Multi-temporal comparisons of ERS-1 and JERS-1 SAR data for land applications

The site chosen for this investigation is Flevoland in the Netherlands. The terrain is very uniform and mostly covered by bare soil surfaces, crops, and forests. ERS-1 SAR precision images in PRI format and JERS-2 SAR data version 2.1 collected during 1993 are used as well as ground data measured by local investigators at the time of the overflight. Due to the excellent stability of the spaceborne satellites, it is possible to make multi-temporal observations over the site. The technique is used to improve the classification of the different crops. Moreover, taking into account that the SAR data are sensitive to temporal changes in surface roughness and soil moisture, the variations of these two parameters are investigated over a period spanning several months. In order to retrieve geo- and bio-physical parameters from microwave remote sensing data, a very detailed description of the interaction of the electromagnetic waves with the earth terrain being imaged is required. In the case of SAR images, both in situ data and radar data are needed in order to validate direct models, i.e. predict the backscattering coefficient as a function of the soil parameters, and retrieval algorithms to derive bio- and geophysical parameters from the SAR data. Some of the latest direct models for rough surfaces are tested and their domain of validity for ERS-1 data is investigated