Assessing the Risk of Rice Production under Drought Conditions in the Jhuoshuei River Alluvial Plain

In Taiwan, water storage grapples with challenges from steep slopes, fast-flowing streams, limited land size, and a dense population, intensifying constraints on water resources. Climate-sensitive agriculture, especially vulnerable to drought, faces threats to Taiwan’s food supply, particularly in the Jhuoshuei River Alluvial Plain, crucial for rice cultivation. To investigate the drought-rice yield relationship there, this study focused on Changhua County, chosen for its comprehensive meteorological data, encompassing maximum and minimum temperatures, precipitation, wind speed, relative humidity, and solar radiation. Given the geographical features of Changhua County, where farmers heavily rely on groundwater, drought events may induce land subsidence due to excessive groundwater pumping. Utilizing AquaCrop, a crop-water productivity model developed by the Food and Agriculture Organization of the United Nations, we simulated historical rice yields from the first crop between 2016 to 2021, configuring model parameters with soil properties, groundwater, water quality, and meteorological data. Rice growth conditions were determined by the growth degree days method, and irrigation was modeled with the first two transplanting stages having a 3 cm ponding depth, followed by a third stage with weekly draining and drying, a nutrient phase with 5 cm water, and final drainage before harvesting. For drought assessment, we calculated consecutive drought days and standardized precipitation index (SPI) to evaluate rice yields risk. A sensitivity analysis of temperature, precipitation, additional irrigation, and harvesting time was conducted to comprehend potential effects on rice yield under drought conditions. Despite drought events in 2017, 2019, and 2021 (SPI≦-1), records showed no reduction in rice yields. AquaCrop simulations revealed a rice yield range of 4.02 to 8.51 (t/ha) with a root mean square error (RMSE) of 2.40 (t/ ha) and mean absolute percentage error (MAPE) of 22.66%. Farmers may have mitigated drought impacts by pumping groundwater. Assessing irrigation water by pumping, adding 1 to 3 irrigation events brought simulations closest to rice production records, with an RMSE of 0.06 t/ha and MAPE of 0.65%. The Sensitivity analysis results showed a strong correlation (R²=0.97) between precipitation change and yield. While no clear linear relationship existed between temperature change and yield, reductions in temperature could increase production. Additional irrigation, up to four times during drought, resulted in the highest yield improvement (2.06 t/ha). Beyond the fifth irrigation, yield slightly decreased to 2.05 t/ha, with no further improvements. Lastly, changes in the harvest period exhibited a high correlation with yield (R²=0.99). These findings provide a valuable reference for developing agricultural policies to ensure food supply and farmers’ livelihoods. Future research will explore the relationship between drought and rice yield under climate change scenarios to understand potential agricultural risks better.