Drought conditions continue over much of the LMRFC area and we are entering the most active portion of the hurricane season. LMRFC has been busy with procedure development efforts to improve our forecast models and to bring new products online.

LMRFC has been active in working with our cooperators participating in coordination meetings with the US Geological Survey, US Army Corps of Engineers, and Tennessee Valley Authority. Our procedure development efforts continue with model calibration, product development with geographic information

systems (GIS), and changes in the way quantitative precipitation forecasts (QPF) are prepared and used by the NWS taking a large portion of our efforts. To help everyone understand the major changes that have taken place in the preparation of QPF, much of this issue will be devoted to documenting this change.

We will continue to improve our forecast techniques and work with our cooperators to provide better products and services. Any comments and ideas are greatly appreciated.

- Dave Reed

The old QPF process at the River Forecast Centers consisted of HPC providing graphical guidance to the WFO. The WFO would use the program WinQPF to make a 24 hour QPF in six hour

Figure 1: An example of NMAP displaying 24 hour QPF values for the LMRFC area. The HAS Forecaster can modify the precipitation values before entering this into our computer models.

Click for larger image

increments using the graphical guidance from HPC and other meteorological tools. The WFO would send the output from WinQPF to the RFC. The HAS forecaster at the RFC would combine QPF from each WFO using a program called HASQPF and create a mosaic of QPF over the entire RFC area. The QPF output from HASQPF was used as input for the hydrologic models at the RFC.

The new QPF process started at the LMRFC during the winter/spring of 2000. HPC started to provide gridded QPF forecasts to the RFC to supplement the graphical QPF forecast. To be able to use the gridded QPF forecasts directly, new software was needed at the RFC. A new software program called NMAP (see figure 1) was installed at the LMRFC and HPC provided NMAP training to the LMRFC staff on 4/13/00. The NMAP software inputs QPF from HPC directly and allows the LMRFC forecaster to manipulate/modify the gridded QPF. After the HAS forecaster modifies the QPF, output from NMAP is formatted for input into the LMRFC hydrologic models.

The NMAP program contains several post processing scripts which are used by the LMRFC. One script produces a text file of six hourly QPF for each sub-basin in the LMRFC area. Another script produces 6 and 24 hour gif files of QPF and sends the files to the web site http://www.srh.noaa.gov/lmrfc/qpfpage.shtml. Finally, NMAP post processing scripts create LMRFC modified gridded QPF files and transmits them through the wide area network for distribution through the NWS.

Before converting officially to the new QPF process, a test and evaluation of the procedures were conducted from 5/15/00 thru 7/31/00. LMRFC forecasters were able to get accustom to the new procedures and use QPF from NMAP for input into the hydrologic models.

WFO's continued to send QPF during the test and evaluation period in the event problems were encountered using the new QPF procedures. On 8/1/00, LMRFC began officially using the new QPF process for daily QPF operations and WFO's were no longer required to issue QPF forecasts.

The new QPF process will continue to evolve over time. In the future, gridded 6 and 24 hour QPF from each RFC will be combined to generate one national mosaic product. The new QPF process will expand to include RFC's west of the continental divide and the National Precipitation Verification Unit (NPVU) will verify QPF using precipitation estimates from NWS radars across the country.

 

- Jeff Graschel

As was stated in the 1999 Spring issue of Crawfish Tales, there are three formats that make up SHEF. Through the use of parameter code characters to identify the data, these three message formats have the flexibility to transmit a wide range of hydrometeorological information. The formats are as follows:

.A format is for single stations and is designed for transmission of one or more hydrometeorological parameters observed a various times. Example: .AR DENL1 20000327 C DH07/HG 11.20/PP 1.50 .A .A format header code (must be in column 1 and 2 of message) R Revised data transmission. This is used to correct previously transmitted data (must be in column 3 of message.) DENL1 NWS Location identifier (3 to 8 characters) 20000327 yyyymmdd (year, month, day). Date of data to be decoded. C Time zone code (in this case Central Local time, absence of a time Zone code defaults to GMT) DH07 Time of observed data (in this case 0700 central local time) HG Physical element (HG is height, river stage - 11.20 ft.) PP Physical element (PP is the 24 hour rainfall - 1.50 inches.)

.B format is for the transmission of one or more hydrometeorological parameters from several stations for which many or all of the parameters are the same and are observed at corresponding times. Example: .BR SHV 20000727 Z DH12/HP/HT/PPP/QTI TXKT2 226.82/ 215.00/ 2.20/ 1.000 : WRIGHT PATMAN .END Line 1 .B .B format header code (must be in column 1 and 2 of message) R Revised data transmission (must be in column 3 of message) SHV (used in .B formats only) Message source is Shreveport WFO. 20000727 yyyyddmm (year, month, day) Z Time zone code (GMT) Default for no time zone code is also GMT. DH12 Time of observed data (1200 GMT for this example) HP Physical Element (Elevation, Pool) HT Physical Element (Elevation, Tailwater) PPP Physical Element (Precipitation beginning at 7am local and ending at the current local time.) QTI Instantaneous discharge at observed time Line 2 TXKT2 Location Identifier 226.82 Data value associated with HP 215.00 Data value associated with HT 2.20 Data value associated with PPP 1.000 Data value in KCFS associated with QTI Line 3 .END Terminator (must be in the first 4 columns)

.E format is used for the transmission of a single hydrometeorological parameter at an evenly spaced time interval for a single station. Example: : OHIO RIVER CAIRO, IL .ER CIRI2 0728 CS DH06/DC199907270908/HGIFF/DIH24/ .E1 22.0/ 22.1/ 22.3/ 22.4/ 22.5 Line 1 : Denotes line that follows is ignored by decoder unless another colon is encountered to turn the decoder back on. Line 2 .E .E format header code (must be in column 1 and 2) R Revised data transmission CIRI2 Location Identifier 0728 valid mmdd (month, day, assumes current year of 1st data value) CS Time zone code, in this case is Central Standard Time DH06 Time of observed data of 1st data value. (6am CST) DC199907270908 DC denotes creation data . Time at which data elements were created, should be use when transmitting forecast data. yyyymmddhhnn (year, month, day, hour, minute) HGIFF HG( Height of River), I (instantaneous), F (forecast data), and F (forecast data with QPF added, a Z in this position indicates forecast data with no QPF.) DIH24 Denotes time interval between data values. In this case, values are 24 hours apart starting at 6am CST. Line 3 .E1 E format continuation line 1 22.0 forecast data value on 0728 at 6CST 22.1 forecast data value on 0729 at 6CST 22.3 forecast data value on 0730 at 6CST

The current version, 1.3, of the SHEF decoder is Y2K compatible and allows for the extension of input beyond 80 characters. The .E1 format continuation line .E1 in the above example was inserted for visual readability. Any agency or cooperator can download a complete copy of the SHEF documentation and/or a copy of the SHEF decoder from the following web site: http://hsp.nws.noaa.gov/oh/tt/soft/hsoft.shtml

- Marty Pope