Estimate faction of photosynthetically active radiation with three-band vegetation indices based on HJ-CDD satellite in wheat

The fraction of absorbed photosynthetically active radiation (FPAR) is a key biophysical variable that can be used to derive agronomic information for crop condition monitoring and yield prediction. Vegetation indices have been proved to be a useful method for FPAR estimation. However, most vegetation indices have their own limitations and can decrease the sensitivity to FPAR. The object of this study is that three new vegetation indices, a characteristic of three-band vegetation indices with green band, were tested to estimate green part of FPAR (FPARgreen) and aim at improving the accuracy of FPAR estimation. These vegetation indices are Revised NDVI (RevNDVI), Added green band NDVI (AgbNDVI) and Mean and standard deviation-band vegetation index (MSVI). The top-of canopy (TOC) reflectance and FPAR were simulated by the PROSAIL radiative transfer model. The relationships between these vegetation indices and FPAR were performed and then the results were compared with the vegetation indices of NDVI, Green NDVI and WDRVI. A global sensitivity analysis (GSA) approach, the Extended Fourier Amplitude Sensitivity Test (EFAST) method, was used to determine the key impact factor of each vegetation index based on simulated crop canopy reflectance (PROSAIL). Validation study was also conducted using the ground measurements in wheat. The results show that the three new vegetation indices had significant sensitive to FPAR with high correlation coefficients. Both AgbNDVI and RevNDVI are more sensitivity to green part of plants and leaf chlorophyll content, and the AgbNDVI was better than RevNDVI. The result indicated that it will have a benefit for increasing accuracy of FPARgreen retrieving by AgbNDVI or RevNDVI.