900 IEEE TRANSACTIONS ON GEOSCIENCE AND REMOTE SENSING, VOL. 44, NO. 4, APRIL 2006 Using A Priori Information to Improve Soil Moisture Retrieval From ENVISAT ASAR AP Data in Semiarid Regions Francesco Mattia, Member, IEEE, Giuseppe Satalino, Laura Dente, and Guido Pasquariello Abstract—This paper presents a retrieval algorithm that es- timates spatial and temporal distribution of volumetric soil moisture content, at an approximate depth of 5 cm, using multi- temporal ENVISAT Advanced Synthetic Aperture Radar (ASAR) alternating polarization images, acquired at low incidence angles (i.e., from 15 to 31 ). The algorithm appropriately assimilates a priori information on soil moisture content and surface roughness in order to constrain the inversion of theoretical direct models, such as the integral equation method model and the geometric optics model. The a priori information on soil moisture content is obtained through simple lumped water balance models, whereas that on soil roughness is derived by means of an empirical ap- proach. To update prior estimates of surface parameters, when no reliable a priori information is available, a technique based solely on the use of multitemporal SAR information is proposed. The developed retrieval algorithm is assessed on the Matera site (Italy) where multitemporal ground and ASAR data were simultaneously acquired in 2003. Simulated and experimental results indicate the possibility of attaining an accuracy of approximately 5% in the retrieved volumetric soil moisture content, provided that sufficiently accurate a priori information on surface parameters (i.e., within 20% of their whole variability range) is available. As an example, multitemporal soil moisture maps at watershed scale, characterized by a spatial resolution of approximately 150 m, are derived and illustrated in the paper. Index Terms—A priori information, ENVISAT, model inversion, retrieval algorithms, synthetic aperture radar (SAR), soil moisture, soil roughness, surface scattering. I. INTRODUCTION S OIL moisture content is a parameter of major importance for land applications at both watershed and regional scale such as hydrology and agriculture (see [1] and [2] for a top- ical review). In the past, the sensitivity of radar measurements to soil moisture content (via the soil dielectric constant and the soil texture) and to soil roughness was demonstrated through ex- perimental and theoretical studies as in the case of [3] and [4]. The inverse problem of retrieving soil moisture and roughness information from the observed radar response of the surface has also been widely investigated (see [5] for an updated review). Manuscript received April 15, 2005; revised July 29, 2005. This work was supported in part by the European Space Agency, European Space Research and Technology Centre, under Contract 17011/03/NL/JA and in part by the Minis- tero delle Politiche Agricole e Forestali under Contract D.M. 209/7303/05. The authors are with the Istituto di Studi sui Sistemi Intelligenti per l’Au- tomazione, Consiglio Nazionale delle Ricerche, I-70126 Bari, Italy (e-mail: mattia@ba.issia.cnr.it). Digital Object Identifier 10.1109/TGRS.2005.863483 Relatively good results (i.e., root mean square errors on volu- metric soil moisture content ranging from between 3% and 7%) were obtained, for instance in [6]–[8] by adopting either purely statistical or semiempirical relationships relating synthetic aper- ture radar (SAR) backscatter and observed soil moisture con- tent. However, the above methods are specific to the site condi- tions where the relationship was observed and cannot be gener- alized to different areas. More general results could be achieved, at least in principle, by inverting theoretical direct scattering models, which should be adaptable to drastically different site conditions. Nevertheless, to date the use of SAR data to retrieve soil moisture content has been generally limited. The main reason is the intrinsic difficulty of estimating more than one unknown (i.e., soil moisture, soil roughness, soil tex- ture, etc.) using single-parameter radar measurements as pro- vided by the first generation of spaceborne SAR systems (i.e., ERS, JERS, RADARSAT). Typically, there exists many com- binations of surface parameters mapping the same SAR obser- vation. In this sense, the problem is “ill-posed experimentally” and the retrieved soil moisture content is characterized by poor accuracy [9]. The recent launch of the new European ENVISAT system, with the Advanced Synthetic Aperture Radar (ASAR) system onboard, which is able to provide C-band SAR data at two polarizations and at different incidence angles, has certainly improved the scenario. In particular, the higher potential of ASAR system for monitoring environmental parameters relies on its ability to revisit the same site with a relatively short repeating cycle (i.e., between three and seven days in Eu- rope). However, the soil moisture retrieval from ASAR data still remains an ongoing problem. In this respect, a valuable method in improving the accuracy of retrieval algorithms is to constrain the set of their possible solutions by assimilating into the inversion scheme a priori information on geophysical parameters [10], [11]. This is the methodology adopted in this study in order to estimate spatial and temporal distribution of volumetric soil moisture content, at a depth of approximately 5 cm, using horizontal (HH) and vertical (VV) multitemporal C-band ASAR backscatter measurements. The a priori infor- mation assimilated in the retrieval algorithm concerns both soil moisture and soil roughness. The former was obtained through simple lumped water balance models, whereas the latter was derived by using an empirical approach. The de- veloped retrieval algorithm finds the “best” solution for this problem by appropriately inverting theoretical direct models. In order to accept a large variability of roughness conditions 0196-2892/$20.00 © 2006 IEEE