murti et al. (2026) demonstrates that dashboard-based systems of this kind — linking data from the product level down to the individual plot — represent critical infrastructure for meeting EUDR requirements while also advancing regenerative agricultural practices more broadly.

An audit-ready evidence pack for coffee under EUDR requires four interconnected layers:

Table 4. Four evidence-pack layers for EUDR coffee due diligence.

This assessment contributes most directly to the third layer: signalling where land history needs to be examined more intensively. It does not replace plot verification, but determines where that verification should be prioritised.

2. Legal Production: Land Legality as an Evidence Requirement

EUDR is often discussed as a deforestation-free regulation, but the compliance test is broader. The commodity also has to be produced in accordance with the relevant legislation of the country of production. For coffee, this makes legality a separate evidence question: a plot may pass a forest-loss overlay, while still requiring confirmation that production is legally grounded under applicable land-use, tenure, environmental, labour, tax, or other national requirements.

In practice, legality evidence should be connected to the same origin record used for geolocation and land-history screening. A supplier file should therefore link the farm or farmer group to a declared plot, the available farmer or cooperative registration record, land-use or supplier declaration, and any clarification needed for disputed or incomplete cases. The purpose is not to overburden smallholders with paperwork, but to make the evidence pack internally consistent and auditable.

3. Inside the Agriplot Dashboard

The Agriplot dashboard is a visual interface designed to translate complex geospatial data into operationally actionable information. The main view displays several functionally integrated components:

In the Vietnamese coffee context, the integration of district-level results from this assessment with a platform like Agriplot creates a two-layer system that reinforces itself: spatial analysis at the district level provides a macro risk map for procurement prioritisation, while Agriplot supplies the micro verification infrastructure to demonstrate that specific plots do not intersect with historical conversion areas and are free of tree cover loss after the 2020 cut-off.

---

Coffee Supply Chain and Export Flow Context

The spatial assessment becomes easier to use when it is read alongside the way coffee moves commercially. In practice, origin is not created by one export record. It is built through a chain of farms, buying points, collectors, cooperatives, processors, exporters, and buyers. This is why shipment-level data is useful for understanding market exposure, but cannot replace farm-level traceability.

For EUDR, the operational question is therefore twofold. First, can a commercial lot be traced back to the farmers and land parcels that produced it? Second, can the same origin record support deforestation screening, legality checks, and supplier follow-up when risk is detected?

1. From Farmers to Exporters

A typical coffee flow begins with farmers producing cherries or dried beans. Collectors and local buyers then consolidate small volumes from many farms. Cooperatives, traders, or processors may sort, dry, grade, store, and prepare coffee for a specification. Exporters arrange the formal shipment, while overseas buyers, roasters, or operators receive the product and make regulatory decisions.

Each hand-off changes the evidence problem. At the farm level, the core questions are plot identity, land history, and legal production. At collection level, the question becomes whether coffee from different farms or villages has been mixed before documentation is complete. At exporter level, the question becomes whether the shipment can be reconciled back to the intake lots and farmer lists behind it.

Industry examples show why this distinction matters. The Global Coffee Platform describes Simexco Daklak as a leading Vietnamese coffee exporter with a farm-gate purchasing network, training and quality-control activities in growing areas, and annual green-coffee purchasing and export capacity of more than 100,000 tonnes (Global Coffee Platform, 2023). Vietnam Agriculture Newspaper also reports that enterprises and cooperatives in Lam Dong and Dak Lak are building raw material areas through farmer linkages, including company relationships with thousands of households and cooperative sales to exporters such as Simexco Daklak and Dakman (Vietnam Agriculture Newspaper, 2023).

The implication is direct: an exporter name is useful, but it is not enough. Compliance evidence has to follow the coffee upstream, especially where a shipment is assembled from many small farms, multiple collectors, or several villages within one commercial lot.

2. Export Flow as a Trade Lens

The Sankey diagram below complements the spatial assessment by showing the main commercial routes in the filtered Vietnam-origin coffee export dataset. It helps identify major shipper-buyer relationships and destination markets. It should not be read as proof of plot-level compliance, because trade-flow data does not show whether each shipment is linked to farm geolocation, segregated lots, or land-history checks.

