Record Low North American Monsoon Rainfall in 2020 Reignites Drought over the American Southwest

In 2020, drought deepened over the American Southwest against the backdrop of two decades of drought damages that exceed $130 billion and caused alarm about potential water delivery shortages in the Colorado River basin. The proximate causes for these droughts include low precipitation and increased evaporative demand in concert with warming temperatures. While there is strong evidence for anthropogenic forcing of warming temperatures, recent work has also pointed to a potential human effect on precipitation. Record low precipitation during the 2020 monsoon season over Arizona, New Mexico, Colorado, and Utah (Four Corners States) re-established drought. Whereas record low precipitation in June-September 2020 over the Four Corner States capped off a three-year stretch of below average rainfall, no significant trend since 1895 is found. Further, no statistically significant change in the frequency of low precipitation is noted. Given the brevity of observations, we use multiple models and large ensembles, controlled in various ways for historical climate drivers, to test the effect of climate change on low precipitation occurrences. The absence of an observed trend is not sufficient evidence against an effect of human-caused drying, given the possibility that internal variability can mask a climate change drying. Most model experiments indicate record low June-September 2020 precipitation in the Four Corners States was made more likely due to climate change, though our confidence in this result is tempered because the models do not perfectly reproduce precipitation statistics in the region and such a change has not been observed. Four of the five models indicate that low decile and percentile June-September precipitation is 1.5-2.5 and 2.5-5.5 times more likely, respectively, due to climate change. The model results are consistent across three widely used experiment types historical simulations using coupled and atmospheric models, and event-attribution simulations which together provide a more robust test of anthropogenic effects than observations alone. Use of these large ensemble experiments allowed evaluations of extreme event probabilities to be directly calculated, which is a strength of the study, even though the models are not perfect representations of the earth system.