Terrestrial Water Storage
Soil moisture status is a key parameter for drought monitoring because it is a primary driver of drought-related vegetation stress when soil moisture levels approach the wilting point. Given that plants respond to soil moisture
Drought condition in July 2011 based on 6-month Standardized Precipitation Index (SPI) generated using the global precipitation climatology project (GPCP) data. Downloaded from the global drought monitoring and prediction system (GIDMaPS; Hao et al. 2014).
conditions and regulate their water consumption by balancing moisture availability versus evaporative demand, soil moisture represents an indicator of early-stage vegetation drought stress. As discussed earlier, in situ measurements of soil moisture are somewhat limited in the United States and lacking or nonexistent in many parts of the world. Microwave remote sensing has proved useful for estimating soil moisture conditions because the microwave emissivity of soil is strongly impacted by the amount of soil water present. This sensitivity has been leveraged for the development of techniques for inferring surface soil moisture content via satellite-based observations in the microwave spectrum. Two decades of field campaign and aircraft studies support the conclusion that the microwave L-band (near 1.4 GHz) represents the preferred frequency band for such retrievals. As a result, the remote sensing of soil moisture entered a new era with the launch of the first two L-band satellite missions designed specifically for soil moisture retrieval: the European Space Agency's soil moisture ocean salinity (SMOS) mission in 2009 and the NASA soil moisture active/passive (SMAP) mission in 2015.
Since 2010, the SMOS mission (Kerr et al. 2010) has produced a global 45-km soil moisture product with approximately 2-3 day revisit. In particular, SMOS level 3 soil moisture retrievals are publicly available (following registration) at http://www.catds.fr/Products/Products-access, and extensive validation of SMOS soil moisture products is described in Kerr et al. (2016). The SMOS mission is based solely on passive microwave radiometry. In contrast, the NASA SMAP mission was designed to merge soil moisture information acquired simultaneously from both passive radiometry and active radar observations (Entekhabi et al. 2010). The SMAP radar failed in July 2015, but the radiometer has continued to function well and produce high-quality soil moisture products. SMAP has benefited from the application of a sophisticated radio frequency interference (RFI) mitigation strategy developed in response to L-band radio RFI discovered during early portions of the SMOS mission. SMAP 36-km resolution soil moisture products are currently available (with ~24 hours latency and 2-3 day update cycle) at https://nsidc.org/data/smap/smap-data.html. Ground-based validation of these products is described in Chan et al. (2017).
Regardless of their source, microwave-based surface soil moisture retrievals suffer from three primary shortcomings: (1) lower spatial resolutions compared to standard VI datasets (typically greater than 30 km), (2) limited vertical sampling depth, and (3) reduced accuracy over heavily vegetated surfaces. A broad range of spatial downscaling strategies is currently being applied for SMOS and SMAP soil moisture products, and, in mid-2017, the SMAP mission will begin operational production of a 9-km soil moisture product based on downscaling observations from the SMAP radiometer using backscattering observations acquired from the ESA Sentinel-1 satellite. The relatively shallow vertical soil penetration depth (~2-5 cm) of these products also represents an obvious limitation for agricultural drought monitoring because the soil moisture information is not indicative of the entire root zone condition, which influences plant stress. However, recent progress has been made in the development of land data assimilation systems that vertically extrapolate surface soil moisture across the entire vegetation root zone. In addition, the simple exponential filtering of remotely sensed surface soil moisture time series has shown promise for effectively recovering agricultural drought information contained in deeper root-zone soil moisture observations (Qiu et al. 2014). Finally, the attenuation of soil signals by the vegetation canopy is minimized by the use of longer microwave wavelengths, which have relatively less scattering and absorption interaction with vegetation than shorter wavelengths. As a result, the relative transparency of the vegetation canopy is greater for L-band SMOS and SMAP soil moisture products than for higher-frequency X- and C-band sensors used in earlier soil moisture remote sensing products.