FSL's Leon Benjamin receiving a NOAA Honor Award from Vice Admiral Lautenbacher, Under Secretary of Commerce for Oceans and Atmosphere

Assimilation of Surface Cloud, Visibility, and Current Weather Observations in the RUC model – An important problem for short-range numerical prediction is initialization of cloud and hydrometeor fields. Forecasts of cloud, fog, ceiling/visibility, and stable and convective precipitation are dependent on accurate initial conditions for these fields. Most mesoscale models now parameterize stable cloud processes with some type of bulk microphysics. The stable cloud microphysics parameterization used in the RUC is explicitly mixed-phase, with prediction of mixing ratios of five different hydrometeor types (cloud water, ice, rain, snow, graupel). The problem for cloud/hydrometeor assimilation is the mapping of disparate, one-sided (cloud decks apparent from space or the earth’s surface with indeterminate depth) observations onto the 3D multihydrometeor mixing ration field.

The information sources for cloud/hydrometeor initialization include background short-range forecasts, satellite- and radar-based observations, and surface-based observations of cloud, visibility, and current weather. The RUC became the first NCEP operational model to introduce modification of initial cloud fields in its data assimilation in 2002.

Case studies and ongoing cycle (retrospective and real-time) testing will be conducted for assimilation of surface-based cloud observations into the RUC. The technique will be modified during this testing to include assimilation of visibility and current weather, both within the logical cloud variable. The local cloud variable will be subdivided into cloud versus hydrometeor components to allow for clearing or not clearing rain/snow hydrometeors from the cloud base to the surface based on the current weather observation. Most importantly, the assimilation for surface-based cloud observations will be combined with previously developed techniques for assimilation of radar reflectivity into the RUC hydrometeor fields.

A comprehensive RUC cloud/hydrometeor analysis including surface-based cloud observations and radar reflectivity assimilation will result in considerable improvement to RUC aviation-specific forecasts of ceiling and visibility, as well as in forecasts of clouds and precipitation important for all users. Implementation of this combined cloud/hydrometeor assimilation technique will be proposed for implementation into the operational RUC late in 2004.

Dr. Stan Benjamin will present recent results on these studies at the 2004 Annual Meeting of the American Meteorological Society in Seattle.

Operational Performance and Recent Improvements of the RUC 3DVAR – The RUC hourly update cycle utilizes a unified analysis framework encompassing data ingest and quality control routines, and interchangeable three-dimensional variational (3DVAR) and optimum interpolation (OI) based analysis solvers. Following several years of research and development, the operational version of the RUC (run at NCEP) switched from the OI analysis to the 3DVAR analysis on 27 May 2003. Several earlier versions of RUC 3DVAR were successfully implemented in FSL in real-time test mode. The 3DVAR-based RUC performs appropriately in operations, but there are several open problems, most of which are related to full variational solutions of moisture/ cloud fields. One of them is the assimilation of precipitable water data from satellites and ground-based GPS. An experimental version of the RUC 3DVAR with precipitable water data assimilation is being tested, but it is not ready for operational use. Currently, satellite radiances are not assimilated in the RUC 3DVAR, but development is under way to include the OPTRAN radiative forward and adjoint operators. A bias reduction method is under investigation to provide appropriate satellite radiance information. Development of radar data assimilation procedures is also underway, with a long term goal of utilizing reflectivity and radial velocity information to modify water vapor, hydrometeor, and velocity divergence fields. Initial work has focused on updating hydrometeor and water vapor fields, using a national composite maximum reflectivity product in conjunction with satellite data and surface cloud observations. In the present formulation, the radar-based updates occur within the outer loop of the moisture minimization, allowing for an iterative solution in concert with the in situ moisture observations. Real-time parallel tests at FSL indicate a modest improvement in short-term (3–6 hour) precipitation forecasts from this technique.

Dr. Stephen Weygandt will present a a status report on this research at the 2004 Annual Meeting of the American Meteorological Society in Seattle.

Modeling Entrainment and Boundary Layer Growth During a Bore Event – The goal of the International H2O Project (IHOP) field experiment in the Southern Great Plains of the U.S. was to obtain an improved characterization of the time varying three-dimensional water vapor field and to determine its importance in the understanding and prediction of convective processes. Understanding the role played by bores in initiating and maintaining nocturnal convection was one of the objectives of the project. Ground-based remote sensing instruments at the Homestead site in the Oklahoma Panhandle included the NCAR Integrated Sounding System and Multiple Antenna Profile, an Atmospheric Emitted Radiance Interferometer, FM-CW radar, Scanning Raman Lidar and aerosol backscatter lidar.

