Short-Range Weather Forecasting (P. Dallavalle):A complete set of statistical guidance developed from the Nested Grid Model (NGM) is now available for 60 stations in Alaska. During this quarter, we implemented equations to predict probabilities of ceiling height, visibility, and obstruction to vision for the Alaska region. Development of forecast equations required to predict precipitation amount in Alaska during the warm and cool seasons was also completed, and these equations were implemented in late August. All equations were produced by applying the Model Output Statistics (MOS) approach to the NGM. As part of the implementation process, we replaced the MOS guidance based on the Limited-area Fine-mesh Model (LFM) and distributed to U.S. Air Force users in Alaska with a guidance package based exclusively on the NGM MOS forecasts. Work is now underway to remove the LFM MOS guidance for the Alaska civilian stations as well.

Development of NGM-based MOS equations to predict precipitation characteristics (drizzle, showers, or continuous precipitation) for stations in the contiguous United States and Alaska is complete. These equations were operationally implemented in late August. Forecasts produced by these equations are available in digital format for use by local forecast applications such as the Interactive Computer-Worded Forecast (ICWF) or the Local AWIPS MOS Program (LAMP).

During this past quarter, M. Antolik, P. Dallavalle, and M. Erickson presented workshops on TDL's statistical guidance at NWS offices in Pittsburgh, Pennsylvania; Greer, South Carolina; and Wilmington, North Carolina, respectively. M. Antolik also attended the NWS Eastern Region Probabilistic Quantitative Precipitation Forecasting (PQPF) Workshop held on the campus of Penn State University. At the workshop, he presented a talk entitled "TDL's NGM-Based QPF Guidance - 'Categorically' More than Meets the Eye", which discussed the NGM MOS probability forecasts. This guidance represents TDL's current synoptic-scale QPF.

Software has been written and implemented to read MOS forecasts from the operational TDL forecast files, analyze the data on a grid, create a GRIB record, and write the resulting information to a sequential dataset for transfer to one of the National Centers for Environmental Prediction (NCEP) CRAY computers. Many of the MOS forecasts for both the short-range and medium-range projections are now available in NCEP's N-AWIPS graphical display system.

Efforts to design, develop, and implement a new MOS system (MOS 2000) continue. Initial development of the regression program is nearing completion. We are beginning work to move the MOS 2000 system from the HP platform, where it was developed, to the CRAY platform. At the same time, we've begun efforts to convert our entire operational and developmental systems from the HDS mainframe to the CRAY platform. Substantial work will be required to modify software, to revise the format of our developmental data sets, and to write the instructions required to execute jobs on the CRAY.

Medium-Range Weather Forecasting (M. Erickson): In July, we presented our proposal to produce statistical forecasts from NMC's ensemble model predictions to TDL staff. We also made a brief presentation at the first N-ACEP (NCEP Advisory Committee on Ensemble Prediction) meeting held in mid-September. We are moving forward with the short-term goal of creating consensus MOS forecasts based on ensemble runs for maximum (max)/minimum (min) temperature and probability of precipitation (PoP). Coordination with NCEP scientists to discuss requirements for output from the ensemble predictions is ongoing. Verifications of the MOS max/min temperature forecasts based on the Aviation (AVN) Model and the Medium-Range Forecast (MRF) Model were completed for the June - August 1994 summer season and the December 1994 - February 1995 winter season. Verifications of the AVN- and MRF-based MOS PoP forecasts were also completed for the 1994-95 cool season. The results indicate that the MOS forecasts improved over climate for all projections from 24 through 192 hours. A consensus forecast, formed by averaging the NGM MOS PoP and the MRF or AVN MOS PoP forecasts, exhibited greater improvements over climate than either system individually. This comparison was made for the 24-, 36-, 48-, and 60-h projections. Finally, a comparison of the MRF MOS max/min temperature and PoP forecasts with the human forecasts issued by NCEP's Meteorological Operations Division indicates the two systems have nearly identical levels of skill.

