Relating multi-incidence angle RADARSAT-2 data to vegetation characteristics in the Lower Paraná River floodplain (Argentina)

NATALIA SOLEDAD MORANDEIRA; MATÍAS BARBER; FRANCISCO GRINGS; FRANK AHERN; PATRICIA KANDUS; BRIAN BRISCO

Montreal

Workshop; Advanced SAR 2019 Workshop; 2019

Canadian Space Agency

The Paraná River floodplain comprises large extensions of freshwater wetlands at subtropical and temperate latitudes. In its lower portion, mosaics of herbaceous vegetation dominate, dotted by shallow lakes, rivers and streams. The resulting backscatter is due to the interaction among leaves, emergent stems and open waters. A total of 11 full polarimetric RADARSAT-2 C-Band scenes were acquired in February 2018, differing in their orbit direction and beams (FQ5 to FQ29, mean incidence angles between 24.4° and 47.4°). Scenes were calibrated to the coherence matrix and geocoded. Single polarization backscattering (σ0HH, σ0HV, σ0VV), the HH-VV phase difference and the contribution of the Freeman-Durden (FD) components (double bounce scatter, volume scatter, odd scatter) were extracted. During the acquisition period, the hydrometric level of the Paraná River was similar and relatively low. Field sampling was conducted in flooded sites dominated by the herbaceous species Schoenoplectus californicus or Ludwigia peruviana (4 and 3 monospecific sites, respectively). These species show contrasting architectures: Schoenoplectus has almost vertical stems and no leaves; while Ludwigia usually has a main stem with ramifications and leaves. Generalized linear models (GLM) were conducted in order to assess the relation between field-obtained biophysical measures and SAR variables.For Schoenoplectus, the overall pattern shows that σ0HH is higher than σ0VV, and σ0HV is 5 to 10 dB lower than σ0HH. According to GLM, σ0HH decreases with increasing incidence angle and increases with above-ground green biomass, mean height, moisture content and mean stem diameter. Models explain up to 72.8% of the total variation of σ0HH. The backscattering decrease with incidence angle was not significant for σ0VV and σ0HV. σ0HV was affected mainly by mean height, mean stem diameter and the number of green stems, with up to 78.7% of explanatory power. σ0VV was mainly explained by moisture content and stem diameter, explaining 68.0%. The mean HH-VV phase difference decreased with biomass, height and moisture content. FD decomposition attributes the majority of backscatter to volume scattering. The contribution of double-bounce scatter is low and can be explained up to 15.9% by moisture content and mean height; while volume scatter increases with all the vegetation variables, with mean height explaining 45.0% of the total variation. Odd bounce scatter decreases with incidence angle. For Ludwigia, σ0VV is only 1-3 dB below σ0HH, which is less than for Schoenoplectus, and σ0HV is still 10 dB below σ0HH. GLM have less explanatory power than for Schoenoplectus. The decrease of σ0HH with increasing incidence angle is also observed. Incidence angle and above-ground green biomass or mean height, explain up to 44.5% of the total variation in σ0HH. Biomass also contributes to explain σ0HV and the number of green stems contributes to σ0VV. Odd bounce scatter decreases with incidence angle.These results would lead to an improved understanding of the backscatter physics between C-Band SAR signal and herbaceous wetland stands.