"*" indicates required fields
Content Type
Persona
From mapping overnight floods to monitoring wildfire spread and supporting conflict response, this blog offers a glimpse into how Earth observation scientists use satellite data to guide humanitarian action around the world.
At the forefront of the world’s most urgent global crises, Earth Observation (EO) scientists use satellite data, remote sensing technologies, and GIS tools to support humanitarian efforts across a wide range of emergencies and disasters. From floods in Bangladesh to wildfires in California and conflict in Ukraine, their work provides insights into when and where aid is needed, how to deliver it, and what future risks remain.
EO science plays an important role in the global effort of humanitarian response. It involves coordination across federal agencies, mutual aid networks, international NGOs, and local response teams.
No single day looks the same, but here’s a glimpse into what one might involve.
Flooding surged through northern Bangladesh last night, an area prone to flash floods during monsoon season.
To explore what happened overnight, the EO scientist downloads Sentinel-1 data from the Copernicus Open Access Hub. This high-resolution Synthetic Aperture Radar (SAR) data provides us with a clearer view because it can penetrate cloud cover and easily detect water through specular reflectance. Additionally, VIIRS data from the Suomi NPP satellite is utilized to confirm local and regional power outages aligning with flooded areas, assisting with identifying stranded communities.
Within hours, maps are sent to humanitarian agencies and local responders to assist with evacuation planning, identifying stranded communities, and directing aid to communities cut-off from resources.
Example of an interactive flood map utilizing NOAA data from the Bangladesh floods in July 2020
While some efforts focus on now, others model what’s next. EO scientists don’t just respond to individual events; they track their echoes through time utilizing satellite imagery. Humanitarian crises rarely operate on a single timeline.
Consider Cyclone Idai, which struck southeastern Africa in 2019. Satellites captured the storm surge, over 20 inches of rainfall, and widespread flooding that devastated communities across Mozambique, Malawi, and Zimbabwe. Imagery revealed the extent of damage to infrastructure, including washed-out roads, collapsed bridges, and inundated buildings. But the crisis did not end when the floodwaters receded. EO data helped track the cascading effects long after the initial event.
In the weeks that followed, displacement maps showed tens of thousands still living in temporary shelters in flood-prone areas. Months later, NDVI analysis revealed widespread crop failure as waterlogged fields could not support the next planting season. A year on, drought indicators and food security assessments pointed to famine risk, and infrastructure in many regions still had not recovered.
Spatial distribution of April Normalized Difference Vegetation Index (NDVI) averages in the Sofala Province of Mozambique. Pre-cyclone (left: Average NDVI over 7 years (2012–2018), and post-cyclone (right: 2019).
Mid-afternoon meetings connect the EO scientist with teams on the frontlines of environmental vulnerability. Through collaborative platforms, EO scientists exchange data and insights with regional analysts.
Today’s calls focus on long-term deforestation patterns in the Amazon and crop stress monitoring in East Africa.
In the Amazon, analyses utilizing Landsat and Sentinel-2 imagery monitor short- and long-term changes to the forest canopy. By running time series analysis using NDVI and random forest classifiers on Sentinel-2 data, the EO scientist can detect areas of deforestation and differentiate between seasonal leaf loss and permanent land cover change.
In East Africa, EO data feeds into agro-climatic early warning systems used by aid organizations like FEWS NET, helping farmers anticipate food shortages before they escalate. For our EO scientist, this means calculating NDVI anomalies and overlaying rainfall data from CHIRPS to anticipate early crop failure signals.
These conversations are a two-way street. Local experts provide context that no algorithm can detect, verifying essential on-the-ground conditions and identifying potential blind spots in ongoing analysis. Some examples include confirming damage to irrigation canals, identifying when forest cover loss is seasonal or from illegal logging, and warn of unexpected threats not yet visible from orbit.
Just before the day winds down, an emergency hits: an unexpected bombing near a civilian corridor in Ukraine. As reports filter in, the EO scientist pulls available pre- and post-event satellite imagery to begin assessing damage.
Using the same methodologies refined in recent crises such as SAR coherence change detection workflows, pixel-wise t-tests, and machine learning on SAR time series models, the EO scientist quickly identifies damaged buildings, estimates the likelihood of destruction, and maps potential evacuation routes. Some tools also incorporate open-source intelligence by cross-referencing satellite signals with ground reports, creating a faster and more complete picture. These approaches apply directly to conflict zones like Ukraine and help responders assess damage and define safe corridors within hours.
Time is critical. Roads must be verified, shelters identified, and responders coordinated before nightfall if possible.
Meanwhile, the morning’s flood event in Bangladesh is still evolving. The EO scientist does not pause that work. Instead, they balance both crises in parallel, navigating different partners, response priorities, and decision timelines. One event is regional, and weather driven. The other is geopolitical and volatile. Both require speed, precision, and deep collaboration.
This is the reality of humanitarian EO. It is layered, global, and constantly in motion.
In addition to formal responsibilities, many EO scientists contribute to open data efforts like the Humanitarian OpenStreetMap Team (HOT). After hours, they log into the Tasking Manager to help map roads, buildings, and infrastructure in crisis-affected regions. These efforts often support areas that remain blank on commercial maps.
This volunteer work helps close data gaps that directly impact humanitarian access. For instance, during Typhoon Rai in the Philippines or the Syria Türkiye earthquakes, OSM contributions often inform logistics planning and disaster response in under mapped regions.
“Map interface from OpenRouteService for Disaster Management, displaying Dinagat Islands in the aftermath of Typhoon Rai. Updated every 10 minutes with OpenStreetMap data
The data EO scientists use is crucial for humanitarian response and recovery, but it also plays a role in preparedness. That’s where predictive EO work comes in.
Before logging off for the day, the EO scientist checks trajectories of wildfires spreading in California, as well as drought indicators in the Southeast United States. Both events may require EO-informed early warning systems, evacuation protocols, or agricultural advisories in the days to come. No model is perfect, but with timely EO data humanitarian agencies can prepare earlier and act with a more informed approach.
NASA’s Terra satellite captures smoke plumes from multiple wildfires burning across California in July 2021. Visible imagery like this supports air quality assessments, disaster response coordination, and early evacuation planning.
While this may not be the everyday routine of an EO scientist, it offers insight into the scope of what their daily work could be like.
Because at the end of the day, the work is not about maps – it’s about the humans represented within them. It’s about life. So, every pixel counts.
