MULTI-SENSOR ASSIMILATION OF AMSR-E SPECTRAL DIFFERENCES AND GRACE TERRESTRIAL WATER STORAGE RETRIEVALS TO IMPROVE MODELED SNOW ESTIMATES

Loading...
Thumbnail Image

Files

Publication or External Link

Date

2020

Citation

Abstract

Snow, a key component of terrestrial water storage (TWS) in many watersheds across the globe, is a significant contributor to the Earth’s hydrologic cycle, energy cycle, and climate system. This study explores multi-sensor, multi-variate data assimilation (DA) using synthetic Advanced Microwave Scanning Radiometer for EOS (AMSR-E) passive microwave (PMW) brightness temperature spectral differences ($\Delta$$T_b$) and synthetic Gravity Recovery and Climate Experiment (GRACE) TWS retrievals in order to improve estimates of snow water equivalent (SWE), subsurface water storage, and TWS over snow-covered terrain. A series of synthetic twin experiments are conducted using NASA Catchment land surface model as the prognostic model. AMSR-E $\Delta$$T_b$ DA using a support vector machine as the observation operator improves SWE estimates, but adds little value to subsurface storage estimates. A physically-informed GRACE TWS DA approach significantly enhances the TWS vertical resolution via discretization into SWE and subsurface components more accurately. When AMSR-E $\Delta$$T_b$ and GRACE TWS are assimilated simultaneously, dual assimilation significantly improves the SWE estimates with a 14.1% reduction of RMSE (relative to the Open Loop without assimilation) and leads to the largest improvement in TWS estimates (RMSE = 66.4 mm) and most reliable subsurface water storage ensemble spread (spread-error ratio = 1.08) as compared to the single-sensor DA scenarios. However, dual DA does not always yield complementary updates, and can at times, lead to conflictory changes to SWE. That is, the assimilation of $\Delta$$T_b$ often generates positive SWE increments whereas assimilation of TWS often removes SWE in the dual DA system, which can ultimately degrade the posterior SWE estimates. This synthetic experiment provides valuable insight for future DA experiments merging real-world AMRS-E/AMSR-2 $\Delta$$T_b$ and GRACE/GRACE-FO TWS retrievals in order to better characterize terrestrial freshwater storage across regional and continental scales.

Notes

Rights