The primary objective of the Techniques Development Program is to develop analysis and forecast techniques which, when implemented, will help improve forecast accuracy and service to the wide range of users of NWS products. These techniques are implemented on NOAA's computer system, when appropriate, and guidance products disseminated via AFOS, facsimile, or other NWS distribution systems. Techniques are produced for basic weather elements used in public and aviation forecasts, such as temperature and visi- bility. Also, special emphasis is given to marine-related forecasts and to those forecasts especially associated with mesoscale processes. For many synoptic-scale forecasts, the output of operational numerical models is used to produce fore- casts of weather elements of interest to users. When dealing with forecasts of shorter time and space scales, more use is made of data sets rich in information on those scales; for example, hourly and automated surface reports and radar, satellite, and profiler data. For marine-related forecasts, numerical and statistical models relate elements of interest, such as storm surge, to atmospheric analyses and forecasts.

This quarter, development of the MOS-2000 system accelerated now that the conversion of HDS software to the CRAY is complete, and an ensemble-based MOS message was placed on the OSO file server. Some graphic products based on the trajectory model were reimple- mented following their demise with the HDS, and others are being made ready. The Thunderstorm Product was integrated into NSSL's Warning Decision Support System for testing at the Sterling Forecast Office. The ICWF/LAMP system was implemented into AWIPS, and improvements are underway for Build 3.1. Incorpora- tion of radar data into the moisture analysis of LAMP is under- way; radar data were used in LAMP development, but the system implemented does not include radar data. Many enhancements are being made to the ICWF concurrently with the integration of ICWF and AFPS into an Interactive Forecast Preparation System. A major enhancement to the tools to be made available for selective use of the ICWF at AWIPS sites is the slider bar technique for aiding in the interpretation of, and interaction with, model output fields. Work began on an extratropical storm surge model for the west coast of the U.S., and work on the parametric wave model continued.

OBJECTIVE WEATHER PREDICTION PROJECT (P. Dallavalle)

Short-Range Weather Forecasting Task (P. Dallavalle): Work continued at the Cooperative Institute at Pennsylvania State University. A fine-scale gridded climatic database representing the monthly mean values of the daily maximum/minimum (max/min) temperature and the dew point and wind speed valid at 3-h inter- vals was completed for the contiguous U.S. This dataset was created for the max/min temperature by using NCDC's 30-year climatic normals and a local regression technique to interpolate to specified gridpoints. The local regression technique related the climatic normal at an observing site to predictors such as latitude, longitude, elevation, and land use. Data from a set of sites surrounding the specified grid point were included in the local regression model by weighting the information according to distance from the point where the climatic normals were to be estimated. Once the regression equations were developed for a given geographical domain, the climatic normals could be estimat- ed for any point within that domain. Analogously, the climatic data for the dew point and wind speed were calculated from TDL's archive of hourly surface observations and interpolated to gridpoints by the local regression approach. Currently, the climatic data are stored on a grid with approximately 1-km resolution. Work is now underway to develop a methodology to produce MOS forecasts of the max/min temperature, dew point, and wind speed at any point in the contiguous U.S. regardless of the availability of a historical observational record. The MOS forecasts at known points, climatic normals, station parameters, and interactive variables will be used as predictors in this scheme.

Prior to the removal of the HDS main-frame in April, we developed an Eta-based MOS system to predict the max/min temperature and the probability of precipitation amount at over 500 stations in the contiguous U.S. Because of the small developmental samples, we developed equations for only two statifications which corre- sponded to the warm (April - September) and cool (October - March) seasons. In limited testing on independent data from October 1996 through January 1997, we found that, overall, the Eta MOS max temperature guidance was more accurate than the NGM-based MOS guidance; the opposite was true for the min temperature guidance. For precipitation amount, we found that the skill of the Eta MOS was approximately the same as that of the NGM MOS system, despite the relatively small size of the Eta developmen- tal sample.

Checking of the MOS-2000 interpolation software on the Cray is now underway. We've written subroutines to compute certain meteorological variables such as dew point, specific humidity, and moisture divergence, that are unavailable in the basic output of the numerical weather prediction models. As part of the MOS-2000 system, we've also instituted a new archive of forecasts from the Aviation (AVN) run of NCEP's Global Spectral Model (GSM). In our new archive, forecast output from the AVN model is collected four times a day (0000, 0600, 1200, and 1800 UTC initial times). AVN forecasts are saved at 3-h resolution from initial time out to the 72-h forecast projection. We're using an NCEP-provided process to interpolate the AVN model data, which are stored on a 1 degree latitude/longitude grid covering the entire globe, to a polar stereographic grid (with resolution of 92.25 km at 60 degrees north) covering the contiguous U.S., Alaska, Puerto Rico, and Hawaii. Presently, we are archiving nearly 1500 fields per forecast cycle from the AVN model.

