Bayesian combined active/pasive (B-CAP) soil moisture retrieval algorithm: a rigorous retrieval scheme for SMAP mission

MATIAS BARBER; CINTIA BRUSCANTINI; PABLO PERNA; FRANCISCO MATIAS GRINGS; HAYDEE KARSZENBAUM

Milán

Simposio; The International Geoscience and Remote Sensing Symposium 2015; 2015

On-going and near-future Synthetic Aperture Radar (SAR)satellite missions are expected to provide meaningful andtimely information about soil condition over vast agriculturallands such as those of the mid-western United States(corn-belt) and of Argentina (Pampas Plain), leading to actualeconomic benefits regarding to seeding dates, irrigationstrategies and crop yield forecasting. NASA?s Soil MoistureActive and Passive (SMAP) [1] and Argentinean SAOCOM(www.conae.gov.ar)missions have been specifically designedto develop surface soil moisture products. These missionswill exploit microwave radar at L-band ( = 23cm) as sensingwavelength, which demonstrated to be less sensitive toresidue cover over soil?s surface and to be more accurate onretrieving soil moisture than other bands.SAR systems offer the added advantages of fine resolution(on the order of 10 m), multiple polarimetric modes andvariety of beam incidence angles, which make them unique todevelop soil moisture products over agricultural lands. However,modeling the scattering processes that relate backscatteringcoefficient 0 to soil properties (moisture and roughness)is hampered by the speckle noise [2] and the difficulty inmeasuring soil roughness in the field [3]. The former mainlyrelates to the SAR imaging system, whereas the latter relatesto heterogeneity of soil properties. Moreover, several combinationsof surface parameters can often lead to the sameSAR observation. Thus, the impact of soil heterogeneity onretrieving soil moisture from SAR should be somehow minimized.In effect, agricultural land is often divided into smallerfields, each one characterized by the same land managementand land use over several growing seasons. Therefore, fieldsusually have the lowest variability of soil properties in relationwith larger areas. In this way, estimates defined on a fieldbasis are best suited for SAR-based retrieval purposes.