These instruments were complemented by the polarimetric SPOL and Dodge City WSR-88D radars and two research aircrafts equipped with the water vapor differential absorption lidar and surface mesonetwork recording temperature, dewpoint, and wind at 5-minute intervals. The most comprehensive set of observations ever collected on structure and dynamics of bores was probably gathered during IHOP. On 4 June 2002, two bores were observed at Homestead. FSL researchers analyzed the "second" bore which developed in the early morning on this day as a result of an interaction of a cold front with a stable boundary layer. This bore was well documented by the IHOP measurements. Numerical simulations with MM5 reproduced the event quite accurately and are used to study turbulence and boundary layer growth in the wake of the bore. This ongoing research will use the Center for Analysis and Prediction of Storms (CAPS) ARPS model at resolutions of the order of tens of meters to evaluate turbulence parameterizations versus simulations with explicitly resolved eddies.

Dr. Mariusz Pagowski will present a paper on this topic at the 2004 Annual Meeting of the American Meteorological Society in Seattle.

RUC Short-Range Ensemble Forecast System – Because of uncertainties in atmospheric model dynamics, physical parameterizations, and initial and boundary conditions, a single deterministic forecast has some degree of error. In this sense, a deterministic forecast is used merely as a reference to the forecast of the true atmospheric states. Therefore, statistical analysis of a sample of forecasts becomes a plausible approach for the improvement of numerical weather prediction. The task of finding the best way to generate such a sample of forecasts poses an important area of scientific challenge and development.

In collaboration with NCEP’s multimodel Short- Range Ensemble Forecast (SREF) initiative, FSL has developed a SREF system based on the Rapid Update Cycle (RUC) model, targeting both operations and research needs. The RUC forecast system is a NOAA operational weather prediction system. One of the unique features of the RUC is that its dynamical core is based on a hybrid potential temperature/sigma vertical coordinate. This feature will certainly add to the model diversity as far as a multi-model ensemble is concerned, and thus possibly increase the ensemble spread. The RUC SREF now runs twice daily with a total of 10 members. The SREF domain encompasses the entire North America including Alaska, Central America including the Caribbean Sea, the western Pacific including the Hawaiian Islands, and the western Atlantic. The forecast is run out to 60 hours with output every 3 hours.

The statistical verification scores show the RUC SREF forecasts compare well against Eta analysis and Eta 12-km operational runs, and yet the forecast spread calculations show that there is significant variability among the forecast members. Future development of RUC SREF is still in order. We plan to experiment with various initial perturbation methods while still using the Eta regional breeding method, and continue to develop posterior analysis, verification schemes, and probability forecast products. The development of an upgrade version for higher horizontal resolution and some research applications are also planned. FSL will work with NCEP/ EMC in running the RUC SREF as part of a retrospective test of current and prospective members of the NCEP SREF.

Dr. Chungu Lu will present this research at the 2004 Annual Meeting of the AMS in Seattle.

A Graphical User Interface to Prepare the Standard Initialization for the WRF Model – FSL has created a graphical user interface capable of accommodating researcher needs when using the Standard Initialization (SI). Since the SI is a necessary first step when using the WRF model, the GUI provides an easy method to prepare an otherwise quite complicated system. More than 100 users have so far downloaded the latest SI version 1.3.2 containing the GUI. These users have provided valuable feedback, which is used in updates. As other researchers use the GUI, we plan to continue requesting their feedback and use it to keep up with user needs and the latest software.

Paula McCaslin will present a paper on this topic at the 2004 Annual Meeting of the American Meteorological Society in Seattle.

Verification of the FSL Ensemble of Mesoscale Models Used for a Winter Weather Application – The LAPS group at FSL has built an ensemble of mesoscale models that runs in real time in support of field projects and demonstrations. One of these projects is sponsored by the Federal Highways Administration (FHWA) and is focused on winter weather. The FHWA Maintenance Decision Support System is an effort to tailor weather forecasts for the purposes of winter road maintenance. FSL generates the mesoscale model forecasts and transmits them to the NCAR/Research Applications Program, where they are used to make point forecasts along roadways. These point forecasts feed pavement temperature and chemical dilution algorithms (developed by the Cold Regions Research and Engineering Laboratory), which are used along with codified rules of practice (developed by MIT/Lincoln Labs) to automatically recommend timing and location for snow plowing and chemical applications. Last winter, the MDSS ensemble consisted of six members: three mesoscale models (MM5, RAMS, and WRF) with two larger-scale models (NCEP’s Eta and AVN) providing lateral boundary conditions. The models were run out to 27 hours to provide a 24-hour forecast service. The grid configuration, centered on the state of Iowa, is the same for all models. For this test the grid increment was 12 km, with no convective parameterizations because of the focus on winter weather. The execution schedule was driven by the update frequency of the NCEP models; thus, all six members were run four times per day, upon receipt of the NCEP model grids. All model runs were initialized with the same LAPS "hot start" diabatic initialization grids.

