Precipitation effects on L-band SAR imaging of tropical forest and non-forest areas

摘要

We discuss the effects of precipitation and drought events on dual polarimetric ALOS-2 L-band imaging of tropical forest and non-forest areas. Using more than 1 billion GSMaP datasets in a dedicated database together with 3 years of PALSAR-2 ScanSAR timeseries data, we calculate the statistical relation between rainfall and gamma naught in a big data analysis. Antecedent rainfall prior to the satellite overpass is accumulated for intervals ranging from 1 to 14 days. The results show that in the dry forest, Pearson correlation coefficients can be higher than 0.9 while the denser humid rainforest show virtually no correlation. Characteristic relations for different forest types can be used as additional information for improved forest classifications.