FSL in Review 2001 - 2002

Cover/Title Page


Organizational Chart


Office of the Director


Office of Administration
and Research


Information and
Technology Services


Forecast Research
Division


Demonstration Division


Systems Development
Division


Aviation Division


Modernization Division


International Division


Publications


Acronyms and Terms


Figures Listing



Contact the Editor
Nita Fullerton


Web Design:
Will von Dauster
John Osborn


Best Viewed With
Internet Explorer

FIR 2001 - 2002 FRD MastHead
Steven E. Koch, Chief
(303-497-5487)

Web Homepage: http://www-frd.fsl.noaa.gov/

Steven C. Albers, Senior Meteorologist, 303-497-6057
Robert L. Anderson, Research Associate, 303-497-6876
Dr. Stanley G. Benjamin, Chief, RAPB, 303-497-6387
Dr. Daniel L. Birkenheuer, Meteorologist, 303-497-5584
Dr. John M. Brown, Meteorologist, 303-497-6867
Dr. Gerald L. Browning, Mathematician, 303-497-6772
Kevin J. Brundage, Computer Specialist, 303-497-7246
Scott D. Buennemeyer, Systems Admin,. 303-497-6894
Dr. Fernando Caracena, Meteorologist, 303-497-6269
Susan C. Carsten, Secretary OA, 303-497-6779
Randall S. Collander, Meteorologist, 303-497-5960
Dr. Dezso Devenyi, Meteorologist, 303-497-6827
Gary Fisher, Systems Admin., 303-497-6754
Nita B. Fullerton, Writer/Editor, 303-497-6995
Dr. Cecilia M.I.R. Girz, Chief, MAB, 303-497-6830
Dr. Georg A. Grell, Meteorologist, 303-497-6924
Brian D. Jamison, Meteorologist, 303-497-6079
Bernadette M. Johnson, Secretary OA, 303-497-7260
Guo-Ji Jian, Guest Worker, 303-497-3734
Dr. Dongsoo Kim, Research Associate, 303-497-6725
Dr. Chungu (Dan) Lu, Meteorologist, 303-497-6776
Dr. Adrian Marroquin, Meteorologist, 303-497-6202
Paula T. McCaslin, Computer Analyst, 303-497-3187
Dr. John A. McGinley, Chief, LAPB, 303-497-6161
Dr. William R. Moninger, Physicist, 303-497-6435
Dr. Steven E. Peckham, Meteorologist, 303-497-7978
Paul J. Schultz, Meteorologist, 303-497-6997
Barry E. Schwartz, Meteorologist, 303-497-6481
Brent Shaw, Meteorologist, 303-497-6100
Thomas W. Shilling, Electronics Engineer, 303-497-6876
John R. Smart, Meteorologist, 303-497-6590
Tanya G. Smirnova, Meteorologist, 303-497-6253
Tracy Lorraine Smith, Meteorologist, 303-497-6727
Edward J. Szoke, Meteorologist, 303-497-7395
Dr. Edward I. Tollerud, Meteorologist, 303-497-6127
Diane I. Vinaske, Secretary OA, 303-497-6629
Dr. Steven S. Weygandt, Meteorologist, 303-497-5529
Linda S. Wharton, Computer Specialist, 303-497-7239
Dr. Yuanfu Xie, Computer Scientist, 303-497-6846

(The above roster, current when document is published, includes
government, cooperative agreement, and commercial affiliate staff.)

Address: NOAA Forecast Systems Laboratory – Mail Code: FS1
David Skaggs Research Center
325 Broadway
Boulder, Colorado 80305-3328


Objectives

The Forecast Research Division (FRD) is home to most of the research in FSL on short-range numerical weather prediction (NWP), development of advanced modeling and data assimilation techniques, diagnostic studies of mesoscale weather phenomena, and applications of NWP to non-meteorological uses. The division emphasizes the assimilation of operational, research, and future meteorological observations for analyzing current atmospheric conditions and the subsequent generation of short-range numerical forecasts. Produced in real time at frequent intervals on national and local scales, these analyses and forecasts are valuable to commercial aviation, civilian and military weather forecasting, the energy industry, regional air pollution prediction, and emergency preparedness. FRD also has supported several large meteorological field experiments and continues to perform this service to the community.

The Forecast Research Division comprises three branches:

The Regional Analysis and Prediction Branch supports the following research programs:

    Rapid Update Cycle (RUC) – A complete analysis/forecast system for hourly assimilation of meteorological observations over the United States into a numerical prediction model, the RUC has been implemented as an operational forecast system at the National Centers for Environmental Prediction (NCEP). The branch develops and tests improvements to the RUC and its research counterpart, the Mesoscale Analysis and Prediction System (MAPS), in the following areas:
    • Data Assimilation – Improved techniques for estimating meteorological parameters on a regular grid, combining information from in situ and remote observations with that from a forecast model, and investigation of uses for new data sources, such as rapid updating using Geostationary Operational Environmental Satellite (GOES) raw radiances and derived products. The latter task is being performed partly in collaboration with other members of the Joint Center for Satellite Data Assimilation (NESDIS, NASA, and NCEP).
    • Numerical Prediction – Design, testing, and implementation of improvements to the RUC/MAPS numerical prediction model, with a major emphasis on improving representation of processes near the surface and in clouds, which exert a strong control on mesoscale forecasts.
    • Analysis and Model Verification – Statistical and subjective evaluations of RUC/MAPS analysis and forecast products for standard atmospheric variables, surface conditions, aviation-impact variables, clouds, and precipitation.
    • Data Sensitivity Studies – Studies using the RUC conducted to determine the effects of different types of observations on short-range numerical forecasts, including wind profilers, GPS, and space wind lidar systems of the future.

    RUC Applications – Development of coupled atmospheric/land surface model capability in support of the Global Energy and Water Cycle Experiment (GEWEX) programs and the NCEP implementation of the RUC, forecasting of aviation impact variables (icing, turbulence, ceiling, and visibility) in support of the Federal Aviation Administration (FAA), wind forecasting applied to wind energy utilization, and real-time support for the NOAA Pacific Landfalling Jets Experiment (PACJET), the International H20 (IHOP) field experiment, and other field projects in which NOAA is engaged.

    Collaborative Modeling Projects – Lead role in the development and evaluation of the coupled MM5/air chemistry model, continued collaboration with NCAR in the advancement of the science of modeling precipitation physics, participation in the development of the Weather Research and Forecasting (WRF) model system and nonhydrostatic generalized vertical coordinate model, and development of a RUC Short-Range Ensemble Forecast system in collaboration with NCEP.

    Well-Posed Model – Development of a multiscale theory and model, and development of a well-posed, open boundary multiscale ocean model.

The Local Analysis and Prediction Branch is engaged in the following efforts:

    Local Analysis and Prediction System (LAPS) – Incorporation of local datasets into numerical models (e.g., MM5, RAMS, WRF) for the production of very detailed analyses of local weather conditions and short-range forecasts. The model is updated with new observations at least hourly. A diabatic initialization procedure known as the "hot start" has been developed for reducing the problem of cloud and precipitation "spinup" in the early hours of model integration. LAPS supports a broad clientele of mostly government and military entities, including the NWS, FAA, FHWA, AFWA, the Central Weather Bureau of Taiwan, and the Korean Meteorological Administration.

    LAPS Observation Simulation System (OSS) – Evaluation of new observation technology or siting of existing observational systems. This system has been employed to assess the potential of new satellite systems for the Department of Defense (DOD), instrument placement around eastern and western space centers of the U.S. Air Force and the National Aeronautics and Space Administration (NASA), and spaceborne wind lidar systems for NOAA.

    Satellite Products – Utilization and evaluation of raw radiances and products derived from GOES atmospheric soundings, for the purpose of developing a complete national-scale moisture analysis useful for high-resolution model initialization. The branch also participates in the Joint Center for Satellite Data Assimilation.

    Weather Research and Forecasting (WRF) Model Support – Development of a Standard Initialization procedure for community use in initializing the WRF model with background fields obtained from other models and static data defining the surface properties. High-resolution local applications of WRF are currently being developed and tested, initially for evaluation during the IHOP field experiment in the Southern Plains.

    WFO-Advanced Support – Full support of an operational version of LAPS on the WFO-Advanced workstation, including both analysis and prediction. The WFO-Advanced forecaster workstation is used to demonstrate Advanced Weather Interactive Processing System (AWIPS) functions in support of future Weather Forecast Office (WFO) operations.

    Local Model Implementations and Demonstrations – Configuring and installing modeling systems that take advantage of local datasets, advancements in affordable parallel computing, and the results of weather modeling research and developments from FSL and elsewhere (Figure 11). Current and upcoming applications of various models on various computing platforms all take advantage of LAPS initialization. Ensembles of local models will be an increasingly useful approach to numerical weather forecasting problems and applications to a broad spectrum of uses ranging from fire weather prediction to ground transportation needs.

Figure 11 - D3D Display of MM5

Figure 11. A D3D display of output from the MM5 model on a domain centered over Colorado. The image shows temperatures (red is warm, blue is cool) with wind vectors at the surface and clouds in the upper atmosphere. The colored ribbons are wind tracers released from one location advected through time. The ribbons are colored according to height, with the lowest level ribbon colored red. Spaced at every 50 mb, the ribbons indicate the flow of the wind at various levels over time.

Research efforts in the Meteorological Applications Branch consist of the following:

    GAINS Project – Investigation of a concept for routinely sounding Earth's atmosphere over oceanic areas, as in the Global Air-Ocean IN-situ System (GAINS) program.

    Diagnostic Turbulence Forecasting – Development, testing, and verification of diagnostic tools using the RUC model for forecasting turbulence in support of the Federal Aviation Administration (FAA) Aviation Weather Research Program (AWRP).

    Research Quality Datasets – Production of quality-controlled hourly precipitation data and meteorological data from commercial aircraft (ACARS and AMDAR) for access on CD-ROMs and the Web. Assessments of and improvements to the set of hourly precipitation measurements are utilized for verification purposes by the Real-Time Verification System (RTVS).

    Mesoscale Diagnostic Studies – Research performed to increase the understanding of weather systems, improve conceptual and diagnostic models of the atmosphere using data from conventional instruments and new state-of-the-art sensors, and investigate the dynamic linkages between mesoscale processes and heavy precipitation in mesoscale systems, including potential vorticity streamers, the structure and dynamics of the low-level jet, and mesoscale gravity waves.

    Websites for FSL Data – Development of Websites for ACARS data, interactive soundings from the MAPS/RUC forecast model, national precipitation data, national mesonetwork data, the NOAA Chemical Weather Research and Development program, the FSL publications Webpage, and others.


Regional Analysis and Prediction Branch
Stanley G. Benjamin, Chief

Objectives

The primary focus of the Regional Analysis and Prediction (RAP) Branch is research for and development of the Rapid Update Cycle (RUC), which provides high frequency, hourly analyses of conventional and new data sources over the contiguous United States, and short range numerical forecasts in support of aviation and severe storm forecasting and other mesoscale forecast users. The RUC runs operationally at the National Centers for Environmental Prediction (NCEP) at the highest frequency among its suite of operational models. The branch works closely with NCEP in developing, implementing, and testing RUC improvements at FSL, and transferring them to NCEP. A variety of model and assimilation development, verification, and observational data investigation activities are carried out under the RUC focus. Applications of the RUC include the Short-Range Ensemble Forecast system being developed at NCEP, contributions to the GEWEX program, forecasting detailed wind fields in collaboration with the National Renewable Energy Laboratory, and support for a number of field experiments.The RUC has a unique role within the NWS in that it is the only operational system that provides updated national scale numerical analyses and forecasts more often than once every 6 hours. It was developed in response to the needs of the aviation community and other forecast users for high frequency, mesoscale analyses and short range forecasts covering the conterminous United States. It is widely used in NWS Forecast Offices, NWS centers for aviation weather and storm prediction, the FAA, and other facilities. Evaluations of the RUC have clearly demonstrated its advantage in providing high frequency, recently initialized forecasts based on the latest observations. The RUC is a key part of the FAA Aviation Weather Program, since commercial and general aviation are both critically dependent on accurate short range forecasts. The RUC will continue to improve over the next few years, perpetuating the successful collaboration between FSL and NCEP, but a gradual shift in primary focus will be to develop a rapid update component to the WRF model by 2006.

