Towards a continuous operational system to estimate the root-zone soil moisture from intermittent remotely sensed surface moisture

Abstract

A study has been carried out to develop and evaluate a system to estimate soil moisture content in the root-zone using active microwaves from the European Remote Sensing Satellite, ERS-1, to measure moisture content in the top 10 cm of the soil profile. Two permanent grass sites in the UK with contrasting soil types, clayey and sandy, were selected for this study. The system consists of an initialization phase, which provides surface and root-zone moisture contents as initial values for the dynamic phase of a soil water balance model. The initial value of surface moisture can be either a remotely sensed or a measured value. The surface moisture value for a given day is in turn used to derive the initial value of the root-zone moisture for the same day. This can be obtained either from an empirical relationship for drying or wetting conditions, or during drying conditions alone. Both types of relationship have been established for each of the sites, with strong coefficient of determination, R 2. The two-layer soil dynamic model requires as input daily rainfall, evapotranspiration and three soil physical parameters—soil moisture at field capacity, wilting point and a pseudo-diffusivity coefficient. The first layer represents the remotely sensed layer, taken as 0–10 cm, and the second represents the root-zone, taken as 0–50 cm, for both sites. The model has been run for 1992 and 1993. The model was not initialized by remote sensing data owing to an insufficient number of microwave backscatter-surface moisture data pairs to produce a relationship with good R 2. It is hoped that the continuing collection of data will improve the relationships. Initial soil moisture contents of both layers were considered to be at field capacity, which is usually the case during winter time. The dynamic model, which offers a good balance between accurate description of the processes and minimum input of data, proved capable of simulating both surface and root-zone moisture content. Despite the fact that the two phases of the system have been tested using ground-truth measurements, it has the potential to produce continuous surface and root-zone moisture contents using intermittent remotely sensed surface moisture to initialize or update the model. Therefore, further tests with remote sensing data and entirely independent soil moisture data sets (under collection) are still required. At present, the system has been tested on two soil types and one vegetation cover, namely permanent grass. For a wider use of the system at a regional scale, further developments are planned to account for more soil types and land covers.