FSL in Review


Scalable Modeling System


Software Approaches

Smart Tools

GFESuite Software

Designing Digital

Deriving Sensible Weather Elements

Rapid Prototype Project

Forecast Methodology

Mt. Washington


Recent Publications

Contact The Editor

Wilfred von Dauster


...From the Director's Office

NAOS Update - Dr. Thomas W. Schlatter, Chief Scientist

During summer 2000, the North American Observing System (NAOS) program decided to refocus efforts during the coming five years on two major goals:
1) Build the infrastructure to process data from a new generation of satellite radiometers and assimilate them into numerical prediction models. The two instruments of greatest interest are the Atmospheric Infrared Sounder (AIRS) containing 2,378 spectral bands, to be launched in 2001 aboard a NASA Earth Observing System (EOS) satellite, and the Geostationary Imaging Fourier Transform Spectrometer (GIFTS), containing about 4,000 spectral bands, to be launched in 2003 or 2004 with the support of NASA and the U.S. Navy. Both instruments have the potential to improve the resolution and accuracy of temperature and moisture soundings in clear air. They will deliver at least two orders of magnitude more information than current polar orbiting and geosynchronous satellites, and thus it is important to prepare now for the processing and assimilation of the data.
2) Design a composite mesoscale observing system that supports a significant improvement in critical weather forecasts made by field meteorologists and numerical weather prediction (NWP) models. Because this is a NAOS goal, the mesoscale forecasting needs of North America, not just the United States, must be considered in the design.

The NAOS Test and Evaluation Working Group (T&EWG) convened a meeting in Silver Spring, Maryland, on 1 and 2 November 2000 to begin addressing Goal 2. The remainder of this report summarizes the outcome of the meeting.

The basic observational requirements for mesoscale forecasting in the coming 5 to 10 years are partly dictated by the expected resolution of numerical models. The group agreed that this would be in the range from 2–10 kilometers. This grid point spacing would capture mesoscale phenomena in the range from roughly 10–50 km in size, including low-level jets, mesoscale convective complexes, squall lines, frontal rainbands or snowbands, lake-effect snows, spiral bands in a hurricane, and many others. It would also capture topographical and surface boundary effects (e.g., soil moisture, land use, vegetative cover, snow cover) in much greater detail than current models. From the standpoint of data assimilation, it is desirable to have observations often enough and dense enough that the corrections to the background (a short model forecast) remain small, as is the case with synoptic-scale data assimilation. This relates directly to the size and lifetimes of the phenomena that are to be captured (tens of kilometers and hours). Because different parts of North America experience different kinds of hazardous weather and because of uneven population distribution, requirements for mesoscale observations will vary across North America.

Many types of observations are not yet exploited in operational NWP, for example,

  • images from geostationary and polar orbiting satellites that could be used to analyze cloud and moisture fields
  • satellite infrared radiance data over land, specifically from the channels that can
  • microwave measurements from polar orbiting satellites over land
  • lightning data.

    For some sources, national data collections are inadequate or even nonexistent, such as,

  • radial winds from scanning Doppler radars
  • radar reflectivity data
  • local pollution observations (particulates, ozone, nitrous oxides, carbon monixide)
  • many surface observations from local mesonets around the country. The group noted that the availability and the utility of these sources differ markedly between forecast offices, where workstations display many of these data and local forecasters rely on them in making decisions, and national NWP centers, where the use is strictly quantitative.

    Even if the above data sources were fully available, there are serious problems in assimilating them into NWP models:

  • ineffective quality control measures
  • lack of forward models (algorithms for estimating an observed parameter from the model variables)
  • imperfect knowledge of error statistics, either for the observation source or the short-term model forecast that is to be corrected by the observations
  • lack of appropriate constraints to link the mass, motion, and moisture fields in the finished analysis
  • what to do when the correction to the background is large (e.g., thunderstorm in the wrong place).
    The last two problems are distinctly mesoscale in character, but the others are a symptom of incomplete information about the data source. The group agreed that a concerted effort to solve these data assimilation problems must be part of any initiative for a composite mesoscale observing system.

