MDL's July 2001 Progress Report

METEOROLOGICAL DEVELOPMENT LABORATORY

MONTHLY PROGRESS REPORT

July 2001

Statistical Forecast Development (P. Dallavalle): After coordination with the NWS Regional Scientific Services Divisions and with the Air Force Weather Agency (AFWA), we've selected 346 new sites to add to the AVN- and MRF-based MOS systems. With the exception of 23 sites selected by AFWA, all of the additional sites report surface weather conditions via METAR observations. We are planning to produce guidance for the new sites in the fall of 2001. Development of equations to predict temperature, dew point, wind speed, and wind components is underway; we've also begun the process of adding the new sites to our regionalized equations.

Development of new MOS guidance packages based on the 0600 and 1800 UTC runs of the AVN model is underway. Equations to predict wind components and speed have been developed for the warm season and have been turned over to our implementation team. Guidance produced from the new equations for April 2001 was evaluated for approximately 330 stations in the United States. As expected, we found that the 0600(1800) UTC guidance was more accurate than the 0000(1200) UTC guidance for the same valid times. We were surprised, however, to see that the 0600(1800) UTC AVN MOS wind guidance was more accurate than the subsequent 1200(0000) NGM MOS wind guidance valid at the same time. We've begun to develop the AVN-based MOS equations to predict temperature, dew point, clouds, ceiling height, wind speed, wind components, and precipitation type during the cool season.

After determining that the changes implemented in the MRF model in May 2001 resulted in a cold bias in the warm season maximum/minimum temperature guidance, we ran a series of tests to see if simple modifications to the MOS system might eliminate part of the bias. After experimenting with different dependent samples and eliminating certain predictors, we found that modifying the temperature fields predicted by the MRF seemed to improve the MOS guidance significantly. This modification was based solely on the error in the 24-h forecast of the MRF thermal fields. We will pursue these experiments further after we redevelop the MRF-and AVN-based temperature and dew point forecast equations for the cool season.

We've developed Eta-based MOS equations to predict the probability of 0.01 inches or more of liquid-equivalent precipitation (PoP). These cool season equations are valid for 6- and 12-h periods beginning at 6 hours after 0000 or 1200 UTC and ending at 60 hours after initial cycle time. Tests on independent data during the 2000-2001 cool season indicate that the Eta-based MOS PoPs are more accurate than the AVN-based MOS PoPs for projections out to 36 hours; beyond that time, the AVN MOS PoPs provide more skillful guidance. Both the Eta- and AVN MOS PoPs are 5 to 10% more accurate than the NGM-based PoPs. We're now working on developing an Eta-based MOS system to predict precipitation amount.

Operations and Software Support (M. Erickson): The new AVN MOS precipitation amount guidance system has now been implemented. Categorical precipitation amount guidance for both 6- and 12-h periods is available in the AVN MOS alphanumeric messages for forecast project tions out to 72 hours in advance. Probabilistic forecasts of categorical amounts are available in digital form in BUFR products disseminated to AWIPS and to Family of Services users. Changes in the operational MOS system were also made to produce probabilistic and categorical guidance for precipitation characteristics (drizzle, steady precipitation, or showers). This guidance is required to support the Interactive Forecast Preparation System (IFPS) and is also available in the BUFR products. Finally, we implemented modified visibility and obstruction to vision forecast equations and probability thresholds that eliminated spurious forecasts of fog in Texas and some areas of the Appalachians.

National Digital Forecast Development (D. Ruth): We prepared and distributed the third version of the NDFD Plan Draft to IWT members for review. This draft incorporates changes that resulted from Corporate Board Operations Committee and Corporate Board briefings, Bob Glahn's 30-day team recommendations, as well as decisions reached at the May meeting of the IWT.

Statistical Update Development (J.Charba): A manuscript titled, A Comparative Verification of Recent Quantitative Precipitation Forecasts in the National Weather Service: Development of a New Methodology for Hydrologic Applications, co-authored by J. P. Charba, D. W. Reynolds, B. E. McDonald, and G. M. Carter has been submitted for publication in the Bulletin of the American Meteorological Society. This paper documents the analysis that led to streamlining of the QPF forecasting process.

Storm Surge and Wave Forecasting (W. Shaffer): We continued to test several deep-water parametric wave models using SLOSH winds for predicting waves generated by tropical cyclones. The estimates of the highest waves based on the input of central pressures compared favor ably with observations. We anticipate that the computations of maximum waves will be useful to marine forecasters, as well as for our use in developing coastal wave impacts.

Forecast Verification Analysis (V. Dagostaro): We provided local and NGM MOS verification statistics for June 2001, to OCWWS for max/min temperature, PoP, ceiling height, visibility, and cloud amount. These data update the existing statistics-on-demand system. In addition, we assisted OCWWS in error-checking a prototype system to compare AVN MOS verification scores with those of either NGM MOS or local forecasters. Finally, we provided monthly comparative verification statistics for the local forecasts, NGM MOS, and AVN MOS for October 2000 - June 2001 for all of the above weather elements. These data will be used to populate the expanded statistics-on-demand system.

We're continuing the development of a prototype web page for disseminating MOS verification information for approximately 330 sites. We're also continuing to investigate the utility of various types of verification diagrams and scores for addition to this web page.