Climate risk management for the rainfed rice yield in Lao PDR using APCC MME seasonal forecasts

The tailored process-based crop model equipped with seasonal climate forecasts have been a useful tool for managing agricultural risks from climatic variability. In this study, we provided a risk management framework for a rainfed paddy rice field in Savannakhet province of Lao People's Democratic Republic (PDR) that allows achieving the national target yield (5.0Mg ha-1) under climatic uncertainty. We collected the multi-model ensemble (MME) tercile hindcasts of 3-month (September to November, SON) and 5-month (July to November, JASON) rainfall for 1991–2010 from the international collaborative archive of the APEC Climate Center (APCC). The monthly probabilistic hindcasts were temporally downscaled by a stochastic weather generator to daily meteorological inputs for the CERES-Rice crop model. Then, the lowland rainfed rice yields at the field scale were simulated by the process-based crop model with different nitrogen (N) fertilizing strategies, and compared the outcomes against the observed rice yields. The results showed that the simulated rice yields from the 3-month hindcasts with the N-fertilizer application rate of level 1 (60 kg ha-1) were closer to the observed than those from the 5-month hindcasts. About 3.05Mg ha-1 of rice yields were estimated under the 3-month seasonal forecasts in 2020 with the N-fertilizer application rate. We found that additional 40 kg ha-1 of N-fertilizer rate may be required to achieve the national target rice yield (5.0Mg ha-1) under climatic uncertainty conditioned by the seasonal forecasts. We also suggest that the proposed crop modeling approach underpinned by seasonal climate forecasts can be useful to reduce impacts of climate variability on rice yield by providing various agricultural management practices.