Table 5. Main destination markets for Vietnam-origin coffee exports in the filtered dataset.

 

Table 6. Selected top shipper-buyer flows used to support the Sankey interpretation.

3. Blending Risk and Lot-Level Traceability

The most material traceability risk in coffee is often not the exporter name itself. It is the possibility that coffee from many farms, villages, or districts is combined into one commercial lot before the evidence is complete. A shipment can have a clear shipper and consignee while still containing coffee from multiple production areas with different land histories.

Under EUDR, that matters because a deforestation-free conclusion must be supported by a traceable link between the shipment, the commercial lot, the supplier intake records, and the relevant farm geolocations. If the lot cannot be reconciled back to its contributing farms, then the shipment has an evidence gap even if the exporter is known and the destination market is clear.

For hotspot districts identified in this assessment, companies should test whether lots were segregated, whether purchase records can be reconciled with farmer lists, whether weight balances are plausible, and whether the declared origin matches the physical flow of coffee. Where traceability is incomplete, the gap should be documented as a data-improvement issue and handled through a time-bound corrective action plan, not treated as affirmative evidence of compliance.

---

Supplier Engagement and Risk Response

The purpose of the assessment is to guide better engagement, not to label whole areas as unacceptable. A high-risk district is a signal that verification should be deeper and documentation should be stronger. It is not, by itself, evidence that every supplier or farmer in that district is non-compliant.

This distinction is important for smallholder-inclusive sourcing. If companies react to risk maps by excluding entire sourcing areas, compliant farmers inside those areas may be unfairly removed from the supply chain. A more proportionate response is to use risk tiers to prioritise data correction, supplier clarification, and targeted verification.

1. High-Risk Areas Are Not Automatic Non-Compliance

Suppliers in higher-risk areas should be asked for stronger geolocation evidence, clearer lot documentation, and written clarification where a plot intersects with historical conversion or recent tree-cover loss. Where the issue is data quality, the response should focus on correction and verification. Where the issue is confirmed post-cut-off deforestation or an unresolved legality concern, the sourcing decision should be escalated and documented.

This approach keeps the burden proportionate. It also creates a practical pathway for improvement: first map the supplier base, then reconcile lots, then test land history and legality, and finally record the sourcing decision with the reasoning behind it.

2. Control Points for Action

Table 7. Supply-chain control points and due diligence focus for EUDR coffee.

The control points show why EUDR evidence should be designed as a chain, not as a single document. The same shipment may require farmer geolocation, lot reconciliation, land-history overlay, legality review, and a supplier engagement record before it can support a defensible due diligence conclusion.

---

Operational Implications for Value Chain Actors

Each actor in the value chain reads the assessment results from a different position. This section translates the assessment findings into role-specific implications.

1. Operators and Exporters

Operators are closest to upstream data. The primary challenge: geolocation quality from farmers — imprecise coordinates, overlapping polygons, or unclear farm boundaries. Assessment-based recommendations:

• Prioritise geolocation data improvement programmes in Dak Song, Dak R’Lap, Di Linh, and Cu M’gar first.
• Ensure supply segregation systems distinguish lots with verified geolocation from those without.
• Build automated overlay processes between supplier coordinates and historical conversion layers as a standard part of lot intake.

2. Traders

Traders require sharp risk segmentation because they manage large volumes from many sources. Assessment implications:

• Group suppliers by origin district and risk tier — not merely by country or province.
• Apply differentiated evidence requirements: suppliers from High-risk districts require polygons and land history; Low-risk districts can be handled with coordinate points and standard documentation.
• Integrate district summaries from this assessment into supplier scorecard systems.

3. Roasters and Brands

Roasters and brands bear the weight of expectations from consumers, regulators, and auditors. Assessment implications:

• Deforestation-free claims must be systematically explainable: where the coffee originates, how its land history was checked, and how sourcing decisions were documented.
• This assessment can serve as the basis for a credible sustainability narrative — demonstrating that risk is read spatially, not merely claimed in general terms.
• For public communication, use precise language: ‘We conduct risk-based due diligence using spatial analysis’, rather than absolute claims that are difficult to substantiate.

