Detecting flash floods in small urban watersheds.

Intense rainfall rates occurring in highly urbanized watersheds can produce deadly flash flooding. Some recent examples of urban flash flooding include Dallas, TX (1995), Fort Collins, CO (1997) and Kansas City, MO (1998). Detecting the heavy rainfall that produces these flash floods is one of the most difficult problems facing the National Weather Service (NWS) forecaster.

The Areal Mean Basin Estimated Rainfall (AMBER) flash flood program can effectively alert the flash flood forecaster to the potential for urban flooding, before flooding begins. Flash flooding in the eastern suburbs of Pittsburgh, PA on May 18, 1999 was successfully detected using AMBER. A flash flood warning was issued almost 2 hours before 10 homes were evacuated in Forest Hills at the peak of the flooding.

This case study will demonstrate how the Average Basin Rainfall (ABR) and ABR rate computed by AMBER can be used to determine the risk and severity of flash flooding. AMBER computes ABR every 5-6 minutes for every defined watershed and urban area. The hourly ABR rate, determined from 5-6 minute ABR, alerts the forecaster to flash flood potential well before flooding occurs.

AMBER has been used effectively at the Pittsburgh NWS office since the spring of 1996. This useful flash flood tool is moving toward national implementation in the NWS by the 2001 flash flood season.