Close Window

Pennsylvania State Univ.: "An early alert flash flood warning system

Final Report


The last six months of this project have focused on completing a comprehensive data base (using MS Access) of flooding throughout the MARFC domain. There has also been the completion of the project into shear climatology using the NCEP Reanalysis Fields and Storm Data from the eastern half of the nation.
The algorithm which ranks the risk of flood threats based on the most closely related pattern recognition variables is being tested in real-time. Several times a day, the risk fields based on the 4 separate runs of the ETA are e-mailed to the principal investigators.


University related:

Nearly a dozen undergraduate and graduate students completed the joint QPF course taught in the Fall 2002. They have gained significant experience in examining flood cases and ranking them by Maddox types. The students finished their semester project which centered on the development of a comprehensive relational database of flooding and severe weather in the Middle Atlantic domain.

NWS related:

Work related to this project resulted in the development of about 10 instructional modules related to forecasting rainfall in general and the QPF problem specifically. Most of these modules are available on the web at:

In addition to the training module, the NWS in conjunction with PSU have developed a 30-year shear climatology. This climatology supplements the existing climatologies relating to height, u-wind, v-wind, and precipitable water developed for past studies. There appears to be some relationship between shear and heavy rainfall. This will be a line of further investigation.

The shear climatology also is related to severe weather and we are examining the use of the shear climatology in its relationship to large-scale severe weather events.


University Partner:

There has been significant spillover into other courses. Several students involved in an intensive writing class on mesoscale processes have chosen flooding events as their case studies and are applying principles and techniques learned in the Flash Flood course offered in the fall. Also, a course on mesoscale prediction has several students enrolled who also took the fall flood class. They have selected flooding issues for their semester projects.

NWS Partner:

The NWS has benefited in a variety of ways. First, the on-going course is attended by several NWS employees, offering them unique training opportunities. Second, the course has had guest speakers, such has Wes Junker of NCEP. Wes' lectures were well attended by both NWS-CTP and MARFC personnel. The course provided many NWS employees a chance to meet and talk with Wes. Additionally, this course gave several CTP staff members experience as co-instructors for a high level Penn State Meteorology course.

Other benefits include a huge work force to help categorize and classify heavy rainfall and severe weather events over our forecast area. With the class now building relational databases from these data, potential spin-off studies will be many. Fortunately, many of these studies are directly related to forecast problems of concern to our office.

One member of the NWS attended the course for credit. The General Forecaster is the newest addition to the CTP staff. This course provided him with a valuable learning experience, not only for forecasting heavy rain events in Central Pennsylvania, but he also participated in the categorizing and classifying severe weather events for our area. This research opportunity provided a wealth of severe weather information that takes most forecasters years to experience. This forecaster now posses a wealth of climatological information, about severe weather, that occurs in Central Pennsylvania. Additionally, this experience has prompted the forecaster to become more involved with StormData and now is part of the StormData group for CTP.


A presentation was given at the Northeast Regional Operational Workshop in Albany, NY on November 5-6, 2002. Also, a formal presentation was given about the project to the Hydrology conference at the AMS annual meeting in Long Beach on February 11, 2003. Our student intern, Justin Brolley, attended the conference and made the presentation.


The pre-print from the AMS Hydrology conference contains the project summary.


There have been no significant problems encountered concerning this project. We intend to further test the algorithm and try a different presentation of the risk assessment before going operational later this spring.