---

Conclusion

The forest-to-coffee conversion assessment for Dak Nong, Lam Dong, and Dak Lak produces a clear picture: coffee-related deforestation risk in Vietnam’s Central Highlands is not uniformly distributed. It is concentrated in specific districts — particularly Dak Song and Dak R’Lap in Dak Nong, Di Linh in Lam Dong, and Cu M’gar in Dak Lak.

The total 36,371 ha of forest-to-coffee conversion identified over the period 2001–2024 should not be read as an indictment of the Vietnamese coffee industry. It is a risk map that helps supply chain actors ask more precise questions: from which district does this coffee originate, what is its land history, and is the available evidence sufficient to meet EUDR due diligence standards.

The primary value of this assessment lies not in the map itself — but in the way the map compels more precise questions, drives more structured documentation, and directs verification resources to locations that need them most. Over the long term, companies capable of reading risk spatially will be better positioned to substantiate deforestation-free claims — and better able to sustain equitable, responsible sourcing.

EUDR compliance for coffee cannot be resolved through a compelling origin narrative. It requires maps, data, and systems that connect the two into auditable evidence.

---

References

1. European Parliament and Council of the European Union. (2023). Regulation (EU) 2023/1115 on the making available on the Union market and the export from the Union of certain commodities and products associated with deforestation and forest degradation. Official Journal of the European Union.
2. European Commission. (2026). Regulation on Deforestation-free Products. Directorate-General for Environment. Retrieved May 2026 from https://environment.ec.europa.eu/topics/forests/deforestation/regulation-deforestation-free-products_en
3. European Commission. (2026). Frequently Asked Questions on the EU Deforestation Regulation, 5th iteration. Directorate-General for Environment.
4. Coffee Deforestation Risk Assessment (DRA), Dak Lak Province, Viet Nam. Technical report on methods, district analytics, and spatial deliverables (internal analytical document, 2025).
5. Spatial Forest-to-Coffee Conversion Analysis, Dak Nong, Lam Dong, and Dak Lak, Vietnam, 2001–2024. Provincial summaries, district summaries, AOI maps, indicative rasters (internal analytical document, 2025).
6. Hansen, M. C., Potapov, P. V., Moore, R., Hancher, M., Turubanova, S. A., Tyukavina, A., Thau, D., Stehman, S. V., Goetz, S. J., Loveland, T. R., Kommareddy, A., Egorov, A., Chini, L., Justice, C. O., & Townshend, J. R. G. (2013). High-resolution global maps of 21st-century forest cover change. Science, 342(6160), 850–853. https://doi.org/10.1126/science.1244693
7. Joint Research Centre (JRC), European Commission. (2023). JRC Global Surface Water Explorer and Forest Cover Monitoring datasets.
8. Global Human Settlement Layer (GHSL). (2023). GHS-BUILT and GHS-POP datasets. European Commission, Joint Research Centre. https://ghsl.jrc.ec.europa.eu
9. NASA Shuttle Radar Topography Mission (SRTM). (2000). SRTM 1 Arc-Second Global Elevation Data. U.S. Geological Survey. https://doi.org/10.5066/F7PR7TFT
10. Food and Agriculture Organization of the United Nations (FAO). (2015). FAO GAUL: Global Administrative Unit Layers. UN FAO.
11. FAO / Forest Data Partnership. Coffee Probability Layer and Regional Land Cover datasets (accessed 2024–2025).
12. General Statistics Office of Vietnam (GSO). (2023). Agricultural Statistics: Coffee Production Area and Output by Province.
13. Vietnam Ministry of Agriculture and Rural Development (MARD). (2023). Annual Forest Inventory and Reporting, Central Highlands Provinces.
14. Breiman, L. (2001). Random Forests. Machine Learning, 45(1), 5–32. https://doi.org/10.1023/A:1010933404324
15. Murti, S., Rahmawati, D. C., Pratama, M., Yazid, H., Ishak, Pamungkas, C., & Rafina, I. (2026). From product to plot: developing a visual dashboard to support beyond deforestation-free to regenerative supply chains. IOP Conference Series: Earth and Environmental Science, 1622, 012016. https://doi.org/10.1088/1755-1315/1622/1/012016
16. Global Coffee Platform. (2023). Meeting Members: Over a cup of coffee with SIMEXCO DAKLAK. https://www.globalcoffeeplatform.org/latest/2023/meeting-members-over-a-cup-of-coffee-with-simexco-dak-lak/
17. Vietnam Agriculture Newspaper. (2023). Robusta Coffee: Expanding raw material areas to meet orders. https://van.nongnghiepmoitruong.vn/expanding-raw-material-areas-to-meet-orders-d357765.html
18. Panjiva. (2025). Panjiva Consolidated Export Coffee 0901 shipment dataset. Shipment-level dataset used for supplementary trade-flow analysis.

