Characterising Satellite Soil Moisture Drying Rates Using a Bivariate Recursive Filtering Approach
Abstract
Physical constraints to acquiring in-situ soil moisture measurements over large spatial domains have promoted satellite remote sensing. However, discrepancies in the measuring depths of soil layers and spatial coverages generate systematic errors in the drying rates of soil wetness. Verifications of consistency between satellite and in-situ soil moisture measurements during the drying phases are often performed without any reference to the hydrological processes commonly perceived in ground measurements. The drying phases are generally termed the drydown events with no infiltration of precipitation into the soil. The drydowns are expressed mathematically with an exponential decay function, characterised by a recession coefficient that indicates how rapidly the drying of surface soil moisture occurs. The present study demonstrates a bivariate recursive filter that attempts to correct the systematic deviation in the drying rates of satellite surface soil moisture compared to in-situ datasets. The study is conducted with overall 525 in-situ global stations and Soil Moisture Active Passive (SMAP) L4 gridded datasets. The approach aims to minimise the mean distance between SMAP and in-situ points in the joint distribution plot of recession coefficient and initial wetness condition. Considerable improvement is achieved in the correction of the drying rates of SMAP observations across spatial heterogeneity primarily defined with dryness index, sand and clay fraction representing climatic and surface attributes. The present approach is extendable to other purely derived satellite SM datasets that are indicative of similar systematic errors in drydowns.
- Publication:
-
AGU Fall Meeting Abstracts
- Pub Date:
- December 2022
- Bibcode:
- 2022AGUFM.H32H..01S