    Ideally, new observing systems should be simulated realistically with all other systems currently available. Prerequisites to a good observing system simulation experiment (OSSE) are a well-calibrated and tested nature run (used to simulate the observations) and the data assimilation machinery to assimilate all current and proposed data sources. These prerequisites are far from being met at the mesoscale. In particular, a state-of-the-art mesoscale model for generating a regional nature run has yet to be identified and its performance carefully calibrated. The effects of lateral boundary conditions (supplied at relatively coarse resolution) on the simulation are unclear. Data assimilation has yet to deal adequately with radial velocity and reflectivity data from Doppler radars, data on surface properties, satellite imagery, satellite sounder data over land, and precipitation initialization. The group concurred that the capability to mount a credible mesoscale OSSE is several years away.

    Jack Hayes, newly appointed Director of the NWS Office of Science and Technology, attended the Test and Evaluation Working Group meeting, as did others from NWS Headquarters. He voiced support for designing a mesoscale observing network that would optimize forecast services and support mesoscale NWP in times of hazardous weather. He will support an initiative of this nature, but he needs information on the types of observations required, their spatial and temporal resolution, whether they are existing or new systems, a plan for network architecture and communications, and cost estimates. Because he would like to consider a NOAA budget initiative for Fiscal Year 2003, he needs a rough outline by February 2001. He would like to consider three scenarios: the optimum network, where cost is a secondary consideration, a middle-of-the road network which blends efficacy of the composite observing system with cost, and a low-cost network that would nonetheless provide a jump in capability. The initial proposal should be based upon expert judgment, but not OSSEs.

    Like many groups before, the T&EWG concluded that the greatest unmet need for mesoscale observations is in the atmospheric boundary layer. In Canada, the need extends higher, to the midtroposphere. Observations of temperature, pressure, wind, and moisture are needed. There is also a requirement for soil moisture and aerosol concentration, in that the former strongly influences the surface energy budget and the latter influences the hydrometeor size distributions in clouds. Satellites can provide some of these measurements in clear air, especially next-generation radiometers with high spectral resolution. In situ and ground-based remote sensors must provide the rest.

    The group recognized that higher resolution alone should improve predictions of those mesoscale weather events that are forced by topography, but that detailed specification of the initial conditions (through mesoscale observations) is vital for prediction of "free tropospheric" events, those forced by the winds under specific conditions of temperature and moisture stratification.

    Another meeting of a subgroup of the T&EWG along with several field forecasters will be held in mid-December 2000. This meeting will consider these questions:
    1) In priority order, what current sources of observations, not presently assimilated, would do the most to improve mesoscale forecasts? Perhaps some sources would be deemed not particularly helpful.
    2) What actions are required to assimilate these sources? (The most basic requirements are to generate forward models for each source and methods of quality control). Example: radial winds from WSR-88D radars.
    3) What new sources are likely to appear in the next five years that already have a hardware budget? Examples: AIRS, GIFTS. What resources are available in other programs to accommodate these sources? What’s left uncovered?
    4) After considering current sources and highly probable new sources and their utilization in mesoscale forecasting, what gaps in the composite observing system will still remain (e.g., observation of standard atmospheric parameters in the boundary layer, perhaps extending up to the midtroposphere)? What instruments could fill the the gaps?
    5) How many units of each new observing system are required, and where should they be located?

    If this meeting succeeds, the next step will be to generate rough cost estimates for a February 2001 initiative.

    ...From the Forecast Research Division Dr. Steven E. Koch, Chief

    The Role of Dual Cold Fronts Aloft in a Major Tornado and Flash Flooding Event – Steven E. Koch will present a paper on this topic at the January 2001 AMS Meeting. This research is performed in collaboration with Jamie Mitchem of North Carolina State University.

    Meso-Eta cross-sectional analyses indicate that dual CFAs (cold fronts aloft) were associated with the strongest lifting, the largest potential instability, and the most moisture, which is why the squall lines formed along the CFAs instead of the surface features. A tropopause fold and the descending dry air within the dry conveyer belt supplemented the midtropospheric dry air behind the westernmost CFA. The first CFA was not identified with a tropopause fold.

    WSR-88 Velocity Azimuth Display (VAD) Wind Profiles (VWP) at ~5-minute intervals and retrieved cross sections of geostrophic thermal advection pinpointed the location, strength, and depth of the CFAs as they crossed each of several radar sites in good agreement with the Meso-Eta model. These data further suggest that the CFA system provided the dry air and momentum source for the rear-inflow jet behind the powerful squall line as it passed through Alabama. This study demonstrates the ability of readily available mesoscale model fields and observational data to detect the presence and vertical structure of an important class of mesoscale phenomena that can produce extreme weather events, such as the one in this case. The results presented also testify to the highly complex, nonconventional structure of some atmospheric frontal systems.