 

Prefer to read it offline or share it with your team?

Download the PDF version of the full article below.

Download the Article

Mato Grosso is Brazil’s largest soy-producing state, responsible for roughly 27% of the country’s soybeans and about 30% of its soy exports – a critical piece of the global soy supply chain (Reuters, 2023). Soy cultivation in this region has long been associated with deforestation, making sustainability and legal compliance a central concern (WBCSD & Guidehouse, 2024).

The recently adopted EU Deforestation Regulation (EUDR) aims to curb global deforestation by prohibiting imports of commodities produced on land deforested after December 31, 2020, and by requiring compliance with relevant local laws. In practical terms, for Mato Grosso’s soy this means exports to the EU must be deforestation-free (post-2020) and legally produced under Brazilian law. The EUDR’s implementation timeline was even extended by 12 months to give stakeholders more time to comply (Council of the European Union, 2024), but soy exporters are already under pressure to meet these strict criteria.

In this case study, we assess Mato Grosso’s soy farms against EUDR criteria to evaluate the state’s readiness for compliance. Our analysis examines whether soy cultivation areas have avoided deforestation since 2020 and adhered to Brazilian legal requirements. We integrated multiple data sources to evaluate every major soy farm in the state for recent forest clearing and legal irregularities. This report presents the findings in an accessible way, highlighting how prepared Mato Grosso is for EUDR compliance and what challenges and opportunities lie ahead.

 

Methodology

 

Our assessment combined geospatial analysis with official data to determine which soy cultivation areas meet EUDR requirements. We gathered high-resolution data on soy planting locations and cross-referenced these with deforestation records, land registries, and protected area maps. Key datasets and tools used include:

• Official land registry data: We used Brazil’s rural land cadastre systems (SIGEF) to obtain soy farm boundaries and ownership status, identifying which soy farms are officially registered versus those that are not. This helps flag any soy plots lacking proper registration or title, since unregistered farms often indicate legal irregularities (ABIOVE & APROSOJA, 2024).
• Baseline forest cover (2020): A Global Forest Cover 2020 map (from the European Commission’s Joint Research Centre) served as a baseline for remaining forest as of the EUDR cutoff date (Köhl et al., 2023). This allowed us to know which areas were forested at end of 2020, so any subsequent clearing in those areas would violate EUDR’s no-deforestation rule.
• Deforestation alerts (2021–2023): We monitored recent forest loss using integrated satellite-based deforestation alert systems, notably RADD and GLAD alerts (Landsat and Sentinel-2) for 2021–2023 (see Figure 1). These near-real-time alerts flag new clearing events across Mato Grosso’s landscapes, enabling us to detect any forest clearance that occurred after the cutoff date. Integrated alerts provide frequent updates and wide coverage to detect deforestation as it happens.
• Tropical Moist Forest Deforestation Year: This dataset was developed by the European Commission’s Joint Research Centre (JRC). It provides the annual records of forest loss across Mato Grosso from 2013 to 2023. It was critical for assessing long-term deforestation trends and identifying the districts most affected by forest clearance. TMF deforestation year helps visualize deforestation changes over the decade, contextualizing recent developments in land use.
• Inovasi Digital’s verified deforestation alerts: To improve accuracy, we leveraged Inovasi Digital’s Verified Deforestation Alerts (2021–2023). These are curated alerts that filter out false positives from the generic satellite alert systems and confirm genuine deforestation events. The Inovasi Digital’s- verified alerts offer superior accuracy compared to raw integrated alerts, ensuring we only count true forest loss in soy areas. This higher-quality alert data strengthens our monitoring of EUDR compliance by avoiding the noise of erroneous alerts (see Figure 1).
• Soy planted area maps: We also used Inovasi Digital’s proprietary soy planted area maps to precisely locate soy cultivation plots across Mato Grosso. This dataset allowed us to focus the analysis on actual soy-producing areas rather than broad land parcels. Using Inovasi Digital’s soy map (as opposed to more general land cover data) ensured no soy farm was overlooked and that we didn’t misidentify non-soy land as soy. It greatly increased the accuracy of our compliance checks by zeroing in on the true soy footprint in the state.
• Protected areas and public forests: We incorporated maps of protected areas, including Indigenous territories, conservation units, and Permanent Forest Reserves (public forests where agriculture is prohibited) from official sources (ICMBio, 2023) (see Figure 2). Any overlap between soy farms and these protected lands would indicate a legal violation. Identifying soy cultivation inside protected forests (so-called PRF overlaps) is crucial, since those areas would fail EUDR’s legality criterion regardless of deforestation timing.

