1996 Summer Olympic Games Support

Overview

One of the functions of the Air Resources Laboratory (ARL) of the National Oceanic and Atmospheric Administration (NOAA) is to provide meteorological services and related research to NOAA and to other Federal agencies, in order to predict the consequences of atmospheric releases of radioactivity and other potentially harmful materials. In addition, the Federal Radiological Emergency Response Plan (FRERP), which outlines a coordinated response by Federal agencies to a peacetime radiological emergency, requires the Department of Commerce (DOC) to provide, among other things, weather forecasts and dispersion estimates to support emergency response activities. These capabilities reside within NOAA, specifically within NCEP and ARL. For radiological sabotage or terrorism the Atomic Energy Act directs the Federal Bureau of Investigation (FBI) to investigate any violations of the Act and to locate and restore any nuclear weapon, device, or material to its rightful custodians. A formal FBI agreement with the Department of Energy (DOE) provides access to DOE's Nuclear Emergency Search Team (NEST) to support the process of locating and removing a nuclear device. The NEST, in turn, calls upon NOAA to provide meteorological support in the form of observational data and dispersion forecasts.

During the 1996 Summer Olympic Games, held in Atlanta, GA, ARL provided meteorological data and dispersion model support to the NEST in the event of nuclear terrorism. The National Weather Service's (NWS) ambitious effort to provide the "world's best meteorological science, skill, service, and technology" during the Summer Olympic Games gave ARL a unique opportunity to use highly-detailed information from NCEP and the NWS's Olympic Weather Support Office (OWSO) in Peachtree City, GA, within it's own products.

Research Summary

ARL regularly retrieves gridded forecast meteorological data directly from NCEP's computer facility located in Suitland, MD, to maintain a state of emergency readiness. During the 1996 Summer Olympic Games several meso-scale modeling tools were available to support the OWSO's operations that are normally not available in an operational forecasting environment. Arrangements were made in advance with the OWSO for ARL to routinely access these special datasets to support its emergency response role during the Games. Table 1 lists most of the gridded model datasets available at ARL during July and August of 1996. During the Olympics ARL was able to access 10 km Eta model data from NCEP and 2 km and 8 km RAMS model data, as well as 8 km LAPS model data from the Peachtree City, GA, Olympic Weather Support Office (OWSO).  These are shown in red in Table 1.

The hourly 8 km LAPS model output provided a diagnostic meteorological dataset for dispersion analysis in addition to the forecast datasets. In addition to providing access to the Real-time Environmental Applications and Display sYstem (READY) to the NEST personnel, meteorologists from ARL's Special Operations and Research Division (SORD) constructed high resolution upper air wind and temperature profiles over Peachtree City, Georgia for input into transport and dispersion models run by the NEST at Sandia National Laboratories. The profiles consisted of winds and temperatures at 100 meter intervals for the first 1000 meters and every 1000 meters to 10 km. These high resolution profiles were created from the Peachtree City observed soundings and the 10 km and 40 km Eta model data sets available within READY.

The HYsplit_4 (HYbrid Single-Particle Lagrangian Integrated Trajectory; Draxler, 1992,1996) model was run operationally during the Games using many of the datasets shown in Table 1.

Dispersion model simulations were conducted for July 27, 1996, with a hypothetical source (uniform between 10 and 500 m AGL) releasing 1 unit of mass over 1 hour beginning at 06 UTC in Atlanta (ATL), GA. The products shown below are an example of the products that would have been available to the NEST during an actual emergency.

The 40 km (Fig. 1b) and 10 km (Fig. 1d) Eta model runs were similar with the plume moving initially toward the southeast and then turning northeast when the Eta predicted the winds to become southwesterly at around 15 UTC. At about 18 UTC the 10 km Eta developed higher wind speeds by about 3-5 ms-2 than the LAPS (Fig. 1a) at levels below 900 hPa, which can account for the large spread of pollutant to the northeast over South Carolina. The 10 km Eta, however, did a better job than the 40 km Eta at modeling the higher exposures (10-8 contour) to the east of Atlanta.

The 8 km RAMS (Fig. 1c) model run was similar in moving the pollutant to the southeast during the first 6 hours resulting in an exposure maximum of 2.2x10-7 sm-3, of the same magnitude and location as the LAPS run. However, beyond the first 6 hours, RAMS moved the pollutant to the southwest, west, and finally, northwest of ATL. A study of the RAMS model results indicates that the model did not properly resolve the winds around the stationary front, located to the north of ATL in the LAPS run, and instead generated north-northeasterly flow over most of Georgia by 12 UTC. By 18 UTC, RAMS correctly developed a southerly wind causing the pollutant to move north again to the west of Atlanta.

a) c)

b) d)

Figure 1. Hysplit_4 simulations of an 18-hour surface to 500 m layer-average exposure (sm-3) based on a 1 hour, surface to 250 m release from ATL, GA, at 06 UTC on July 27. Input meteorological data from the a) LAPS, b) Eta 40 km, c) RAMS, and d) Eta 10 km models.

Click HERE for a VIS5D simulation of this case (320 Kb)

The different results found among these models presents a challenge to emergency personnel who need to provide guidance quickly to local, State and Federal agencies during an emergency. The use of meso-scale models in an operational environment requires that the atmospheric modeler know when better results are more likely with a synoptically-driven model. In this example, the driving force acting on the wind field was a stationary front and was best handled by a synoptic-scale model. Whereas, a typical nuclear power plant, situated in a river valley or along a bay, where terrain- induced flow is important, may best be handled with a finer resolution meso-scale model. Having a variety of meteorological datasets at hand does not always make the emergency modeler's work easier, especially when they differ greatly, but can provide a degree of confidence in the forecast when all are similar.

List of Publications

Rolph, G.D., J.T. McQueen, J.B. Sanders, and D.A. Soule, 1997: The use of NWS Summer Olympic Games resources in NOAA's environmental emergency response program. Preprints 13th Intl. Conf. on IIPS for Metr., Ocean., and Hydr. (Long Beach, CA), AMS, 76-79.