Simulating cultivar variations in potato yields for contrasting environments

Potato (Solanum tuberosum L.) is a major food commodity becoming increasingly important for food security, especially in the developing world. The rising demand for potatoes combined with yield gaps and potential adverse impacts from climate change call for strategies for yield improvement and environmental adaptation. Crop growth models can help identify and assess such strategies. In this study, the SUBSTOR potato model was used in a systematic assessment of all cultivar parameters in the model in a range of environments including temperate, subtropical, and tropical regions to identify options for future crop improvement and to develop strategies for climate change adaptation in potato production. Our results show that yield responses to changes in cultivar parameters are specific to the environment. Some changes are less effective in subtropical and tropical environments and more effective in increasing yields in temperate environments. Solar radiation, day length, and temperature are the environmental factors that constrain the effectiveness of cultivar parameters in changing yields. The simulated variation in yields among environments was larger than the variation from changes in cultivar parameters. The impact of cultivar parameter changes on yield also varied with the cultivar parameter. The potential tuber growth rate was the cultivar parameter with the strongest effect on yields. Changes in the potential tuber growth rate parameter can lead to large yield changes in tropical highlands and temperate environments that have high solar radiation to ensure sufficient assimilate production for a larger sink. Results also suggest that improving crop management (e.g., N input) is more important for increasing yields than the potential in cultivar improvement. The study showed that crop modeling can help assess alternative strategies of yield improvement and support targeting and prioritization of efforts to improve crop productivity across different environments, based on an improved understanding of genotype by environment by management interactions. Results also showed how crop models can yield insights relevant for climate change adaptation even when only using weather data of the current climate.