Amazon Deforestation Detection and Forest Loss Mapping

Track clearing events and degradation fronts across the Amazon basin in near real time. Klarety agents execute monthly change detection on fused Sentinel-1 SAR and Sentinel-2 optical to flag new clearings by area, identify illegal logging corridors, and classify degradation type. Conservation organizations receive monthly forest loss reports with polygon-level exports.

Monitor deforestation

Forest loss detection for conservation enforcement teams

Klarety AI

Monthly clearing reports from fused satellite data

Klarety agents fuse Sentinel-1 SAR and Sentinel-2 optical to detect and classify new clearing events with polygon-level reports.

SAR + Optical

All-weather forest loss via SAR-optical fusion

Agents use SAR to detect canopy structure loss through cloud cover and optical to validate clearing extent and degradation type.

GIS Output

Forest loss polygons for REDD+ and conservation GIS

Export clearing polygons with area, date, and degradation class for REDD+ monitoring systems and conservation enforcement GIS.

SAR-optical forest loss detection and degradation classification

Klarety agents compute NDVI loss from monthly Sentinel-2 composites and cross-validate with Sentinel-1 SAR backscatter decrease to confirm canopy removal. The SAR layer enables detection through the persistent cloud cover over Amazonia. New clearings are classified as complete clearing, partial degradation, or edge effect using spatial pattern analysis. Output is a monthly GeoJSON with area, date, and degradation class per polygon.

Klarety AIMonitor deforestation in your Amazon region

analysis/amazon_deforestation_detector.pyAgent code

python

import eeimport numpy as np
ee.Initialize()
# Para state deforestation frontier, Brazilaoi = ee.Geometry.Rectangle([-51.5, -6.5, -49.5, -4.5])
def s2_ndvi(start, end):    return (ee.ImageCollection('COPERNICUS/S2_SR_HARMONIZED')            .filterBounds(aoi)            .filterDate(start, end)            .filter(ee.Filter.lt('CLOUDY_PIXEL_PERCENTAGE', 30))            .map(lambda img: img.normalizedDifference(['B8','B4']).rename('ndvi'))            .median())
def s1_vv(start, end):    return (ee.ImageCollection('COPERNICUS/S1_GRD')            .filterBounds(aoi)            .filterDate(start, end)            .filter(ee.Filter.eq('instrumentMode', 'IW'))            .select('VV').mean())
# Baseline vs current periodndvi_before = s2_ndvi('2025-06-01', '2025-08-31')ndvi_after  = s2_ndvi('2026-01-01', '2026-03-31')
s1_before = s1_vv('2025-06-01', '2025-08-31')s1_after  = s1_vv('2026-01-01', '2026-03-31')
# Forest loss: NDVI drop >0.25 (primary forest: NDVI ~0.75; cleared: ~0.15)ndvi_loss = ndvi_before.subtract(ndvi_after).rename('ndvi_loss')optical_loss = ndvi_loss.gt(0.25)
# SAR confirmation: VV increase in former forest = canopy removedsar_change = s1_after.subtract(s1_before).rename('sar_delta')sar_clearing = sar_change.gt(2.0)  # >2 dB increase = canopy loss
# Confirmed clearing: both optical NDVI loss + SAR changeconfirmed_loss = optical_loss.And(sar_clearing).rename('confirmed_loss')# Optical only: partial degradation / cloud contaminationoptical_only = optical_loss.And(sar_clearing.Not()).rename('degradation')
def area_ha(mask):    a = mask.multiply(ee.Image.pixelArea()).reduceRegion(        ee.Reducer.sum(), aoi, 10, maxPixels=1e9    ).values().get(0).getInfo()    return round((a or 0) / 10000, 0)
print(f"Confirmed clearing (SAR+optical): {area_ha(confirmed_loss):>8.0f} ha")print(f"Degradation (optical only):       {area_ha(optical_only):>8.0f} ha")

output/amazon_deforestation_report.mdOutput report