Dynamical Flash Flood Risk Forecast

Abstract

Flash floods represent very dynamical natural phenomenon. Mostly, they are the result of torrential rains which can rise suddenly in any part of a country and are tough to predict. Of course, there are many weather forecasting systems, but their spatial and temporal resolution is usually insufficient for these purposes. There are also monitoring systems which can either register precipitation over the ground (a network of rain gauge stations) or runoff in riverbeds (a network of hydrometric stations). Again, spatial (and possibly temporal) resolution is not sufficient, and in the case of runoff monitoring, there is a substantial delay between actual rainfall and registration of runoff in riverbeds. And, of course, when the hydrometric station registers higher runoff than the flash floods is running or even over. From the point of early warning, all these systems reveal disadvantages. Aside from these systems, there is one which provides us with timely information about the spatial and temporal distribution of precipitation intensity over the ground. That is weather radar. We will demonstrate possible usage of these data for dynamic prediction of flash flood risk distribution in space and time over the monitored area. Proper processing of these data in combination with soil saturation indicator established using Flash flood guidance methodology developed by the US Hydrologic Research Center can generate timely information usable for early warning with a substantially reduced level of false warnings.