Following statistical evaluation of the models’ performance during the 2003 Demo, we have begun experiments with alternate configurations of the ensemble modeling system. The pertinent lessons learned were that the use of two different models (AVN and Eta) for lateral boundary conditions did not provide much diversity, the models did not provide much added value beyond 18 hours, the RAMS model routinely had large errors in precipitation and temperature, and the WRF model generated too much precipitation. In light of this experience, we have developed an alternative strategy to take better advantage of what these models do best, which is exploit more of the available observations (particularly radar and satellite) to improve precipitation forecasts in the range of 1–12 hours. This configuration consists of running MM5 and an improved version of WRF every hour, and using "time-lagged" ensembling techniques. For example, a 6-hour ensemble forecast uses the current 6-hour forecast, the previous 7-hour forecast, and the 8-hour forecast from the cycle before that, all forecasts valid at the same time. It is expected that such practice will reduce the cycle-to-cycle "shock" in the MDSS forecast services that was sometimes caused during the 2003 demo when the models updated.

Paul Schultz will present this research at the 2004 Annual Meeting of the American Meteorological Society in Seattle.

Comparisons Between Observations Made during NEAQS and Air Quality Forecasts from MM5 and WRF Chemistry Models – The 2002 New England Air Quality Study (NEAQS) was an intensive effort to investigate the chemical and meteorological factors that contribute to poor air quality in the New England region. The campaign combined efforts of numerous educational institutions as well as federal, state, and local agencies. Atmospheric chemical forecasts and retrospective simulations have been produced using the MM5/Chem and WRF/Chem numerical models, respectively. The forecasts using MM5/Chem took place between July and August of 2002 and coincided with the New England Air Quality Study. The retrospective simulations using WRF/Chem were conducted for the same region and time period. Initial analysis of the numerical model results indicates that both models are capable of producing the observed chemical structure of the lower troposphere. Differences between the observations and simulation results appear to be a product of the relatively large grid spacing used in the model as well as the surface emissions data. Future simulations using WRF/Chem will examine the use of smaller horizontal grid spacing and improved surface emission data. In addition, the impact of including the feedback between aerosols and shortwave radiation will be examined.

Dr. Steven Peckham will present this research at the 2004 Annual Meeting of the American Meteorological Society in Seattle.

Fully Coupled "Online" Chemistry within the WRF Model – The simulation and prediction of air quality is a complicated problem, involving both meteorological factors (such as wind speed and direction, turbulence, radiation, clouds, precipitation) and chemical processes (such as emissions, deposition, transformations). In the real atmosphere the chemical and physical processes are coupled. The chemistry can affect the meteorology, for example, through its effect on the radiation budget, as well as the interaction of aerosols with cloud condensation nuclei (CCN). Likewise, clouds and precipitation have a strong influence on chemical transformation and removal processes, and localized changes in the wind or turbulence fields continuously affect the chemical transport.

Until recently, the chemical processes in air quality modeling systems were usually treated independently of the meteorological model (i.e., "offline") except that the transport was driven by output from a meteorological model, typically available once or twice per hour. Because of this separation of meteorology and chemistry, there can be a loss of important information about atmospheric processes that quite often have a time scale of much less than the output time of the meteorological model, for example, wind speed and direction, rainfall, and cloud formation. This may be especially important in air quality prediction systems, in which horizontal grid sizes on the order of 1 km may be required. In addition, the feedback from the chemistry to the meteorology – which is neglected in "offline" approaches – may be much more important than previously thought.

Over the past few years, several research institutes have collaborated in the development of a new state-of-the-art Weather Research and Forecast (WRF) model (http://www.mmm.ucar.edu/wrf/users/document.html). WRF is nonhydrostatic, with several dynamic cores as well as many different choices for physical parameterizations to represent processes that cannot be resolved by the model. This allows the model to be applicable on many different scales. The dynamic cores include a fully mass- and scalar-conserving flux-form mass coordinate version, which represents a major improvement over commonly used nonhydrostatic models. Similar approaches have recently been implemented in the Operational Multiscale Environment Model with Grid Adaptivity (OMEGA) as well as the Japanese numerical weather prediction model. A fully conservative flux-form treatment of the equations of motion may be especially important for air quality applications. This makes the WRF model ideally suited to be the cornerstone for a next generation air quality prediction system. Fully coupled, "online" chemistry has been implemented into the WRF model. The resulting WRF/Chem model has been evaluated in comparison to MM5/Chem model with a testbed dataset.

Georg Grell will present a summary of statistical comparisons of atmospheric composition predicted by WRF/Chem and MM5/Chem at the 2004 Annual Meeting of the American Meteorological Society in Seattle.