National Verification Processing (V. Dagostaro): In support of the AFOS-Era Verification (AEV) program, 1994-95 cool season verification summaries have been distributed to the regional Scientific Services Divisions (SSD's). For max/min temperature, probability of precipitation (PoP), cloud amount, and 42-h significant wind speed, the summaries contain comparative verification results for local and MOS forecasts for approximately 96 stations in the contiguous U.S. and 6 stations in Alaska. For snow amount and precipitation type, the summaries contain comparative verification results for approximately 86 stations in the contiguous U.S. For the aviation weather elements (i.e., ceiling height, visibility, and wind), local and MOS forecasts are not directly comparable because the forecast valid times generally do not match. Therefore, for ceiling height and visibility, comparative verification results for local forecasts and persistence observations were produced for approximately 96 stations in the contiguous U.S. and 6 stations in Alaska. Summaries for wind speed and direction contain local forecast verifications for the same stations. In addition, we produced verification summaries comparing the MOS guidance and persistence observations for ceiling height and visibility and verification summaries for just the MOS wind speed and direction. For all weather elements, NGM-based guidance was verified for the contiguous U.S., and LFM-based MOS was verified for Alaska. For max/min temperature and PoP only, AVN-based guidance was included in a separate verification comparing the two MOS guidance systems (NGM and AVN) and the local forecasts.

We've also produced and distributed to the SSD's verification results for the April - June 1995 period. For this 3-month season, only the max/min temperature and PoP forecasts were verified. Comparative verification results for local forecasts and NGM-based MOS guidance were produced for 95 stations in the contiguous U.S. Because of missing precipitation amount observations, verifications for 6 stations in Alaska included only results for the local and LFM-based MOS max/min temperature forecasts. For all stations in the contiguous U.S., the AVN-based MOS max/min temperature and PoP guidance were included in a separate comparative verification. For the stations in Alaska, only the AVN-based temperature guidance was included in the separate comparative verification.

Severe Local Storms Forecasting (R. Reap): Procedures and software to generate the new NGM-based MOS probability forecasts of nonconvective clear-air-turbulence (CAT) for the contiguous U.S. were operationally implemented in mid-August. The probability forecasts are valid for projections of 2-8 h, 8-14 h, 14-20 h, and 20-26 h after 0000 and 1200 UTC initial data times. Separate forecasts are available for the unconditional probability of category 3 (light to moderate) or greater and category 5 (moderate to severe) or greater CAT for 48-km grid blocks covering the contiguous U.S. The forecasts are also valid for low-band (below 15,000 ft) and high-band (above 15,000 ft) CAT. The probability forecasts are stored in GRIB format on a 40-km Lambert conformal grid (AWIPS grid 212). During operations, the GRIB records are moved to the NCEP N-AWIPS system where the forecasts are decoded, and graphic images displaying the probability forecasts over the contiguous U.S. are generated. The graphics are then transmitted to the National Aviation Weather Advisory Unit (NAWAU) in Kansas City, Missouri, which is responsible for issuing operational AIRMETS and SIGMETS for use by the aviation community. The graphical forecasts are displayed on the National Severe Storms Forecast Center's N-AWIPS system. The same products can also be displayed at NCEP and at TDL where the forecasts are available for operational use and quality control, respectively.

Technical Procedures Bulletin No. 430 titled "Probability Forecasts of Clear-Air-Turbulence for the Contiguous U.S." by R. Reap was prepared and will be distributed following review. A paper with the same title was also written and submitted for publication in the preprint volume of the 13th Conference on Probability and Statistics to be held February 21-23, 1996, in San Francisco, California.

LOCAL TECHNIQUES DEVELOPMENT PROJECT (W. Seguin)

Radar Applications (D. Kitzmiller): We began our development of a refined 0-1 h QPF algorithm by making use of the Stage III precipitation analyses derived from radar and raingage data from the Twin Lakes, Tulsa, and Wichita radars. As expected, the initial results show that the Stage III rainfall estimates are lower than those used in the development of the original rainfall forecast algorithm. The original algorithm was based only on radar reflectivity data which can be contaminated by hail or melting snow. This can cause the algorithm to overestimate the amount of precipitation. The extrapolation algorithm's skill at forecasting the precipitation is also lower. We plan to improve the algorithm by using a multiple motion vectors rather than the single motion vector, and we will refine the precipitation estimates by accounting for the varying beam elevations. A TDL Office Note describing the 0-1 h QPF algorithm was drafted and is now being reviewed.

0-3 Hour QPF and Severe Weather Applications (S. Smith): A prototype of the Eastern Region thunderstorm identification technique was implemented on the AWIPS Government Development Platforms (GDPs) this quarter. Using archived radar and lightning data, the technique enabled the computation of storm cell displacements and the display of the lightning and radar data. The prototype application is capable of displaying an alphanumeric product identifying the cell, its direction and speed of motion, and its distance from nearby airports.

