Space agency's satellite observation helps with wildfires

Space agency's satellite observation helps with wildfires

• Print
• Email

As climate change and increased land use continue to increase the frequency and intensity of wildfires, it is critical that South Africa's capabilities to deal with wildfires be improved to protect lives, property and the environment.

Since its inception in 2011, the South African National Space Agency (SANSA), an entity of the Department of Science, Technology and Innovation, has been at the forefront of Earth observation. SANSA's geospatial services transform raw satellite data into actionable information that can be used support environmental monitoring, disaster management, urban planning and agriculture.

According to Dr Abel Ramoelo, SANSA's Executive Director for Earth Observation, Earth observation satellites that deliver near-real-time information enable the detection and monitoring of active wildfires, inform operational decision-making during fire events and enable post-fire impact assessments.

Satellite data with varying spatial and temporal resolutions are integrated to use in all phases of the fire disaster management cycle, from early warnings and emergency responses to recovery and rehabilitation.

"By leveraging a combination of open-access and commercial satellite imagery, SANSA enhances wildfire detection capabilities, monitors fire spread, and supports the assessment of fire origins and associated damage. These capabilities provide essential situational awareness to disaster management authorities and land managers, enabling faster response times and more effective allocation of firefighting resources," said Ramoelo.

However, as wildfire behaviour becomes increasingly complex and less predictable, reactive monitoring alone is no longer sufficient. There is a clear need to strengthen fire detection and monitoring systems by incorporating predictive capabilities that can anticipate where and when fires are most likely to occur. In response, SANSA is advancing the integration of artificial intelligence and machine-learning-based predictive models into its wildfire monitoring framework.

"These AI and machine-learning models integrate historical fire occurrence records with satellite-derived variables such as vegetation condition, fuel moisture, land surface temperature, rainfall anomalies, wind patterns, topography and land-use characteristics," Ramoelo explained.

He said that, by learning from past fire behaviour and evolving environmental conditions, the models generate probabilistic fire-risk forecasts and short-term fire spread predictions. "This approach shifts wildfire management from a largely reactive process towards a proactive, risk-based system".

The introduction of predictive wildfire intelligence supports early warnings, enables the pre-positioning of firefighting resources, and improves preparedness at local, provincial and national levels. "By integrating AI-driven fire-risk forecasting with near-real-time satellite monitoring, SANSA strengthens its capacity to support evidence-based decision-making, reduce wildfire impacts, and enhance resilience in the context of a changing climate," said Ramoelo.

Predictive fire risk maps and early warning indicators allow disaster management authorities, land managers and conservation agencies to implement preventive measures, pre-position resources, and plan controlled burns more effectively. Ultimately, integrating predictive capabilities into SANSA Earth Observation's fire monitoring framework would strengthen national disaster preparedness, reduce response times, and minimise the environmental, economic and societal impacts of wildfires.

Areas in the Eastern and Western Cape affected by recent wildfires are shown in the maps below. Some of the images are not in true (natural) colour, but in false colour, which uses different colours to represent different intensities in radio emissions not visible to the human eye.

The false-colour image shows active fires and fire burn scar around Franschhoek and Wemmershoek Forest Reserve areas as captured on 9 January 2026. The area affected is displayed in brown-orange tones with a black outline.

The natural colour image shows active fires and fire burn scar around the Garden Route National Park on 8 January 2026. Smoke can be observed around the areas that were burning when the satellite image was acquired.

The false colour image on the map shows old and new fire burn scars around Humansdorp on 8 January 2026. The areas affected by fire in January 2026 are located in the red circle.

The mapped burn scar reveals a continuous and elongated fire footprint extending parallel to the coastline, with the highest burn intensity concentrated in the central and eastern portions of the affected area. The spatial pattern suggests wind-driven fire spread, with the scar widening in areas of continuous natural vegetation and narrowing near infrastructure corridors. Although the fire perimeter approached the outskirts of Pearly Beach, the burnt area largely remained outside the urban footprint, indicating limited direct impact on residential zones. The burn scar geometry and alignment are consistent with regional topography and prevailing coastal wind conditions during the January 2026 fires.