As part of the MOS-2000 system, we are converting our historical archives to a new data storage format. At this point, the hourly observational data from 1990 through 1996 have been written in the new format. We are writing software to convert our archives of model data to the new format as well.

Medium-Range Weather Forecasting Task (M. Erickson): An experimental ensemble-based MOS message has been placed on the OSO file server for interested users. Ensemble MOS forecasts are obtained from the 12 runs of the Medium-Range Forecast (MRF) model in-cluded in the 0000 UTC ensemble suite each day. The model forecast from each of the ensemble runs is used as input to the operational MRF-based MOS prediction equations. The experimental message contains forecasts of max/min temperature and the probability of precipitation (PoP) for both 12- and 24-h periods. Forecasts are valid from 12 to 192 hours after 0000 UTC. For each element and projection, the message contains the operational MRF MOS forecast, a mean forecast computed by averaging the 12 separate ensemble MOS forecasts, the standard deviation of the forecasts, and the highest/lowest MOS forecast included in the set of 12. Verification of the forecasts contained in this message is in progress.

A new archive of the MRF model forecasts has been established. Model forecasts of 51 fields will be saved in MOS-2000 format for each of the projections 0 through 384 hours after 0000 UTC in 12-h increments. For the 0- through 168-h projections, the data are being converted from a 1 degree latitude/longitude grid covering the globe to a 95.25 km polar stereographic grid cover- ing the same area as the AVN archive described above. For the 180- through 384-h projections, the data, available on a 2.5 degree latitude/longitude grid, are being saved on a 190.5-km grid.

The investigation of data available from the NCEP/NCAR Reanalysis Project has gotten underway. This project involves using a frozen analysis/forecast system (the T62 version of the GSM) to reanalyze and generate forecasts for historical data from 1957 to the present. In the Reanalysis Project, forecasts are generated every 5th day for projections out to 384 hours from the 0000 UTC start time. To date, forecasts are available from 1973 to the present. The data are available on the same 2.5 degree grid discussed above and for the same fields TDL is now archiving from the MRF. We have obtained the necessary software and procedures to read these data, and will determine the practicality of expanding our historical database of MRF forecasts. The new MRF archive will be used in redeveloping the medium range forecast guidance as well as in future efforts to develop guidance for week two.

National Verification Processing Task (V. Dagostaro): In support of the AFOS-Era Verification (AEV) program, we continued our effort to convert the data processing system to run on a UNIX-based platform. Existing software to read or extract data from the AEV score archive was compiled and modified as necessary to run on the Cray. In addition, new software to process and temporarily store the unedited AEV data was modified to improve its efficiency. We are beginning efforts to save the AEV data in the new MOS-2000 format. Finally, drafts of TDL office notes documenting our new ASOS precipitation amount quality control algorithm and the verification results for previous warm and cool seasons were prepared.

Severe Weather Prediction Task (R. Reap): New NCEP graphics routines were incorporated into the operational NGM-based trajectory model program in order to process the three-dimensional parcel trajectories for display as graphical products on AFOS. The individual graphics contain plotted 24-h trajectories that terminate at the surface, 850-mb, and 700-mb levels. These products were designed to replace existing products that have not been available to field forecasters since the deactivation of the HDS main-frame system. The new code was implemented in June.

Work is currently underway to replicate the trajectory model 4-panel chart that is transmitted on facsimile (DIFAX) in slots D074 (0000 UTC) and D218 (1200 UTC). This chart has not been transmitted since the demise of the HDS system. The 4-panel chart contains parcel trajectories combined with various con- toured fields from the trajectory model. Thus, one panel dis- plays the surface trajectories and the surface dew point field, a second panel displays the 850-mb trajectories and the 850-mb temperature field, and a third panel displays the 700-mb trajec- tories with the 700-mb dew point field. A fourth panel contains the K stability index combined with the 700-mb net vertical displacement over a 12-h period. At present, test charts are being generated daily with plotted parcel trajectories and the graphics legends. NCEP is currently working to combine these charts with the contoured fields to create the 4-panel display. We anticipate that these charts will be available on DIFAX in the near future.