In collaboration with other government agencies (e.g., NCAR, NCEP, NESDIS) and universities (e.g., University of Miami, University of Oklahoma), RAP branch scientists develop improved data assimilation and modeling methods for use in the RUC. Techniques for assimilating new observational datasets are developed toward the goal of the best possible current estimate of current atmospheric and surface conditions, as well as the best possible short-range foercast. The branch also interacts with other FSL staff in implementing optimal computing methods with RUC software, making the model as efficient as possible on modern computing platforms.

A second primary focus of the branch is the development, real-time implementation, and evaluation of a fully coupled atmospheric/air chemistry mesoscale model prediction system. An MM5-based, fully coupled system has been run in support of the 2000 Houston experiment, and during the 2002 NOAA Temperature and Air Quality Pilot Project in New England. Regional air pollution studies are becoming an increasingly important focus of research in the branch.

A third thrust of research is the multiscale theory and model development. A well-posed, open boundary multiscale ocean model is being developed for the Office of Naval Research.

Accomplishments

Development and Implementation of the 20-kilometer RUC

NCEP Implementation – Culminating a four-year development and testing activity, a major revision to the RUC system including 20-km horizontal resolution was implemented at NCEP on 17 April 2002. Scientists in the RAP Branch completed final development activities and extensive testing of the 20-km RUC (RUC20) over the last 12 months. The higher horizontal resolution of the new version takes advantage of the improved computing capability at NCEP on its IBM SP computer. This new version has four key aspects: finer (20-km) horizontal and vertical (50 levels) resolution (requiring about 10 times the computations of the 40-km version for the forecast model), an improved version of the RUC forecast model, assimilation of GOES-based cloud-top pressure to improve the initial RUC cloud and precipitation fields for each forecast, and use of an improved version of the RUC optimal interpolation analysis.

Resolution and Domain – The increased horizontal resolution (20-km) provides considerable improvement in accounting for the effects of topography and land-surface variations on wind and precipitation. In addition to much improved orographic precipitation forecasts, the smaller grid volumes in the 20-km RUC allow improved depiction of cloud and more representation of mesoscale convective cloud/precipitation systems at the resolved grid scale. These smaller grid volumes also improve the ability of the RUC to resolve clouds and areas with supercooled liquid water with potential for icing. The 20-km resolution allows the RUC to better delineate areas with potential for turbulence, whether of clear air, mountain wave, or convective origin. The 20-km RUC uses 50 vertical levels, with 7 levels added in the upper troposphere and 3 in the lower troposphere. It uses the same hybrid isentropic/terrain-following coordinate used successfully in previous versions. In the 20-km version, the isentropic spacing is 2–3K for reference potential temperature from 270–355K. The top level is now at 500 K (approximately 40–60 hPa). The spacing near the surface is 2-, 5-, 8-, and 10-hPa in the first 4 layers, with an explicit model calculation level at 5 m above the surface.

Improvements to the 20-km RUC Forecast Model – The 20-km RUC forecast model has incorporated many improvements that, even without the change in horizontal resolution, result in better RUC forecasts:

  • Improved convective (subgrid-scale) precipitation from an ensemble closure/feedback convective parameterization by Grell and Devenyi. The Grell/Devenyi scheme currently uses 10 closure assumptions and 9 feedback assumptions, as implemented in the 20-km RUC model. This scheme also detrains cloud water and ice directly to the RUC cloud microphysics, a feedback absent in the 40-km RUC.
  • Revised version of explicit mixed-phase cloud microphysics used in the RUC and MM5 in collaboration with NCAR/RAP. The key changes in the RUC microphysics are improved representation of supercooled liquid water, reduced exaggerated amounts of ice/graupel, and improved forecast precipitation type at the surface.
  • Improvements to the land-surface/vegetation/snow model, including provision for frozen soil and a 2-layer representation of snow, and much more detailed land surface data. The previous land use and soil datasets used in the 40 km RUC were from 1-degree resolution data, whereas the 20-km datasets are aggregated from 30-second resolution data (Figure 12). The RUC land surface model has been tested extensively in long term one dimensional simulations, which show that the frozen soil and snow model changes will decrease surface temperature biases in transition seasons. The prescribed values for thermal conductivity are also changed, leading to a more accurate diurnal cycle for soil temperature.
Figure 12 - Land Use Data - L

Figure 12 - Land Use Data - R


Figure 12. Land-use data for a, top) RUC40 (1 degree latitude/longitude resolution) and for b, bottom) RUC20, from the U.S. Geological Survey (USGS) 24-class, 30-second dataset.

These changes in the 20-km RUC forecast model have led to some significant improvements in near-surface and precipitation forecasts. The 40-km RUC has had some underestimate of the diurnal temperature cycle, but this problem is much reduced in forecasts from the 20-km RUC due to the improved land surface model, improved cloud fields, and an improved diagnosis of 2-m temperature in the 20-km RUC. Precipitation predictions of heavier amounts are important for severe weather and hydrological forecasting, but have been underestimated in the 40-km RUC. The 20-km RUC also provides substantial improvement in warm-season and cold-season precipitation forecasts, with a much improved equitable threat score and bias score at all thresholds. An example of this improvement is presented in Figure 13a, b showing 12-hour forecasts of 3-hour accumulated precipitation from the RUC40 and RUC20 respectively, for comparison to a radar image in the verifying period in Figure 13c. In this case the RUC20 has accurately forecast much more intensity than the RUC40 along the southern end of a convective line. Not only is the intensity improved in the RUC20 forecast, but also the position of the line is more accurately forecast to be farther east than the RUC40 forecast.

Figure 13 - Precipitation Forecasts - L

Figure 13 - Precipitation Forecasts - R


Figure 13. Precipitation forecasts from RUC40 and RUC20 and verification. Forecasts are initialized at 0000 UTC 26 March 2002 and are predicted 3-hour accumulation (in) between 0900 – 1200 UTC. a, top) RUC40, b, middle) RUC20, and c, bottom) Unisys radar summary (observed) for 1115 UTC.

RUC forecasts of upper-level wind and temperature also are improved with the RUC20, as shown in Figure 14 (verification against rawinsonde observations for period in early 2002). This improvement is apparent for both 3-hour and 12-hour forecasts.

Figure 14 - Verifications - RUC 40 and RUC 20 - L

Figure 14 - Verifications - RUC 40 and RUC 20 - R


Figure 14. Verification of RUC40 and RUC20 3- and 12-hour forecasts against rawinsonde observations for a, top) wind, and b, bottom) temperature for
period 22 January – 8 February 2002.

Data Assimilation Changes – Two major changes in data assimilation for the 20-km RUC were introduction of an initial cloud analysis using GOES cloud top pressure to modify RUC 1-hour hydrometeor forecasts, and improvement of the optimal interpolation (OI) analysis procedure used to initialize the RUC model.

Development of a National-scale Moisture Analysis – The 20-km RUC includes a cloud analysis in which GOES (sounder) single field-of-view cloud top pressure data are used to clear and build clouds/hydrometeors using the previous 1-hour RUC 3D hydrometeor forecast as a background. This technique has been shown to improve short range cloud forecasts, even out to 12-hour duration. The introduction of the GOES cloud assimilation is new to the 20-km RUC. In cloud cleared areas, the water vapor mixing ratio is also decreased below saturation, along with setting hydrometeor mixing ratios to zero. If cloud building is required, either an ice or water cloud or both may be built, depending on temperature, in a manner consistent with the RUC/MM5 microphysics. Lower troposphere cloud top pressures are rederived from RUC 1-hour forecast temperature profiles. Procedures have been developed to identify possible subfield-of-view cloud being detected by the GOES data in order to prevent erroneous cloud-building.

Current research on national moisture analysis for the RUC is focused on introducing national radar reflectivity and lightning data to improve analyses in the presence of precipitation. National 2-km resolution reflectivity mosaic fields and 10-km Radar Coded Message fields are being inserted into the RUC20, since the former provides higher resolution but the latter provides beam blockage information to help differentiate "no echo" from "no coverage." Lightning data are used as an indication of convection. Initial experiments in a test real-time RUC cycle are giving promising results.

Improved Optimal Interpolation Analysis – The optimal interpolation (OI) analysis in the RUC20 is an improved version of that used in the RUC40. It continues to use a native isentropic sigma hybrid coordinate, thus preserving the advantage of confining the influence of in situ observations to within the air mass with similar isentropic properties. The improvements of the RUC20 optimal interpolation over the RUC40 version include improved preprocessing of observations, resulting in more accurate use of near-surface observations and precipitable water observations. Quality control of observations is also improved with the RUC20. Finally, it has been tested extensively with four new observation types at FSL and to be incorporated at NCEP later in 2003:

  • GPS ground-based precipitable water values (now over 100 in the United States)
  • 915-MHz boundary-layer profilers (about 25 in the RUC domain)
  • RASS temperature low-level virtual temperature profiles from selected 405-MHz and 915-MHz profilers
  • Mesonet surface observations.

Support of the Operational RUC at NCEP – During work to implement the new 20-km RUC system on NCEP's computing system, it has been necessary to develop expertise on their SP computing system and maintain a close, long-term collaboration with many groups in NCEP. FSL also monitored the current operational 40-km RUC and worked with NCEP to make necessary modifications. FSL supports a related major ongoing task, that of running in real time a backup version of the RUC in a "hardened" computer environment to assure high-level reliability. During NCEP outages, RUC grids from FSL are substituted through NWS distribution channels to support all real-time RUC users. This task involves both the RAP Branch and FSL's Information Technology Services, along with NCEP and other parts of the National Weather Service. A backup for the RUC40 has run through last year, and a backup for the RUC20 has been implemented. Ongoing enhancements continue on the RUC Website, http://ruc.noaa.gov, including products from the test version of the 20-km RUC, and the use of the 20-km RUC grids in the FSL interactive sounding program.

Applications of the RUC

Development of Improved Atmospheric/Land-Surface Coupled Model Capability and Production of Integrated Datasets for GEWEX/GCIP/GAPP – FSL continues to participate in the multiyear Global Energy and Water Cycle Experiment (GEWEX) and the GEWEX American Prediction Project (GAPP) by providing data from the RUC/MAPS model/assimilation system to the GCIP (GEWEX Continental-scale International Project) archive of gridded datasets. The goal of GCIP is improved understanding of the continental scale hydrological cycle components, and ultimately, improved climate prediction capability. Ongoing improvements to all aspects of RUC/MAPS, especially to its land surface component, contribute toward meeting this goal.

