| Chandran Subramaniam | Richard T. Jesuroga |
| Cooperative Institute for Research | NOAA/ERL/FSL, R/E/FS5 |
| in Environmental Sciences | 325 Broadway |
| University of Colorado/NOAA | Boulder, CO 80303-3328 |
| Boulder, CO 80309-0216 | (303) 497-6936 |
| (303) 497-6015 |
SUMMARY
The National Weather Service (NWS) is in the midst of a multibillion dollar modernization program that will provide very timely, accurate mesoscale weather data in the Weather Service Forecast Offices (WSFOs) nationwide. While local area high-resolution weather data can be extremely valuable for emergency preparedness, emergency management agencies typically lack the meteorological expertise and computing capabilities to process advanced weather information in real time. Thus, in accordance with its modernization plans, the NWS has implemented the NOAA Emergency Management Weather Dissemination Project (EMWDP) within the Forecast Systems Laboratory (FSL). Experiments are conducted at FSL to determine the use of advanced meteorological information by local government operations.1,2 Local emergency preparedness agencies including sheriff and police departments, can greatly benefit from appropriate high resolution information about local weather hazards. Using the FSL-developed Experimental Dissemination System (EDS) four weather hazards of particular importance to emergency preparedness are emphasized: flash floods, fire danger, severe weather, and disruptive winter storms. The EDS uses high-resolution weather datasets produced by numerical weather models such as the Local Analysis and Prediction System (LAPS)3 and Mesoscale Analysis and Prediction System (MAPS)4,5 and the Weather Surveillance Radar 1988-Doppler (WSR-88D)6. LAPS and MAPS surface analyses provide hourly updates at 10 and 60 km resolution, respectively. WSR-88D 5-minute updates at 2-km resolution provides the user with mesoscale detail about rainfall distribution that is not available from raingauge networks.7,8 In addition to images of weather and geographic data, the system computes and displays assertions, weather characteristics related to spatial (regions) and temporal (periods) objects such as river basins and storm evolution.9
I. INTRODUCTION
In July 1991, Denver was the location of the costliest hail storm in the history of the United States. This storm was tracked as it moved roughly southeast through Boulder County, across central Denver and exited the southeast corner of Denver. It was an intense hailstorm that was depicted very well on the new computer systems at FSL and the Denver WSFO. Meteorologists at the Denver WSFO were at their stations tracking the storm and sending out watches and warnings in a timely and accurate manner. These forecasters assumed that the accurate predictions of the path of the storm and the timely broadcast of watches and warnings would be used by the local Emergency Management (EM) offices to effectively mitigate the danger.
However, that night and the following day the media provided graphic images of injured children being helped off a Ferris wheel at a local amusement park where they had been stuck during the storm. The investigation that followed found that although the warnings were timely and were received by the EM offices, an experienced dispatcher at a local office disregarded the warning and did not take the necessary actions.
This example of previous lack of communication between the WSFOs and the emergency management community illustrates the purpose of EMWDP. Its primary goal is to build a bridge between the weak link in the above scenario. Certainly the dispatcher would have given more weight to the warning if it had been more effectively conveyed using graphical displays rather than a simple text message, albeit a very accurate one. The maxim "a picture is worth a thousand words" is followed closely in this project. The EDS uses images, graphs, tables, text and even sounds to effectively provide the emergency manager with good weather information. The system is designed for easy use by providing a lot of information in an uncluttered manner, where the important information is highlighted and brought to the attention of the EM.
The system architecture is described next, followed by a brief history and the assessment of the prototype systems. Sections IV and V cover the current status of the system and future plans.
II. SYSTEM ARCHITECTURE
The EDS consists of a server and a set of Emergency Management Decision Support (EMDS) workstations located at local Emergency Operations Centers (EOCs). The server has four primary purposes: data interpreter, database, file server, and messaging center. The server, called the Community Server (CS), was designed for and run by the community. It is the sole link to the local WSFO and retrieves data from the local WSFO to be distributed to and displayed by the EMDS workstations. In addition to serving the general emergency preparedness users, the system will also serve expert users from organizations that have expertise in particular fields of weather hazard. For example, the Denver Urban Drainage and Flood Control District (UDFCD) will be an expert user with expertise in flash flooding. These users will use their expertise to interpret and add value to the weather data that reside in the CS. The improved product is then sent to the general users via the CS. At present, FSL simulates the role of the local WSFO as the weather data source.
EMWDP follows a simple technique to achieve its objective. It takes weather information from the local WSFO, geographic information from local sources and action rules from the EM's warning plans and combines them to generate a set of user-friendly graphical displays that are condensed, coherent information to the EM. The EMWDP development methodology is to develop and install experimental weather decision support systems at various evaluation sites, get feedback from real users, and repeat the cycle again until the best possible workstation is in place.
III. HISTORY
The first prototype of the general user system was designed, built, and installed in the Boulder EOC in the autumn of 1992. Five months later an improved version replaced the first. The second version was decommissioned in April 1994. The meteorological information and depiction methodology was studied and assessed for its potential use by university researchers and City and County users.