LOCAL TECHNIQUES DEVELOPMENT PROJECT (R. Reap) 0-3 Hour QPF and Severe Weather Task (D. Kitzmiller): obtained a new national archive of radar mosaics for use in developing the 0-3 h QPF system and other algorithms. The mosaics are prepared commercially by WSI and are distributed to academic and government researchers by the Global Hydrology Resource Center at the Global Hydrology and Climate Center, Huntsville, Alabama. They consist of radar data on 2-km grids for 16-levels at 15-minute intervals. The new archive presently covers the period November 1995 to March 1997; more data will be ordered shortly.

The developmental sample for the extrapolative-statistical 0-3 h QPF algorithm was expanded by interpolating the new 2-km radar data to the operational 10-km Radar Coded Message mosaic grid. Experimental results suggest that there is little differ- ence in quality between extrapolative forecasts of radar reflec- tivity based on lag-correlation pattern matching and those based on advection following the 700-mb wind vector. It is possible that this finding is specific to the Plains region, where the forecasts were verified. In any case, the results appear to indicate that convective storm velocity is influenced more by large-scale air flow than by storm-scale effects, such as bound- ary-layer interactions, over time scales greater than one hour.

TDL's 0-1 h QPF forecast algorithms are currently being run at NSSL in real time with base reflectivity data from the WSR-88D at Twin Lakes, Oklahoma. The severe weather probability, VIL, and large-hail probability algorithms have been also been implemented within this system, prior to incorporation within NSSL's Warning Decision Support System (WDSS). This represents the first real-time test of these algorithms as produced from base rather than graphic product data. By August, the TDL algorithms should be available at selected NWS field sites where WDSS has been in- stalled.

Thunderstorm Identification and Forecasting Task (S. Smith): Coordination continues with PRC on implementing the AWIPS Thunderstorm Product code within the D2D display of AWIPS Build 4.x. The code itself underwent a major overhaul in order to improve overall efficiency.

Work continues on the implementation the Phase-II Thunderstorm Product on the GDP system for the 1997 convective season. In this phase, GOES infrared satellite imagery is incorporated to provide a more robust thunderstorm detection algorithm, to estimate the minimum cloud-top temperature of radar-detected storms, and to improve the detection of ground clutter and anonmalous propagation.

The Thunderstorm Product code was integrated into NSSL's WDSS for the 1997 SCAN field test at Sterling, Virginia. NSSL is current- ly testing the code using real-time radar data from the Twin Lakes WSR-88D radar in Norman, Oklahoma. A homepage for the world-wide web was designed and implemented for SCAN. The page includes a description of the general SCAN effort and goals, the minutes from the regular SCAN meetings, and the status of the 1997 SCAN field test at the Sterling forecast office. Three teleconferences were also held in preparation for the SCAN 1997 field test at the Sterling forecast office.

Coordination was completed with the Storm Prediction Center to provide a Day-3 convective outlook for the SCAN field test area. The outlook is used primarily in the travel planning of SCAN participants outside of the Washington D.C. area. The outlooks are made available via the SCAN homepage. An automated routine to display graphics of the MOS Day-1 and Day-2 general thunder- storm and severe storm outlooks on the SCAN homepage was imple- mented. SPC's official outlooks are also available on the homepage for quick comparison with the MOS outlooks.

Local AWIPS MOS Program (LAMP) Task (J. Ghirardelli): Routine quality control of the LAMP forecasts revealed a problem with the warm-season forecasts of temperature and dewpoint, primarily in the southwestern United States. We thoroughly examined this issue and found that the problem was related to two interactive predictors. On very rare occasions, the predictors had non-zero values in the developmental warm-season sample. As a result, inclusion of the rare data caused overfitting in the final forecast equations. We identified those stations and times which were affected by the inclusion of data from this rare event. To correct the problem, we changed the equations to not include the value of these predictors at those sites and times where there were too few cases. Approximately 8000 temperature and dewpoint equations were corrected. We continue to monitor the forecasts and will soon test our corrections to the equations and their impacts on the forecasts. The corrected equations will be operationally implemented when our testing is satisfactorily completed.

During this quarter, we started our effort to include radar data into the operational LAMP system. We began examining the 10-km Radar Coded Message (RCM) data for quality and timeliness. We examined various ways of transferring the 10-km RCM reflectivity data to an 80-km LAMP grid and converting the data to a "radar factor" representing the percentage of the surrounding area in which precipitation was not occurring. We have transferred the data to an 80-km grid and plotted the results. The next step is to transfer the data to the 80-km LAMP grid. Finally, we will evaluate the effect of the inclusion of the radar data into the LAMP forecast system. We also began displaying both the LAMP grid fields and the radar data by using GEMPAK graphical display software, which was an invaluable tool in examining the quality of the transformed fields.

