A Spatial Downscaling Method for Solar-induced Chlorophyll Fluorescence Product Using Random Forest Regression and Drought Monitoring

Abstract Solar-induced chlorophyll fluorescence (SIF), an electromagnetic signal that has been proven to be an efficient tool for monitoring and assessing gross primary productivity (GPP) and drought. To solve the problem of sparse resolution of SIF based on satellites, researchers have developed different downscaling algorithms. Recently, the SIF products frequently used is 0.05° spatial resolution. This study selected global ‘OCO-2’ SIF, normalized difference vegetation index, and land surface temperature products, obtained downscaled SIF data with 1km resolution using random forest method. Using the downscaled SIF results with 1km resolution, SIF anomaly index was established and calculated to monitor drought. Results showed that the RF-based downscaled SIF result had the same trend with GOSIF value. Subsequently, the correlation coefficients among SIF and GPP were calculated. The downscaled SIF had a higher correlation with GPP from MODIS than 0.05° GOSIF. The correlation coefficients were 0.74 and 0.68 in May 2018. Moreover, SIF anomaly index had positive correlations with crops yield, the correlation coefficients were 0.93 (wheat) and 0.89 (maize) respectively. Drought index had negative correlation with areas affected by drought and correlation coefficient was − 0.58. The SIF index based on RF method will be helpful for studying regional drought.