An open-source tool for improving on-farm yield forecasting systems

Abstract

IntroductionThe increased frequency of extreme climate events, many of them of rapid onset, observed in many world regions, demands the development of a crop forecasting system for hazard preparedness based on both intraseasonal and extended climate prediction. This paper presents a Fortran version of the Crop Productivity Model AquaCrop that assesses climate and soil fertility effects on yield gap, which is crucial in crop forecasting systemsMethodsFirstly, the Fortran version model - AQF outputs were compared to the latest version of AquaCrop v 6.1. The computational performance of both versions was then compared using a 100-year hypothetical experiment. Then, field experiments combining fertility and water stress on productivity were used to assess AQF model simulation. Finally, we demonstrated the applicability of this software in a crop operational forecast system.ResultsResults revealed that the Fortran version showed statistically similar results to the original version (r2 > 0.93 and RMSEn < 11%, except in one experiment) and better computational efficiency. Field data indicated that AQF simulations are in close agreement with observation.ConclusionAQF offers a version of the AquaCrop developed for time-consuming applications, such as crop forecast systems and climate change simulations over large areas and explores mitigation and adaptation actions in the face of adverse effects of future climate change.