Soil Moisture (SM) is recognized as one of the essential climate variables required for understanding the interactions between the Earth’s surface and the atmosphere through energy and water cycles. Nevertheless, computation of the spatiotemporal distribution of land soil moisture content, the moisture contained in the unsaturated soil, is a complicated task conventionally achieved through ground sensors and satellite sensing. The former provides high-accuracy point-based measurements, but it has a limited spatial coverage, high deployment costs and very limited global availability. The latter provides low-cost spatiotemporal SM data but at coarse spatial and temporal coverage. VAIS Global Soil Moisture Engine (GSME) is a satellite-based product that provides SM data for any point on Earth by using both active and passive microwave sensing, near infrared, and thermal infrared data. GSME employs the advanced analytic technologies used in our Virtual Field Probing technology for fusing multi-modal remote sensing data acquired from tens of satellites to deliver high-accuracy at up to 20-meter resolution surface level and root zone soil moisture information in near real-time. Unlike existing soil moisture products, our high-resolution GSME enables us to deliver value and address the needs and challenges faced by many farmers all over the world, especially smallholder farmers residing in arid and semi-arid regions.
Currently available remotely-sensed soil moisture is referred to as skin or surface level soil moisture with depths ranging from only few millimeters for optical and thermal bands to few centimeters for microwave sensors in the X-, C- or L-bands. A plethora of approaches is found in the literature that estimate root zone soil moisture based on surface level soil moisture. The connection between root-zone and surface-level soil moisture is achieved through basic data assimilation techniques that extrapolate surface level information to lower depths via flow models, empirical models that associate brightness temperature and in-site soil moisture time-series, or models that use a variety of weighting functions to correlate soil temperature between upper and deeper layers. Majority of these methods need vertically dense soil profile measurement or simulation, rely on prior knowledge that is hard to acquire in practice, or make assumptions regarding soil temperature gradients along the optical depth. To solve this complex problem, the Global Soil Moisture Engine offers a unique approach that enables accurate estimation of root-zone soil moisture. A deep data assimilation model is utilized to combine surface level moisture data with soil information, in-situ measurements, elevation data, hydrological models, and meteorological models to compute actual soil moisture at depths up to 100 centimeters.