Analysis and Modeling

This is the unofficial, WWW version of the VORTEX-95 Operations Plan. It may differ from the published operations plan which is available by 15 March from the National Severe Storms Laboratory.

(This chapter is extracted from an article that has been accepted for publication in the June 1994 Bulletin of the American Meteorological Society that gives an overview of VORTEX. For bibliographical information, see the article.)

To address the problem of predicting severe storms, scientists with the Center for Analysis and Prediction of Storms (CAPS), a NSF Science and Technology Center, are collaborating with investigators from VORTEX in an effort to begin making storm scale numerical weather predictions for parts of the Southern Plains utilizing very high-resolution (one to three kilometer) mesoscale numerical models. The collaboration between CAPS and the NOC will enable the exchange of new knowledge between researchers and local forecasters concerning short-term forecasting of the initiation and evolution of severe convective storms.

A hypothesis to be tested by CAPS during VORTEX is: Given a region of high-resolution data from a multitude of sensors, a data ingest and analysis system capable of producing gridded fields of the required variables, a state-of-the-art mesoscale/cloud numerical prediction model and an appropriate data assimilation scheme, then quantitative short-range predictions of convective storm initiation, location, type and evolution are possible. From this, the general objectives of the storm-scale numerical weather prediction experiment during VORTEX are:

(i) Determine the feasibility of high-resolution (meso-gamma) numerical prediction of convective storms and convective storm systems;

(ii) Verify the quality of numerical prediction of, e.g., convective storm structure, severe storm indicators and convective rainfall amounts;

(iii)Assess the value of the information content from a variety of mesoscale datasets to the analyses and forecasts; these include but are not limited to Doppler winds from WSR-88D and research radars, polarimetric radar data (for the initialization of precipitation fields), wind profiler and RASS data, commercial (ACARS) and research aircraft data, and detailed surface mesonetwork data.

The primary focus during the next two years will be on predicting the time, location, type and evolution of the initial convection, with an emphasis on storms that evolve into supercells. There also will be attempts to initialize the models while a storm is in progress and to predict the remaining evolution of the storm system.

The ingest, quality-control and analysis of all the conventional and research data sets will be accomplished by the Local Analysis and Prediction System (LAPS) developed by scientists at NOAA\qs Forecast Systems Laboratory (McGinley et al. 1991). Their analysis system is being adapted by CAPS and is centered over the VORTEX data collection region during the field experiment (see the accompanying map ). In addition to the usual datasets utilized by LAPS (see McGinley et al. 1991), we will be able to add information from the 108 station Oklahoma Mesonet (Crawford et al. 1993), Doppler wind data from eventually up to eight WSR-88D radars, wind and polarimetric data from NSSL\qs Cimarron radar, NSSL\qs Mobile-CLASS soundings, mobile mesonet, wind and thermodynamic data from the sensors associated with the ARM (Atmospheric Radiation Measurements) site in northern Oklahoma (Department of Energy 1990) and research aircraft data. The resulting analysis system will be known as O-LAPS (Oklahoma LAPS).

The models to be used for real-time storm-scale prediction experiments include the Straka Atmospheric Model (SAM) (Straka 1994), and the Advanced Regional Prediction System (ARPS) (Droegemeier et al. 1993). Both have shown skill at simulating convective-scale phenomena. VORTEX will provide the first opportunity to obtain the data base necessary to initialize all model variables with different values at each grid point over a state-scale domain.

The O-LAPS analyses, which could be produced as often as every five minutes, will be used in four-dimensional data assimilation (4DDA) procedures being developed for the two prediction models. While some forecasts will be attempted from unaltered O-LAPS fields, it is anticipated that the O-LAPS analyses will primarily serve as background fields for several retrieval and 4DDA techniques being developed at CAPS (e.g. Liou et al. 1990; Brewster 1991; Sun et al. 1991; Straka and Zrnic 1993; Shapiro 1993; Qiu and Xu 1993). These techniques, which range from simple nudging methods to full adjoint approaches, are intended to bring about four-dimensional consistency between the wind, temperature and buoyancy fields even if only single-Doppler radial winds are used as input. Another important component to the initialization phase is to test methods to incorporate polarimetric data to initialize cloud and precipitation fields.

In the first year of the experiment our focus will be to produce real-time O-LAPS analyses, and begin making real-time, and hind-sight storm-scale numerical predictions of convective storm formation (or lack of formation), type, evolution, and movement. Additional products that will be produced include mean 0-6 km winds, storm-relative helicity (based on modeled storm movement as well as mean lower tropospheric wind speeds), Bulk Richardson Number, Convective Available Potential Energy (CAPE), and Convective Inhibition (CIN). Forecasts maps of weather features such as jet streaks, fronts, boundaries, and areas of focused forcing (low-level convergence) also will be produced. Finally, operations will be run every day for a specified period during the experiment (one month) so that a first attempt at statistically determining if there is any skill in making high-resolution storm-scale numerical predictions will be possible. During the second year of the experiment, after the initial components of the prediction system are considered stable, we will begin testing model initialization cycles with single and dual Doppler radar data where and when they are available.

Return to Table of Contents