Recent development of the soil/snow/vegetation model incorporated into RUC20 includes refinements in the parameterization of processes in the frozen soil. It was found that the earlier approach to parameterize the phase changes in the frozen soil was not representing adequately the springtime situation when the soil is warming up and, simultaneously, moistening up from the melted snow. Under such conditions the volumetric content of the soil ice in reality is not increasing because the amount of new ice from incoming moisture is cancelled out by the amount of old ice melted due to the temperature increase. In this situation only the structure of the soil ice was changed, while in the model the soil ice was increasing, and the springtime thawing of soil was delayed. Necessary changes were made to improve the performance of the frozen soil algorithm under such conditions and tested in the one-dimensional multiyear simulations using the Valdai dataset (Figure 15).

Figure 15 - Skin Temperatures

Figure 15. Daily averaged skin temperature observed (dots) and simulated in RUC with the original version of the frozen soil physics (red line) and with the improved frozen soil physics (blue line) April and May 1980, Valdai, Russia.

Further development of the snow model was also performed. The improved two-layer snow model takes care of the problem for the situation when the snow layer is very shallow, which had caused excessive cooling at night under clear skies and low wind conditions. In this case, the shallow snow layer was combined with the top soil layer in the improved snow model, and the energy budget was then applied to this combined layer. This procedure increases the heat capacity of the layer and prevents excessive cooling. Also a simple dependence of the snow albedo on the snow depth was introduced for the established snow pack. The performance of the improved soil/snow model incorporated into the RUC20 showed better verification against the METAR data than the former model. An analysis of skin temperature differences demonstrates major changes in the forecast of skin temperature with the improved snow/soil model for the areas with shallow snow cover (Figure 16). The sampled point is more than 7oC warmer in the 21-hour forecast valid at 1200 UTC 5 March 2002 from the experimental run compared to the operational run, and the resulting forecast surface air temperatures were much more in agreement with observations (not shown).

Figure 16 - Skin Temperature  Differences

Figure 16. Skin temperature difference between 21-hour forecasts from RUC20 with improved snow/soil model (name extensions end in ".0") and operational RUC20 (name extensions end in ".1"). The forecasts are valid at 1200 UTC 5 March 2002. The values are sampled at the point with the shallow snow cover. (SOILT: skin temperature; TA: 2-m atmospheric temperature; SNHEI:
snow height, in meters.)

The RUC/MAPS system was the first regional model to cycle its multilevel soil moisture and soil temperature fields, and continues to be the only one to cycle snow water equivalent depth and multilevel snow temperature. Recent additions of frozen soil physics and a multilevel snow model to the RUC/MAPS land surface model provide improved seasonal transitions. The precipitation forecasts constrained by hourly observational data assimilation have been sufficiently accurate to allow an accurate evolution of these generally poorly observed fields. Paradoxically, although the RUC is used primarily as very short-range forecast guidance, the cycling of these surface fields requires considerable robustness from the land surface model for what is, in effect, a simulation spanning months to years. In turn, the cycling of soil moisture and snow water equivalent in the RUC leads to improved short-range forecasts.

Several modifications were made to the initialization of land-use properties in the MAPS/RUC with 20-km horizontal resolution. Soil and vegetation types initialized from high-resolution (1-km) datasets were further improved and smoothed to exclude singular points with particular vegetation and soil types and to represent more accurately the coastline. The dataset for initializing albedo in the model was also improved.

Collaborative work on intercomparison of land-surface schemes continued with the GCIP/GAPP community. A RUC-based Coupled Data Assimilation System (CDAS) utilizing optimal combination of observed and predicted precipitation and cloud fields is undergoing testing. The surface fields from the CDAS will be validated against available observations and compared to the LDAS (Land Data Assimilation System) surface fields depending only on observed precipitation. The fields from the CDAS are expected to show improvements over the LDAS fields in the areas with a sparse net of observations. Finally, the branch continues to evaluate the MAPS/RUC land-surface scheme in the Project for Intercomparison of Landsurface Parameterization Schemes (PILPS) and the Snow model Intercomparison Project (SNOWMIP).

Use of RUC Wind Forecasts for Estimated Wind Power Potential – FSL continues a collaborative wind energy study with the National Renewable Energy Laboratory (NREL, Department of Energy) using 20-km RUC/MAPS forecasts. Time-lagged ensembles produced from 20-km RUC forecasts out to 36 hours are used to estimate near-surface wind power potential, while variance among forecast ensemble members provides a measure of uncertainty. The high vertical resolution in the RUC near the surface and frequent update cycle makes it well suited to these types of studies.

Special RUC Forecasts for NOAA Pacific Landfalling Jets Experiment (PACJET) – FSL ran a special high-resolution version of the RUC model and distributed forecast fields to the NWS Western Region Headquarters for real-time AWIPS display at local offices in support of PACJET 2002. The special PACRUC configuration consisted of a 10-km grid covering all of the NWS Western Region, nested within the CONUS 20-km RUC domain. The 20-km RUC utilized a 1-hour assimilation cycle to ingest all conventional observations. PACRUC 24-hour, 10-km forecasts were produced every 6 hours, and AWIPS files containing selected surface fields were transferred to NWS via automated LDM scripts. Additional RUC fields were provided to the forecasters through a Webpage, and a Web-based forecast evaluation form was used to obtain forecaster feedback. The 10-km forecasts provided considerably enhanced mesoscale detail compared to the larger scale operational models. Figure 17 shows an example in which a 6-hour PACRUC forecast captured a band of weak convection over Southern California. As expected, the depiction of terrain-related temperature, wind, and precipitation features was quite good in the PACRUC forecasts, as illustrated in Figure 18 for mesoscale wind variations in a March 2002 case.

Figure 17 - AWIPS PACRUC 6 Hour Display

Figure 17. AWIPS display of 6-hour PACRUC predicted surface winds (cyan barbs) and convective precipitation (thick red contours) over southern California, valid 0000 UTC 29 March 2002. The model forecast is overlaid upon surface METAR observations and the observed radar composite indicating close agreement between the observed and model predicted surface fields.
Figure 18 - AWIPS PACRUC 9 Hour Display

Figure 18. AWIPS display of 9-hour PACRUC predicted surface winds (cyan barbs) and surface wind speed (color shading) over north-central Colorado, valid 1500 UTC 28 March 2002. The model forecast is overlaid upon surface METAR observations illustrating the close agreement between the observed and model predicted surface wind field.

During PACJET 2002 the conventional observations were supplemented with experimental satellite-derived winds obtained as part of the GOES rapid-scan Winds Experiment (GWINDEX). The real-time PACRUC forecasts were complemented by a retrospective data impact test designed to quantify the impact on forecast skill from assimilating the rapid-scan satellite wind observations obtained in GWINDEX. Initial results from two 3-day periods showed that inclusion of the rapid-scan satellite wind observations improved short-range (3–12-hour) wind predictions by up to 10%.

Following the completion of the PACJET project, the RUC 10-km nest was relocated to the south-central United States to support field operations during the International H20 Experiment (IHOP). FSL provided real-time RUC and WRF model guidance to forecasters at the IHOP operations center and the Storm Prediction Center.

Observation Sensitivity Experiments Using RUC to Examine the Impact of GPS Precipitable Water Observations – In collaboration with the Demonstration Division, 60-km RUC parallel cycle experiments with and without assimilation of GPS precipitable water observations have continued. Positive impact (leading to more accurate forecasts) of GPS precipitable water observations on short-range relative humidity forecasts seen in earlier sensitivity experiments has continued to increase as more stations have been added to the network. Results from a new set of 2-week retrospective experiments for both cold and warm season periods with the 20-km version of the RUC show still greater impact from the addition of GPS precipitable water observations than when using the 60-km RUC model. A similar set of experiments to assess the impact of wind profiler data on RUC forecasts shows considerable improvement in short-range forecasts when profiler data are assimilated.

Collaborative Modeling Projects

Air Chemistry Experiments and Real-Time Forecasts from a Coupled Weather/Air Chemistry Prediction Model – Real-time forecasts were produced from the NOAA FSL air quality modeling system twice daily for 48-hour runs from 8 July–31 October 2001. This modeling system is based on a nonhydrostatic meteorological model (MM5) that was coupled online with the RADM2 chemical mechanism. Biogenic emissions, deposition, tracer transport by convection and turbulence, photolysis, and transport by advection are all treated simultaneously with the meteorology ("online"). These runs were initialized with RUC20 analyses over a 12-hour preforecast period to improve the chemical initial field. Runs were made with 27-km horizontal resolution over a domain that covered the eastern and central United States (3600 km x 2970 km). In addition, some very high-resolution (grid length of 3 km and 1 km) runs have been performed in the New Hampshire area to prepare for high-resolution forecast runs performed in real time during 2002 over New England in support of a field experiment. The various domains that were run are shown in Figure 19.
Figure 19 - NOAA Air Quality Domains

Figure 19. Domains over which the NOAA FSL air quality modeling system
was run.

Output from the large domain of the model was produced hourly and included three-dimensional meteorological fields as well as 41 chemical species in three dimensions. The Website http://www-frd.fsl.noaa.gov/aq/ was designed to display model forecasts. Figure 20 shows an example of a forecast for the large domain displaying near-surface wind fields and carbon monoxide (CO) concentrations (a) as well as ozone concentrations (b). Near the East Coast, in the predominant southwesterly flow, polluted air from the large metropolitan areas is transported northeastward along the coast. Slight sea-breeze effects can be seen even on this relatively coarse grid resolution, with the wind pushing the polluted air inland in a plume from the Hudson Valley into southern Maine.

Figure 20 - Air Quality Prediction - L

Figure 20 - Air Quality Prediction - R


Figure 20. Air quality prediction for 23 July 2001. Near surface wind fields are shown from the 27-km mesh model (barbs), a, top) carbon monoxide (CO) concentrations, and b, bottom) ozone (O3) concentrations (ppm) in color, and terrain elevations with white contours.

For the high-resolution runs, however, much more of the local detail is resolved, resulting in a more realistic and detailed wind flow. Figure 21 shows a comparison of CO concentrations at different times in the morning of 23 July. Figure 21a is a snapshot of the fields before the sea breeze sets in; note that the polluted air from the Boston area is transported out over the ocean. In Figure 21b, the sea breeze has started to develop, and the polluted air is being brought back on shore over the New Hampshire and Maine area. Also note the up-valley transport of carbon monoxide along the Connecticut River Valley.

Figure 21 - Air Quality Prediction

Figure 21. High-resolution (3-km grid) air quality prediction for New Hampshire. CO concentrations are displayed in color, wind fields (barbs on every second grid point), and terrain elevations (white contours) before the onset of the sea breeze (a, left), and just after the onset (b, right).

In collaboration with the Aeronomy Laboratory, FSL also started preparation of datasets for retroruns, which allows rerunning the model (for example, to test sensitivity to changes or development work) over a 2-month summer period, including model evaluation and intercomparison. Figure 22 shows an example of a model comparison with observations for one particular model run and one observing station. Development of capabilities to rerun the model over an extended and extensively tested time period is a powerful tool for further model development.

Figure 22 - Initial Comparison - Ozone

Figure 22 - Initial Comparison - Carbon Monoxide


Figure 22. Initial comparison of predicted versus observed ozone (top two plots) and carbon monoxide (bottom two plots) concentration for one surface observing station (Thompson Farm in New Hampshire). Compared with observations are 0 – 12-hourly (blue), 12 – 24-hourly (green), and 24 – 36-hourly (red) model predictions. Scatterplots (top and third from top figures) and time series (second from top and bottom figures) are shown. This evaluation was done by Wayne Angevine of the NOAA Aeronomy Laboratory.

