Improving soil moisture assimilation efficiency via model calibration using SMAP surface soil moisture climatology information

Authors: ZHOU, J. - Hohai University; Crow, Wade; WU, Z. - Hohai University; DONG, J. - Massachusetts Institute Of Technology; HE, H. - Hohai University; FENG, H. - Hohai University

Submitted to: Remote Sensing of Environment
Publication Type: Peer Reviewed Journal
Publication Acceptance Date: 7/1/2022
Publication Date: 7/18/2022
Citation: Zhou, J., Crow, W.T., Wu, Z., Dong, J., He, H., Feng, H. 2022. Improving soil moisture assimilation efficiency via model calibration using SMAP surface soil moisture climatology information. Remote Sensing of Environment. 280. Article 113161. https://doi.org/10.1016/j.rse.2022.113161.
DOI: https://doi.org/10.1016/j.rse.2022.113161

Interpretive Summary: The accurate tracking of soil moisture availability is critical for a number of agricultural applications including drought monitoring, large-scale yield forecasting, and irrigation scheduling. The current state-of-the-art approach for such tracking is a land data assimilation system where background water balance model estimates of soil moisture are periodically updated by relevant remote sensing observations. However, to date, these systems have been tasked only with improving the temporal tracking of soil moisture. This is a key limitation, because many agricultural applications (e.g., numerical weather prediction or drought monitoring) would also potentially benefit from an improved representation of soil moisture spatial variability. This paper describes the development of a new data assimilation system that, for the first time, can improve both the temporal and spatial skill of soil moisture estimates for agricultural monitoring applications. Once fully developed, this approach will be used by operational agencies within USDA to improve their ability to track crop soil water availability within large geographic regions

Technical Abstract: Prior to data assimilation (DA), satellite remotely sensed (RS) soil moisture (SM) retrievals are typical-ly rescaled to mitigate systematic differences relative to the SM climatology of a land surface model (LSM). This preprocessing implicitly discards spatial information contained in the RS SM climatology and, as a result, potentially degrades SM spatial patterns generated by a land surface DA system. Here, we leverage the RS SM climatology information in a DA framework based on a previously proposed model calibration method and investigate its potential DA benefits. The SM climatology is derived from L-band Soil Moisture Active Passive (SMAP) L3 surface SM retrievals, and the DA framework is built based on the Ensemble Kalman Filter (EnKF) and the Variable Infiltration Capacity (VIC) model. After calibrating the VIC model against the SMAP mean SM climatology, both Advanced SCATterometer (ASCAT) and SMAP SM retrievals are assimilated. Results show that model calibration utilizing SMAP SM climatology information not only directly improves the spatial pattern of daily and time-averaged SM estimates in the DA SM analysis, but also corrects the response of model error variance to model perturbations and thereby yields better Kalman gains that improve the SM analysis time series. In addition, SM improvements are effectively propagated into improved streamflow estimates. Overall, RS SM climatology information is shown to be valuable for SM DA and model calibration methods that retain such information will improve future land surface DA systems.