Natural disasters, such as floods, (wild)fires, volcanic eruptions, droughts, hurricanes, tsunamis, earthquakes or landslides are often so large in scope, it's difficult and expensive to monitor from the ground. Aerial and satellite imagery is invaluable in mapping large areas, helping people assess the extent of a disaster, damages and possible future developments.
Some disasters are caused by humans - to learn more about how satellite imagery can be used to monitor armed conflicts, read our guest blog post by Wim Zwijnenburg.
To explore how satellite imagery can be used to monitor disasters yourself, visit EO Browser, and choose a theme you're interested in (such as volcanoes, wildfires, or floods and droughts) in the dropdown menu. Then select highlights, where we have prepared interesting locations for you to explore. Read the descriptions, and try out different premade visualizations, that are commonly useful for each event.
Several useful custom scripts can be found on our custom scripts github page, where we have a category Disaster management and prevention algorithms for Sentinel-1 and Sentinel-2. Each of them has a description and a direct link to EO Browser.
Wildfires occur naturally in summer, when the vegetation is dry due to a drought, and the ground is hot due to high temperatures. In such conditions, even a single spark can ignite a devastating wildfire. Sometimes, fires are ignited by the sun, but more often, they are ignited by humans, who are either burning the land for agriculture or wood, or by accident, such as with a discarded cigarette or a campfire. Once a fire started, it can spread rapidly, if the wind is strong enough, resulting in a catastrophe.
Monitoring land surface temperature and the state of the vegetation due to drought is important in preventing wildfires, and looking at wildfires from space makes it possible to assess their extent, helping us predict their behaviour and assess damages. Satellite imagery is commonly used to detect and map wildfires. Sentinel-2 satellite short wave infrared bands 11 and 12 are heavily reflected by heated areas and are capable of penetrating the atmosphere, thus giving us valuable information on what is happening under thick smoke. Additionally, a near infrared band 8 of Sentinel-2 can be used to map vegetation, as NIR wavelengths are heavily reflected by dense and healthy vegetation. We can thus use this information to construct indices and visualize wildfires.
On the left image below, you can see the Oregon wildfire on September 9, 2020, in natural colors, emmiting thick smoke. On the right image, a custom composite using SWIR and NIR bands was used to penetrate the smoke and highlight the fires and the area burned.
As wildfires emit harmful gases, such as carbon monoxide, carbon dioxide, methane and nitrogen dioxide, satellites capable of detecting these, such as Sentinel-5P, can be of great help in monitoring such atmosphere changes. On the image below, you can see a clear increase in carbon monoxide gas emitted by the same wildfire.
Satellite imagery can assist environmental monitoring by detecting changes in the Earth's vegetation, atmospheric trace gas content, sea state, ocean color, and ice fields. By monitoring vegetation changes over time, droughts can be monitored by comparing the current vegetation state to its long term average.
Meteorological droughts are defined by rainfall deficiencies over an extended period of time and can turn into agricultural droughts, which are characterized by soil water deficiency and subsequent plant water stress and reduced yield. Agricultural droughts can then turn into hydrological droughts, which refer to deficiencies in surface and subsurface water supplies. Different drought definitions imply that several parameters are used to monitor droughts: precipitation, temperature, evapotranspiration, soil moisture, and vegetation. These parameters can be observed from space; the popular approach, due to its simplicity, is using vegetation indices. Vegetation indices like NDVI or EVI are freely available in near-real-time (cf. Data Application of the Month on vegetation indices). In addition, calculating the anomalies or the Vegetation Condition Index indicating the state of the greenness of vegetation in relation to a time-series is straight-forward. For further information visit this website.
On the images below, Sentinel-2 moisture index was used to compare the area before (left) and during (right) the drought, showing the desiccation of the Red Hills Reservoir over the heat period in 2019 with a clearly retreating water line.
