Satellite remote sensing for surface soil water content estimation

Abstract

Satellite remote sensing is a useful source of observations of land surface hydrologic variables and processes and could be a practical substitution of conventional in-situ monitoring. Most of hydrological dynamic processes change not only throughout the years but also within weeks or months and their monitoring requires frequent observations. The most prominent advantage of the remote sensing technologies is that they offer a synoptic view of the dynamics and spatial distribution of phenomena and parameters, often difficult to monitor with traditional ground survey, with a frequent temporal coverage. Many of the variables in the land surface water balance can now be observed with satellite techniques thanks to an extensive development over the last decades. Often the problem connected to the use of remotely sensed data is their accuracy that, according to the sensor used and to the application considered, can ranges from moderate to excellent. The objective of this thesis has been to evaluate the use of satellite remote sensing techniques for the monitoring of two variables useful for hydrology applications: water body extension and soil moisture monitoring. The capability to map water surface is important in many hydrological applications, in particular accurate information on the extent of water boundary is essential for flood monitoring and water reservoir management. Often, this information is difficult to retrieve using traditional survey techniques because water boundaries can be fast moving as in floods or may be inaccessible. In this PhD thesis, an artificial basin for which in-situ information about the water extension are available is used as case study. The area extension recorded daily by the dam owner is compared to the one retrieved by using satellite images acquired from SAR and TM/ETM+ sensors. The outcomes of the analysis show that satellite images are able to map water body surfaces with a good accuracy. The analysis also highlighted the factor to be taken into account while using types of sensors. Soil moisture is recognized as a key variable in different hydrological and ecological processes as it controls the exchange of water and heat energy between land surface and the atmosphere. Despite the high spatial variability of this parameter it has been demonstrated that many satellite sensors are able to retrieve soil moisture information of the surface layer at catchment scale. Among other sensors, the Scatterometer is very useful for climatic studies and modelling analysis thanks, respectively, to the temporal frequency, global coverage and to the long time series availability. Even though the ERS Scatterometer has been designed to measure the wind over the ocean surface, in recent years it has been pointed out that backscattering measurements have high potentiality for soil moisture retrieval. The second task of this PhD thesis, concerning the use of satellite data for soil moisture monitoring, has been developed at Serco S.p.A. in the framework of the Advanced Scatterometer Processing System (ASPS) project developed by ESA (European Space Agency) to reprocess the entire ERS Scatterometer mission. Since the beginning of the ERS-1 Scatterometer mission in 1991 a long dataset of C-band backscattering signal from the Earth surface is available for studies and researches. This is a very consistent dataset, but in particular for climatology studies it is important to have high quality and homogeneous long term observation as also stated in the key guidelines included in the Global Climate Observing System (GCOS) from the World Meteorological Organization (WMO). The main goal of this task has been the generation of the new Scatterometer ASPS products with improved data quality and spatial resolution. This achievement required a long preparation activity but represents an important contribution to the C-band Scatterometer dataset available to the scientific community. In order to evaluate the usage of the re-processed Scatterometer data for soil moisture estimation, the backscattering measurements derived in the new ASPS products have been then compared to in-situ volumetric soil moisture data and the relationship between radar backscattering and soil moisture measurements has been investigated under different conditions: angle of incidence, angle of azimuth, data measurements resolution, season of the year. Analysis results show that a relationship between the C-band backscattering coefficient and the in-situ volumetric soil moisture exists and takes into account the incidence and azimuth angles and the vegetation cover. [edited by author]