Radar teledetection for estimating the superficial humidity of soil

This is the conclusion of Jesús Álvarez Mozos in his PhD thesis, Evaluation of the applicability of radar teledetection in estimating the superficial humidity of agricultural basins, defended recently at the Public University of Navarre.

Roughness of the terrain

Jesús Álvarez Mozos analysed photographs taken by two currently active radar satellites, RADARSAT-1 and ENVISAT/ASAR, that respectively belong to the Canadian Space Agency and the European Space Agency. In concrete, seven images were studied and which correspond to the Navarre Government experimental agricultural basin known as the Tejería, located near Villanueva de Yerri.

From this analysis, Álvarez Mozos has concluded that a good knowledge of surface rugosity is a key factor in order to make a correct estimate of humidity based on photographic images. Roughness is based on random variations of the terrain and depends on the crops grown and the tasks carried out on the soil. For example, a field on which a furrow plough has worked will be rougher than one that has been sown with seed. In other words, the high spatial variability of the field, on the one hand, and, the sensitivity of the coefficient of retrodispersion to its parameters, on the other, behoves a more detailed characterisation.

Given the importance of this parameter and with the aim of being able to measure it, a laser profilometer was designed. This proved to be a very valuable tool for the study of rugosity. Thus, based on measurements taken with this instrument, the behaviour of various existing parameters was analysed for rugosity, as well as the influence of tilling the soil on such parameters, its variability or scale of measurement which proved to be suitable in order to characterise the rugosity of agricultural surfaces in an optimum manner.

Reliablity of the scale of a geographical/agricultural basin

Once the rugosity data for a specific terrain was measured, estimates for humidity – on the scale of a geographical basin – were obtained from the high-reliability images. The reliability of the estimates improves with the level of aggregation or scale with which the humidity is estimated. In this research, geographical basin-scale estimates were obtained which had an error of 0.06 cm3cm-3, comparable to that obtained with humidity measurement methods in situ.

The results of the research pointed to the fact that another parameter that bears on the measurement of humidity through teledetection is vegetation. However, it is possible to correct the influence of this in quite a simple way. Thus, values or characteristic variables of wheat, for example, given that the zone analysed is a dry farming one (cereal crop and its biomass), may be corrected and representative values for the soil obtained.

Radar satellites

Knowledge of soil humidity can be of great utility in, for example, the detection of water stress conditions or to set out needs for crop irrigation, as well as to evaluate damage due to drought and to administer the concomitant remedial measures.

Over the last fifty years, with the aim of estimating soil humidity, a number of techniques have been developed which can be classed into two types: precise estimates and those based on teledetection. The latter enable the estimation of the humidity content of the surface based on the measurement of electromagnetic radiation emitted and/or reflected by it. In general, the measurement of this radiation is undertaken by means of electronic sensors installed on space platforms.

In concrete, the satellites studied for this thesis incorporate synthetic aperture radar sensors (SAR), which are being increasingly used for evaluating natural resources. Amongst its advantages, the fact that photographs can be taken whether there are clouds or not, is particularly important. Moreover, the main space agencies (Canadian, Japanese, etc.) plan to launch a number of satellites of this type in 2006 and 2007, all with enhanced characteristics with respect to those studied for this thesis.