FSL's Mesoscale Model Ensemble Supports the 2005 MDSS Demonstration

NOAA initiated the Surface Weather Program in 2004 to develop and implement new weather products and services to support the nation's surface transportation system. This program follows a five-year project, the Maintenance Decision Support System (MDSS), conducted by the Federal Highway Administration, to which the Forecast Systems Laboratory was a key contributor. Under the leadership of Paul Schultz, FSL is collaborating with the Department of Transportation and other research laboratories to develop a prototype of the MDSS. FSL runs an ensemble of mesoscale models which is used to generate point forecasts along targeted roadways. These forecasts are then used to inform pavement condition modules to predict pavement temperature and chemical concentrations; in turn, these predictions are used to apply codified rules of treatment, which recommend plowing and chemical application frequency to snowplow garage supervisors.

During demonstrations in Iowa over the winters that ended in 2003 and 2004, multimillion-dollar savings were realized as a result of efficient scheduling of snowplow drivers, reduced overapplication of salt and other deicing chemicals, and less chemical damage and sand clogging in roadside drainages. The 2005 MDSS demonstration taking place in Colorado will continue through May, focusing specifically on the E-470 roadway in the Denver vicinity. The MDSS Web page provides links to program information as well as graphics from the real-time modeling services used in MDSS.

The ensembling system has several unique features. Two different models, MM5 and WRF, are kicked off every hour using the Local Analysis and Prediction System's "hot start" method of diabatic initialization, which facilitates very-short-range precipitation forecasts of unprecedented accuracy. There is no "spin-up" time required, which in most modeling systems causes underforecasting of precipitation amounts in the first few forecast hours. Another key feature is much-reduced latency in the delivery of forecast services: the 1-hour forecasts are delivered about 1 hour after the valid time of the radar and satellite data that were used to initialize the models. This traces back to the efficiency of the hot-start initialization method, which opens the door to a broad range of tactical (very short range) applications of numerical weather prediction.

Safer roadways around the country can be expected when the technology supporting MDSS is made available to private sector weather service providers who support state departments of transportation.

Contact information
Name: Paul J. Schultz
Tel: 303 497 6997