    ...From the Regional Analysis and Prediction Branch Dr. Stanley G. Benjamin, Chief

    RUC News – A 20-km version of the Rapid Update Cycle (RUC) is in real-time testing now, and scheduled for implementation at the National Centers for Environmental Prediction (NCEP) by spring 2001. Real-time products and a description of the 20-km RUC are available from the RUC Website, http://ruc.fsl.noaa.gov. The 20-km RUC includes a three-dimensional variational (3D-VAR) analysis with assimilation of GOES cloud-top pressure and a revised version of the model, including the latest upgrades to the NCAR/FSL Reisner microphysics, and a Grell ensemble cumulus parameterization.

    The spring issue of the FSL Forum will feature a series of articles on the RUC upgrades.

    An Initial RUC Cloud Analysis Assimilating GOES Cloud-Top Data – Dr. Dongsoo Kim will present a paper on this topic at the January 2001 AMS Meeting.

    A cloud analysis technique for the RUC has been revised and incorporated into a parallel 1-hour assimilation cycle. From 17 May–30 June 2000, a cycle with hourly assimilation of GOES cloud-top pressure using this revised technique (along with other observations) was run in parallel with a control cycle without GOES cloud data. Verification results showed a strong positive impact from the GOES cloud assimilation on subsequent cloud-top forecasts for 1-hour and 3-hour forecasts, and a weaker positive impact out to 12 hours. The effect of the GOES cloud assimilation on 3-hour MAPS relative humidity forecasts was also found to be positive, correcting an earlier deficiency.

    The GOES sounder-based cloud-top pressure data are timely and provide good coverage for the RUC domain, but this initial RUC cloud analysis still does not take complete advantage of the full-resolution satellite information. For example, the total number of cloud-top data processed for the RUC grid from both GOES platforms (8 and 10) is about 104, but the number of data used in either cloud building or cloud clearing is about 3*103. This reduction occurs in large part because of the difficulty in assimilating fractional cloud data within a grid volume. However, this problem will be improved with the upcoming 20-km RUC, which will be able to resolve many of the cloud areas considered fractional at 40-km resolution. Methods to merge GOES sounder and imager data for cloud analysis are under development.

    ...From the Meteorological Applications Branch Dr. Cecilia M.I.R. Girz, Chief

    Global Air-ocean IN-situ System (GAINS) Pump Tests – A critical component of the GAINS balloon is the pump. The pump moves ambient air into and out of the balloon’s air cell, changing the density of the balloon and thus the balloon’s float altitude. Vertical control in this manner allows the balloon to take advantage of shear in the horizontal wind in order to maintain an appropriate spacing in the dynamic network of floating balloons.

    A series of tests has been completed on an experimental pump being developed for GAINS by Advanced Engineering in Lakewood, Colorado. Between January and November 2000, the Climate Monitoring and Diagnostics Laboratory (CMDL) environmental chamber was used to characterize the pump’s performance for superpressure and power consumption with a number of impeller and controller designs. With the completion of these tests, Advanced Engineering has delivered a candidate pump that is able to meet the GAINS superpressure specification in the laboratory (i.e., 15 mb superpressure at an ambient pressure and temperature of 100 mb and -55oC).

    Due to the small volume of the environmental chamber and the high flow rates of the pump, an accurate measurement of flow rate was not possible in the chamber. The next step, then, is to confirm pump performance at altitude and to determine flow rates at altitude. A test flight is planned for early spring in eastern Colorado. For this test, a zero pressure balloon will lift the pump to altitude. During ascent, the pump will fill and release air from an anchor balloon at 50-, 60- and 70-kft altitudes.

    Thanks to Tim Shilling, Advanced Engineering, Dave Latsch, Brian Jamison, and Randy Collander, CIRA, and Bob Anderson, Basic Automation, for the successful completion of these tests.

    D3D: A Potential Three-Dimensional Visualization Tool for the National Weather Service – Edward Szoke will present a paper on this topic at the 2001 AMS conference. The research is performed in collaboration with Herb Grote, Patrice Kucera, Paula McCaslin, Philip McDonald, and William Roberts.

