Polarimetric Radar Signatures Associated with Flash Floods in the Phoenix, Arizona Area
In the summer convective seasons of 2012 and 2013, the Phoenix metropolitan area experienced several heavy rain and flash flood events. The Phoenix area is characterized by a high degree of urbanization and a large number of small (usually dry) tributaries known as washes. Flash flood events typically occur when convective rainfall overwhelms the washes and other man-made drainage systems. This process is enhanced by the urbanized nature of the area, as a relatively small amount of precipitation soaks in and a high percentage is converted into runoff. Urban flash flood events are often caused by rainfall that is of short duration, localized areal extent, and high intensity.
The staff at the National Weather Service Forecast Office in Phoenix has several tools available to assist in monitoring and warning for flash floods. These include the Flash Flood Monitoring and Prediction (FFMP) system; a database of flood average return interval data incorporated into FFMP; and an extensive ALERT system maintained by the Flood Control District of Maricopa County (FCDMC). Each of these tools is beneficial; however, the tools have potential shortcomings. FFMP's basin-averaging technique may sometimes underestimate the localized, intense rain events found during the convective season. The FCDMC's network, while beneficial, may not fully document the intensity of a rain event if the maximum rainfall does not occur over one of their gauges.
This presentation will review the flash flood events of 31 July 2012; 23 August 2012; and 21 July 2013. We will investigate the FCDMC ALERT network data, the precipitation accumulation data from the non-polarimetric precipitation processing algorithm, the polarimetric precipitation accumulation data, and the DPR data. In each of these flood events, the DPR showed areas of intense rainfall rates (greater than 100 mm hr-1), persisting over small-scale drainage basins for two to four WSR-88D volume scans. Rainfall areas having these characteristics may produce low impacts over rural areas, but they appear to be sufficient to produce excessive runoff and flooding in the Phoenix area.
Future work will investigate additional flash flood events to see if the DPR parameters are consistent with those seen in the events documented thus far, and will compare flash flood events with other non-flooding rain events to look for statistically significant differences in the DPR characteristics.