 

Using these data layers, we evaluated every major soy plot in Mato Grosso against EUDR-like rules: (1) no forest clearing after 31 Dec 2020, (2) valid land tenure/registration, and (3) no farming inside protected forests. By cross-validating multiple sources (satellite alerts, official records, and Inovasi Digital’s refined data), our methodology minimized errors and provided a robust compliance status for each soy farm in the state.

 

 

Figure 2: Protected Areas and Conservation Units in Mato Grosso. The maps illustrate officially designated Protected Areas (left, yellow) and Federal Conservation Units (right, green) distributed across the state. These zones are legally restricted from agricultural activities, highlighting areas critical for assessing compliance with EUDR legality requirements. Data sourced from ICMBio (2023).

Results and Findings

 

Deforestation Trends

Our analysis of deforestation data (2013–2023) in Figure 3 shows that annual clearing in Mato Grosso spiked around the EUDR cutoff and then sharply declined. The state saw a peak in forest loss at nearly 500,000 hectares in 2020, one of the highest levels in the past decade (Köhl et al., 2023). Deforestation remained high into 2022 (around 500,000 ha that year as well), but by 2023 the annual clearing dropped to around 200,000 ha. This recent decline in deforestation may reflect stronger enforcement efforts and regulatory anticipation leading up to EUDR implementation. In short, Mato Grosso’s deforestation trend moved from a worrying surge to a significant reduction by 2023, which bodes well for future compliance if the trend holds.

Figure 3: Annual deforestation in Mato Grosso, 2013–2023. The bars show total area deforested each year (in thousands of hectares, “ha”). After moderate levels (~100–300 thousand ha) earlier in the decade, deforestation spiked to around 500,000 ha in 2022. A sharp decline is observed in 2023 (well under 200,000 ha), suggesting recent improvements in control measures.

 

Geographic Distribution of Deforestation

Deforestation linked to soy is not uniform across Mato Grosso. A small number of districts in the state’s northern and northwestern region account for a disproportionately large share of forest loss. For example, the districts of Colniza, Aripuanã, and Marcelândia each saw over 100,000 hectares of deforestation from 2013 to 2023, topping the list in forest clearing. Many of these top deforesting districts experienced notable spikes in particular years (such as Colniza in 2022, and Poconé with a large spike in 2020) as soy and other land uses expanded (see Figure 4). This means that enforcement and monitoring efforts may need to concentrate on a handful of hotspot regions, even though state-wide averages are improving. Overall, most other districts had much lower deforestation totals over the last decade, indicating the problem is concentrated rather than widespread.

Figure 4: Top 10 districts in Mato Grosso for deforestation (2013–2023), with breakdown by year. Each horizontal bar represents a district, ordered by total deforestation over the 11-year period. The length of the bar corresponds to total deforestation (ha), and colored segments show the contribution of each year (legend on right, 2013–2023). Colniza, Aripuanã, and Marcelândia (top bars) had the highest totals, each exceeding 100,000 ha of clearing. Colniza in particular shows a pronounced spike in 2022 (brown segment), and Poconé shows a very large spike in 2020 (teal segment). (Data source: TMF Deforestation Year for Mato Grosso).

