Angola contains the second-largest area of miombo woodland in Africa, an expansive belt of dry deciduous forest stretching across the central and eastern provinces from Bié and Moxico through Cuando Cubango and into Lunda Sul. These woodlands, covering approximately seven hundred thousand square kilometres, represent one of the most ecologically significant biomes on the African continent. They are also among the most threatened. Charcoal production, slash-and-burn agriculture, illegal logging, and expanding settlement patterns are driving deforestation at rates that have accelerated markedly since the end of Angola’s civil conflict in 2002. Satellite remote sensing, particularly through the European Space Agency’s Sentinel-2 mission, has become the primary tool for monitoring this vast and largely inaccessible landscape.
The Scale of the Challenge
Understanding the deforestation challenge in Angola requires first appreciating the sheer scale of the territory involved. Angola’s total forest cover, including miombo woodlands, tropical moist forests in Cabinda, and gallery forests along major river systems, is estimated at approximately sixty-nine million hectares. This places Angola among the top ten most forested countries in Africa. However, the Food and Agriculture Organization estimates that Angola lost approximately three hundred and eleven thousand hectares of forest annually between 2010 and 2020, a rate that has shown no signs of deceleration.
The drivers of deforestation in Angola differ significantly from those in other tropical forest countries. Unlike Brazil or Indonesia, where industrial-scale agriculture and plantation expansion are the primary causes, Angola’s deforestation is predominantly driven by small-scale subsistence activities. Charcoal production for domestic cooking fuel is the single largest driver, accounting for an estimated forty to fifty percent of total forest loss. The majority of Angola’s twenty-three million rural and peri-urban residents depend on charcoal as their primary energy source, creating a structural demand that current electrification programmes have been unable to offset.
Slash-and-burn agriculture, locally known as “queimada,” accounts for another thirty to thirty-five percent of deforestation. This traditional farming practice involves clearing forest plots, burning the vegetation to release nutrients into the soil, and cultivating the cleared land for two to three seasons before abandoning it and moving to a new plot. While sustainable at low population densities, the practice becomes destructive when population growth reduces fallow periods below the threshold required for forest regeneration.
Sentinel-2 Time Series Analysis
The Sentinel-2 mission, comprising twin satellites in polar orbit with a combined revisit time of five days at the equator, provides the ideal platform for large-scale deforestation monitoring in Angola. Each satellite carries a Multispectral Instrument (MSI) with thirteen spectral bands ranging from visible blue through shortwave infrared, at spatial resolutions of ten, twenty, and sixty metres depending on the band. This combination of temporal frequency, spectral richness, and spatial detail enables the detection of forest disturbance events at scales ranging from individual clearing plots of less than one hectare to landscape-level conversion patterns spanning hundreds of square kilometres.
The fundamental analytical approach involves computing the Normalised Difference Vegetation Index (NDVI) for each pixel in every cloud-free Sentinel-2 scene acquired over the monitoring period. NDVI, calculated as the normalised ratio of near-infrared to red reflectance, is a well-established proxy for vegetation density and photosynthetic activity. Forested pixels exhibit high NDVI values, typically between 0.6 and 0.9, while cleared or degraded pixels show markedly lower values.
By constructing a continuous NDVI time series for each pixel, analysts can identify the precise timing and magnitude of vegetation loss events. A sudden drop in NDVI from forested to non-forested values indicates a clearing event, while a gradual decline may indicate progressive degradation through selective logging or canopy thinning. The Sentinel-2 archive, which extends back to June 2015 for the first satellite and March 2017 for the second, now provides more than a decade of continuous monitoring data for the Angolan landscape.
However, cloud cover presents a significant challenge in tropical and subtropical environments. During Angola’s wet season, from October through April, persistent cloud cover can obscure the surface for weeks at a time, creating gaps in the optical time series that complicate change detection. Several approaches have been developed to mitigate this limitation. Temporal compositing, which selects the clearest pixel from multiple acquisitions within a defined window, reduces cloud contamination at the cost of temporal precision. Fusion with SAR data from Sentinel-1, which penetrates cloud cover, provides complementary information during cloudy periods. Machine learning algorithms, particularly recurrent neural networks trained on historical time series, can interpolate missing values and flag probable disturbance events even in the presence of gaps.