In general the system was received very well. The systems graphical user interface (GUI) was easy to learn and use. Since the system included the NWS Automation of Field Operations and Services (AFOS) forecasts, watches, and warnings that the users had depended on in the past, the transition from the old technology to the new was relatively smooth. One premise of the prototype systems was that images of the weather would be very useful; it was assessed that although images were useful, they were still limited. The users found that multi-modal displays which combined an image, graph, text and table (shown in Figure 1) to be much more valuable.10
IV. CURRENT STATUS
A. System Architecture
Presently the first demonstration version is being installed at the Boulder EOC, Denver UDFCD, Denver WSFO, and Colorado Department of Transportation (CDoT) in Denver. The CS is a PC running Windows NT Advanced Server. Communication to the remote display PCs (running Windows NT) is accommodated through NT's Remote Access Service (RAS) software and modems on a business telephone line.
A Monitor performs the messaging center function of the CS. The messaging center has three subfunctions: passing messages between the CS and remote systems, passing messages between the remote systems, and keeping track and statistics of the remote systems. We plan to use the statistical information to assess EMWDP's system and communication technology.
B. Functionality
Currently, the user can select from two spatial scales. The local scale, of primary interest to the user, covers most of the state of Colorado and parts of Wyoming, Kansas, Oklahoma, and New Mexico. It corresponds to the LAPS scale which is the main data source for the local scale. The national scale uses MAPS surface analysis and corresponds to the MAPS, scale which covers the continental United states.
The display workstation has two display modes: Picture mode and Probe mode. The Picture mode is used to view the weather and geographical information in various combinations. For example, a user interested in winter road hazards may wish to view an image of the snow accumulations with overlays of visibility, wind, and highways as shown in Figure 2. This will help users locate roads where the driving conditions are hazardous, so they can decide whether to close or reroute traffic from that particular road or road section.
The Probe mode, on the other hand, lists all weather conditions within the area of interest. The user selects a region of interest and the system calculates the weather characteristics for that region. The information is presented in a tabular format along with attribute information about the region such as the population affected and contact person (see Figure 3). For example, when the emergency managers see that a particular region such as a medical facility is located in an area of hazardous weather, they can automatically access information regarding the number of people to be evacuated, a contact person at the site, and a telephone number.
V. FUTURE UPGRADES
The system will have two additional modes in the near future: a Surveillance mode and a Text mode.
A metro scale will complement the local and national scales. This scale will primarily use high-resolution data from the WSR-88D radar and information from the users of the Experimental Dissemination System. Another important feature is an application toolkit that will allow users to:
Finally, we will be moving the CS into in the community, specifically, the Denver UDFCD office.
ACKNOWLEDGMENT
We are grateful to Nita Fullerton for the numerous suggestions that improved the quality of the paper.
REFERENCES
1. D. Small, "The Dissemination Project." FSL in Review Fiscal Year 1992, Ed. N. Fullerton, p. 5-7, (1993).
2. D. Small, "The Dissemination Project." FSL in Review Fiscal Year 1993, Ed. N. Fullerton, p. 5-9, (1994).
3. J.A. McGinley, S.C. Albers, and P.A.Stamus. "Validation of a Composite Convective Index as Defined by a Real-Time Local Analysis System." Weather and Forecasting, 6, 337-356, (1991)
4. S. Benjamin et al., "An Isentropic Three-Hourly Data Assimilation System Using ACARS Aircraft Observations." Monthly Weather Review, 119, 4:888-906, (1991).
5. R. Bleck, and S. Benjamin, "Regional Weather Prediction with a Model Combining Terrain-Following and Isentropic Coordinates. Part I: Model Description." Monthly Weather Review, 121, 6:1771-1785, (1993).
6. E. Rasmussen and J. Smith, "Real Time Precipitation Accumulation Estimation using the NCAR CP-2 Doppler Radar." Preprints, 24 Conf. on Radar Meteor., Tallhassee, Florida, March 27-31, 1989, American Meteorology Soc., Boston, Mass., 236-239, (1989).
7. J.K.Smith and R.C. Lipschutz, "Performance of the NEXRAD Precipitation Algorithms in Colorado during 1989." Preprint, Conference on Hydrometeorology, Kananaskis Park, Alta., Canada, October 22-26, 1990, p. 184-188 American Meteorological Soc., Boston, Mass, (1990).
8. M. Kelsch, "Estimating Maximum Convective Rainfall Rates for Radar-Derived Accumulations." Proc. Fourth Workshop on Operational Meteorology, Whistler, B.C., Canada, September 15-18, 1992, (1992).
9. S.Kerpedjiev, "MeteoAssert: Generation and Organization of Weather Assertions from Gridded Data," Proc, Tenth International Conference on Artificial Intelligence for Applications, San Antonio, Texas, March 1-4, 1994, p. 4275-4281, IEEE Computer Soc. Press, Los Alamitos, California (1994).
10. C. Subramaniam, and S. Kerpedjiev, "BRMS: Monitoring Basin Rainfall and Potential Flooding." Proc., International Emergency Management and Engineering Conference, Hollywood Beach, FL., April 19-21, Eds. J.D. Sullivan and S. Tufecki, p. 353-358, (1994).
FIGURES.
Figure 1. Detailed information about a single sub-basin (region). It provides the user with weather information affecting the region and prescribed actions to be taken.
Figure 2. Emergency Management Decision Support workstation display for winter road hazard.
Figure 3. Emergency Management Decision Support workstation in Probe and Picture Mode displaying weather and regional information.