Passive and Active Multiple Scattering of Forests Using Radiative Transfer Theory With an Iterative Approach and Cyclical Corrections

In this paper, a unified framework of vegetationscattering using radiative transfer theory for passive and activeremote sensing of vegetated land surfaces, especially thoseassociated with moderate to large vegetation water contents(VWCs), e.g., forest field, is presented. The framework allowsfor modeling passive and active microwave signatures of thevegetated field with the same physical parameters describingthe vegetation structure. Radiative transfer equations are solvedby a numerical iterative approach for both passive and activeconfigurations. This approach allows including higher orderscattering which represents multiple scattering. In fields likeforests with large VWCs, associated with large scattering albedoand optical thickness, multiple scattering effects are critical.In the active iterative approach, cyclical terms are identifiedand backscattering enhancement is included by doublingcontributions from cyclical terms. The method is applied to aspentrees in forest fields to compute the brightness temperatures andbackscattering coefficients for passive and active remote sensingconfigurations, respectively. In passive configuration, for forestfield with VWC of 15 kg/m2, the deviation between zeroth orderbrightness temperature, i.e., the tau-omega model results, andmultiple scattering results around 40 observation angle, can beas large as 50 K for vertical polarization and 35 K for horizontalpolarization. In active configuration, the deviation between firstorder results, which is identical to distorted Born approximation,and the multiple scattering results around 40 incidence angle, isabout 1.6 dB for V V and 0.7 dB for HH polarization. Multiplescattering is shown to be crucial for accurate forward modeling,especially over forested areas. The proposed approach is thussuitable for vegetation scattering with large VWCs. Furthermore,the proposed model is validated with the Passive and ActiveL-band Sensor (PALS) acquired in SMAPVEX12 measurementsin 2012, which demonstrates the applicability of this model.