Summary: In this dissertation addresses the issue of estimation of the surface soil moisture through remote sensing radar. The surface soil moisture is a variable that is involved in many processes that take place on the Earth's surface. His estimate from space-based sensors would be very attractive to disciplines such as hydrology, agriculture, meteorology, etc.. Remote sensing radar sensors emit a pulse of radiation into the soil surface and receive the same proportion of returning to the sensor, thereby permitting calculation of the coefficient of the backscatter s0 of surface. This coefficient depends on the dielectric properties of the soil surface, which in turn are closely related to moisture content. However, there are other variables that influence the radar observations, such as surface roughness, thus complicating the estimation of surface soil moisture from such images. Various techniques have been proposed for estimating the soil surface moisture from radar images. These include the implementation and investment models backscatter is the most suitable option. Other techniques, such as linear regression models or techniques for detecting changes require homogeneity roughness and angle of incidence, so its applicability is smaller. Among the various models that have been proposed backscattering the Integral Equation Model (IEM) is the most appropriate. There are others who may be of interest because of its simplicity and because it incorporates the simpler description of the roughness, but the IEM has a solid theoretical foundation and has been successfully validated in laboratory conditions. In this dissertation, evaluates the applicability of these methods for the estimation of surface soil moisture in a basin agricultural Navarre. It has acquired comments by two space-based sensors, RADARSAT1 and ENVISAT / ASAR in two rounds of experiments conducted in the years 2003 and 2004. In these experimental campaigns were conducted field measurements of soil moisture and the surface roughness. These measurements are used to estimate the parameters necessary roughness in the model and to provide backscatter measurements of the moisture of reference with which to compare the estimates are made. The existence of vegetation complicates the study of moisture through this technique. In this thesis correcting the influence of grain a deck using a model semi-empírico called Water Cloud Model (WCM). This method provides a simple and useful tool to correct the attenuation exercising vegetation. Among the different models evaluated backscattering, the IEM is the one that gives results more appropriate. The empirical model Oh et al. (
1992) does not work correctly in terms of angle of incidence and low bit rough surfaces. The model of post-Oh (2004), by contrast, provides adequate results. The ability of the estimates improves the greater the level of aggregation or scale at which it is estimated moisture. In this paper we have obtained estimates wide basin with an error of 0.06 cm3cm-3, comparable to that obtained with methods of measuring moisture in the field. Estimates at the spot are not so appropriate because of the influence of the roughness space and spatial variability. The surface roughness is the main sticking point of the estimated moisture through remote sensing r 8 adar. His a62 high spatial variability, on the one hand, and the sensitivity of its backscattering coefficient s parameters on the other, make it necessary to characterize in great detail. In this context, were used iterative schemes based on the method proposed by Pauwels et al. (2002) allowing both to estimate the parameters of surface roughness and moisture from two points gained in homogeneity roughness. These schemes are based on the combined use of two models backscattering forming a system that is solved in an iterative fashion. While the foundation of this methodology is interesting, the results vary with the roughness, angle of incidence and humidity making such schemes are not generalizable. On the other hand, sse has tried to estimate the parameter roughness correlation length l, as measured in the field is more complicated paths through expressions based on the work of Davidson et al. (2003) and Baghdadi et al. (2002, 2004). The results obtained by these methods are interesting because they show the possibility of estimating the parameter of the standard deviation of the heights of the surface parameter considerably easier to determine. Being a variable surface roughness knowledge of which is vital to estimate soil moisture from radar remote sensing, this thesis explores in depth characterization. In the context of this issue has designed a perfilómetro laser has proven to be a valuable tool for the study of roughness. Based on measurements obtained using the same Analyzing the behavior of various existing parameters for the characterization available for the characterization of the roughness, the influence of tillage on such parameters, its variability or scale of measurement that is appropriate to adequately characterize the roughness on farmland.