Compliance Assessment

 

Soy Compliance Status

When we overlay all the criteria – deforestation cut-off, legal registration, and protected area exclusions – the analysis shows that the overwhelming majority of Mato Grosso’s current soy farmland meets EUDR compliance criteria (see Figure 5). Approximately 97.21% of the assessed soy plot area can be categorized as “compliant.” In practical terms, these compliant plots have not cleared any native vegetation after 2020 and they also did not fail the legal checks (they are properly registered and not farming inside a protected reserve). This finding is encouraging: it suggests most soy exported from Mato Grosso would be considered deforestation-free (post-2020) and legally produced, thereby meeting EUDR’s core requirements.

On the other hand, about 2.79% of the soy area was flagged as “non-compliant” by our assessment. These plots failed one or more of the criteria – typically either they showed evidence of deforestation after the cutoff date, or they were unregistered farms found within a protected forest area (an illegal situation). We examined this non-compliant subset to understand the main risk factors behind it.

Recent Deforestation

A number of the non-compliant plots had detectable forest clearing in the years 2021, 2022, or 2023. In fact, thousands of hectares were cleared across soy farms in those years – with the largest cleared areas being roughly 2,400 ha in 2021, 3,000 ha in 2022, and 2,400 ha in 2023. Any soy planted on land that was deforested in those years violates the EUDR cutoff, rendering those plots non-compliant. These cases likely represent farmers expanding their cultivation into forest areas or clearing portions of their legal reserves after 2020.

Land Legality and Protected Areas

The vast majority of soy-producing land in Mato Grosso appears to be legally compliant with respect to land tenure. Approximately 97.1% of the state’s soy cultivation area is officially registered in SNCI/SIGEF, indicating those farms have proper land title and environmental registration. Only about 2.9% of the soy area remains unregistered or of unclear legal status. We also found a few cases of soy encroachment into protected forests: soy plots overlapping Permanent Forest Reserves (PRFs) were identified in eight districts of Mato Grosso (ICMBio, 2023). These instances represent clear legal infringements, as farming is not allowed in those reserve areas. While the number of overlaps is limited, any soy coming from those plots would be automatically non-compliant due to illegal location. Aside from these isolated cases, however, most soy farms are operating on legally authorized land.

Protected Area Overlaps

We identified 36 soy plots that lie inside designated Permanent Forest Reserves (spread across the eight districts mentioned earlier). Soy from these areas is automatically non-compliant because the land itself is off-limits for agriculture. These instances are relatively rare, but each is a serious breach of land-use law.

Unregistered Farms in Protected Areas

A subset of non-compliant cases may involve plots that are both unregistered and in a protected area. Our criteria were somewhat conservative – if a plot was unregistered but not overlapping a protected area (and had no post-2020 deforestation), we did not immediately mark it as non-compliant. We assumed lack of registration alone (without other red flags) might not violate EUDR if the farm is otherwise legal and deforestation-free, since EUDR doesn’t explicitly mandate formal land titles but rather compliance with local laws. This cautious approach means the 2.79% non-compliant figure is likely a lower-bound estimate. If we had applied stricter rules (for example, considering any unregistered plot as non-compliant), the share of non-compliant soy area could rise to around ~5% of the total. We focused on the most clear-cut violations for this analysis.

Figure 5: EUDR Compliance Status of Soy Plots in Mato Grosso. The map displays the classification of soy cultivation areas based on deforestation and legal status. Green areas represent compliant plots with no post-2020 deforestation and valid land registration. Red areas are non-compliant due to either recent clearing or unclear tenure.

District-Level Risk Assessment

 

Beyond looking at individual farms, we also evaluated overall compliance risk by district (see Figure 6). Out of Mato Grosso’s 141 districts (municipalities), none were categorized as “High Risk” in aggregate, which is a positive sign. We classified 97 districts as Low Risk and 44 as Medium Risk for EUDR compliance. Even though no district was flagged as high risk, the Medium Risk category highlights where attention is needed. For example, certain districts like Porto Estrela have known cases of soy fields overlapping protected forests (hence some risk), and Colniza has a notable number of unregistered soy plots contributing to risk.