J. Ghirardelli and S. Kelly authored an abstract titled "A General Overview of Methodology and Applications of the Local AWIPS MOS Program (LAMP), a Short-Range Forecast Guidance Prod- uct." The abstract was submitted for presentation at the joint 16th Conference on Weather Analysis and Forecasting and the 12th Conference on Numerical Weather Prediction, to be held at the 78th Annual AMS Meeting in January 1998 in Phoenix, Arizona.

LAMP testing continued with live data on AWIPS equipment. LAMP data files for Pleasant Hill, Missouri; Tulsa, Oklahoma; and Wichita, Kansas, were transferred to PRC for testing and inclu- sion in AWIPS Build 2.1.

Heavy Precipitation Forecasting Task (J. Charba): A paper titled, "Gridded Monthly Climatic Frequencies of Precipi- tation Amount for 1-, 3-, and 6-H Periods over the Conterminous United States," was accepted for publication in the AMS Journal of Weather and Forecasting. The paper documents the development of a detailed climatological analysis of precipitation and its application as one of the predictor inputs into the LAMP QPF model.

At NSSL's request, a portion of the LAMP precipitation climato-logy dataset was provided to them for comparison with a precipi- tation analysis performed by NSSL in connection with the SWAMP96 project conducted over the southwestern U.S.

The LAMP QPF model continues to be run in an experimental mode on TDL workstations. Graphical products from the model are currently available to NWS field offices for operational evalua- tion. Post-processing of the products has also undergone some changes. Specifically, the ordering of the smoothing and normal- ization of the forecast probabilities was reversed so that the final probabilities now retain exact normalization. A change was also made in the directory structure used for storage of the forecasts on the OSO server to assure consistency with other TDL products on the server. This change necessitated a modification to the shell script used by field offices to download and display the forecasts.

Data and software transfers from NCEP's HDS 9000 to the CRAY were completed. Software codes that process and archive climatic hourly precipitation data from the National Climatic Data Center have been converted, as have a number of other codes that access and display various LAMP model data that were originally archived on the HDS 9000. Finally, TDL library software that performs verification analysis of archived forecasts was converted for use on the Cray.

MARINE ENVIRONMENTAL PREDICTION PROJECT (W. Shaffer)

Hurricane Storm Surge Forecasting Task (W. Shaffer): We examined Holland's hurricane wind model and compared it to our SLOSH wind model. The structure variations and maximum wind estimates are closely correlated, although the two approaches are different. SLOSH specifies a wind profile and uses the radius of maximum wind central pressure to determine the wind structure. Holland anchors his scheme to the central pressure and assumes a pressure profile. We were able to generate a relationship between Hol-land's parameters and our radius of maximum wind.

Extratropical Storm Surge Forecasting Task (W. Shaffer): We began work on an extratropical storm surge model for the West Coast of the U.S. The grid chosen covers the northwest coast, with emphasis on Puget Sound and the Columbia River basin. We are beginning to extract bathymetric data for the grid area. These data will be based on data from NOAA's National Geodetic Data Center.

Coastal Wave Forecasting Task (C.S. Wu): We applied our deep-ocean parametric wave model to hurricanes Opal (1995) and Fran (1996). Observations of winds and waves compared favorably for hurricane Fran. However, for hurricane Opal, the waves were overestimated due to overly strong winds produced by the SLOSH model's wind field. Opal was undergoing rapid intensification as it passed to the right of the wave buoy in the Gulf. Data were collected for hurricane Felix (1995) for input into the SLOSH wind model. We are continuing to verify various numerical models for breaking waves using various nearshore depth profiles and incident wave conditions.

As a NOAA partner with the University of Delaware in NOAA's Sea Grant program, we had a water level sensor installed at Duck, North Carolina, for the SandyDuck '97 experiment.

WFO Application Development Task (D. Ruth): Design, Development, and Testing teams continued to respond to Deficiency Reports (DRs) against hydrometeorological (HM) applications. AWIPS Release 1.2 distributed in January 1997 included contour setup data fixes, improved functionality of the NEXRAD image and graphic application, and improved performance of the grid pro- cessing applications. AWIPS Release 1.3 released in March included the addition of boundary layers for some derived parame- ters and the modifications necessary to accommodate changes to the NCEP data stream. Final user guides were completed for the grid processing, NEXRAD image and graphic, contouring, and parameter derivation applications.

