Modeling cascade for flood forecasting

Floods are among the most devastating natural hazards in the world, affecting more people and causing more property damage than any other natural phenomena. Most flood events are driven by rapid development of runoff caused by intense precipitation or/and snowmelt. Therefore, quantitative estimation of precipitation and snowmelt is a crucial point in flood forecasting. To assess these phenomena hydrologists traditionally rely on conventional in-situ observations that were later expanded by meteorological radars and satellite-based precipitation estimates. The use of meteorological models was very limited due to their coarse resolution that allowed it to be used only for macroscale hydrological studies. However, with the progress in regional weather modeling and rapid increase of computation power it is become possible to utilize such models as additional source of precipitation.
To predict flooding parameters such as rivers stage/discharge and extents of flooded areas we use cascade of simulation models: regional numerical weather prediction (NWP) model, hydrological model and hydraulic model.
To obtain quantitative estimates of precipitation and other meteorological forcing in the Space Research Institute WRF (Weather Research&Forecasting) regional NWP model is used. WRF model is a joint development of a number of USA agencies and universities (http://wrf-model.org). This model was configured and adapted to the territory of Ukraine to run with spatial resolution of 10 km. Currently we routinely produce 72-hours weather forecasts every 6 hours. To drive regional model the additional weather forecasts from global NWP model are used. These data are required to specify external meteorological forcing as boundary conditions for regional weather model. Currently we use forecast frames produced by GFS (Global Forecast System) model operated by NCEP.
10 km ground spatial resolution is generally enough to support regional applications while it is clearly insufficient to represent small-scale features like precipitation or temperature distributions in mountainous of the Carpathian region. Ground resolution of about 1 km is required to cope with such life-threaten events such as flash floods. Regional models such as WRF are designed to operate at such resolutions, however direct increase of spatial resolution results in enormous increase of required computation performance. To solve this antinome problem nesting approach is used. Within this method a set of nested model domains is used, each subsequent domain “zooms in” parent domain thus providing better resolution.
Two approaches are used to simulate hydrological processes in the Carpathians watersheds: more traditional one using dedicated hydrological model and the second using coupled modeling. As dedicated hydrological model we use robust industry proven HSPF (Hydrological Simulation Program – Fortran, http://water.usgs.gov/software/hspf.html). HSPF is lumped model, i.e. within this type of models the whole watershed is divided into a number of sub-catchments that are considered hydrologically homogeneous. To run HSPF model several meteorological forcing parameters are required (at least precipitation and estimation of potential evapotranspiration). These data are obtained from high-resolution weather forecasts. Another approach we utilize is a coupled modeling. Within this approach we use Noah Land Surface Model (LSM) that is fully coupled with WRF model. Noah LSM is a distributed physically-based model that works with the same resolution as WRF model does.
Both hydrological model (HSPF and Noah) are able to compute quick response runoff to precipitation events. To route computed runoff and derive river stages and extent of flooded areas appropriate hydraulics models are involved.
This work is supported by ESA CAT-1 project “Wide Area Grid Testbed for Flood Monitoring using Spaceborne SAR and Optical Data” (No. 4181); joint project of INTAS, the Centre National d’Etudes Spatiales (CNES) and the National Space Agency of Ukraine (NSAU), “Data Fusion Grid Infrastructure” (Ref. Nr 06-1000024-9154); joint project of the Science & Technology Center in Ukraine (STCU) and the National Academy of Sciences of Ukraine (NASU), “Grid Technologies for Multi-Source Data Integration” (No. 4928).