Potential Predictability of Two-Year Droughts in the Missouri River Basin

Abstract Potential predictability of two-year droughts indicated by low runoff in consecutive April-September seasons in the Upper (UMRB) and Lower (LMRB) Missouri River Basin are examined with observed estimates and climate models. The majority of annual runoff is generated in April-September, which is also the main precipitation and evapotranspiration season. Physical features related to low April-September runoff in both UMRB and LMRB include a dry land surface state indicated by low soil moisture, low snowpack indicated by low snow water equivalent, and a wave train across the Pacific-North American region that can be generated internally by the atmosphere or forced by the La Niña phase of the El Niño-Southern Oscillation. When present in March, these features increase the risk of low runoff in the following April-September warm seasons. Antecedent low soil moisture significantly increases low runoff risks in each of the following two April-September, as the dry land surfaces decrease runoff efficiency. Initial low snow water equivalent, especially in the Missouri River headwaters of Montana, generates less runoff in the subsequent warm season. La Niña increases the risk of low runoff during the warm seasons by suppressing precipitation via dynamical-induced atmospheric circulation anomalies. Model simulations that differ in their radiative forcing suggest that climate change increases the predictability of two-year droughts in the Missouri River Basin related to La Niña. The relative risk of low runoff in the second April-September following a La Niña event in March is greater in the presence of stronger radiative forcing.