6B.1 A Real-Time Hydrometeorology Research Testbed for Heavy Rainfall and Streamflow Prediction

Monday, 28 August 2017: 1:30 PM
Vevey (Swissotel Chicago)
Rita Roberts, NCAR/RAL, Boulder, CO; and J. Wilson, D. Megenhardt, J. Sun, D. Gochis, A. Rafieeinasab, and B. Brown

This paper describes an experimental, hydrometeorology-based, heavy rainfall and streamflow prediction system developed by NCAR’s Short Term Explicit Prediction (STEP) research program and demonstrated for the past three summers along the Colorado Rocky Mountain Front Range. Storms along the Front Range often produce flash floods that can have significant impact on metropolitan regions, mountain towns and streams.

The STEP system, called the Hydromet Heavy Rainfall Prediction System, ingests data in real-time from three operational NEXRAD radars, GOES satellite imagery, radiosondes, surface stations, rain gauges, and streamflow gauges. As a fully-integrated system it includes quantitative precipitation estimation (QPE), forecasting (QPF), and nowcasting (QPN), streamflow prediction and object-based verification. Two versions of QPE are run in the system (NOAA’s MRMS and NCAR’s dual-polarization Hybrid QPE) with rain gauge measurements used to check the quality of the radar estimates. Two state-of-the-art NWP models (WRF 3DVar) are run with and without radar data assimilation (DA) to assess the impact of DA on forecast performance. The models are run with frequent (1 and 3 h) update cycles. Cell-based radar echo extrapolation (TITAN) and a cross-correlation tracker (CTREC) are run to obtain storm characteristics and storm motion vectors. The AutoNowcaster/Trident heuristic nowcast systems ingest both versions of QPE and produces storm initiation, growth, decay, and 10 min – 1 h heavy rainfall nowcasts. High-resolution 4-D winds and buoyancy analyses are produced from the VDRAS system. Streamflow prediction on a spatially-continuous 100 m resolution grid is provided by the WRF-Hydro coupled atmosphere and hydrology model. Fractional Skill Scores, Model Evaluation Tools (MET), and the Method for Object-Based Diagnostic Evaluation (MODE) tool are used in verification of model performance.

Each year, upgrades have been made to the system components to increase the accuracy of the overall system. Results show there is still much room for improvement in providing location-specific and time-specific prediction of heavy rainfall, flash floods and peak streamflow and in providing useful products for the end-user. Improvements incorporated into the system for the upcoming 2017 demonstration will be discussed during the conference and examples of the real-time performance of the system for selected Colorado heavy rainfall (>1 inch per hour) events will be presented.

- Indicates paper has been withdrawn from meeting
- Indicates an Award Winner