While ESA’s water mission was built to advance our understanding of Earth, it continues to show how well it’s suited to delivering information for numerous applications that improve everyday life. Taking this a step further, soil moisture data products are now available within three hours of measurement, which is essential for many applications.
SMOS was launched in 2009 to provide global observations of soil moisture and ocean salinity – two important variables in Earth’s water cycle.
The satellite captures images of ‘brightness temperature’, which correspond to radiation emitted from Earth’s surface and can be used to gain information on soil moisture and ocean salinity.
As well as being used to study how Earth works as a system, SMOS’ readings of brightness temperature have proved to be a completely new source of information for tracking hurricanes, measuring thin ice floating in the polar seas, for assessing fire risk, and more.
However, for SMOS to benefit society even more, its data need to be available fast – in what is termed ‘near-real time’, which means within three hours of sensing.
To accommodate this, the process of translating brightness temperature measurements into soil moisture products has been completely redesigned. It involved developing an artificial ‘neural network’, akin to the vast network of neurons in a brain.
After being trained with old soil moisture data, this neural network is now able to compute values of soil moisture from the satellite’s observations within seconds.
ESA’s SMOS mission scientist, Matthias Drusch, said, “Short latency and fast access to data products are very important for many applications such as weather prediction and flood forecasting.
“The neural network approach, developed at CESBIO, has allowed us to integrate state-of-the-art science into operational processing, opening the door for operational agencies.”
The operational data processing is being done at the European Centre for Medium-Range Weather Forecasts and the final data products can be obtained through Eumetsat’s Eumetcast system.
The fact that soil moisture data are available within three hours of sensing also makes it easier to combine SMOS data with similar information from other satellites.
In fact, SMOS and NASA’s Soil Moisture Active Passive satellite can provide accurate coarse-resolution soil moisture information. Measurements from the Copernicus Sentinel-1 satellite can then be applied to improve the resolution to ‘field scale’.
By combining measurements from different sensors the spatial resolution is increased from 25 km x 25 km to 100 m x 100 m.
VanderSat, a Dutch company that focuses on adding value to satellite data products, produces these images regularly, furnishing more than 3000 users with essential information.
VanderSat’s Richard de Jeu said, “The new data fusion method provides cost-effective and information-rich soil moisture information.
“This means that more informed decisions can be made – whether you are monitoring crops, predicting the weather, performing predictive analysis or preventing forest fires.”
Susanne Mecklenburg, SMOS mission manager, said, “A single satellite cannot provide high-accuracy datasets, high spatial resolution and fast global coverage. Therefore, a constellation of satellites with complementary instrumentation is needed to address the needs of agriculture, hydrology, weather forecasting, and climate applications.”