Comparative analysis of forest biomass retrieval from water cloud model (WCM) of polarized SAR data with different wavelengths

  Objective  Water cloud model (WCM) is a semi empirical model using SAR data to retrieve forest aboveground biomass (AGB). The objective of this study is to explore the capability of introducing different wavelengths at different polarization channels into WCM for forest AGB inversion. And through the exploration, it is expected to provide scientific reference for improving the accuracy of forest AGB retrieval.

  Method  In this paper, firstly, we applied WCM in forest AGB estimation at X-, C-, L- and P-band with HH, HV, and VV polarizations, respectively, and their results were compared and analyzed. Then a parameter named the ratio of surface scattering power and volume scattering power was constructed based on polarization decomposition components and embedded in WCM, here we named it PolWCM. The potential of PolWCM on forest AGB estimation was explored by X-, C-, L- and P-band polarimetric decomposition components.

  Result  (1) HV backscattering coefficients showed best performance in forest AGB estimation using WCM at C-, L- and P-band, among them, L- and P-band performed better than X- and C-band (R² = 0.46, RMSE = 18.0 t/ha for L-band and R² = 0.43, RMSE = 21.18 t/ha for P-band). (2) PolWCM performed better than WCM for forest AGB estimation at X-, C-, L- and P-band, respectively. Their RMSE values for X-, C-, L- and P-band were 24.90, 24.71, 17.70 and 18.08 t/ha, respectively.

  Conclusion  The forest AGB estimation of WCM shows obvious dependence on wavelength and polarization, HV backscatter coefficients at L-band perform best in forest AGB estimation. Polarimetric information embedded in WCM through the ratio of surface scattering power and volume scattering power can improve the forest AGB estimation accuracy.