Joint Center for Satellite Data Assimilation (JCSDA) activities – To more efficiently transfer new research findings and satellite assimilation techniques to operations, the Joint Center for Satellite Data Assimilation was established. In cooperation with NESDIS, NCEP, ETL, and NASA, FSL has made progress in comparing GOES-8 sounder channel radiances to computed radiances using a forward model applied to RUC20 temperature and humidity profiles. Similar comparisons have been performed using imager channel radiances from the GOES-8. In particular, the 11-micron "window channel" and the two longwave infrared channels are used to estimate cloud-top pressure for comparison with the NESDIS cloud-top product used operationally in the RUC20. This approach is designed to take full advantage of the raw radiances. The results are being verified against vertically pointing cloud-profiling radar at the ARM/CART site in northern Oklahoma.

Participation in development of the Weather Research and Forecast (WRF) Model System – The overall goal of the WRF model project is to develop a next-generation mesoscale forecast model and assimilation system that will advance both the understanding and prediction of important mesoscale weather, and promote closer ties between the research and operational forecasting communities. The model and associated system are being developed as a collaborative effort among NCAR, NCEP, FSL, the Center for the Analysis and Prediction of Storms (CAPS), and other research institutions, together with the participation of a number of university scientists. In collaboration with NCAR, FSL has worked on the development of physics packages and the development of a three-dimensional variational (3DVAR) analysis. FSL has also developed the standard initialization package for the Weather Research and Forecasting (WRF) model (discussed below), and has worked with the University of Miami on development of a quasi-isentropic variant of the WRF nonhydrostatic model. FSL will adapt WRF assimilation and model systems over the next several years to include an advanced rapid update capability.

The RUC land-surface model (LSM) code was rewritten and significantly optimized to satisfy the programming requirements for WRF. The new Fortran-90 module of the RUC LSM can be easily coupled to the atmospheric schemes and was incorporated into the test version of WRF. The implementation of the MAPS/RUC LSM into the official version of WRF is planned for July and August 2002.

GPS Slant Water Vapor Analysis experiments – Ground-based GPS signal delay is caused by the refractivity of water vapor and the atmosphere. Results from an Observing System Simulation showed that it is possible to recover three-dimensional water vapor fields from a high-resolution network of GPS receivers. A new 3DVAR system and multigrid techniques were developed to assimilate the refractivity from slant water vapor measurements (as well as other types of observations) and tested this package using a set of analytic solutions. The Quasi-Nonhyrdostatic Model (QNH) and the MM5 model were used for the nature runs. The GPS tomography procedure to derive slant water vapor was published in Monthly Weather Review.

Numerical Experiments on 3DVAR and 4DVAR – There are generally two ways to incorporate multivariate correlations into a 3DVAR data assimilation system. One can use a form of linear balance to correlate background errors in the mass field with errors in the wind field. Alternatively, only the spatial correlations in the background error-covariance matrix can be specified using a recursive or digital filter, and then the mass and wind fields can be penalized so that they are balanced under a weak constraint formulation. Application of Fourier analysis to 3DVAR systems shows that the penalization of geostrophic imbalance can be used for mesoscale data assimilation by significantly constraining the large-scale flow without damaging the mesoscale features. A 4DVAR system for MAPS/RUC40 is under development.

Multiscale Theory and Model Development – Traditionally it has been assumed that the majority of the atmospheric energy is contained in balanced large-scale motions. Under the hypothesis that the period of time before the energy in small-scale storms has a significant impact on the total atmospheric energy is longer than the timescale for balanced large-scale atmospheric motions, the forced hyperbolic system that the latter flow should satisfy was derived and published in the Journal of the Atmospheric Sciences. As a first step in ascertaining the appropriate large-scale forcing that should be included in the hyperbolic system, the relative contributions of the various physical parameterizations included in a typical operational weather prediction model to the large-scale forecast accuracy have been determined. When a simple boundary layer drag is used as the sole physical parameterization, the large-scale forecast accuracy is comparable to that of the operational model for several days. In order to determine the most efficient method to obtain initial conditions for the hyperbolic system from the current sparse observational network, the relative contributions of various observational data sources to the operational data assimilation process have also been determined. In agreement with the theory of the periodic updating of hyperbolic systems with multiple timescales, the periodic insertion of wind data alone provides mid-latitude initial conditions for the winds essentially identical to those of the operational assimilation system.

Projections

The Regional Analysis and Prediction Branch will continue to work with scientists at NCEP, NCAR, and other organizations to improve the RUC and WRF models over the next few years. The primary near term tasks follow.

Implementation of Three-Dimensional Variational Analysis (3DVAR) in the 20-kilometer RUC – With tuning to ensure sufficient accuracy for short-range wind forecasts, the RUC 3DVAR analysis will be added to the 20-km RUC running at NCEP. The 3DVAR implementation in the recently implemented RUC20 model was deferred for additional testing, but its addition to the RUC20 and replacement of the current optimal interpolation analysis will provide smoother analyses, more accurate forecasts, and a framework for assimilation of radial wind observations from radar in the future. Development and real time testing will continue on the 3DVAR analysis, both for incorporation into the RUC and toward the development of the 3DVAR analysis for the WRF model.

Continued Development of a National-Scale Cloud/Hydrometeor Analysis – Development and real time testing will continue for further improvements to the RUC national scale cloud analysis, with the addition of radar, lightning, and surface observations to satellite cloud top data. Experiments will be carried out testing assimilation of a GOES imager-based multilevel cloud product, as described below under the JCSDA plans.

Refinement and Testing of Improved Physical Parameterizations for Soil/Vegetation Processes, Turbulence, Convective Clouds, and Cloud Microphysics – Some of this work will be done in collaboration with NCAR, since the RUC model uses some of the MM5 parameterizations, which will be options for the WRF model.

National Observing System – FSL has formed a team working on an initiative to develop a national mesoscale observing system consisting of tropospheric and boundary layer profilers, ground GPS receivers and radiosondes with ground tracking systems. This is an initiative with great potential impact for mesoscale forecasting. An observation simulation system experiment (OSSE) with a practical observation network design and numerical model to verify the budgets and applicability is under development.

Data Assimilation – Work will commence to modify the WRF 3DVAR system to use velocity or vorticity fields as control variables instead of streamfunction and velocity potential.

High-Resolution Experiments Using RUC for the IHOP and New England Temperature Air Quality Experiments – RUC forecasts will be made at 10-km resolution in support of these upcoming experiments. In addition, the WRF model will be run over the same domains, also at 10-km resolution, initialized from the RUC. This will allow initial intercomparisons between the current RUC hydrostatic model and the WRF non-hydrostatic model including RUC physical parameterizations, the RUC land-surface model, and the Grell/Devenyi ensemble-closure cumulus parameterization. Objective verification of the model forecasts will be performed as part of these studies.

Participation in GEWEX – Collaboration will continue on the GEWEX/GAPP program, with the focus on development of a coupled atmospheric/land-surface assimilation system that uses an optimal combination of radar and satellite observations to modify clouds and precipitation along with model forecasts in regions where observations are unavailable.

Contribution of RUC 63-Hour Forecasts to an NCEP Short-Range Ensemble Forecast – As part of an expansion to NCEP's Short-Range Ensemble Forecast (SREF) project, FSL will complete its effort begun this last year to set up an ensemble version of the RUC model running out to 63 hours on a 48-km grid over the Eta domain. The RUC SREF is spawned from a set of five members bred from the NCEP Eta model, and is a candidate for inclusion in the NCEP SREF set currently composed only of Eta and Regional Spectral Model bred members. Tentative results from the RUC SREF show substantial spread in the ensemble, but it needs to be determined whether the forecast skill for the ensemble mean exceeds that of the operational RUC20 model, and statistical techniques must be developed to assess the results.

Joint Center for Satellite Data Assimilation (JCSDA) Activities – Future work will first involve running and testing the OPTRAN radiative transfer model to replace the ECMWF RTTOV code that has been used in all RUC forward model calculations thus far. OPTRAN has been chosen as the community radiative transfer model by the JCSDA. Outgoing radiances from the RUC will then be subjected to OPTRAN forward model calculations to compare with the GOES radiances. The imager data will be used to determine clear-air radiances with greater resolution than using the sounder estimates. Eventually, the goal is to incorporate the adjoint of the forward calculations into the RUC 3DVAR and to begin using this to rapidly update the radiance data in the RUC and, later, the WRF models.

Multiscale Theory and Model Development – Theoretical work on a remaining issue in multiscale atmospheric dynamics will be conducted. The feasibility of accurate and stable open boundary conditions for a model based on the well-posed reduced system for oceanography will be completed.

Return to top of Forecast Research Division Section


Local Analysis and Prediction Branch
John A. McGinley, Chief

Objectives

The Local Analysis and Prediction (LAP) Branch responds to the needs of many government agencies and the private sector in the areas of local and mesoscale data analysis, data fusion, data assimilation, quality control, three-dimensional display and visualization, and numerical modeling.

The branch is charged with the research and development of the Local Analysis and Prediction System (LAPS) and the implementation of mesoscale forecast models. The primary objective is to provide real-time, three-dimensional, local-scale analyses and short-range forecasts (0 – 24 hours) for operational weather offices or facilities. Activities cover four broad areas:

    Data Acquisition – Includes identifying, collecting, and quality-controlling any kind of atmospheric or earth surface measurement, such as those provided by satellites, radars, mesonets, aircraft, GPS, balloons, and profilers. This activity also includes developing interfaces to "national" datasets, such as the gridded data services provided via the Satellite Broadcast Network (SBN) data feed and similar military systems. LAPS is coupled with the Local Data Acquisition and Dissemination (LDAD) system, which stores portable applications that retrieve and render AWIPS weather data into images and graphical displays for dissemination.

    Data Analysis – Accomplished using an integrated software package containing well-documented objective analysis schemes that spatially represent atmospheric conditions, perform spectral filtering, and ensure vertical consistency. This analysis capability has been demonstrated in National Weather Service forecast offices as part of the AWIPS Application Software Suite, and in global windows to support Department of Defense operations. A prototype system has been developed in conjunction with Lockheed and Raytheon Systems Corporation for installation at the space launch centers at Vandenberg AFB and Cape Canaveral.

    Mesoscale Model Implementation – Accomplished using an expanding variety of mature nonhydrostatic modeling systems, such as the Regional Atmospheric Modeling System (RAMS) developed at Colorado State University, MM5 (developed jointly by NCAR and Pennsylvania State University), the hydrostatic version of Eta developed at NCEP, and the most recent addition, the WRF model under joint development by FSL, NCAR, and NCEP. These models have been configured to be initialized by LAPS analyses and with time-dependent boundary conditions furnished by all operationally available gridded datasets (RUC, Eta, Aviation, and the U.S. Navy Operational Global Atmospheric Prediction System). Implementation of the LAPS system at some NWS forecast offices has demonstrated the portability and effectiveness of running models locally. The collocation of FSL with the Denver-Boulder National Weather Service Forecast Office has demonstrated the effectiveness of locally run models throughout the past few years, as LAPS-initialized mesoscale models have been run on their local AWIPS hardware and on FSL's High-Performance Computing System for operational evaluation.

    Dissemination – Includes delivery of weather products and basic fields developed from LAPS to users in operational forecast offices and state and local government agencies, including emergency managers and other users specializing in fields such as winter highways operations, fire weather, aviation and space operations, and military operations such as those mentioned later.

