Abstract:Wetlands are important ecosystems as they are known as the "kidney of the earth". Particularly, urban wetlands play an important role in providing both natural and social beneficial services. However, urban wetlands are suffering from various human impacts, such as excessive land use conversion, air and water pollution, especially those in developing countries undergoing rapid industrialization and urbanization. Therefore, it is of great necessity to derive timely biomass information for optimal design, management and protection of urban wetlands. In this paper, we develop a set of models for estimating above ground biomass (AGB) in Xixi National Wetland Park in Hangzhou, China by using optical images and Synthetic Aperture Radar (SAR) images. A series of vegetation indices (VIs) derived from optical data is introduced along with spectral data. The modeling methods consist of (1) curve estimation; (2) linear regression for multivariable model; (3) Back Propagation Neural Network (BPNN) modeling. Curve estimation is a combination of linear and nonlinear regressions. It is applied to generate AGB models from a single variable including Normalized Difference Vegetation Index (NDVI) and radar backscatter coefficient. The models are then compared via three accuracy metrics. According to the results, SAR models generally show better accuracy than optical models and BPNN models show the greatest accuracy among all the models. The BPNN model from the combination of Terra Advanced Spaceborne Thermal Emission and Reflection Radiometer (ASTER) and European Remote-Sensing Satellite-2 (ERS-2) SAR (Synthetic Aperture Radar) image has the least root mean square (RMSE, 0.396 kg/m 2 ), least mean absolute error (MAE, 0.256 kg/m 2 ) and the greatest correlation coefficient (0.974). The results indicate that AGB can be estimated by integrating optical and SAR imagery. Four maps of AGB are derived to illustrate the distribution of AGB in the study area. The total AGB in the study area is estimated to be between 165,000 and 210,000 kg/m 2 .