Dryspells and Minimum Air Temperatures Influence Rice Yield and its Forecasts in Rainfed Systems
Abstract
Crop yield predictions on an inter-seasonal to inter-annual horizon are subject to a diverse set of uncertainties associated with climate forecast scenarios. Here we evaluate a novel approach of producing probabilistic forecasts of seasonal rice yields that uses a coupled hydrologic-crop modeling framework. At a provincial scale, the crop model extensively captured the uncertainties in yield whilst significantly complementing observations over the growing season. In addition, we investigated the information mutually exchanged between yield and hydrologic/drought variables over different time frames within a season. We found a higher synchronization of information transfer between yields, dryspells and minimum air temperatures from planting to harvest, with the prevalence of strong explicit links between these variables and crop yield towards the end of season. These insights are expected to aid crop yield forecasting as well as nowcasting, especially in data-poor regions.
