
|
In this Issue:
|
From the Hydrologist In Charge...
By Billy Olsen
Return to Menu
|
W e received several positive comments about our inaugural edition of The Gage in June. Thank you for such great encouragement. It looks like we will possibly have the upcoming Winter edition of The Gage available online. There will be more details on this next time.With exceptions for the extreme eastern and extreme western portions of the ABRFC area of responsibility, we have had very dry conditions prevail during the past two to three months. Based on the Palmer index (August 4), all portions of the ABRFC area are classified in the mild to moderate drought categories. However, the fall season is normally fairly active and we are certainly hoping that will be the case this year. Speaking of dry weather reminds me that the ABRFC features a drought information section on our web pages. Visit us at http://www.srh.noaa.gov/abrfc. We provide gridded precipitation estimates, normal precipitation values (PRISM) and percent of normal precipitation for the past 3, 7, 10, 30, 60, 90 and 180 days. The information is also provided for the current month, current year and current water year. In addition, we have archived information back several years. These data have proven useful in providing general precipitation information, such as that needed by the WFO in the preparation of the E-5 reports. It has also provided customers with detailed precipitation estimates for specific locations.I would also like to highlight one other feature of our web pages. This is what we call the Raw River Forecast Model Output. This section provides both a general overview of forecast river conditions as well as detailed forecast hydrographs for every ABRFC forecast location. This output is exactly what the river forecaster uses as guidance in preparing our river forecasts. |
Therefore, our customers can
get a quick "heads-up" on expected river conditions at every
forecast point very soon after each model run (routinely every 6 hours
or more frequently). These raw model pages are password-protected, so
give us a call or email if you require access.
We will be introducing two new hydrologic products in the next few months. On September 5th, the ABRFC and the other three NWS Southern Region (SR) RFCs will begin preparing River Flood Watch products. The SR WFOs and the RFCs will jointly issue these products. These products will be in both text and graphic image format. They will highlight the areas and the magnitude of flooding that is possible in the next 48 hours based on forecast rainfall. The other new product to be implemented later in the year will be the Flood Potential Outlook. This will be a national graphic product that will highlight ongoing flooding, as well as possible flooding which may occur during the next five days. Based on customer feedback, these new products are expected to meet the demand for critical "heads-up" type information concerning the potential for flooding.
Thank you,
Billy Olsen Hydrologist In Charge-ABRFC
|
Know Your WFOs
Norman, Oklahoma WFO
By Janet McCormick
Return to Menu| The
Norman Weather Forecast Office (WFO) will be featured in this edition of
The Gage’s continuing series, "Know Our WFO’s". This
series will describe the hydrology of each of the 13 WFOs within ABRFC’s
geographic area of responsibility.
Figure 1. Norman WFO Area of Responsibility. (Source: Norman, Oklahoma National Weather Service Forecast Office) Figure 1 shows the Norman WFO’s area of river forecasting responsibility. This large area includes more than 50 river forecast locations and consists of the western two-thirds of Oklahoma and a portion of northcentral Texas. The Norman WFO is responsible for four major river systems from where they enter western Oklahoma and end in eastern Oklahoma in flood storage pools of three large reservoirs. Starting in the northwest corner, these rivers are the Cimarron and Salt Fork of the Arkansas Rivers, which drain into Keystone Lake. Next is the Canadian River that feeds Eufaula Lake. Finally, the Red River runs along the WFO’s southern border and enters Lake Texoma. Together these rivers provide drainage for more than 135,000 square miles of Oklahoma, northeast New Mexico, southeast Colorado, southwest Kansas, and the Texas Panhandle. Major rivers in the Norman WFO area begin in the high plains or mountainous regions of northeast New Mexico and generally flow eastward across the area.Upper reaches of the drainage basins in New Mexico and Panhandles of Texas and Oklahoma have soils with high infiltration rates, and therefore are non-contributing to the floods in central Oklahoma. The fall of the rivers gradually decreases from about 6 feet per mile in western Oklahoma to less than 2 feet per mile as they enter eastern Oklahoma. Channels are broad, meandering and sand-filled with the active water courses only occupying a small portion of the flood plain. Low water channels are continually shifting with variations in stream flow. Compared to eastern Oklahoma, the flood control capacity of central Oklahoma is limited. Other than a few large dams in far western Oklahoma, most of the flood retaining structures in the Norman WFO area are small dams with drainage areas of one to ten square miles. More than two thousand of these |
dams have been built in Oklahoma, with about eleven hundred in the Washita River tributary of the Red River alone. Unfortunately, many of the dams are approaching the end of their 50-year life span and have limited capacity due to filling with sediment.
Figure 2. Normal Annual Rainfall for ABRFC. (Source: PRISM) As shown in Figure 2, average annual rainfall amounts across the Norman WFO increase from about 20 inches in the west to more than 35 inches in the east. In late spring or early fall, storms can often produce locally heavy rains which create a high potential for uncontrolled flooding in most of central and western Oklahoma. The Enid Storm of 1973 produced 15-20 inches of rainfall over a 100-square-mile area, with 12 inches falling in three hours. Flooding may become widespread when storms repeatedly develop over the same terrain or move slowly eastward along the drainage courses of the major rivers. This type event occurred in October 1986 when the remnants of Hurricane Paine dropped from 10 inches of rain in southwestern Oklahoma to almost 20 inches in north-central Oklahoma to southern Kansas with nearly a 10-inch basin average for the area above Keystone Lake. Norman WFO Service Hydrologist, Steven Kruckenberg, identified non-supported river gage stations as the major forecasting challenge for his area. These are gages that may only have a wire-weight and require an observer to make a reading or an automated gage where flows are not measured. Without measurements, stage-discharge ratingsare difficult to determine accurately. At wire weight stations, the data is usually only collected during a significant flood event and even then may be sporadic. Additionally, some of the stations are located in regions where flood events are infrequent, making the limited data even more important. Models used at the ABRFC for river forecasting are impossible to calibrate accurately without this important data. Fortunately, the USGS has developed a plan to provide reliable stream flow data for all river forecast locations, but so far the funding only allows for a few stations to be added to their system every year. Two stations are to be upgraded by this plan in FY 2002. They are Cimarron River near Buffalo, which is a forecast point, and Skeleton Creek near Lovell which will aid forecasting the Cimarron River at Perkins.