Accomplishments

Basic Analysis System Development

Three-Dimensional Variational Methods – LAPS has made progress in applying variational methods in its analysis. The variational approach to the LAPS moisture analysis remains the method of choice to integrate GOES sounder radiances, GOES derived products, GPS, boundary layer moisture, cloud information, and new for this year, radiosonde and profiler data. In LAPS the variational step was previously used only with GOES sounder radiances. Other moisture variables were analyzed separately and either merged with the variational result or with the background field prior to the variational step. Recently, the variational adjustment using GOES radiances was expanded to include GOES three-layer precipitable water vapor, GPS total column water vapor, and cloud information from the LAPS cloud analysis in one variational formulism.

An ongoing data denial experiment begun late last year provides insight into the impacts that the various data sources have in the analysis. In addition, this assessment tool can be used to gauge the strength and weakness of the different data sources, in order to optimize their respective weights in the variational equation. The statistics used in this study are a comparison of analysis output to radiosonde data (taken as referenced truth). This is possible since radiosonde data are not typically used in the operational LAPS system due to their latency. The real-time approach to data impact lends itself well to analysis adjustment since it is not prone to "case study bias," and is truly a real-time picture of performance.

Water in all Phases (WIAP) Analysis – A goal for LAPS is to provide a complete national-scale product that describes the atmospheric water distribution from vapor to cloud droplets to precipitation, both liquid and frozen. The Water in all Phases (WIAP) analysis will improve model initialization, the goal of which is 15-minute, 5-km resolution analyses of water species in all phases over the continental United States. This analysis aims to utilize all conventional data, along with satellite, radar, and GPS data. The routine is based on the LAPS cloud analysis, but then seeks to quantify all water substance. This year the analysis is moving toward the goal of producing a reliable 10-km version that will become the showcase of the LAPS analysis system. A secondary function will be to use this larger-scale analysis to replace the traditional Regional Operating Cooperative (ROC) domain (centered over Colorado) that has been the mainstay of our development system. The advantage of this change is to enhance heterogeneous terrain and surface type in analysis development. An example of the WIAP moisture analysis running at 10 km over a CONUS domain is shown in Figure 23. Potentially the data can be subdivided and extracted for supporting the initialization of a local-scale model set up anywhere in the interior of this larger domain. Conceivably this product could be offered via the web to supply any local-scale model operation data that is required for initialization, thus saving the modeler that effort.

Figure 23 - Water in all Phases

Figure 23. An example of the Water in all Phases (WIAP) moisture analysis running at 10 km over a CONUS domain 1300 UTC 6 April 2002. (LAPS total precipitable water in centimeters.)

LAPS quality control using the Kalman Filtering Scheme – Quality control of observations is a continuing focus of LAPS analysis development. The Kalman filtering scheme allows users to optimally exploit local model output and past station trends and buddy trends to produce check values for surface stations. In conjunction with LAPS support of a three-dimensional 30-minute analysis cycle, the Kalman scheme can provide a continuously updated and accurate set of observations at times when data density is poor. This is an appropriate approach for instances when a user requires good product time continuity but has high variability in observation count from cycle to cycle.

LAPS "Hot Start" Procedure – The LAP Branch continued to improve upon a procedure, "Hot Start," for diabatic initialization of mesoscale models. The Hot Start initialization scheme is designed to develop initial conditions for mesoscale models such as the MM5, RAMS, and (soon) the mass-coordinate version of the Weather and Research Forecast (WRF) model. This scheme is unique in that it runs on small PC clusters with Linux operating systems and is ideal for applications in local weather offices where accurate short-term cloud and precipitation forecasts are needed. This system depends greatly on the accuracy of the background modeling system, currently the NCEP versions of the RUC and Eta models. The Hot Start scheme uses estimates of vertical motion and cloud water and ice mixing ratios from the LAPS cloud analysis. A variational analysis that applies both mass continuity and mass-momentum balance makes small adjustments to the wind and temperature field to accommodate and sustain the clouds in the first few time steps of the model integration. Recent improvements to the cloud retrieval algorithm have allowed inclusion of a broader range of microphysical species, better estimates of cloud vertical motion, and saturation of the cloud environment.

Verification with MM5 continues to show that the Hot Start outperforms other initialization techniques in the 0 – 9-hour time frame in forecasts of surface station precipitation, most state variables, and 3-D cloud and radar fields. It is used to initialize MM5 over a variety of domains, including the local Denver forecast area, where forecasters use it as an operational tool; the U.S. Air Force Western Range at Vandenberg Air Force Base as part of the Range Standardization and Automation (RSA) implementation discussed below; and IHOP field operational areas at 12- and 4-km grid resolution. The Hot Start system will soon be coupled to the WRF-mass coordinate model for testing in a local weather service office in conjunction with the Coastal Storms Initiative carried out with the National Weather Service and National Ocean Service. It is also being coupled to the MM5 modeling system at the Central Weather Bureau of Taiwan.

GOES Improved Measurements and Product Assurance Plan – The GOES Improved Measurements and Product Assurance Plan (GIMPAP) project was initiated a few years ago as part of the LAPS moisture algorithm development for integrating the high spatial structure of GOES imagery and sounder data into the LAPS system. GIMPAP has been expanded to include NESDIS products of cloud-top pressure and layer-precipitable water. These activities have meshed with the IHOP objectives of supporting the assimilation of these products in real time. This has prompted focus on data latency — a critical hurdle that has previously prevented the inclusion of satellite product data of such short-term analysis systems (30-minute and less cycle time). The development of the data processing strategies to reduce latency from 1 – 2 hours to under 30 minutes (75% reduction) has been a major success. The shorter latency permits assimilation into small-scale analyses with fast cycle updates while causing no negative impact to model initialization application.

An example of the total precipitable water analysis over the IHOP 4-km inner nested domain as generated by the LAPS analysis system is presented in Figure 24. This analysis incorporates real-time NESDIS product data with greatly reduced latency and GOES radiance data from the FSL satellite acquisition system. It also uses indirect NESDIS cloud-top product data via the LAPS cloud analysis.

Figure 24 - Total Precip. Water - IHOP

Figure 24. An example of the total precipitable water analysis over the
IHOP 4-km inner nested domain as generated by the LAPS analysis system
for 1330 UTC 6 April 2002.

Joint Center for Satellite Data Assimilation (JCSDA) activities – The latest OPTRAN code was obtained from NCEP and compiled and tested. Immediate plans are to get this code to run under IBM AIX, Linux Dec Alpha and Linux PC, and Sun OS, and then to write the interface between OPTRAN and LAPS.

Applications of LAPS

LAPS in AWIPS – The LAPS package has long been an integral element of the WFO-Advanced workstation, running as an application within AWIPS to produce a variety of gridded fields that may be combined with satellite imagery and radar on state- and local-scale displays. The LAPS in AWIPS serves the LDAD system operating outside the AWIPS network. The independent LAPS quality control system will supplement the LDAD quality control system to ensure that local data are monitored and properly employed. In preparation for AWIPS 5.0, running at all Weather Forecast Offices, staff developed the capability to create and install a grid domain centered on each forecast office, and to automatically execute LAPS functions using the local datasets available at each location.

The WFO-Advanced workstation in Boulder receives 10-km resolution MM5 model output from twice daily model runs on exactly the same grid and projection as the LAPS analysis. This permits the display of mesoscale model output in a fully integrated fashion, along with radar, satellite, and surface data. Forecasters can check the quality of a model run by directly comparing model output with observations. The model is running experimentally in the Denver National Weather Service Forecast Office on a system within the AWIPS application hardware suite. The model runs in automated mode with little intervention required, and is diabatically initialized using the Hot Start procedure discussed above.

The Range Standardization and Automation (RSA) Project – Much progress was made last year on a project involving collaboration with Lockheed Martin Mission Systems (LMMS) to install an integrated local data assimilation and forecasting system at the two U.S. Space Launch facilities, located at Vandenberg Air Force Base, California, and Cape Canaveral Air Station, Florida. Upon FSL's recommendation, LMMS purchased two Linux "Beowulf" clusters from IBM and installed one at FSL for use in system development and one at Vandenberg AFB as the first operational configuration. The clusters consist of 8 dual-processor Pentium III nodes and 1 dual-processor front-end node, totalling 18 processors. A Myrinet interconnect is used for high-speed message passing between nodes. Late last year, FSL delivered and installed (at Vandenberg AFB) the first version of the RSA Data Assimilation and Forecast system, based on LAPS coupled with the NCAR fifth-generation Mesoscale Model (MM5). The delivered system produces hourly analyses and a new forecast run every 6 hours on a triple-nested domain with 10-km, 3.3-km, and 1.1-km grid spacing, respectively. These analyses make use of the AWIPS Local Data Acquisition and Dissemination (LDAD) interface to incorporate data sources unique to the launch facilities in addition to the radar, satellite, and other datasets available via the AWIPS data feed. Every 6 hours, these analyses are used to diabatically initialize an MM5 forecast run. The forecast model outputs hourly forecast fields out to 24, 12, and 9 hours for each of the 3 grids, respectively, using 2-way nested feedback. The entire system is integrated with the Linux version of AWIPS installed at the Air Force ranges.

These RSA capabilities represent the first operational installation of a local modeling system completely integrated with AWIPS in a WFO-like environment, the first operational installation of the LAPS diabatic initialization–Hot Start method, and the first operational use of a Linux-based AWIPS system. FSL will continue work on ingesting and optimizing the use of all local meteorological datasets, and incorporating new capabilities, such as online verification of the forecast grids, improved cloud analyses through enhanced utilization of the satellite data, and improvements to the LAPS diabatic initialization method.

High-Performance Computing – The LAP Branch was one of the first group users of the Jet supercomputer when it became available for use in the spring of 2000. Since then, Jet has been a critical resource for all of the branch's numerical modeling activity as well as the national-scale Water In All Phases (WIAP) analysis project. This experience provided important feedback to the system developers and maintainers regarding configuration issues and future upgrade plans for the FSL High-Performance Computing System. Also, benchmarks developed through the use of this system were used to design smaller Linux PC cluster systems that will run the LAPS analyses and model forecasts for the RSA system and other projects.

Collaboration with the Taiwan Central Weather Bureau – Scientists continue an active collaboration with the Central Weather Bureau of Taiwan. The branch has hosted long-term visitors from Taiwan, forming working relationships that are very beneficial in improving the real-time data preprocessing for LAPS, the analyses, and the modeling components. Software has been added for processing more types of surface and upper-air data, including rawinsonde, mesonet, cloud drift wind, METAR, and synoptic observations. LAPS now runs with data ingest from a CWB model background, several different types of surface stations, and a narrow-band radar (at Wu-fen-shan, Taiwan).

The CWB-LAPS configuration feeds into the Central Weather Bureau's local WINS (Weather Information and Nowcasting System)/AWIPS workstation. Reliability of the onsite and shadow (running at FSL) systems were improved with incremental changes to system/scripting software. Additional improvements include ingest of wide-band Doppler radar, satellite imagery, rawinsonde, cloud-drift wind, and ACARS data into the Taiwan LAPS analyses. ACARS wind and temperature observations are now used. Wide-band reflectivity and velocity data from the Wu-fen-Shan radar are undergoing testing, and some remapping issues need to be resolved before operational use. The surface analyses have been upgraded to run more robustly with different data densities and clustering, resulting in improvements in the use of CWB surface pressure observations in the mean-sea-level-pressure analysis. The interface of the Taiwan "NF" model backgrounds into LAPS has been improved, and work continues on developing interfaces to CWB radar and GMS satellite data. Satellite imagery is being preprocessed for infrared and visible bands, with infrared being tested in the cloud analysis on the FSL shadow run. Satellite cloud-drift winds are used in test mode, though timing considerations dictate that they only be used in a displaced real-time mode.

