Rice growth parameters retrieval in Central China in a complex rice cropping system using multi-temporal and quad polarization Radarsat-2 data

摘要

The objective of this study was to estimate rice growth parameters (LAI, crop height) in a timely and precise manner with FQ Radarsat-2 images. The proposed approach is based on the water cloud model, five Radarsat-2 images and coincident ground measurements obtained in Chizhou, in the Anhui province of Central China, during a single-cropping rice growth stage in 2014. This study investigated the comprehensive relationship between backscattering coefficients (VH, VV, HH, HV) derived from a high temporal frequency of Radarsat-2 images and rice growth parameters (LAI, crop height) during the vegetative stage, reproductive stage and entire growth period. Based on a correlation analysis and constructed models, the VV model (R2>0.8) was clearly better than HH model, and a VV image can be used to retrieve the LAI (vegetative stage, entire growth season); The crop height can also be derived from the VV image using the water cloud model with a determination coefficient above 0.6. The applicability of the models and SAR data from a different year (2013) was also assessed for Chizhou. The LAI time series and the crop height maps represented the rice growth status. The study demonstrated that the C-band SAR data and models would be useful for rice growth parameters retrieval independent of the region and year in regional scale.