In addition to the GRACE TWS products, data assimilation techniques have been applied for other emerging drought monitoring tools, particularly in the assimilation of surface soil moisture retrievals derived from microwave remote sensing into the water balance component of a land surface model. Data assimilation refers to mathematical techniques designed to integrate sparse observations (in both space and time) into more continuous dynamic models. Relative to the use of observations alone, a data assimilation analysis has three main advantages: (1) it produces estimates that are more temporally and spatially continuous, (2) it provides for the optimal merger of observations and models such that the impact of independent errors in both are minimized, and (3) it allows for the efficient extrapolation of information between observed and unobserved land surface states. The third advantage is particularly relevant for the assimilation of microwave soil moisture because the vertical support of microwave soil moisture retrievals is limited to only the first few centimeters of the soil column. As a result, land data assimilation systems are commonly employed to produce (deeper) root-zone soil moisture predictions that are constrained by a time series of surface soil moisture observations (Kumar et al. 2009).
Recently, a number of these data assimilation systems have been implemented operationally to produce soil moisture information, which is relevant for drought monitoring. Collectively, these systems currently provide the best available representation of continuous global variations in root-zone soil moisture availability. A primary example is the 9-km SMAP surface and root-zone soil moisture products (http://nsidc.org/data/docs/daac/smap/ sp_l4_sm/index.html) based on the assimilation of the SMAP radiometer brightness temperature into the NASA global modeling and assimilation office catchment model (Reichle et al. 2016). In addition, near-real-time SMOS retrievals are currently being assimilated into the USDA's 2-layer Palmer model to produce a global 0.25-degree root-zone soil moisture analysis for large-scale crop conditions assessment by the USDA Foreign Agricultural Service (Bolten and Crow 2012). Resulting root-zone imagery from this analysis is regularly posted at http://www.pecad.fas.usda.gov/cropexplorer.