Feet in the water and hands on the keyboard: A critical retrospective of crop modelling at AfricaRice

Rice is cultivated throughout Africa in a vast array of environments. Crop growth modelling at AfricaRice seeks to develop an understanding of genotype, management and environment interactions to inform research and development. This paper reviews progress made over thirty years of modelling, as well as the knowledge gaps remaining. Major advances were made in modelling phenology and heat- and cold-induced sterility. This crucially took into account the crop-generated microclimate via transpirational cooling in irrigated rice. On this basis, the RIDEV model and its successors provided effective support to applied breeding, genetics, agronomy and cropping systems research. As a major learning, rice very effectively avoids heat stress if it can transpire water abundantly. For water-limited systems, ORYZA2000 based yield gap, climate-change impact and drought mapping projects gave direction to AfricaRice's applied research agenda. But large gaps remain in modelling capabilities and underlying knowledge, particularly regarding biotic stresses, inland valley hydrology, and rice-based cropping sequences, e.g. including vegetable crops. In terms of understanding the physiology, more research is needed to accurately model spikelet number, thermal acclimation, photosynthesis response to extreme temperatures, and variation in rooting depth. This will require enhanced collaboration between AfricaRice and advanced research centers to resolve the scientific and technical bottlenecks in crop modelling.