The investigation into the relationships between the time of the first lightning flash and initial rapid cloud top cooling as observed by geostationary satellite in severe thunderstorms is continuing. This work will be integrated with the thunderstorm identification technique to improved the warning lead times for thunderstorms.

Local AWIPS MOS Program (LAMP) (W. Seguin): Emphasis on the LAMP equation development is continuing with a goal of completing all equation development this fall. The 1700 UTC warm equations were rederived because of problems discovered during their evaluation, the 2300 UTC cool equations were completed, and most of the 0200 UTC cool season equations were also derived. The latter represents the last full set of LAMP equations to be derived.

A verification study of LAMP ceiling heights was completed for the 1400 UTC cool seasons of 1993 and 1994. The results indicate that threat scores for LAMP forecasts improved on persistence forecasts for all projections for inland regions and most projections along the coast. LAMP also outscored NGM MOS forecasts although NGM MOS generally had higher Heidke skill scores due to LAMP's having difficulty forecasting unlimited and high ceilings in the northern U.S. An unmatched comparison to Local forecasts produced similar threat scores with LAMP having a greater probability of detection along with a higher false alarm ratio. Verification of LAMP visibility forecasts for the same start time and sample has begun.

Field Applications Assistance (R. Beasley): Updated programs for decoding surface observations (SAODECII, version 11.00), collecting and collating data for the National Verification Program (VERIFY, version 10.00), and retrieving surface observations from collectives (MOBSEP, version 2.00) were completed. A new program, RDALN, which reads a message containing the Automated Surface Observation System (ASOS) software load number, was also completed. All of these programs were extensively field tested and then mailed to the regions for distribution in September.

The aviation monitoring program, MONITR, was updated to allow it to swap to the new SAODECII. This will facilitate going to METAR messages in the future. The program is now being modified to swap to the terminal forecast message decoder.

The new program, KEYSADD, which will facilitate adding keys to the database in preparation for METAR, was given to the Integrated Testbed Team for testing. A final draft of the documentation was also prepared.

Substantial progress was made on the development of the new METAR and TAF decoders. Testing of both programs will begin next month. A quarterly conference call with the Local Applications Working Group was held, and a Quarterly Report was sent to printing and distribution.

New dual drive, 330-mb hard disks were installed on the two machines in the development facility. This substantially eases the maintenance of the AFOS systems used for development because all three systems now have identical databases, and PILEDITSs/EDITMERGEs can be accomplished at one time.

Quantitative Precipitation (J. Charba): Like the other LAMP equation sets described above, the emphasis this quarter was on completing the development of all remaining (eight) QPF start times for the warm season. The cool season equation sets were completed last quarter, and testing of these equations was carried out this quarter. Results from subjective examination of forecasts for individual cases shows that LAMP adds considerable spatial and temporal resolution to centralized MOS QPF, particularly in mountainous areas. Quantitative scoring statistics confirm the qualitative results, with LAMP showing slight nationwide improvement on MOS at all LAMP projections (which extend to 22 h from cycle time). Regionally, the improvement was substantial in mountainous areas.

The development of the implementation system for this product on the AWIPS Government Development computers continued this quarter as well. Software to change some predictor identifiers in the regression equations was prepared to accommodate the implementation system. A program was written to ingest NGM 850-b winds into the QPF implementation system.

MARINE ENVIRONMENTAL PREDICTION PROJECT (W. Shaffer)

Hurricane Storm Surge Forecasting (W. Shaffer): Work continues on the two basin updates for Florida - the Biscayne Bay (Miami) and the Florida Bay (Keys) basins. In addition, the grid for the Pamlico Sound (Cape Hatteras) SLOSH basin was extended to cure some boundary problems that were noted by runs at the National Hurricane Center. A numerical wave, generated as a storm left the northern boundary of the basin, amplified, propagated southward, and contaminated the results along the outer banks of North Carolina. NHC found that the extended basin helps cure that problem.

Extratropical Storm Surge Forecasting (W. Shaffer): The extratropical storm surge model for the east coast is being converted to run on NOAA's Cray C90 computers. This conversion allows us to input Aviation model wind and pressure forecasts at 3-h intervals. This finer temporal resolution has resulted in better water level forecasts for cases we compared.

S. Kim and W. Shaffer were invited by the Alaska Region to present their work on extratropical storm surge modeling, especially as it applies to the areas near Nome. The Nome office was able to arrange a helicopter overflight of Norton Sound, allowing Kim and Shaffer to adjust the model bathymetry to this shallow bay and to better appreciate the threat storm surge poses to Norton Sound residents. Seminars were presented at both the Anchorage and Fairbanks offices. While in Nome, they discussed their work at a meeting of the city council, and were interviewed by the local media. In addition, discussions with forecasters on the use of the model were held in Anchorage, Nome, and Fairbanks.