Volcanic Ash Coordination Tool – FSL announces the successful delivery of a new realization of the FX-Collaborate (FXC) system, referred to as the Volcanic Ash Coordination Tool (VACT). This valuable technology has been installed at three locations: the Anchorage Center Weather Service Unit (ACWSU), the Alaska Aviation Weather Unit (AAWU), and the Alaska Volcano Observatory at the U.S. Geological Survey. The Aviation Division will work with these organizations as well as the NWS Alaska Region Headquarters to develop and demonstrate the VACT applied to a rules-based approach to collaboration in volcanic ash advisory preparation.

A first step in this multiagency effort was to train nine users of the VACT: three ACWSU forecasters, three AAWU forecasters and the MIC, and two USGS geophysicists at the AVO. Now AD staff will work to enhance and refine the VACT to support collaborative decision-making. A key motivating factor for developing the VACT was an analysis by Simpson et al. (2002) of the operational response to the eruption of Mt. Cleveland, Alaska, in 2001. They found that SIGMETs generated for the Anchorage Oceanic FIR (Forecast Information Region) and the Oakland FIR (which are adjacent) called for ash plume heights that were very different.

Future enhancements to VACT include additional satellite imagery displays including polar orbiters; prototype volcanic ash products developed by the FAA Product Development Team for Oceanic Weather; output generated by volcanic ash dispersion models; and radar observations. A software tool will also be developed that will enable Center Weather Service Units (CWSUs) to efficiently generate Center Weather Advisories for volcanic ash.

This work is funded by the NWS Alaska Region and the FAA Aviation Weather Research Program. For more information, contact Dennis Rodgers at Dennis.M.Rodgers@noaa.gov, or 303-497-6933.

Examples of GFE Use in Operations During 2003 – In the last few years, the Graphical Forecast Editor and supporting software (called the GFESuite or GFE) has become the primary tool that operational forecasters at the National Weather Service (NWS) offices use to create and edit their gridded forecast fields. The GFESuite provides a wide range of tools and capabilities for this purpose, but it has been left up to the NWS regions, individual forecast offices, and ultimately individual forecasters to decide what approach to take to generate and maintain these forecast fields. Along with maintaining an internally consistent gridded forecast database, forecasters must also consider the gridded forecasts generated by surrounding offices in order to maintain a level of spatial and temporal consistency over the large national domain.

Over the past several years, the FSL Evaluation Team has evaluated most aspects of the NWS modernization and new operational components including AWIPS. We have developed several evaluation metrics that have been successfully used to evaluate these changes and help direct future development activities. The goal of this study is to determine how the GFE is currently used operationally at NWS offices. Specific objectives are to find out what GFE components are being used, when the GFE is used, how the grids are initialized and modified, how does the GFE fit into the NWS operational framework, and what is the impact of the GFE on the forecast process?

Operational GFE computer logs have been the primary source of information for this study, along with interviews and observations conducted at some of the NWS offices. A survey has also been developed, but has not been administered at the offices. The GFE logs record status information, which tools and capabilities are used, and a time stamp indicating exactly when tools are used or when specific actions are performed. Week-long log "snapshots" were collected in 2003 from 5 randomly selected forecast offices at varied geographical locations and during a variety of weather conditions. These snapshots were examined in order to see the range and frequency of GFE use by a number of forecasters with a variety of forecast responsibilities.

Woody Roberts will present a summary of these results at the 2004 Annual Meeting of the American Meteorological Society in Seattle.

GPS Water Vapor Observation Errors – FSL has been carrying out research related to the observation errors associated with retrieving integrated or total atmospheric column precipitable water vapor (IPW) from Global Positioning System (GPS) signal propagation delays caused by the neutral atmosphere. Another aim of the project is to show how occasional discrepancies between operational National Weather Service (NWS) radiosonde soundings and GPS precipitable water estimates impact a numerical weather prediction model assimilating both measurements. Although GPS water vapor-observing systems provide no direct information about the vertical distribution of water vapor in the atmosphere, they have several advantages over other moisture sensing systems. Some of these advantages include high measurement accuracy; arbitrary temporal resolution; all weather operability (i.e. they provide data under conditions when other observations fail or provide degraded data); no requirement for calibration; high reliability; and low acquisition and maintenance costs. FSL has quantified the observation errors associated with estimating the GPS radio signal propagation delays caused by the neutral atmosphere, and retrieving integrated (total atmospheric column) precipitable water vapor from these delays.

Comparisons of GPS water vapor retrievals with other observing systems, especially radiosondes, have been carried out for 10 years. Though uncertainties exist in the absolute water vapor estimation accuracy of any one system, it is fairly certain that radiosondes and GPS are capable of providing total column precipitable water estimates with 1–2 mm level accuracy under ideal circumstances.

Seth Gutman (Seth.I.Gutman@noaa.gov) will present a status report on this research at the 2004 Annual Meeting of the American Meteorological Society in Seattle.