Additionally, major soybean production hubs such as Sorriso might carry traceability risks (mixing of soy from different farms at aggregation points). The Medium Risk areas are those with some red flags – be it recent deforestation, legal registration gaps, or proximity to remaining forests that could tempt future clearing. These areas will require enhanced due diligence and monitoring, even though they are not in outright violation at the moment. By contrast, the Low Risk districts show stable, compliant conditions with no significant deforestation after 2020 and strong legal adherence.

Figure 6: District-Level EUDR Risk Classification in Mato Grosso. The map categorizes the state’s 141 districts based on deforestation, legality, and traceability factors. Districts outlined in green are classified as Low Risk, while those in orange are Medium Risk.

Challenges and Opportunities

 

Implementing EUDR in the soy supply chain does pose several challenges that stakeholders must navigate:

• Traceability: Achieving full farm-to-port traceability for soy is difficult, yet it is crucial under EUDR. Soy supply chains often mix produce from many farms, making it hard to trace each batch back to its origin. Ensuring every shipment’s origin is known and verified as deforestation-free requires overhauling traditional commodity trading systems (WBCSD & Guidehouse, 2024).
• Legal Verification: Verifying that all soy is produced legally (in line with Brazilian environmental and land laws) demands rigorous and continuous monitoring of legal permits, environmental licenses, and land titles. Any changes in a farm’s legal status (e.g. a license expiring or a farm expanding into an unregistered area) could affect compliance (ABIOVE & APROSOJA, 2024). Keeping tabs on these aspects for thousands of farms is a significant administrative challenge.
• Continuous Monitoring: Even if a farm is compliant today, ongoing surveillance is needed to catch any new deforestation or encroachment that might occur. Regular satellite monitoring and alert systems are critical to detect deforestation in near-real-time and to ensure that no new illegal clearing goes unnoticed (DLA Piper, 2024). In a state as large as Mato Grosso, maintaining this continuous vigilance over millions of hectares is challenging.

 

On the other hand, EUDR also brings opportunities and benefits that can be leveraged:

• Transparency and Market Advantage: Complying with EUDR can push companies to improve supply chain transparency and documentation. In a market that increasingly demands sustainable products, this transparency can become a competitive advantage. Soy exporters who can prove their product is deforestation-free and legal may access premium markets and foster trust with buyers (WBCSD & Guidehouse, 2024).
• Technological Innovation: The stringent requirements spur innovation in monitoring and traceability tools. Advanced technologies like satellite imagery, AI-driven deforestation detection, blockchain for supply chain traceability, and big-data platforms are being adopted to improve accuracy and efficiency. These innovations not only help with EUDR compliance but also modernize agricultural supply chain management for the better.
• Alignment with Sustainability Goals: By enforcing no-deforestation and legality, EUDR compliance efforts align with broader environmental and sustainability goals. Strengthening land-use governance and protecting forests contribute to climate change mitigation and biodiversity conservation. This alignment enhances Mato Grosso’s reputation as a responsible soy producer committed to sustainable development (Trase, 2020).

 

Role of Technology in EUDR Compliance

 

One promising avenue to tackle the above challenges is the deployment of dedicated technological solutions for monitoring and due diligence. For instance, MosaiX’s Agriplot Due-Diligence System (DDS) is a platform designed to help companies seamlessly comply with regulations like EUDR. Such a system provides end-to-end visibility and control over the supply chain by integrating various data streams. Key features of the Agriplot DDS include accurate geolocation of soy farms and real-time tracing of supply chains, automated compliance checks against deforestation alerts and protected area databases, and risk assessment tools that flag high-risk suppliers or regions in the chain. The platform can also generate standardized due-diligence reports to fulfill EUDR reporting requirements with ease.

Notably, systems like Agriplot DDS come with extensive global data coverage. In the case of MosaiX’s platform, it builds on a rich database of soy production worldwide (MosaiX, 2023). In Latin America alone, the system includes data on roughly 51.1 million hectares of soy fields and tracks operations of over 100 soy-related facilities across Brazil, Argentina, Bolivia, and Paraguay. Additionally, it includes data on about 1.6 million hectares of soy plantations in Canada. This breadth of data enables cross-border analysis and benchmarking, which is valuable as traders source soy from multiple countries. A dedicated due diligence platform, hosted on secure servers, allows businesses to customize and privately manage their supply chain data. By leveraging such advanced tools, soy exporters and buyers can more efficiently achieve EUDR compliance, manage their deforestation and legality risks, and promote sustainability across their operations. In short, technology is acting as an enabler – turning a daunting compliance exercise into a more streamlined, data-driven process.