LOCAL TECHNIQUES DEVELOPMENT PROJECT (D. Ruth)

Interactive Techniques Development (D. Ruth): Work proceeded this quarter on Build 7.0 of the ICWF. Magnification was built into the grid modification (GMOD) software. This feature enables forecasters to enter detailed spatial forecasts in an area of concern without introducing unwanted discontinuities elsewhere on the forecast grid. Reference grids were added to GMOD. Reference grids allow forecasters to overlay their working forecast with contours from several sources of guidance. ICWF contouring software was replaced with hydromet contouring utilities being developed for AWIPS.

A feature which animates the working forecast using station models was incorporated into the ICWF matrix editor. Zone background colors in the loop correspond to the probability of precipitation. We continued to improve the zone combination recommendation. Instead of comparing forecasts projection by projection, a pattern matching scheme is now used.

The Watch, Warning, and Advisory (WWA) Interface was connected to the new WWA phrase server. The software was improved to graphically display multiple events forecast for the same zone.

Product Generation (M. Peroutka): Build 6.0 of ICWF was provided to the Olympic Weather Support Office (OWSO) in Peachtree City, Georgia, in July. This build contained several features which were tailored to support the Olympic games. For much of July and August, forecasters at the OWSO used the ICWF to generate specialized forecast products for Atlanta Sport '95-a sports exhibition. A minor update (Build 6.1) was furnished in August to each office which uses ICWF.

ICWF initialization and formatting routines were adapted to support AFPS which had its first end-to-end runs in Boulder during a meeting of the AFPS Forecasters' Working Group. M. Oberfield and M. Peroutka participated in this meeting.

M. Peroutka also travelled to the NWS Training Center in Kansas City to present instruction to a class of meteorological interns. He spent one morning teaching them about ICWF as part of a Forecaster Development Course (FDC). Routines which produce watch, warning, and advisory products as well as aviation products continued to be developed for Build 7.0. We also began developing software to phrase local effects in zones. A design review of this important capability was conducted in September.

WFO Application Development (D. Ruth): Design, Development, and Testing teams continue to develop hydrometeorological (HM) applications for the early builds of AWIPS. This work is being carried out in part by TDL and its application support contractor, GSC, and in part by the AWIPS Prime Contractor, PRC. Integration efforts became more focused this quarter as PRC's design for the AWIPS infrastructure gradually emerged. TDL and PRC worked together to define dozens of APIs which initiate HM applications, access and store HM data objects, render HM data on the screen, and handle log and error messages. HM applications were reconfigured to run under PRC's Developers Work Bench (DWB). Build 1 applications for contouring and NEXRAD display are being restructured to answer callbacks from the AWIPS single viewer.

A draft Test Plan for all DDT development was prepared and a final Software Development Plan was printed. Test procedures based on software requirements were developed for the GRIB decoder. An early version of the HM configuration management plan was drafted. Configuration management software (PCMS) was purchased for the headquarters GDP to support HM applications development for AWIPS. We expect to receive training in PCMS and complete the Configuration Management Plan early next quarter.

The fifth DDT team kicked off in early July. The team is led by D. Shelton of the Office of Hydrology. They are responsible for integrating the WFO Hydromet Forecast System into AWIPS Build 2. A startup meeting for the graphics portion of the NEXRAD display application was held in August. The design briefing for the NEXRAD image display was conducted in September.

Techniques Specification (E. Mandel): TDL and General Sciences Corporation (GSC) continued to support the AWIPS WFO Application Development. This included preparing a Requirements Traceability Document (RTD) that provides traceability of the set of HM requirements embodied in the Technique Development Packages (TSPs) down to the HM Computer Software Components (CSCs) and AWIPS build. These requirements will be used by the application developers to ensure that the appropriate requirements have been considered in the design and testing of the hydromet application.

TDL and GSC are continuing to identify and track changes to the WFO Application Requirements Database, and the System/Segment Specification (A-Spec). This included eliminating PASCAL as a language to be supported by AWIPS.

TDL and GSC continued to work with the AWIPS User Interface Working Group and participate in a data modelling effort as subject area experts for the Joint Application Development (JAD) group. Subject areas include point and gridded data, graphic and text products, and cartographic data.