Monitoring and evaluation of restoration and conservation projects has long relied on fragmented field measurements, self-reported indicators, and project-specific analyses that are difficult to compare or scale. This fragmentation weakens accountability, limits learning across projects, and slows the translation of evidence into policy – at exactly the moment when global restoration commitments (the United Nations Decade on Ecosystem Restoration, Land Degradation Neutrality, national climate pledges) demand comparable, reproducible monitoring at scale.

MonitorEO is a standardized Google Earth Engine workflow that brings scientific rigor to Earth Observation-based monitoring of restoration and development projects. It combines three satellite-derived biophysical indicators – vegetation greenness (NDVI), surface and vegetation moisture (NDWI), and land surface temperature (LST) – with a quasi-experimental evaluation design built around two complementary techniques:

• Difference-in-differences (DiD) analysis, comparing biophysical change inside a project site against ecologically matched, non-intervened control areas – isolating the change plausibly attributable to the intervention from natural or regional background variability.
• Trend testing using the Hamed-Rao modified Mann-Kendall test, a statistical method that detects genuine long-term ecological trends while correcting for the serial autocorrelation common in satellite time series – reducing the risk of mistaking noise for a real signal.

Control areas are identified through three complementary approaches – a 2 km geographic buffer around each project, algorithmically-matched areas selected for land-cover similarity, and expert-designated comparison sites – giving each project multiple, cross-checked counterfactuals rather than a single, potentially fragile comparison.

Validated at global scale

MonitorEO has been applied across 300 project sites in 55 countries, drawing on a global database of verified nature-based restoration projects spanning avoided deforestation, reforestation, afforestation, and REDD+ interventions. The results demonstrate that open, standardized Earth Observation monitoring can be applied credibly at scale:

• 62% of projects showed statistically significant ecological trends in at least one indicator.
• 92% of projects with a significant vegetation trend showed increasing NDVI – consistent with successful vegetation recovery.
• Land surface temperature declined significantly in a meaningful share of cases, most notably in Africa – a conservative but robust signal of surface cooling associated with vegetation recovery.
• Monitoring duration matters most: projects observed for more than 10 years were far more likely to show statistically significant trends than those monitored for under 5 years – underscoring the value of sustained, long-term monitoring over one-off assessments.

Case studies illustrate the pattern clearly: afforestation projects in Uruguay and the southern United States showed strong, aligned gains in vegetation cover and surface cooling, while a REDD+ project in West Africa's Pendjari and W-Benin National Parks showed vegetation recovery without a corresponding gain in surface moisture – a reminder that restoration outcomes are best read through multiple indicators together, not any single metric in isolation.

An evolving framework

MonitorEO is now being extended to meet specific country needs – for example under GreenMEX – incorporating additional indicators such as actual evapotranspiration (AET/ET), water use efficiency (WUE), and net primary productivity (NPP) as a productivity signal, alongside a synthetic control module for even more rigorous impact comparisons. This makes MonitorEO a flexible, evolving tool that can be tailored to the monitoring and evaluation needs of individual restoration programmes and national policy frameworks alike.

While the MonitorEO platform provides a standardized workflow for EO-based impact evaluation, users may experience variability in performance depending on the selected region and data availability. In some cases, vegetation index charts may take time to load, or certain control areas (e.g., randomly generated polygons) may not be immediately visible due to processing delays or data constraints.

Users are advised to:

• allow sufficient time for EO layers and charts to load; 
• zoom and re-center the map if control areas are not displayed; 
• verify layer selection and legend labels when interpreting outputs; 
• repeat the analysis for alternative sites if outputs are incomplete. 

Step 1 – Select country and restoration project

• In the left panel, use the Select a country dropdown to choose a country that has an uploaded project footprint (e.g. Ghana).
• Use the Select a restoration area dropdown to choose a project (e.g. VCS987).

Step 2 – Load and review the project area

• Click Zoom to restoration area & controls.

The map will frame three overlays:

• Restoration (treatment) area – shown in green
• Control Area 1 – 2 km buffer, shown in red
• Control Area 2 – randomly selected comparison area, shown in pink/orange

Review the project information panel: project name, start date and end date.

Step 3 – Select and explore an indicator

Under Select an indicator to visualize, choose one of:

• NDVI – vegetation greenness and condition
• LST – land surface temperature
• NDWI – water availability and moisture

Observe the MODIS raster on the map and use the colour legend in the lower-left corner. Toggle layers on/off using the Layers icon (upper-right of map).

Step 4 – Examine the time-series chart

Scroll down in the left panel to view the time-series chart for the selected indicator, there are three charts displayed mean time series, mean differences and cumulative differences.

The chart compares the restoration area against both control areas over time. Look for:

• Trends in the 3 years before the intervention start date
• Trends from the start date to the end date
• Divergence between the treatment and control areas

Tip: click the expand icon (upper-right of chart) to open the chart in a new browser tab for easier interpretation. You can hover over data points, zoom in, and download data.

Step 5 – Review the difference-in-differences (DiD) coefficient

• Locate the DiD coefficient section in the left panel.
• The DiD coefficient quantifies the estimated impact of the restoration intervention after accounting for changes seen in the control areas.

Step 6 – Repeat for all indicators

• Return to Step 3 and repeat for each remaining indicator (NDVI, LST, NDWI).
• Compare DiD coefficients across indicators to assess evidence of improvement in vegetation, water availability and temperature.

Quick interface tips

• Left panel: contains all dropdowns, project info, DiD results, and charts. Scroll down to see everything.
• Map panel: shows the satellite base layer, coloured area overlays, and indicator rasters. Use the Layers icon (upper-right) to toggle overlays.
• Expand charts: click the expand icon (upper-right of any chart) to open it in a new tab for full-resolution viewing, point-level hover details, and CSV download.