Short-Term Explicit Prediction (STEP)
Short-term forecasting of High-impact weather
Short-Term Explicit Prediction (STEP)
Short-term forecasting of High-impact weather
While our capability in prediction large-scale flow structure for the medium range has experienced marked advances in the last few decades, the short-term forecast (less than one day) of high impact weather at the county/city scale lags seriously behind. The Short-Term Explicit Prediction (STEP) program was established in 2005 to tackle the challenging problem of accurate (location and timing specific) prediction of hazardous weather via a collaborative effort across several NCAR labs. In the recent few years, the main focus of STEP is to improve heavy precipitation and flash flood prediction by developing an integrated hydro-meteorological system that is able to produce quantitative streamflow forecast with improved rainfall estimate and nowcast/forecasting. That objective is being supported by four research topics: optimal design and utilization of observations, multi-scale data assimilation, utilizing high-resolution observations, convection-permitting modeling and ensemble prediction, and atmosphere-land coupling to improve rainfall and flood prediction. The main outcome of these research topics will be the development of advanced short-term local-scale modeling capabilities, which will further support technology transfer and real-time demonstration, as well as education and outreach.
High-resolution, short-term forecasts of high-impact weather provide critical information for a wide range of users, including the aviation community, ground transportation, urban emergency and water resources management groups, recreation facilities, construction industries, and the military, that assists them to safely and efficiently deploy resources. However, achieving reliable and accurate convective weather forecasts remains a scientific challenge due to uncertainty in grasping initial conditions, shortcomings in model physics and computational capabilities, and limitations of our understanding of convective weather. The forecast skill of observation-driven expert systems decreases rapidly with increasing lead-time, while numerical weather prediction models exhibit a limited forecast ability within the first few hours after initialization primarily due to spin-up problems. For explicit flash flood forecast, the coupled hydro-meteorological model requires accurate precipitation forecast that is beyond the capability of the current operational models can provide. STEP’s mission is to address the challenge of short-term high impact weather prediction through a broad end-to-end approach and a broad across-NCAR collaborative effort.
STEP currently funds seven projects with participants from three research Labs: RAL, MMM, and EOL. These projects, listed below with PI and lead lab, are being conducted to tackle various research problems in the four STEP research areas:
- Short Term Explicit Prediction: Increasing the value of QPN/QPF for flood prediction through optimal use of observations and model coupling (Ethan Gutmann, RAL)
- Evaluation and improvement of model microphysics parameterizations (Sarah Tessendorf, RAL)
- Advancing Real-time Hydrologic Predictions as Part of the Short-Term Explicit Prediction Experiment (Dave Gochis, RAL)
- Improving short-term local-scale high-impact weather prediction by assimilating unconventional observations (Jenny Sun, RAL & MMM)
- Convective-scale ensemble analysis and high-impact weather prediction (Glen Romine, MMM)
- Physical Processes Influencing Convection Initiation and Mesoscale Prediction in the WRF Modeling System (Stan Trier, MMM)
- Use of micropulse DIAL (Differential Absorption Lidar (MPD) toward advancing basic knowledge and understanding of short-term, local-scale, high-impact weather prediction (Tammy Weckwerth, EOL)
As a cross-lab science program, the key to measure STEP’s success is collaboration between NCAR labs and scientists. Quarterly meetings are conducted to update progress, identify issues, foster collaborations. Annual workshops are conducted to review the projects and discuss future directions.
Representative Projects
Key Projects
- Improving rainfall forecasting and nowcasting through advanced data assimilation and ensemble prediction Developed rapid cycle radar data assimilation techniques to improve rainfall prediction.
- Improving WRF Physics: Improved Thompson microphysics by developing multi-moment ice physics and evaluated it in multiple case studies.
- Quantitative Precipitation Nowcasting (QPN): Investigating the impact of assimilating storm updraft, extracted from radar reflectivity using a machine learning technique, on precipitation nowcasting.
- Streamflow Prediction with WRF-Hydro: Real-time, high-resolution hydrologic streamflow predictions were produced using the community WRF-Hydro® modeling system.
- Unifying global and regional model systems: Evaluating the capability of regional MPAS against WRF regional model toward the goal to transition from WRF to the global variable-resolution MPAS
Contact
Jenny Sun
Senior Scientist