Assessment of spaceborne SAR remote sensing for monitoring soil moisture | |
| Author | Ugsang, Donald M. |
| Call Number | AIT Diss. no.SR-00-2 |
| Subject(s) | Soil moisture--Remote sensing Synthetic aperture radar |
| Note | A dissertation submitted in partial fulfillment of the requirements for the degree of Doctor of Engineering. |
| Publisher | Asian Institute of Technology |
| Abstract | For many years, Earth Observation Satellites (EOS) have gained wide importance in understanding and monitoring changes to our environment. Recent advances in remote sensing have seen the development of radar satellites (e.g. ERS-2, Radarsat and Envisat). Their capability to obtain information independent of weather conditions and external illumination source have made them potentially important particularly in the tropics where persistent cloud cover limits the use of optical spacebome satellites like SPOT and LANDSAT. Monitoring of soil moisture is important to a number of applications, from climate and hydrological modeling at the global scale to local applications in agriculture and water management. Conventional method of soil moisture measurement by groundbased sampling can not provide a reliable estimate of soil moisture over a large area. Therefore, the principal advantage of remote sensing over conventional data collection methods is its capability to make frequent observations of soil moisture over a large area. Although remote sensing of soil moisture can be accomplished to some degree by all regions of the electromagnetic spectrum, only the microwave region offers the potential for truly quantitative measurement of soil moisture from space. The theoretical basis for measuring soil moisture by microwave technique is the large contrast between the dielectric properties of water and dry soil. This large contrast provides a means for the remote sensing of the moisture content in surface layer about 5 cm thick. However, the sensitivity microwave response to soil moisture variation is affected by surface roughness and vegetative cover. The uncertainties in soil moisture assessment brought about by these factors may be accounted for using ancillary data. This, however, should be demonstrated with further experiments particularly in rainfed paddy fields. This study evaluates the potential and limitations of using ERS-2 C-band SAR data for monitoring soil moisture in rainfed paddy fields in the West Sukhothai District of North-western Thailand. In particular, this study aims to establish the relationship between soil moisture and radar backscattering coefficient, and to identify the limitations in the soil moisture-radar backscattering coefficient relationship considering such factors as vegetation height and biomass, and surface roughness. Ground measurements of soil moisture and vegetation parameters were conducted concurrent to each image acquisition. This is to insure minimal anomalies in the establishment of soil moisture-radar backscattering coefficient relationship. Surface roughness measurements were conducted during the fallow period, where the root mean square height deviation was assumed to be at minimum, and during the land preparation where surface roughness is assumed to be at maximum. The ERS-2 C-band SAR data used in this study are three-look precision images (PRI) processed at the ESNESRIN Central Facility (EECF) in Italy. The ERS-2 C-band SAR data were corrected for geometric distortions and projected to Universal Transverse Mercator (UTM) coordinates. Speckle-reduction filters were applied to the ERS-2 C-band SAR data to reduce the effects of speckle on the images. Finally, the ERS-2 C-band SAR data were fully calibrated, using Laur's Method, to allow direct comparison of the radar backscattering coefficients of the images acquired in different dates. Results of the study show that ERS-2 C-band SAR data can be used to monitor soil moisture (top 5-cm layer) in rainfed paddy fields with good accuracy. Soil moisture can be estimated from ERS-2 C-band SAR data using the soil moisture (SMvo1)-radar backscattering coefficient (cr0 ) relationship expressed in the function SMvot = 5cr0 + 63.75, with an accuracy of R2 ~ 0.9 1. The equation was established under two conditions. First, when the rainfed paddy fields were bare until they are covered with rice crop at the early vegetative growth stage - 38 cm in height or 425 and 102 g/sq.m. in wet and dry biomass, respectively. Second, when the change in surface roughness throughout the growing period is small. A slight improvement in the accuracy (R2 = 0.913) was observed in the soil moisture-radar backscattering coefficient relationship in speckle-filtered images (e.g. Lee 5x5 and Gamma 3x3 filters). It was found out that the attenuating effect of vegetation on the radar backscattering coefficient from the soil is not significant on the early stages of the vegetative growth of rice, thus allowing the inference of the moisture content of the soil underneath. Beyond these early vegetative growth stages (i.e. when rice plant exceeds 38 cm in height or 425 g/sq.m. in wet biomass), the radar backscatter from the vegetation and its effect on the soil moisture-radar backscattering coefficient needs to be studied. The root mean square height variation of the soil surface, 2.21 cm during land preparation and 1. 1 cm during fallow period, is the typical range of surface roughness of the rainfed paddy fields in the study area. This small variation in surface height allows monitoring of the temporal variation of soil moisture in rainfed paddy fields. |
| Year | 2000 |
| Type | Dissertation |
| School | School of Advanced Technologies (SAT) |
| Department | Other Field of Studies (No Department) |
| Academic Program/FoS | Space Technology Application and Research (SR) |
| Chairperson(s) | Honda, Kiyoshi; |
| Examination Committee(s) | Murai, Shunji;Apisit Eiumnoh;Tokunaga, Mitsuharu;Musiake, Katumi; |
| Scholarship Donor(s) | Government of Japan; |
| Degree | Thesis (Ph. D.) - Asian Institute of Technology, 2000 |