Fires release heat, and some of that heat is in the form of infrared energy that can be picked up even hundreds or thousands of km away by remote sensing instruments operating onboard Earth orbiting satellites.
This method of fire detection can be used to identify the location of landscape fires burning anywhere on Earth, and also to measure their strength, which relates for example, to how much material they are burning and releasing into the atmosphere in the form of smoke containing both gases and particulates.
The European Sentinel-3 satellite series carries an instrument called the Sea and Land Surface Temperature Radiometer (SLSTR), which can be used for this application.
Scientists at NCEO-Kings College London have designed and wrote the algorithm for the European Space Agency (ESA) and EUMETSAT that produces this information in near real time from the SLSTR data, to support for example, global and regional air quality forecasting.
The same team are also partners in the C3S project that takes these daily global data and makes them into simple, easy to use, consistent near-monthly products focused on quantifying the amount of fire on Earth. Such products can for example be used to identify how the amount of fire is changing over time – for example, in response to changes in climate or human use of the land.
The four planned Sentinel-3 satellites are scheduled to operate for around 15 years, and will then be replaced by even more capable versions being designed currently to generate an even longer-term dataset.
The overpass time of Sentinel-3 is very similar to that of the MODIS instrument onboard NASAs Terra satellite – and hence the SLSTR active fire detection dataset is likely to be used in conjunction with the Terra MODIS record stretching back to the early 2000s to deliver a very long term record capable of being mined to understand how and where patterns of fire are changing, and ultimately the drivers of this. Such information can then be used to evaluate and improve the type of Earth system models used to forecast future climate, which themselves need generally need to include the modelling of landscape fires and the carbon they release to the atmosphere.