 

Summary of Findings

 

Our assessment of Mato Grosso’s soy supply chain reveals an overall positive outlook for EUDR compliance. Over 97.21% of the state’s soy-producing area already meets the EUDR criteria, meaning the vast majority of farms are deforestation-free (post-2020) and in adherence with Brazil’s legal requirements. This high rate of compliance indicates that most soy exporters in Mato Grosso are well-positioned to fulfill their due diligence obligations. Only a very small fraction of production – about 2.79% of the area – is flagged as potentially non-compliant or “at-risk” under the EUDR.

These non-compliant cases are concentrated in a few localized areas and are not representative of the state as a whole. The primary risk factors identified include post-2020 deforestation on certain farms, overlaps of soy fields with protected forests, and a handful of farms lacking proper registration. Importantly, no major region or municipality in Mato Grosso was categorized as “High Risk” for EUDR; the compliance issues tend to be isolated exceptions rather than widespread problems.

Overall, our findings highlight that EUDR-aligned production is already the norm in Mato Grosso’s soy sector. The small number of non-compliant cases can be addressed through targeted interventions, bringing the entire state’s production up to the required standard.

 

Conclusion

 

Mato Grosso’s soy supply chain appears largely prepared to meet the EU’s deforestation-free and legality standards. With well over 97.21% of soy area in compliance, the state stands as a strong example of how large- scale agriculture can align with strict environmental regulations. Addressing the remaining few percent of at-risk areas will be critical to achieve full compliance and maintain market access to the EU. Targeted actions should focus on the handful of problem cases identified – for instance, enforcing protections in zones where soy still encroaches on public forests, and accelerating land regularization for the few farms on unregistered properties. Strengthening on-the-ground enforcement (aided by enhanced alert verification systems like Inovasi Digital’s) and supporting farmers in rectifying any legal irregularities will help close the gaps.

By taking these steps, Mato Grosso can safeguard its vital EU export market and also solidify its reputation as a sustainable, deforestation-free soy producer. The case of Mato Grosso demonstrates that with robust data-driven monitoring and decisive enforcement, even a major agricultural frontier can achieve a high level of compliance with global anti-deforestation standards. This marriage of productivity with environmental compliance serves as an encouraging model for other regions. Moving forward, continued vigilance and innovation will ensure that Mato Grosso not only remains compliant with EUDR but also contributes positively to global sustainability goals.

 

 

Disclaimer: This article is intended for research and informational purposes only. It is not intended to serve as a legal basis for EUDR submissions or regulatory compliance.

 

References

 

• ABIOVE & APROSOJA. (2024).Impact of EUDR on Brazilian soy exports. FeedNavigator. Retrieved from https://www.feednavigator.com
• Council of the European Union. (2024). EU deforestation law: Council extends application timeline. Retrieved from https://www.consilium.europa.eu
• DLA Piper. (2024). EU Deforestation Regulation implications. Retrieved from https://www.dlapiper.com
• ICMBio. (2023). National System of Conservation Units (SNUC). Retrieved from https:// www.icmbio.gov.br
• Köhl, M., Bourgoin, C., et al. (2023). JRC Global Forest Cover and Deforestation Dataset. European Commission, Joint Research Centre.
• MosaiX. (2023). Agriplot Due-Diligence System Features. Retrieved from https://mosaix.earth/solution/
• Reuters. (2023). Brazil soy output and trade statistics. Retrieved from https://www.reuters.com
• Reuters. (2024). Brazil Supreme Court suspends Mato Grosso deforestation law. Retrieved from https:// www.reuters.com
• Trase. (2020). Soy and Deforestation Dynamics in Brazil. Retrieved from https://trase.earth
• WBCSD & Guidehouse. (2024). Understanding the EU Deforestation Regulation. Retrieved from https://www.wbcsd.org