Session 14B.1 Developing Hurricane Related Inland Flooding Forecasts at EMC

Thursday, 4 June 2009: 10:30 AM
Grand Ballroom West (DoubleTree Hotel & EMC - Downtown, Omaha)
Yihua Wu, NOAA/NWS/NCEP and I.M. Systems Group, Camp Springs, MD; and M. Ek, K. Mitchell, V. Tallapragada, R. Tuleya, and Y. Xia

Presentation PDF (723.0 kB)

In the 1970s, '80s, and '90s, inland flooding was responsible for more than half of the deaths associated with tropical cyclones in the United States. A hurricane-related inland flooding forecast technique is under development at Environmental Modeling Center (EMC) by enhancing the HWRF (Weather Research and Forecast system for hurricane prediction) forecast capability with a comprehensive and advanced land surface model (LSM), and linking the HWRF model with the EMC streamflow routing scheme. The HWRF is a new computer model for hurricane forecasting developed at EMC. The model is a nested grid system with an outermost domain and a nested grid with resolutions of 27 and 9 km respectively and 42 vertical levels. The HWRF uses a modified 6 hour forecast as the first guess and regional GSI 3DAR data assimilation for the hurricane vortex initialization. The HWRF is coupled to a high-resolution version of the Princeton Ocean Model for the Atlantic Basin. The ocean initialization system uses observed altimeter observations to provide a more realistic Loop Current and Gulf Stream conditions. The HWRF is running operationally at EMC to produce hurricane forecasts every six hours for up to four tropical storms at a time. One of the goals of EMC is to use HWRF model output as input to hydrology and inundation models to forecast hurricane related inland flooding through its land surface component. However, the operational version of HWRF uses the GFDL (Geophysical Fluid Dynamics Laboratory) Slab LSM to model land-atmosphere interactions. In the Slab LSM, only one layer soil temperature is predicted while the initial soil moisture remains fixed in time during the HWRF forecast. Additionally, the Slab LSM does not predict the runoff response to HWRF precipitation forecasts. Hence the Slab LSM is obviously unable to serve the hydrology goals cited above. Therefore, the Noah LSM, a more comprehensive and advanced LSM, was added to HWRF while the Slab LSM was kept as an option. The Noah LSM uses: 1) multiple soil layers with a one-layer vegetation canopy, 2) spatially varying root depth and seasonal cycle of vegetation cover, 3) frozen soil physics for cold regions, and 4) improved soil and snowpack thermal conductivity. The Noah LSM predicts soil moisture, soil temperature, latent heat and sensible heat flux, and total runoff which accounts for sub-grid variability in precipitation and soil moisture. The runoff prediction can then be used as forcing input to EMC's Streamflow Routing Scheme (Lohmann et al., 2004). Additionally, the HWRF-Noah forecasts of soil moisture and runoff are good spatial indicators of soil moisture saturation (water logging) and flooding. In the streamflow routing scheme, the concentration time for runoff reaching the outlet of a grid box and the transport of water in the channel system is computed, water can leave the grid cell through (at least) one of the eight directions, the runoff transport process is linear and time invariant, the causality and the impulse response functions are nonnegative. Several flooding cases caused by hurricane landfall have been investigated. Preliminary evaluation indicates that rainfall appears more realistic using the Noah LSM and GFDL/GFS surface physics. While predicted streamflow is comparable to the observational measurements, the stream flow prediction would be more realistic using more accurate initial soil moisture from the regional mesoscale model (NAM) and/or offline-NLDA.
- Indicates paper has been withdrawn from meeting
- Indicates an Award Winner