Integrating sub-pixel SIF extraction with a mechanistic model for maize yield estimation in north China

Against the background of global food security, as well as China’s carbon peaking and carbon neutrality goals, accurate and timely monitoring of crop yields is essential. However, coarse-resolution satellite solar-induced chlorophyll fluorescence (SIF) data suffer from mixed-pixel effects, limiting their use in refined yield estimation. To isolate crop-specific signals, a sub-pixel decomposition framework was adopted based on high-resolution vegetation indices. These signals were incorporated into a mechanistic light response (MLR) model and combined with a calibrated harvest index (HI) to estimate summer maize yield in the North China Plain (NCP). Validation with statistical data showed that the framework effectively simulated GPP at the site scale and reproduced spatial yield patterns at county and provincial scales. In major maize producing provinces, the R2 value reached 0.72, while in regions with scattered cultivation, the model still explained over 50% of the yield variance. These results demonstrate that the MLR model is applicable to maize production systems in China, and when combined with sub-pixel SIF extraction, it enables accurate yield estimation at regional scales.