Effect of Rain Gauge Proximity on Rainfall Estimation for Problematic Urban Coastal Watersheds in Virginia Beach, Virginia

Abstract

In urban areas, it is important to have spatially and temporally dense rainfall measurements for flood modeling, monitoring, and prediction. The objective of this paper is to quantify the effect of rain gauge proximity on area-averaged rainfall estimation for small (<1 km2) problematic urban watersheds in Virginia Beach, Virginia, United States. To achieve this, rainfall was estimated by including and then excluding nearby rain gauges from a Kriging analysis. This was done for seven focus watersheds draining to flood-prone areas in Virginia Beach for the 20 days with the highest total rainfall depth during the period 07-02-2013 to 01-01-2016. Results show that including local rain gauges in the Kriging analysis resulted in an average percent difference in area-averaged rainfall of close to 60% at a 15-min time step (24.5 mm max) and 20% at a daily time step (55.6 mm max) across the focus watersheds. The results suggest that a rain gauge within 0.5 km of the target watershed would be needed for flash-flood warning applications in Virginia Beach. Radar-derived rainfall estimations were used for comparison and showed significant differences from the rain gauge rainfall estimates greater than 20 mm over an hour [r=0.72, root mean square error (RMSE)=9.72 mm].