    In its exploration of 3D visualization, FSL has taken the University of Wisconsin's Vis5D software program and enhanced its capabilities and changed its graphical interface to appear very much like the AWIPS D2D system. This workstation application is being developed to examine the value of 3D visualization in an operational forecast setting. Our current version of D3D displays gridded output (analysis and output fields in AWIPS netCDF format) from several national-scale models: the Rapid Update Cycle, Eta, Nested Grid Model, Aviation (Global Spectral Model), and the Medium Range Forecast Model, as well as analysis fields from the Local Analysis and Prediction System. D3D is also capable of displaying output from a local-scale model that is run at FSL. Test versions of radar displays also have been demonstrated, but are not yet in the real-time framework. So far responses to the demonstrations and exercises have been positive.

    ...From the Aviation Division Dr. Lynn Sherretz, Chief, Requirements, Applications, and Quality Assessment Branch

    The Aviation Digital Data Service (ADDS) has been selected as one of the winners of the 2000 Government Technology Leadership Award. The award recognizes federal organizations that have demonstrated extraordinary leadership in using information technology to improve service to the public, lower costs of government, and improve the ability of agencies to meet their mission requirements.

    The award was presented at a ceremony at the Reagan International Trade Center in Washington, D.C. on 28 November 2000. The ceremony was held in conjunction with the Government Technology Leadership Institute, a two-day postgraduate program, which convened 200 senior federal managers to learn about effective use of technology in agency programs. ADDS is a joint development effort among FSL, NCAR Research Applications Program (RAP), and NWS Aviation Weather Center (AWC). More information about ADDS is available at http://adds.awc-kc.noaa.gov. http://adds.aviationweather.noaa.gov (link updated 03-11-2004)

    A Verification Approach for Assessing the Quality of Model-based Precipitation Forecasts during Extreme Precipitation Events – Andrew F. Loughe will present a paper on this topic at the January 2001 AMS Meeting. The research is performed in collaboration with Judy Henderson and Jennifer Mahoney.

    Accurate forecasts of precipitation are exceedingly important to the public. Emergency management teams rely on these forecasts to plan when and where disasters might strike, and the general public is concerned about whether precipitation will affect their many outdoor activities, especially if heavy precipitation is possible. In either case, accurate precipitation forecasts are largely dependent upon numerical model predictions. Therefore, a verification approach which assesses a model's ability to accurately predict precipitation at specific locations, and over relatively short (~3 hour) time periods, has been developed at FSL. Although numerical models are not prepared to directly address the problem of point-specific precipitation forecasting, this stringent verification approach will serve to track the progress of models over time as they evolve to meet this high expectation to the public.

    Convective Intercomparison Exercise: Baseline Statistical Results – Jennifer Mahoney presented a paper on this topic at the Ninth Conference on Aviation, Range, and Aerospace Meteorology. This research is performed in collaboration with Barbara Brown, Cynthia Mueller, and Joan Hart.

    Various convective forecasts, which are often used by air-traffic decisionmakers, were verified in near real time from 1 June–31August 1999 using the Real-Time Verification System (RTVS) to develop a statistical baseline for the quality of convective forecasts, and to illustrate the underlying differences in the forecasts. The forecasts used in this convective forecast intercomparison fell in two categories. One category includes forecasts that are issued frequently and produce short 1–2 hour forecasts, such as the National Convective Weather Forecast (NCWF) and C-SIGMETs (Convective Significant Meteorological Advisories). The other forecasts extend from 2–6 hours and are issued less frequently, such as the Collaborative Convective Forecast Product (CCFP) and the C-SIGMET Outlooks (Outlooks). Basic statistical results of the evaluation and the underlying differences in the forecast products are presented.

    ...From the Systems Development Division U. Herbert Grote, Chief

    Porting AWIPS to Linux – Darien Davis will present a paper on this topic at the January 2001 AMS Meeting. She and Herb Grote are collaborating on this project.

    FX-Linux software can be run on more than one computer platform. It has been thoroughly tested by FSL on the Hewlett-Packard running HP-UX 10.20 and PC/Linux hardware. The ability to run the software on another platform has made it possible to achieveAWIPS functionality and better performance on a desktop PC and a laptop computer. Greater hardware independence will make it possible to take advantage of new technological developments by the computer industry. FSL has so far ported and tested only the AWIPS display software and basic system infrastructure; the remaining portions of the system will be ported in the future.

    FSL Staff