SR/HSD 97-3

8-15-97

Technical Attachment

A History of QPF in Southern Region

Prepared by:

Hydrologic Services Division

NWS Southern Region Headquarters

Fort Worth, TX

Other than a mention in somewhat qualitative terms as part of the agricultural forecast package, Quantitative Precipitation Forecasts (QPF) have not been routinely issued by Southern Region (or other offices) until recent years. In the late 1970s, spurred by increased emphasis on flash floods and land falling tropical systems, a handful of Southern Region forecasters became involved in forecasting QPF. This initial emphasis was mainly on forecasting heavy rain. Members of this "cadre" included Alan Johnson, Gary Grice, JimmyDon Ward, Edward Mortimer, and Jim Bellville, among others. The group began publishing papers, Technical Memoranda, and Southern Region technical attachments on the subject which led to an increased interest in QPF throughout the Southern Region. Among these was the 1978 NOAA Technical Memorandum titled, "A Flash Flood Aid-The Limited Area QPF," written by Jim Belville, Alan Johnson, and Jim Ward.

By the early 1980s, annual Southern Region QPF workshops were being held. Southern Region Technical Memoranda were published in 1985 and 1986 containing papers presented at the workshops. With increased national emphasis on heavy rain forecasting, these early regional workshops became NWS National Heavy Precipitation Workshops which are now held every other year. Most recently the 5th NWS National Heavy Precipitation Workshop was held on September 9-13, 1996, in State College, Pennsylvania.

Meanwhile, a few Southern Region offices experimented with preparing (and even issuing) QPFs. The Lubbock WSFO began issuing maximum point QPFs for the International Boundary and Water Commission (IB&WC) during the middle 1970s. In the middle 1980s the IB&WC requested a new approach be used. This involved the WSFO forecasting QPF for areas in the Rio Grande Basin having a 20 percent chance of receiving an inch or more of rain in 24 hours. The NWSFO Lubbock may hold the distinction of being the first Southern Region office to routinely issue QPF.

After the disastrous Pearl River Flood of 1979, the WSFO in Jackson, Mississippi, began using QPF to assist in forecasting river floods. The process involved the WSFO calling the Lower Mississippi River Forecast Center (LMRFC) with a forecast or "most-likely" QPF and a worst-case-scenario QPF. The LMRFC then ran their models with these values and either telephoned or transmitted the results over AFOS. This as-needed QPF method continued for about 15 years. However, in October of 1994, WSFO Jackson began using WinQPF to This forecast quantitative precipitation.

During the middle 1990s, the Tulsa Risk Reduction Exercise was taking place as part of the NWS modernization. This effort involved the Arkansas-Red Basin River Forecast Center (ABRFC) in Tulsa, Oklahoma, and the Southern Region offices they serviced. Instead of using the WinQPF software (developed by OHRFC Sr. HAS Mark Fenbers), ABRFC used a UNIX-based program run on a LINUX computer. The Risk Reduction Exercise provided valuable lessons on how to implement QPF in the field forecast environment. At the completion of the project, however, a choice had to be made about which QPF software application would be used. With the nationally supported WinQPF clearly becoming the standard, Southern Region chose it.

In time, it was also decided that the implementation date for region wide QPF issuance would be July 1, 1996. Before this implementation date, two test markets were established at NWSFOs Lubbock (Texas) and Jackson (Mississippi). These "experienced" test markets served to troubleshoot the operational (WinQPF) and verification (Noreen Schwein, CR HSD) software packages. As the Southern Region QPF implementation date approached, HSD planned to help the WFOs enter this new era with as much information, training, and education as possible. In April and again in July of 1996, Southern Region HSD distributed a bibliography and several technical papers concerning QPF, heavy rain forecasting, heavy rain climatologies, decision tree applications, and so forth. These materials were designed to help WFO training officers better acquaint their staffs with, among other things, the most difficult concept involved in QPF; basin averaged precipitation.

Additionally, four QPF workshops were organized by Southern Region HSD and members of the four collocated WFOs/RFCs during the Summer of 1996. These workshops, held at each of the four collocated RFC/NWSFOs, were attended by forecast staff from every WFO in Southern Region and forecasters from Central Region offices (in the ABRFC domain). The workshops were designed to give field forecasters training and education concerning QPF and its implementation across Southern Region. The training included time spent at the RFCs to see how the HAS forecasters mosaicked WFO-produced QPF, and how RFC hydrologists used QPF to generate river model forecasts.

Presently, all Southern Region WFOs (except San Juan) produce QPF at least twice a day with issuances at 12Z and 00Z. These regular issuances include four 6-hour forecasts comprising a 24-hour time frame. Updates and inconsistencies between adjacent offices are coordinated with the help of HAS forecasters at the appropriate Southern Region RFCs. Updates to QPF are expected when valid WFO-produced QPFs become unrepresentative. The San Juan WFO does not yet produce a QPF because the WinQPF software grid does not accommodate the longitude and latitude of Puerto Rico. This problem is being resolved.

Several Southern Region RFCs have begun utilizing the Internet to help WFO forecasters. ABRFC and WGRFC presently post their mosaicked QPF results on their homepages. They each include sections explaining the concept of Mean Area Precipitation (MAP). ABRFC is posting flash flood risk graphics on its homepage. These are derived from WFO-produced QPF combined with flash flood guidance values.

In the year since the Southern Region QPF implementation began, significant strides have been made. While the problems of over forecasting small rain events and under forecasting extremely heavy rain events still exist, a study by staff members of the ABRFC (NOAA Technical Memorandum NWS SR-187) has shown a 20 percent increase in overall river forecast accuracy with the use of QPF compared to not using QPF for the same forecast run. ABRFC is also routinely issuing RVFs (river forecast products) both with and without QPF; the "without QPF" forecast is for information only. This practice has been deemed an unqualified success by the offices they support.

With the successful implementation of QPF throughout the Southern Region, continued improvement in river forecast and QPF accuracy becomes the goal. This will be attained through the experience of field forecasters with daily issuances of QPF, the production of precipitation climatologies for each HSA, and the continued hydrologic model enhancements made by RFC personnel.