Evaluating the impacts of drought on rice productivity over Cambodia in the Lower Mekong Basin

Recurring drought in the Lower Mekong countries has inflicted enormous pressure on the natural ecosystem, rice productivity, and water resources. A regional scale assessment over Cambodia was carried out to examine the linkages between rice productivity and meteorological/hydrologic drought variability from 2000 to 2016. We implemented a comprehensive drought and crop yield information system, the Regional Hydrologic Extremes Assessment System (RHEAS) framework, that couples a hydrologic model with a crop growth model to capture the subtle, intrinsic nature of drought, and assess the impact on inter-seasonal and intra-annual rice yields. Simulations based on RHEAS show good agreement with observations (R-2 similar to 0.65 for soil moisture from the hydrologic model; R-2 similar to 0.84 for crop model). Using a suite of standardized drought indices, the onset and prevalence of dry and wet periods throughout the study period were examined at multiple temporal scales. The temporal variability in drought intensity exhibited higher water stress during the initial months (Mar-May), indicating prevalence of medium to severe dry conditions prior to the planting season. However, the onset of monsoon at the beginning of the growing season (June) resulted in the prevalence of normal to moderate wet conditions. A linear trend analysis for the period 2000-2016 showed a consistent increase (similar to 2900 kg/ha in 2000 to similar to 3550 kg/ha in 2016) in rice yields, although drought-stricken provinces showed lower yields (similar to 1650 kg/ha) throughout the study period. Overall, a continuous increase in annual rice yields irrespective of the stress conditions was noted with no clear pattern linking drought parameters with crop yields on a regional scale. The application of chemical-based fertilizers has steadily increased over the years since 2008 and the consistent increase in observed rice yields correlated with increased fertilizer use (R-2 similar to 0.84). Information from the hydrologic and crop model components within RHEAS enables development of critical regional and local thresholds, reflecting the increasing levels of risk and vulnerability towards drought.