Combining sun-induced chlorophyll fluorescence and seasonal climate forecast for 8-day dynamic in-season maize yield prediction in northeast China

Satellite-based solar-induced chlorophyll fluorescence (SIF) has shown a promising skill for end-of-season crop yield prediction due to its close linkage with photosynthesis. However, the benefits of SIF have rarely been examined for in-season crop yield forecasts, which would depend on current-phase crop growing status and unknown-stage climate conditions. By leveraging SIF, seasonal climate forecasts and machine learning, we build an in-season maize yield forecast system at the 8-day scale in Northeast China (NEC). The value of SIF is demonstrated by comparing it against traditional vegetation indices (VIs). Overall, reliable yield forecasts can be achieved two months before the harvest (jointing–tasseling) in NEC, with an average bias of less than 2.5%. Assimilating SIF into the yield forecast system exhibits a better performance than VIs except in the medium-growing stage. The added value of SIF is more pronounced in the dry and hot years, especially under the early and early-medium growth phases.  Attribution analysis reveals that the absorbing radiation signal carried by SIF is the main driver for its advantage over VIs under the early and early-medium phases, while its outperformance under the medium-late and late stages is related to both the reflection of photosynthetic rate and the absorbing radiation signal. This study provides a valuable framework for weekly yield predictions, which has great implications for early warning of yield loss risk in China.