The Stanford fire scar displays a large, continuous burnt area with pronounced spatial heterogeneity. High burn impact zones are concentrated in the central and western sections of the scar, while lower-intensity or patchier burn patterns are observed toward the periphery. The fire affected both natural vegetation and cultivated lands, reflecting fuel continuity across land-use types. The burn perimeter intersects multiple transport corridors, indicating potential disruption to infrastructure during the fire event. Overall, the spatial extent and configuration of the scar suggest rapid fire propagation under favourable burning conditions on 13 January 2026.

The Pearly Beach wildfire produced an extensive and largely continuous burn scar spanning coastal and inland landscapes. The most severely affected areas are concentrated within the central portion of the fire footprint, where strong reductions in vegetation cover are evident. More heterogeneous burn patterns are observed toward the edges of the scar, suggesting variability in fire intensity and fuel availability. The proximity of the burn scar to the coastline and nearby settlements underscores the vulnerability of coastal environments to wildfire events.

The map above illustrates the burn scar and burn severity pattern associated with the wildfire that occurred in the Soetkraal Nature Reserve in the Eastern Cape between 8 and 10 January 2026. The fire extent is delineated by the outlined boundary, within which burnt and unburnt areas are mapped using satellite imagery.

Burnt areas are shown in red to orange tones, indicating extensive vegetation loss caused by the fire, while green areas represent vegetation that remained unburnt or was minimally affected. The spatial distribution of the burn scar highlights a largely continuous fire footprint across mountainous terrain, with limited unburnt patches occurring primarily along drainage lines and topographic breaks, suggesting localised variations in fire intensity and fuel conditions.

The proximity of the burn scar to key landmarks, including the Soetkraal Nature Reserve and nearby settlements such as Krakeel, emphasises the ecological and socio-environmental significance of the event. The mapped fire footprint covers approximately 144 km², indicating a large-scale disturbance with potential implications for biodiversity, soil stability and post-fire erosion processes.

The burn scar imagery was acquired shortly after the event, enabling clear differentiation between burnt and unburnt surfaces based on their spectral response. This satellite-based assessment provides an objective and spatially consistent representation of fire impacts, supporting post-fire damage assessment, rehabilitation planning, and environmental monitoring.

This map depicts the burn scar associated with the wildfire that occurred near Greytown in the Western Cape on 10 January 2026. The fire extent is delineated by the outlined boundary, within which burnt and unburnt areas are distinguished using satellite-derived information.

Burnt areas are shown in red to orange tones, indicating varying degrees of vegetation loss and surface disturbance caused by the fire. Areas displayed in green represent vegetation that remained unburnt or experienced limited fire impact. The spatial pattern of the burn scar reveals a largely continuous fire footprint across mountainous terrain, with pockets of unburnt vegetation persisting along drainage lines, agricultural boundaries, and areas of lower fuel continuity.

The fire scar covers an estimated 138 km², affecting predominantly natural fynbos ecosystems and adjacent land-use areas. The proximity of the burnt area to the town of Greytown highlights the interface between natural vegetation and human settlements, underscoring the potential risks posed by wildfires in the wildland-urban interface.

The burn scar was mapped using Sentinel-2 satellite imagery, which enables clear discrimination between burnt and unburnt surfaces based on post-fire spectral responses. This satellite-based assessment supports rapid damage evaluation, environmental impact assessment and post-fire rehabilitation planning.

The map below shows the extent and severity of a wildfire near Mossel Bay along the southern coast of South Africa. Burnt areas are highlighted in red to orange and unburnt areas in green, overlaid on satellite imagery. The fire affected parts of Asandrif, Island View and Vleesbaai during January 2026, with the mapped burn scar covering roughly 6,4 km². Nearby urban areas and infrastructure are visible, showing how close the fire came to settlements.

SANSA's Earth Observation division continues to analyse the wildfires to assist government and private sector with insights to manage the fires and to support post-fire recovery efforts. To access detailed reports, maps or data for other areas, please contact Customer Services at This email address is being protected from spambots. You need JavaScript enabled to view it..

• tweet