A Tale of a Flash Flood: WRF Modeling of Record Rainfall and Flash Flooding in Arizona

Perhaps the most regular and predictable weather pattern in North America is the North American (NA) or Mexican monsoon. Occurring in summer, it delivers about 35% and 45% of Arizona’s and New Mexico’s annual precipitation, respectively, and about 60% of northern Mexico’s. While recent studies have linked strong Arizona monsoons to summer drought in the U.S. Midwest, the sequence of events which produce the NA monsoon remains unclear. An important question for the predictability of the summer Monsoon in Arizona, and its related weather patterns, is: what are the factors determining the favorable circulation and boundary layer configuration for the monsoon onset and widespread precipitation resulting in flash flood events to occur?

In early September 2014, remnants of Hurricane Norbert brought record-setting rainfall that swept across the Southwest U.S. Flash flooding in Phoenix caused major damage to infrastructure, roadways, and many human casualties including two fatalities. As the storm bands cross the core Phoenix metropolitan area, a record 5.51 inches of rain fell over the Chandler area in just under seven hours. This was the highest recorded rainfall amount since 1895 for the Chandler area.

The Phoenix flash flood of September 7-8th, 2014 resulting from the hurricane Norbert is investigated in this Weather Research and Forecasting (WRF) modeling study. Our goal is to simulate the general features of the boundary layer in Arizona prior and during the flash flood events and to study the related hazardous aviation weather patterns. The WRF mesoscale model simulations of the monsoon and flash floods in AZ are analyzed and the output files for the surface winds, u- and v- components of the winds at different pressure levels, circulation and vorticity components, mixing ratio, potential temperature, convective available potential energy (CAPE), convective inhibition (CIN), and precipitation are discussed. The WRF accommodates four-dimensional data assimilation, multi-nested domains and other physical parameterizations, which allow us to simulate the general features of the boundary layer during the flood period. We research what supports the widespread precipitation resulting in flash flood and the favorable changes in the boundary layer and the circulation that lead to it. Factors like monsoonal tropical moisture, sea surface temperatures (SSTs), terrain, soil/vegetation cover, and the parameterizations used during the numerical simulations are examined. . Being able to confidently predict the occurrence of these major flash flood events, especially during the Pacific hurricane season when tropical cyclones can serve as a catalyst, can greatly improve flood forecasts.