Development activities this quarter centered around the integra- tion and implementation of the fog product and ICWF/LAMP into AWIPS Release 2.1.

Other areas of work include the preparation of an AWIPS Applica- tion Integration Framework Manual which will support local application development for Build 3 and beyond, preparation of a thunderstorm product for Release 3.1 and forecast verification applications for Build 4.

Product Generation Task (M. Peroutka): TDL completed the delivery of the ICWF to PRC. The last few bug fixes were implemented, installation scripts were developed, tests were performed at TDL and at PRC, and configuration data were devel- oped for Pleasant Hill, Missouri; Wichita, Kansas; Tulsa, Oklaho- ma; and Taunton, Massachusetts. The ICWF Users Guide was re-written as part of the AWIPS documentation set, and a hypertext (HTML) version was developed to be used as on-line documentation. TDL developers also worked with the NWS Training Center to develop a 5-day course to train Interactive Forecast Preparation (IFP) Focal Points. This included participation in the pilot version of the course as well as the actual training.

ICWF source, executables, and documentation for the ICWF were made available for FTP downloading at the end of April. Several commercial vendors took advantage of this opportunity. This version of ICWF was also installed at the ICWF test sites in Boise, Idaho, and Peachtree City, Georgia.

Design began to allow IFP implementation to follow an evolution- ary path throughout the NWS. This means that selected parts of the ICWF will be configured and used at all AWIPS sites, with the intent that they would migrate toward using the full digital database at a later date. Those parts which will be useful at all sites must be identified, made readily accessible, and documented.

Design work was performed to produce messages for voice synthesis via the NOAA Weather Radio (NWR) Console Replacement System. The technique will support the generation of messages from grids, digital forecast matrices, and from existing text forecasts. Forecasters will be able to select "remapping patterns" which control how zone-based forecasts are applied to the NWR Service Area.

Three TDL developers traveled to Boulder to continue work on the consolidation of ICWF and AFPS. Discussions focused on the database server, the flow of data throughout the system, and internal data representations. TDL worked to provide ICWF GRIB messages to FSL for display by the Graphical Forecast Viewer. This will allow ICWF forecast offices to display experimental gridded forecast products via a link from their homepage.

Interactive Techniques Development Task (D. Ruth): Work continued on the development of model interpretation techniques using slider bars for the National Centers. The intent of this capability is to allow forecasters at the National Centers to prepare "value-added" forecast gridfields without loss of model detail. Threshold adjustments on national grids are interpolated in space, as well as time. The development of a prototype system is being coordinated with NCEP's Hydrometeorolgical Prediction Center.

An ICWF Development Homepage was drafted. The purpose of this document is to keep ICWF users up-to-date on recent enhancements, known problems, and proposed enhancements to the ICWF. Develop- ment information specific to each of 16 ICWF components is provided. A similar document for IFPS development is also being prepared.

TDL and FSL jointly submitted four abstracts on IFP to the 14th International Conference on Interactive Information and Process- ing Systems for Meteorology, Oceanography, and Hydrology, to be held during January, 1998, in Phoenix, Arizona. The abstracts were entitled: "Interpretation and Editing Techniques for Interactive Forecast Preparation," "The Generation of Products in Interactive Forecast Preparation," "Implementing Interactive Forecast Preparation Nationwide," and "Creating NOAA Weather Radio Broadcast Scripts using Interactive Forecast Preparation."

AFOS Field Applications Assistance Task (R. Beasley): TDL rewrote major portions of the MONITR and TAFDEC software which are used to compare aviation forecasts (TAFs) with current observations (METARs) to support a continuous weather watch. The display of the amendment/alert messages was improved and enhanced to include a discrepant element summary. The new software was sent to all field offices in early June.

The new TAFDEC software was also incorporated into the national verification program (VERIFY). This will again allow forecast offices to include aviation forecasts in their forecast verifica- tion process. The software was released for testing at two sites in each region in mid-April. Additional corrections and enhance- ments are being made to address field concerns which arose during testing. TDL expects to release the new software to all field sites during July.

TDL continued to support the ASOS program by providing reports on a routine basis to the ASOS Program Office and the Observing Systems Branch of missing observations, observations with missing elements, incorrectly formatted observations, and garbled obser- vations. This information was helpful in determining the cause of garbled observations from several PACE systems. Output was provided by TDL's CHKSFCOBS and PROCAES ASOS quality control software. These programs were also used to provide the FAA with data needed for an ASOS quality control assessment study con- ducted from January through June.