Branch staff visited with Central Weather Bureau forecasters and researchers in Taipei to conduct training on the potential use of LAPS in operational forecasting. A Webpage (http://laps.fsl.noaa.gov/szoke/taiwan/taiwan_lapstrainingpage.html) was created to display the training sessions online and provide a subjective evaluation of the LAPS performance at the Taiwan CWB.

Collaborative Modeling Projects

Ensemble Modeling of Winter Road Conditions – In collaboration with NCAR, LAP Branch scientists began designing an ensemble of mesoscale models to support a Federal Highways Administration (FHWA) road weather project. The ensemble will include multiple models (MM5, RAMS, and WRF) with lateral boundaries provided by multiple large-scale models (AVN, Eta, and RUC) that will run at relatively high spatial resolution. Although ensemble techniques have been applied before on grids with approximately 25-km resolution, for this experiment will run on 10-km grids and make site-specific (road) probabilistic forecasts.

Developments for the Weather Research and Forecast Model – The branch is involved in three key areas of the WRF modeling system. The Standard Initialization (SI) software creates model start-up grids from the NCEP national AVN or Eta models. The land-surface module of WRF now uses "static" fields (such as vegetation greenness, albedo, land use, terrain height, land/water fraction) which were assembled and reformatted, along with efficient interface software. An example of land-use data mapped to a 10-km CONUS domain is presented in Figure 25. Substantial progress was made on development of a graphical user interface (Figure 26) to simplify use of the model (setting up model domains, grid resolution, parameterization selections, etc.) and reduce user error. These software components–developed by the LAP Branch and sponsored jointly by the Air Force Weather Agency, FHWA, and FAA–are released routinely with the WRF model itself.

Figure 25 - Land Use Data

Figure 25. An example of land-use data mapped to a 10-km CONUS domain on the WRF model.
Figure 26 - Land Use Data

Figure 26. WRF model graphical user interface, designed to simplify use of the model and reduce user error.

Lidar OSSE Studies – If deployment of a new system for atmospheric measurements is costly, it is prudent to simulate the new observing system in advance and determine whether the additional observations will lead to better weather forecasts. This is typically done using the observing systems simulation experiment (OSSE). The data must be simulated to represent what the observational system would sample. Those involved in numerical weather prediction have long argued that wind observations at multiple levels over the oceans would lead to better forecasts downstream over the continents. A strong interest has evolved over the last 20 years in the possibility of inferring atmospheric winds from Doppler lidar measurements aboard a polar orbiting satellite. Two satellites in appropriate orbits could provide daily global wind coverage, at least where the pulses of energy from the lidar are not blocked by clouds.

Cost is not the only issue with this observing system. The technology for obtaining radial wind velocities with good signal-to-noise is still being perfected, and more than one lidar has been proposed to do the job. For this reason, FSL has been cooperating with the Environmental Technology Laboratory (ETL), NCEP, and NCAR to study the impact of Doppler wind lidar data on numerical models. NCEP has been responsible for the global OSSE to estimate the potential impact of the windfinding lidar on global model forecasts. FSL is responsible for the impact of the data on forecasts over the U.S. using a regional OSSE, which to our knowledge has not been attempted before. For the case study, a long-duration, pure forecast from February 1993 was generated using the global model developed at the European Center for Medium Range Weather Forecasts (ECMWF). This basic atmospheric state, the so-called "Nature run," was assumed to be without error at observation points. Error characteristics in this study were based on years of experience from observation quality control procedures. In a post-processing procedure, the error model (interpolator) converts the Nature run values to true observations by applying gross and Gaussian errors.

FSL has provided output from a Regional Nature Run (RNR) using the PSU/NCAR mesoscale model (MM5) initialized from the ECMWF Nature Run to simulate onboard lidar data (clouds, winds, vertical velocity). The huge MM5 domain of the RNR covers most of North America with a 10-km horizontal (740 x 520) grid spacing and 43 vertical levels, a computational requirement that was handily met by the FSL super-Linux cluster "Jet." This domain was tailored to include the present domain of the RUC model, chosen to assimilate the extracted observations from the RNR over an 11-day period.

Scientists completed the RNR and wrote the software for extracting a suite of simulated observations from the RNR for inclusion into the RUC data assimilation system. Because the purpose of the OSSE is to investigate how the new observations interact with those already in the operational mix, it is essential to simulate the entire suite of operational and future lidar observations; thus, synthesized rawinsonde, ACARS/MDCRS, surface, METAR, wind profiler, and VAD winds must all be extracted from the RNR fields with appropriate error characteristics. The lidar line-of-sight winds from the satellite positions in four-dimensional space were determined by ETL from the Cartesian winds and cloud hydrometeor fields provided from the RNR. The last remaining task is to conduct a variety of sensitivity experiments with the mesoscale data assimilation system (RUC) with and without the lidar data, with various combinations of boundary conditions, and with and without liquid and/or ice clouds present. We expect that the impact of lidar observations on short-range forecasts will be difficult to discern over the data-rich U.S.

Projections

During 2002, the Local Analysis and Prediction Branch plans to:
  • Continue to support LAPS in AWIPS, interacting with the AWIPS contractor, Litton PRC Inc., and the NWS to achieve this goal. LAPS will support Alaska, Hawaii, and Puerto Rico by ingesting AVN model grids as background information.
  • Continue the cooperative effort with Raytheon Inc. and Lockheed in developing the RSA weather support systems for the Space Flight Centers at Cape Kennedy and Vandenberg Air Force Base.
  • Support the International H2O Project (IHOP) via high-resolution forecast model runs, field operations services, and postexercise data processing and research.
  • Demonstrate LAPS capabilities on the new high-performance multiprocessor, continue investigating hot start techniques, and perform an assessment on the use of 3DVAR and adjoints for local analysis.
  • Continue development of the multimodel ensemble using the three models being run: WRF, RAMS, and MM5; determine the optimum configuration for best forecasts and user-friendly products.
  • Complete the WRF Standard Initialization and contribute to 3DVAR/4DVAR development.
  • Adapt the new OPTRAN radiative transfer code to run with the LAPS moisture and temperature profiles and compare to raw GOES radiances. Improved understanding of the background error characteristics will be obtained from comparing GOES sounder and imager radiances with the special moisture measurement systems deployed in the IHOP experiment. Staff will also continue exploring the use of the 3.9-micron imager channel data to determine cloud phase and help to infer characteristics of low clouds.

Return to top of Forecast Research Division Section


Meteorological Applications Branch
Cecilia M.I.R. Girz, Chief

Objectives

The Meteorological Applications Branch performs diagnostic studies of weather-related phenomena, including mesoscale convective systems and clear-air turbulence. A springboard of these studies is the development of diagnostic tools that are applicable to routine observations, data from experimental networks, or model gridpoint data, and that utilize statistical methods, fundamental dynamical relationships, and derived parameters relating to unobserved variables. These studies often result in products of value to forecasters and are transferred to the operational sector, such as the National Weather Service (NWS). Research quality datasets of operational sounding and precipitation data and of commercial aircraft atmospheric data are assembled to support FSL modeling and diagnostic activities, and are shared with other NOAA laboratories and NWS research groups. The branch also conducts field tests and computer simulations to study the impact of balloon-based, in situ observing systems on atmospheric and oceanic monitoring for environmental prediction and climate observations.

Accomplishments

Global Air-ocean IN-situ System (GAINS)

GAINS is an Earth observing system of 400 regularly spaced platforms from which in-situ sounding of the atmosphere and ocean, and in-situ collection of air chemistry samples can be performed. It is anticipated that several kinds of vehicles, operating between 18 and 23 km, will make up the GAINS low-Earth-orbit constellation. Superpressurized, shear-directed balloons and remotely operated aircraft (ROA) are presently under consideration. GAINS vehicle development currently emphasizes the superpressure balloon and appropriate altitude control mechanisms. A number of forecast tools have been refined for planning and support of GAINS field tests. Objectives for the GAINS project were to (1) continue preparations for a 48-hour flight of the PIII balloon, (2) continue development and testing of a pump for altitude control, (3) begin development of the full-scale balloon, (4) assess trajectory forecasting tools, (5) begin to quantify the impact of GAINS observations on numerical weather prediction, and (6) assess the length of record required for in-situ data to determine climate trends in temperature.

Balloon Vehicle Development – Concerns on the weight of the payload, which came in at 50% over budget for the PIII flight, were addressed. Outside of the gondola, the greatest mass was the flight's power system. Alternatives were designed and will be implemented in 2002. The 48-hour flight is being temporarily delayed until the new power system is in place. In the meantime, efforts were stepped up on the development of the next critical system of the GAINS balloon – altitude control. In laboratory testing in an environmental chamber, the performance of the pump was measured for volumetric flow at nominal operating pressure and temperature. The candidate pump successfully met the GAINS flow and power specifications. In order to confirm these measurements at altitude in the almost infinite plenum of the atmosphere, the pump was interfaced with a 2.4-m PI superpressure balloon (Figure 27), and telecommunications were developed for flight testing the pump above 15 km.

Figure 27 - GAINS - Cec and Tom

Figure 27. GAINS leader Dr. Cecilia Girz and Thomas Shilling of Advanced Engineering preparing for a flight test of the GAINS superpressure balloon's altitude control system.

At 33.5 m in diameter, the full-scale GAINS balloon potentially contributes a significant mass to the entire system. Minimization of the vehicle mass allows a larger payload mass for in-situ and on-board remote sensors. Under a Phase 1 SBIR grant, GSSL, Inc., of Hillsboro, Oregon, began investigating alternate materials and construction methods that would retain the strength of the SpectraTM balloon, and perform well for long periods in the hostile environment of the lower stratosphere, while showing a savings in weight. GSSL has proposed a new construction method, and evaluated several materials appropriate to this new method for their weight, UV, and temperature properties.

Since 1998, software has been developed that uses observations and model wind data for prediction of balloon trajectories for GAINS. The four versions currently in use are Version 7 (rawinsonde data), Version 8 (objectively-analyzed rawinsonde data), Version 10 (global AVN model winds), and Version 11 (RUC-2 model winds). Each version produces predicted balloon positions in 1-minute increments. Diagnostic information about the input wind data (percentage of sounding stations available, identification of missing sounding levels or missing model forecasts) was added to the trajectory prediction page that allows users to evaluate the probable accuracy of the predicted trajectories based on quantity of the input winds. Efforts were begun to evaluate these trajectory forecasts. Prior to 2001, balloon position data from experimental balloon flights were compared to predicted trajectories solely on a case-by-case basis. A preliminary verification study of predicted trajectory accuracy was performed. Since resource constraints have not permitted twice daily balloon launches (from which actual balloon trajectories can be obtained), a verification system was developed using hourly analyses from the MAPS RUC-2 model (Version 11) as a baseline to examine differences between baseline and predicted trajectories from Versions 7, 8, and 10 for the period 1– 30 May 2001. Differences in balloon position were computed at 1-hour intervals (scatterplot not shown). Comparison was made of latitude differences at 3 hours between the RUC-2 baseline and the 0000 UTC rawinsonde forecasts of Version 7 during May 2001 (Figure 28). Scatterplots were run to determine directional differences for 3-hour flights predicted by Verson 7 from the 0000 UTC rawinsonde data compared with the RUC-2 baseline (Figure 29). These methods will be used in 2002 to examine additional time periods as well as data from actual flights conducted in the future.

Figure 28 - Comparisons of Latitude Differences

Figure 28. Comparison of latitude differences at 3 hours between the RUC-2 baseline and the 0000 UTC rawinsonde forecasts of Version 7 during May 2001.
Figure 29 - Scatterplot of Directional Differences

Figure 29. Scatterplot of directional differences for 3-hour flights predicted by Version 7 from the 0000 UTC rawinsonde data compared with the RUC-2 baseline during May 2001. Median difference shown in red square.