In the last couple of years we have experienced record-breaking extreme weather. In terms of lives lost and property damaged, floods are among the most devastating natural disasters. No matter what type of flooding occurs, in most cases it leaves devastating consequences. To assess the damages experts use different approaches, among which the use of satellite data is becoming more and more common. Copernicus Sentinel data, delivering near-real-time images, are a good source of information for authorities coping with the crisis during and after flooding. They provide images of a wide area, showing a clear picture of the overall extent of the flood. Part of this constellation of satellites is also Sentinel-1, a radar satellite which has the ability to show the difference between waterlogged and dry land.
Read our blog post on crisis management after flooding here.
Below, we can observe the striking difference between the before (left) and after (right) Sentinel-2 images of Bangladesh flooding in 2017, using a SWIR visualization.
On the image below, multitemporal Sentinel-1 imagery was used to map the flood in Aghghala, Iran on March 23, 2019. A different date was used as a refference and input into the blue channel, while the image on the day of the flooding was input into the red channel. The difference between the two displays flooded areas in red color.
Called hurricanes when they develop over the North Atlantic, central North Pacific, and eastern North Pacific, these rotating storms are known as cyclones when they form over the South Pacific and Indian Ocean, and typhoons when they develop in the Northwest Pacific.
Hurricanes begin as tropical disturbances in warm ocean waters with surface temperatures of at least 26.5 degrees Celsius (80 degrees Farenheit). Those low-pressure systems are fed by energy from warm seas. A storm with wind speeds of 61 kilometers (38 miles) an hour or less is classified as a tropical depression. It becomes a tropical storm — and is given a name - when its sustained wind speeds top 63 kilometers an hour.
Hurricanes bring destruction ashore in many different ways. When a hurricane makes landfall, it often produces a devastating storm surge—ocean water pushed ashore by wind—that can reach six meters (20 feet) high and move several kilometers inland. Storm surges and flooding are the two most dangerous aspects of hurricanes, accounting for three-quarters of deaths from Atlantic tropical cyclones. A hurricane’s high winds are destructive and may spawn tornadoes. Torrential rains cause further damage via flooding and landslides, which may occur many kilometers inland.
The best defense against a hurricane is an accurate forecast that gives people enough time to get out of the way. Read more here.
Unless aircraft reconnaissance is available, surveillance relies almost exclusively on spaceborne observations. Sentinel-1 satellite can be used to discern hurricane wind speed (learn more), and optical satellites, like Sentinel-OLCI are used to view the hurricanes, map their size, and interpolate their path. They are also commonly used by press. Images below were taken by Sentinel-3 OLCI satellite.
Sentinel-2 satellite can be used to map latent heat release inside deep convective clouds, as it plays a crucial role in intensification of hurricanes and cyclones in general (left image below, see a custom script). Sentinel 5P can provide information about cloud thickness and pressure, giving us additional information on the cyclone's characteristics (right image below).
Earthquakes occur when two blocks of the earth suddenly slip past one another, which happens around tectonic plate boundaries. The energy produced by such event travels in all directions, shaking the earth. Earthquakes are very common, as they happen several times per day. However, some of them are devastating; powerful earthquakes can cause many casualties from collapsed buildings, debree, or landslides. They can also trigger tsunamis, which often cause additional devastation on top of the earthquake. Learn more.
Earthquakes and tsunamis are nowadays mapped using aerial and satellite imagery. Large scale high resolution imagery makes it possible to assess the extent of the destruction top down. Using different acquisitions, a difference between before and after the disaster is looked at, resulting in a map of changes, corresponding to destruction.
On 28 September 2018, the Indonesian island of Sulawesi was struck by a 7.5 magnitude earthquake. The epicentre was on the island’s northwest coast – 77 km north of Palu, which lies at the head of a long narrow bay. The quake triggered a tsunami that swept huge surges of water – as high as 10 m – along the bay and swamped the city. The combination of the earthquake, tsunami, soil liquefaction and landslides claimed well over 2000 lives, destroyed homes, buildings, infrastructure and farmland in several districts.
European Space Agency mapped the aftermath using Sentinel-2 and Sentinel-1 data, providing people with the information on ground deformations, which are difficult to assess from the surface. This information is essential for Indonesia to rebuild effectively. Learn more.