Provincial Hotspot Analysis
Satellite time series analysis has identified several provincial hotspots where deforestation rates significantly exceed the national average. Moxico Province, Angola’s largest by area, consistently records the highest absolute forest loss, with an estimated forty-five thousand hectares cleared annually. The province’s sparse population and vast distances from administrative centres make ground-based monitoring and enforcement essentially impossible, rendering satellite surveillance the only viable oversight mechanism.
Cuando Cubango Province, in southeastern Angola, presents a different pattern. Here, deforestation is concentrated along river corridors and transport routes, creating linear clearance patterns that are clearly visible in satellite imagery. The construction of new roads connecting previously isolated communities has opened forest frontiers to commercial charcoal production and mechanised clearing, accelerating loss rates by an estimated thirty percent since 2020.
Bié Province, in the central highlands, exhibits the most complex deforestation dynamics. The province sits at the ecological transition between miombo woodland and Afromontane grassland, and forest loss here is intertwined with fire regime changes, agricultural expansion, and natural savannisation processes. Disentangling anthropogenic deforestation from natural vegetation dynamics in this transitional zone requires multi-temporal analysis spanning at least five years and careful calibration against ground-truth data.
Ground-Truth Validation and the Data Gap
The value of satellite-based deforestation monitoring ultimately depends on the accuracy of the algorithms used to convert spectral signals into land cover change estimates. Validation requires ground-truth data — direct field observations confirming that locations flagged as deforested by satellite analysis have indeed been cleared. This is where Angola’s monitoring system faces its most significant challenge.
Angola’s ground-truth data infrastructure is sparse by international standards. The country’s last comprehensive national forest inventory was conducted in the 1970s, during the Portuguese colonial period, and the data is now obsolete. A new national inventory was initiated in 2024 with support from the United Nations Food and Agriculture Organization, but fieldwork has been slow due to logistical challenges, security concerns in former conflict zones, and limited technical capacity. As of early 2026, ground-truth validation plots have been established in only three of Angola’s eighteen provinces.
The consequence of this data gap is significant uncertainty in national deforestation statistics. Different satellite-based estimates of Angola’s annual deforestation rate vary by as much as forty percent, depending on the algorithms, training data, and minimum mapping units employed. This uncertainty propagates through national carbon accounting, REDD+ programme design, and international reporting obligations, undermining Angola’s capacity to participate effectively in global climate finance mechanisms.
Integration with National Policy
Despite these challenges, satellite deforestation monitoring is increasingly integrated into Angolan environmental policy and governance. The Ministério do Ambiente has established a satellite monitoring unit within its Direcção Nacional de Biodiversidade that receives weekly Sentinel-2-based deforestation alerts for priority conservation areas. These alerts, generated by an automated processing chain developed with European technical assistance, identify clearing events larger than two hectares within seven days of their occurrence, enabling targeted enforcement responses.
The monitoring system has already demonstrated its value. In 2025, satellite alerts identified a previously unknown industrial-scale charcoal operation in the buffer zone of Cangandala National Park, Angola’s smallest and most endangered protected area. The operation, which had cleared approximately one hundred and twenty hectares of primary forest, was shut down following a joint enforcement action by environmental police and park rangers. Without satellite monitoring, the clearing would likely have continued undetected for months or years.
Angola’s National Determined Contribution under the Paris Agreement commits the country to reducing deforestation by forty percent from the 2015 baseline by 2030. Achieving this target will require a dramatic scaling of both monitoring and enforcement capabilities. Current plans envision the deployment of a national deforestation monitoring system covering all eighteen provinces by 2028, with automated alert generation, provincial-level reporting dashboards, and integration with the ANGOGEO spatial data infrastructure.
The Path Forward
The technical foundations for comprehensive deforestation monitoring in Angola are now in place. Sentinel-2 provides free, open-access imagery with sufficient spatial and temporal resolution to detect forest disturbance at operationally relevant scales. Cloud computing platforms, particularly Google Earth Engine and the European Space Agency’s openEO, enable processing of petabyte-scale satellite archives without the need for local computing infrastructure. And machine learning algorithms continue to improve the accuracy and timeliness of automated change detection.
What remains is the harder work of building institutional capacity, securing sustained funding, closing the ground-truth data gap, and connecting satellite intelligence to effective enforcement on the ground. Angola’s miombo woodlands are a globally significant carbon store, a critical habitat for biodiversity, and the livelihood foundation for millions of rural Angolans. Their fate will be determined not by the availability of satellite data, which is now abundant, but by the political and institutional will to act on the intelligence that satellites provide. The orbital perspective reveals the problem with unprecedented clarity. The response must come from the ground.