New tools for creating real-time plots on the Web of the positions of the balloon and the tracking crews were implemented for flight operations. A script was developed that aids the Boulder base crew to enter data for creating real-time updates on balloon trajectory during flight and a descent vector during flight termination. A list of experiments performed in 2001 was added to the public-accessible experiments Webpage (http://www-frd.fsl.noaa.gov/mab/sdb).

Impact Studies – Additional in-situ data from 400 locations across the globe should have a measurable effect and be applicable to current problems in meteorology. Two studies performed outside of FSL addressed this issue. A partial observations system simulation experiment (POSSE) of the effect of additional winds from a GAINS network on the forecasts of the United Kingdom Meteorological Office (UKMO) Unified Model was performed using the 3DVAR scheme. Positions of a network of 410 balloons for four times per day during the month of January 2001 were generated at FSL from the NCEP/NCAR Reanalysis Data and provided to the UKMO for assimilation into the model. Forecasts were initialized from every 1200 UTC analysis to run for 10 days with output valid at 1, 2, 3, 4, 5, and 10 days. The model was run for four constellations ranging from 52 to 410 balloons; an additional constellation with 410 static balloons was also run. Winds for only the 30-mb level were assimilated, which is a more limited set of data than the GAINS in-situ soundings will provide. Despite this limitation, preliminary results comparing a nature run and these simulations show a significant global improvement in the horizontal wind speeds in the analyses at 20 and 30 mb.

A second study conducted in the Air Resources Laboratory determined the requirements on datasets used to confirm predicted climate trends. This analysis of a 50-year record of temperature from radiosonde data taken at Topeka, Kansas, showed that trends of a fixed magnitude are easier to detect in the free troposphere than in either the boundary layer or the lower stratosphere. The free troposphere may be an important region for climate monitoring, especially given general circulation model projections of a nearly constant warming throughout the troposphere. Geographic regions that may be critical to climate monitoring are important considerations for the GAINS network.

Forecasting Clear-Air Turbulence

Field Studies – FRD conducted a field program to characterize CAT through observations and to improve forecasts of clear air turbulence (CAT). Since 1996, experimental forecasts of turbulence in the vicinity of jets and upper-level fronts have been produced at FSL (http://www-frd.fsl.noaa.gov/mab/tke). Diagnosed from the winds and temperature data of the RUC model and plotted at 1,000-foot intervals between 2,000 and 42,000 ft, 3–12-hour forecasts of turbulence for the 48 states are updated every three hours. These forecasts are also part of the Integrated Turbulence Forecast Algorithm (ITFA) appearing on the Aviation Digital Data Service (http://adds.aviationweather.noaa.gov/) at NCEP's Aviation Weather Center in Kansas City. These products have allowed aviation users to access experimental turbulence forecasts that are more timely and detailed than what has been previously available. However, there is room for improvement in these products, as real-time verification statistics indicate.

Funded in part by the FAA's Aviation Weather Research Program and operating in the Pacific Ocean in collaboration with NOAA's Winter Storms Reconnaissance Program, SCATCAT01 (Severe Clear-Air Turbulence Colliding with Air Traffic 2001) observed turbulence above 20,000 ft from NOAA's Gulfstream IV aircraft. Of particular interest was the relationship between the internal wind shears and associated gravity waves of fine-scale, stable, laminar structures that were seen in the SCATCAT99 and NORPEX98 datasets, and clear air turbulence. Dropsondes at approximately 40-km intervals were deployed from 41,000 ft in regions where turbulence was anticipated to occur, and the aircraft subsequently confirmed the presence of turbulence by sampling the air mass at altitudes between 31,000 and 37,000 ft. Flights occurred on five days where turbulence intensity ranged from spotty, light turbulence to moderate turbulence through a deep layer to severe turbulence confined to a 1,000-foot layer.

Diagnostic Algorithm Development – The forecast skill of ITFA and its component algorithms has been evaluated both objectively and by forecasters. These studies show that the best of the algorithms display similar probability of detection (POD) curves, and that there is considerable room for improvement. Research conducted at FSL indicates that these algorithms also typically predict patterns that are similar to one another, and that moderate or greater (MOG) PIREPs of turbulence often fall in the margins of the predicted ITFA regions. The best of these algorithms are fundamentally based on the destabilizing dynamics of vertical wind shear.

FSL has developed an experimental turbulence prediction scheme based on the established fact that mesoscale gravity waves are generated as an unbalanced jet streak propagates toward an inflection axis in the upper-level height field. We have diagnosed gravity waves and flow imbalance in several detailed case studies. These analyses show that MOG PIREPs and gravity waves analyzed from bandpass-filtered, time-to-space converted 5-minute surface observations consistently occur directly downstream of the region of diagnosed flow imbalance. The flow is considered to be unbalanced when there is a pronounced residual in the computed sum of the terms in the nonlinear balance equation (NBE) from the RUC model. Imbalance often occurs in essentially the same region as that where mesoscale gravity waves typically develop, but can also be due to mountain waves generated by strong flow over rough terrain like the Rocky Mountains. A recent example from our real-time evaluation system (Figure 30) shows an outbreak of moderate-to-severe turbulence that occurred on 27 March 2002 above 15,000 ft primarily over the Rocky Mountains at and downstream of a pronounced trough in the upper troposphere. Flow imbalance as diagnosed from the NBE residual field correlates quite well with the MOG PIREPs of turbulence, as well as other scattered reports across the western and southeastern U.S. Meanwhile, the IFTA paints a very broad picture of predicted moderate turbulence (green), and no prediction of severe turbulence (yellow). Another experimental predictor being tested at FSL that is based on the unbalanced flow concept is shown in Figure 30d. This latter field, which does not relate well to the observed turbulence in this event, is an unbalanced height field derived by inverting the NBE and solving for the balanced height field. Further refinement of forecast turbulence regions might be obtained by adding the requirement that an efficient wave duct must be present downstream of the region of diagnosed flow imbalance to retard the vertical leakage of wave energy, thus allowing coherent waves to persist. This idea will be examined.

Figure 30 - Comparison of Diagnostic Fields A

Figure 30 - Comparison of Diagnostic Fields B

Figure 30 - Comparison of Diagnostic Fields C

Figure 30 - Comparison of Diagnostic Fields D


Figure 30. Comparison of different diagnostic fields for predicting the occurrence of moderate-or-greater (MOG) turbulence. MOG PIREPs encountered from 1900 – 2240 UTC 27 March 2002 are depicted in (a); 15,000 – 45,000 ft composite ITFA prediction from a 9-hour RUC-2 forecast valid at 2100 UTC is shown in (b); the new FSL experimental turbulence predictor field based on the Nonlinear Balance Equation (NBE) method valid at 1800 UTC appears in (c); and another experimental method being developed at FSL is shown in (d).

While mesoscale models like RUC can be useful for diagnosing the flow imbalance regions and where generally gravity waves are likely to form, they do not reliably predict the details of the gravity waves themselves, such as their wavelength, phase speeds, and so forth. The new predictive scheme being developed at FSL not only produces patterns systematically different from the current ITFA algorithms but also predicts turbulence regions missed by those methods. The next steps are to define the optimal thresholds and objectively verify the new unbalanced flow fields against PIREPs, before including this algorithm within the ITFA.

Climate Station Monitoring Project

A new project sponsored by the National Climatic Data Center has begun at FSL to develop an automated system for monitoring precipitation observing sites in the cooperative network. During this past year, research conducted in collaboration with NCAR established effective procedures to monitor the performance of gauge observing sites. Initial research centered on determining the most effective measures of spatial correlation between gage sites, under the assumption that because of the large temporal and spatial variability of rainfall, monitoring changes in these measures of spatial correlation might better identify inhomogeneities than monitoring the precipitation values themselves. In addition to standard correlation coefficients applied to the rainfall values themselves, other candidate measures include such scores as frequency and magnitude bias, and equitable threat score. To simplify initial analyses, we chose to first concentrate on daily observations during a 51-year period at coop sites in Iowa. Scores were computed by season using daily observations at target stations paired up with the 12 closest neighbors. Time series of these scores were then produced for display and future evaluation using change-point statistical methods or other procedures. To test the sensitivity of these scores to data inhomogeneities, various bogus data were introduced into the analyses. Figure 31 shows an example in which the magnitudes of precipitation were doubled for the first half of the time period; the scores show a marked and consistent low value before 1975, the year that the introduction of bogused data was ended. A real-life example in Figure 32 reveals a station where one year's observations (1989) are markedly inferior to scores during the rest of the time series. The data themselves revealed that observations during this year suffered degradation from an undetermined source in which all the observations were several times too small. The bogus time series and other time series plots are suggesting utility in the application of these scores to a monitoring procedure.
Figure 31 - Median Bias Scores

Figure 31. Time series of rank of the median bias score for station with bogused data until 1975. See text for explanation.
Figure 32 - Median Bias Scores II

Figure 32. As in Figure 31 except for the logarithm
of the magnitude bias score for a real station.

Research Quality Datasets

Assessing the Quality of Real-Time Precipitation Gauge Observations – Research on precipitation data quality at FRD has concentrated principally on the network of hourly automated raingauges (part of the Hydrometeorological Automated Data System (HADS). FSL's interest in these data derives from several sources. The Real-Time Verification System (RTVS) ingests them to verify subdaily forecasts of precipitation. They are also displayed alongside daily gauge reports on a real-time precipitation Website (http://esrl.noaa.gov/precip/hourly_precip.html). This year the U.S. Weather Research Program (USWRP) is supporting additional research at FRD on HADS data quality in an effort to refine the observations used in analysis and verification at the National Centers for Atmospheric Prediction (NCEP).

For RTVS, the HADS datastream is routinely monitored using the FRD precipitation Website. Work has also proceeded on the assessment of the extent and effect of known types of HADS data problems. Part of this effort has been directed at the set of stations that the individual River Forecast Center's data screening procedures remove from the HADS datastream. Often this screening appears to be too restrictive, resulting in the removal of many satisfactory observations and unnecessary degradation of sampling for the RTVS. At FSL, visual displays of the time series of observations in regions badly sampled because of this tendency were used to improve sampling in the central United States in support of the IHOP Field Experiment.

ACARS/AMDAR Quality Control System – A computer program to flag and in some cases correct weather data from automated sensors on commercial aircraft (ACARS and AMDAR) was upgraded. The quality control system uses temporal and spatial consistency checks along each flight track and altitude-adjusted climatological consistency checks to discover errors. It also interpolates locations and times for high-resolution observations taken during ascent and descent enabling these data to be used in numerical weather prediction model research and weather forecasting. The quality control system was upgraded to ingest and display data from Lufthansa, which flies primarily over Europe and Southeast Asia. The system was also used to develop data rate statistics, which are being used to support initiatives to develop a mechanism for public funding of weather-related communication charges. An error in calculated aircraft heading (a variable that had not previously been used in subsequent analysis) was discovered and corrected. The corrected heading variable is being used to develop error statistics for longitudinal and transverse winds.

FSL Websites

GAINS Website (http://www-frd.fsl.noaa.gov/mab/sdb/) – To aid flight planning in the 1 – 10 day time frame, a Flight Briefing page was added to the GAINS Website. Updated hourly, this page contains the current surface observation and upper-air winds for Tillamook, Oregon, as well as links to satellite, radar, numerical model, surface and upper-air charts and other data. Upper-air charts for mandatory pressure levels above the height of 100 mb covering the continental United States, Pacific Northwest, and Colorado regions are generated in-house and added to the Briefing page. Since the Briefing page is Web-based, it is accessible to crews at any location in the field wherever a phone line exists.

