A hydrometeorological investigation of two flash flood events at Caprock Canyons State Park in northwest Texas

This study will examine the causes of two flash flood events that impacted Caprock Canyons State Park (CCSP). The first flood occurred on 26 September 2012. In this event, a supercell thunderstorm on the northern flank of a quasi-linear convective system underwent a “collapse phase” as it moved across the park. Torrential rainfall from this storm peaked at a rate of nearly 15 cm per hour. Approximately 4 cm of rain fell in a 30 minute period, which caused significant flash flooding, including the buckling of an asphalt park road. The second flood occurred on the morning of 19 June 2013, as a result of excessive rainfall produced by a mesoscale convective complex moving across the park. Ground-based reports of rainfall rates and amounts in this case were even more intense than the first, with a station of the West Texas Mesonet (WTM) at CCSP recording approximately 10 cm of rainfall, and a rainfall rate of nearly 30 cm per hour for a brief period. This flood also produced more extensive damage from water and debris flow. The small stream basins within CCSP are susceptible to excessive runoff and damaging flash floods due to the steep and constrictive topography along this portion of the Caprock Escarpment. Additionally, the orientation of the canyons along the escarpment affects the mesoscale wind fields. Under certain flow regimes this can increase boundary layer mass and moisture convergence, and thus enhance convective development and precipitation efficiency as moist air is forced upwards by the terrain.

In this study we examine the radar structure of the two storm systems primarily utilizing the KAMA (Amarillo) and KLBB (Lubbock) WSR-88D radars. We compare the reflectivity and radar precipitation estimates to in situ rain-gauges of CCSP and the WTM. In addition, a detailed surface mesoanalysis is performed for these cases using five-minute data collected from the WTM. Preliminary results indicate that mass and moisture convergence are enhanced at times in vicinity of CCSP. Finally, we evaluate the ability of a convection-allowing mesoscale model to predict both the convective activity and the topographically enhanced near-surface mass and moisture fields that accompanied these two events. The Weather Research and Forecasting Model (WRF) is employed, utilizing both a discrete Advanced Research WRF (ARW) core mode, and a 50-member Ensemble Kalman Filter (EnKF) mode.

The results of this study suggest that forecaster's prior knowledge and consideration of small basins susceptible to flooding can improve the accuracy and response time of flash flood related products and information. Additionally, the high-temporal resolution of the WTM rain-gauge data may alert the forecaster of rainfall rates that are sufficient to produce flash flooding.