|
River Forecast Verification
By Bill Lawrence
Return to Menu| The
goal of hydrologists working at the ABRFC is precise and timely
forecasts, especially in times of flooding. However, until recently,
there was no method available for our hydrologists to quantify exactly
how well their forecasts were verifying. In the first issuance of The
Gage, a previous river forecast verification program used at the
ABRFC was described. This system proved its usefulness by showing that
overall, the river stage forecasts with QPF had lower root-mean square
error (RMSE) values than forecasts with no QPF, especially during the
cool season. Unfortunately, there was also a direct correlation of RMSE
values with the amount of rain that fell. Rainy months yielded higher
RMSE values, while dry months resulted in lower RMSE values. We wanted
to devise a system that would be less reliant on the amount of rainfall,
and would show trends in forecast accuracy.
In early 2000, hydrologists from both the ABRFC and West Gulf RFC (WGRFC) in Fort Worth, Texas joined forces and created a categorical river forecast verification system. This new system, created after countless hours of development and testing, promises to provide useful statistics to RFCs about how well they are forecasting rivers when forecasts are needed the most, during flooding. With the help of all service hydrologists in the Southern Region of the NWS, criteria have been set at all forecast points for varying degrees of flooding. They include minor, moderate and major flooding and floods of record. The new verification program produces statistics in textual and graphical form for each category, effective lead times, average error, probability of detection and false alarm ratio. |
While hydrology is far from an exact science, it is hoped that statistics from this verification program will help us zero in on our problem areas, and help us increase our forecasting skill.
Figure 1. ABRFC Forecast Lead Times We currently anticipate running verification statistics every three months, and then a yearly summary. Many of these statistics will be posted on the internet. Figure 1 is a sample graphic that was created using output from the past six years of data at the ABRFC. The graphic shows lead times for all forecasts issued for the different categories of flooding. Higher numbers are better. |
New River Flood Watch Products to be Implemented
By Mike Boehmke
Return to Menu| A new
program to standardize flood potential information is scheduled to
commence within the Southern Region of the NWS. The NWS has long
provided this information in the form of flood potential outlooks and
similar products. This new initiative will provide a more uniform look
to products issued across the south central and southeastern United
The new River Flood Watch program will consist of both a traditional text product and a new graphical product. Both products will provide essential details such as forecast rainfall
|
and the
possible magnitude of any flooding. In addition, the graphical
representation, which will be available via the Web, is expected to help
bring the information to life by clearly displaying at-risk areas on a
map, and linking this display to a map of forecast rainfall.
The software to extract the flood forecast data from the river forecast model, produce the text products, and provide a summary of areas threatened is being written here at the ABRFC. The software to produce the graphical products is being written jointly by the Southeast RFC (SERFC) in Atlanta, GA and the Lower Mississippi RFC (LMRFC) in Slidell, LA.
|
ABRFC Archives
By James Paul
Return to Menu| Several
years ago, the ABRFC came to the conclusion that maintaining an archive
of both operational data and issued products would benefit the office.
An archive could be used for case studies, comparisons to current
events, calibration efforts and other potential applications.
The ABRFC archives consist of two archiving systems, data stored in a relational database and data/products stored in a typical file/directory structure. The ABRFC began archiving river stage, discharge, precipitation and temperature data in a relational database in March of 1993. Initially, only synoptic time (00Z, 06Z, 12Z, 18Z) data were archived. This database continued to grow with time through the spring of 1999. In April of that year, the ABRFC decided to move this data onto its new AWIPS platform. At that time, the format of the data was altered to more closely match the format of the operational data also stored in an AWIPS relational database. More data types were archived, including forecast river stages. Additionally, hourly data were being archived, as opposed to the six-hourly frequency of the previous design. In May of 1999, the ABRFC duplicated this database structure on a PC running SQL Server. |
This allows the
hydrometeorologists another avenue to obtain data and provides another
level of backup. Also, the ABRFC staff has written several applications
to retrieve this data and display it both graphically and in text form.
The second archiving system, the file/directory structure, also had its beginnings in 1993. Late in that year, the ABRFC became the first RFC, and one of the first offices in the NWS, to become an Internet presence. We started Gopher and FTP services and began putting our precipitation products (Stage III netCDF and image files) on the server. More products, including hydrometeorological discussions (HMDs) and river stage forecasts, were added in early 1994. In May of 1994, the ABRFC created its first website and began making most of its products available to the Internet community. It quickly became apparent that this information would be useful for the local office and others users, so a system was devised to separate data into a long-term archive and a short-term archive, generally less than two weeks old. The archived files are saved according to year, month and data type, and most of the files have a timestamp in their filename to indicate creation time. Having the archived data on the web has proven to be very beneficial for both local and outside users, including emergency managers and researchers. The ABRFC archived data can be found at http://www.srh.noaa.gov/abrfc/archivepage.html. |