Chemical Weather Research and Development Website (http://www-frd.fsl.noaa.gov/aq) – In support of a major NOAA initiative to improve temperature and air quality forecasting, staff developed a Website that will present real-time and retrospective results from air quality models. The region of interest is primarily New England, and the intensive experiment period will take place in the summer of 2002. A related Website that focuses on high-resolution temperature forecasting is http://www.temp-aq.org.

National Hourly/Daily Precipitation Website (http://esrl.noaa.gov/precip/hourly_precip.html) – Development work continued on a Website that displays hourly and daily precipitation data from NCEP. Data are displayed on a national map that optionally shows rivers and county boundaries. Moving the cursor across the map reveals the available data, and a mouse click provides a daily time series of the data. Users can zoom and roam on the map for detailed local structure of precipitation events. Several years of precipitation observations are available at the site. A precipitation "day" was redefined as 1300 UTC – 1300 UTC, rather than 1200 UTC – 1200 UTC to allow a better match between sites that report once per hour and those that report once per day. This site is one of the primary tools used by NCEP staff to quality control the National Stage II Precipitation product.

ACARS/AMDAR Website (http://amdar.noaa.gov/) – The following upgrades were made to this site, which displays weather data from automated sensors on commercial aircraft: 1) Data from Lufthansa airlines were added; 2) A minimum data spacing may now be specified, allowing data to be seen more clearly; 3) Users can now specify an initial default map so that the website immediately opens to their preferred region of interest; 4) Wind barbs in the southern hemisphere are now correctly displayed with the barbs on the lower-pressure side of the shaft; 5) Information at the cursor no longer runs off the screen when the cursor is near the edge of the map; 6) U.S. VORs are now an available overlay; 7) Color coding of vertical gust data is changed to better reflect turbulent conditions; 8) The java code was made more efficient, and modularized into java packages, which enables staff to share code among multiple web sites; and 9) An error in plotted wind direction, small and most noticeable near the equator, was discovered and corrected. In a recent month, the site was accessed by 97 sites, such as airlines, United States and foreign forecast offices, and research institutions. These sites requested more than 2,300 data loads, and looked at more than 4,500 soundings. Figure 33 shows the upgraded display, with the minimum data spacing set to 15 pixels so that wind barbs show more clearly.

Figure 33 - ACARS/AMDAR Website

Figure 33. An upgraded display from the ACARS/AMDAR Website, http://amdar.noaa.gov/, with the minimum data spacing set to
15 pixels so that wind barbs show more clearly.

Interactive Soundings Website (http://www-frd.fsl.noaa.gov/mab/soundings/java/) – This upgraded Website interactively displays past and forecasted soundings from a variety of sources. An error in wind direction for MAPS and RUC soundings was discovered and corrected. Wind directions near the east and west edges of the RUC/MAPS domain were in error by as much as 18 degrees. Figure 34 shows a RUC2 forecast at 2100 UTC 16 August 2002. This page was accessed 25,000 times from 249 sites in January 2001, and continues to be very popular. The easily adaptable Java code that runs this site has been requested by several organizations, and has been released to them under FSL's new open-source software license/disclaimer.

Figure 34 - Interactive Soundings Website

Figure 34. A screen from the Interactive Soundings Website, http://www-frd.fsl.noaa.gov/mab/soundings/java/, showing a RUC2 9-hour forecast sounding for Omaha, Nebraska. It indicates potential for severe thunderstorms. If a forcing mechanism occurs, such as a weather front or higher-than-expected surface temperatures, the region of convective inhibition (the green area) can be overcome, and the convective available potential energy (CAPE, the large red area) will be released as severe weather.

National Mesonet Website (http://madis.noaa.gov/sfc_display/) – Using Java, a national mesonet Website was developed to interactively display observations from six mesonets, maritime buoys, and the METAR network. The site (no longer restricted) is available to the public, and displays weather data and quality control information from FSL's Meteorological Assimilation Data Ingest System (MADIS). The Java code has been modularized into packages, which are shared with other FSL Websites, allowing easier code maintenance.

FSL Publications Website (http://www.fsl.noaa.gov/publications/) – This site has been revamped to provide easier data entry and easier search, particularly for conference publications. For instance, conference and journal names are now available in drop-down menus. This avoids ambiguity in conference names (which have many variations), both for those who enter data and those who search for it. Search results are formatted to more precisely match American Meteorological Society standards (FSL's primary publisher). The publications database is updated through several years, so those interested in gaining access to information about FSL publications can be assured of obtaining the latest information. Also, electronic versions of publications are being produced for many of the latest contributions.

Projections

During Fiscal Year 2002, the Meteorological Applications Branch will be involved in the following activities and studies.

GAINS

Two flights are planned for 2002: the 48-hour flight of the GAINS PIII balloon and a 2-hour demonstration of the experimental pump. The PIII balloon will be launched from Tillamook, Oregon, to demonstrate launch and recovery operations, and the operation of the mechanical systems (balloon vehicle, helium valve, helium cell rip line, Balloon Envelope Recovery System) over two day-night cycles. Initial efforts will concentrate on achieving a T-24-hour launch condition for the payload, and launch and recovery equipment. The balloon will be launched under appropriate weather conditions: scattered to no clouds, a trajectory forecast that keeps the balloon within the borders of the U.S. away from major airports, and a landing location on relatively flat terrain outside major cities. For the second test, the pump will fly on a commercially made zero-pressure balloon and demonstrate its ability to create a 15% superpressure in an 8-ft diameter anchor balloon at three altitudes: 50, 60, and 70 K ft. It will be launched in Colorado in collaboration with the Edge of Space Science organization. Weather conditions for the 2-hour flight to test the pump are the same as for the 48-hour flight.

The feasibility and materials assessment for the full-scale GAINS balloon will be completed by GSSL, Inc. under the Phase I SBIR grant, and a proposal for a Phase II grant will be presented to NOAA. GAINS trajectory plans include development and implementation of Webpages to allow users to generate upper-air charts and predicted trajectories based upon 1946–1999 climatological averages of the North American rawinsonde data. Numerical model-based trajectory prediction software will be updated to match increases in model grid resolution. The UK Meteorology Office POSSE will complete its analysis of the impact of GAINS-like winds on the Met Office Unified Model. Benefits of these data will be assessed as a function of forecast length, and vertical effect by geographic region. The study will also quantify whether a minimal (54-balloon network) or maximum (410-balloon network) is optimal. Climate trend studies will focus on determining geographic regions that are optimal for detecting climate change and on quantifying the effect of the temporal frequency of observations. Both results will guide development of the operational GAINS network.

Forecasting Clear-Air Turbulence

Analysis of the SCATCAT cases will begin with the objective of deriving a comprehensive picture of the atmosphere producing turbulence. This picture will be built using the dropsonde data, meteorological data from aircraft flight level, ozone data from the Aeronomy Laboratory's experimental sensor flown during SCATCAT, satellite-derived winds, and model analyses. In addition to the case study analyses, modeling studies will be run with the 20-km RUC to determine whether large-scale and mesoscale features captured in the aircraft data are resolved in the model, where and with what intensity the model develops turbulence from both diagnostic and prognostic routines, how this model turbulence compares with that measured by the aircraft, and how tropopause folding in the RUC compares with the onboard ozone measurements.

The residual of the nonlinear balance equation and other methods will be further investigated to arrive at the optimum method for diagnosing imbalance and determining the appropriate threshold values. Real-time evaluation of these approaches will continue to be performed in preparation for eventual implementation and full evaluation within ITFA. Staff will perform idealized modeling studies to develop a basic understanding of the nonlinear scale contraction process by which mesoscale gravity waves may steepen and saturate, leading to turbulence production at smaller scales.

Research Quality Datasets

Real-time Precipitation Gauge Observations – Quality control procedures will be automated to identify, among other systematic effects, stations that routinely fail to report zero precipitation, stations that chronically report small values on "non-rain" days, and other stations that are intermittently susceptible to very large values. These plans also involve collaboration with NCEP in the real-time automated assessment of gauge quality that is timely enough to assist in the production of NCEP precipitation analyses.

NCEP is hopeful of developing automated quality control routines that operate in near real time, and thus allow screening of HADS observations in time for analyses and verification for which the daily RFC screening occurs too late. To help accomplish this, FRD has been enlisted to identify HADS stations that fall prey to several known types of error (sticking sensors, for example), blacklist them on a monthly time frame, and migrate them off the blacklist when their performance improves.

ACARS/AMDAR Quality Control – Using the "heading" variable calculated by the ACARS/AMDAR quality control software, the longitudinal and transverse component of the horizontal wind at flight level can be calculated. These wind components will be compared with corresponding values from the RUC 1-hour forecast, in order to identify wind errors, and stratify them by longitudinal and transverse directions. These stratified errors will be used by new RUC data ingest/quality control software.

IHOP Field Experiment – Under sponsorship of the U.S. Weather Research Project (USWRP), FSL will develop plans for direct involvement in the field phase of IHOP during May and June 2002. Aircraft missions that utilize dropsondes, Doppler lidar, and differential absorption lidar to describe in unprecedented detail the evolving moisture structure within the low-level jet (LLJ) will be designed. These missions also have potential to provide observations from which a moisture budget of the LLJ can be diagnosed. Postfield phase analyses will be primarily intended to examine how accurately existing operational analyses and models describe the LLJ location and structure. Since the LLJ is the major factor in the modulation of precipitation in the central U.S. Great Plains during the spring and summer, knowledge gained should be very useful for assessment of Quantitative Precipitation Forecasts.

Monitoring NCDC – Under sponsorship of the "Health of the Network" initiative at the National Climatic Data Center (NCDC), FSL will complete development of a system to monitor the cooperative precipitation network. The principal intent is to identify inhomogeneities in observations at stations as soon as possible and prevent them from propagating further into climate analyses. To accomplish this, FSL will determine the best algorithms that relate station precipitation time series to their neighbors by applying potential candidate algorithms to historical gauges and examining the resulting time series to see how well they identify past inhomogeneities.

FSL Websites

National Hourly Precipitation Website – In support of a new collaborative project with NCEP to improve rain gauge QC and assess the Stage IV precipitation product, this Website will be upgraded as follows.
  • Identify on the display those gauges that have been rejected or flagged as suspect by various quality control algorithms. A similar technique for displaying "bad" and good data upon request has already proven very helpful in the Meteorological Assimilation Data Ingest System (MADIS) developed at FSL for mesonet data acquisition, integration, quality control, and transmission to NCEP.
  • Provide detailed quality control information about each gauge station in a separate window that will pop up upon clicking on the station with the mouse.
  • Provide the ability to easily find stations by name or other attribute. (To find a particular station now, one must know its location, and then look among all nearby stations until the desired station name appears near the cursor.)

National METAR Website – New mesonets will be added as they become available. Wind gust and precipitation amounts will be shown for those sites that support them. Data loading will be sped up.

Soundings Website – This site will be converted to employ the new modularized Java packages used by other FRD applets. This change will make the code easier to maintain, and will decrease the time to load the website. An error in the calculation of interactive Convective Available Potential Energy (CAPE) will be corrected. By definition, CAPE should consist only of buoyant energy above the lifted condensation level (LCL). The current code includes buoyant energy at all levels.

Return to top of Forecast Research Division Section


FSL Staff FSL in Review (Other Years) FSL Forum Publications