Computing Clout

Recent advances in PC-based software capabilities give environmental managers better tools for dealing with an increasing range of problems

• By Erwin T. Prater
• Apr 01, 2005

Environmental managers face a growing number of responsibilities, from air-quality compliance issues to emergency response planning and accident mitigation. Increased concerns about terrorism have created additional responsibilities. Personal-computer (PC)-based air dispersion, fire/explosion, and accidental release software has evolved rapidly during recent years. Advances in computer hardware and software assist environmental managers in successfully managing their growing responsibilities. Recent innovations in managing and leveraging computing resources will allow environmental managers to efficiently tackle even more complicated problems in the future. This article describes the recent developments in model capabilities for environmental managers and suggests a general direction of future capabilities.

Growing Responsibilities, But Better Tools
Despite environmental managers' growing responsibilities related to environmental compliance, emergency response planning, and terrorism threats, increasingly sophisticated and powerful modeling tools are allowing them to address these challenges. Increases in computational technology now deliver "horsepower" that would have been only dreamed about 10 years ago with a typical laptop computer. Using current laptops and handheld pocket PCs, emergency managers can simulate complex problems such as explosions, fires, and accidental releases in order to predict the impacts on affected communities.

Accidental Release Modeling
Environmental managers may also need to predict the impact of potential accidental releases of an airborne toxic chemical. Recent advances in computer hardware have made this easy to accomplish in real time. Emergency managers can use this information to establish evacuation zones, determine level of protective dress for entering potentially contaminated areas, and aid post-accident analysis.

Explosion Modeling
Advances in computer processing capabilities now allow emergency managers and planners to apply advanced explosion and damage prediction programs, such as the HEXDAM (High-order EXplosion Damage Assessment Model) and VEXDAM (Vapor EXplosion Damage Assessment Model), to predict the damage resulting from an explosive blast wave on user-specified structures. Originating from work with nuclear blast research, HEXDAM is designed for high-order explosions (such as bomb blasts), while VEXDAM is designed for vapor cloud explosions. These programs allow the user to fully specify the size, shape, and orientation of structures and are particularly useful to emergency planners and managers charged with the responsibility of protecting people and buildings from the effects of explosions.

Fire Modeling
In some situations, emergency managers are concerned about a potential fire resulting from a cloud of flammable gas. In such a case, the emergency manager needs to predict radiation levels and distances for estimated gas concentration levels within the flammability limits of the particular gas in question. For these threats, a fire model that predicts thermal radiation hazards and an air-dispersion model that predicts flammable gas concentrations would be appropriate.

The Gas Technology Institute's (GTI) LNGFIRE3 is an example of a model that can perform these tasks. It can predict the thermal radiation from a burning pool of liquefied natural gas (LNG). The model generally runs in less than a minute, producing the immediate results needed for emergency response and response planning. The LNGFIRE3's output includes a map of the predicted thermal radiation pattern and a plot that shows the thermal radiation dose as a function of distance from the fire.

Regulatory Air-quality Modeling
Models suitable for regulatory air-quality applications have also evolved rapidly during recent years. Earlier generation models presented the Earth as flat and used hourly meteorological data consisting of a single wind speed and wind direction reading at a single vertical level from one weather station. Those models were enhanced to account for terrain height variations and vertical variations of turbulence. Current computational capabilities allow model results to be displayed in three-dimensional GIS (geographical information system) displays. AERMOD, a new generation air-quality modeling system used to support both regulatory and non-regulatory modeling requirements worldwide, can handle thousands of different air pollution sources and display results in a user-friendly display. Depending upon the complexity of the scenario being modeled, a single AERMOD simulation can be run in minutes or hours on a laptop PC.

More Complex Models for More Complex Problems
Although poised to be a widely-used regulatory model in the future, AERMOD still uses only annual weather information from one weather station. More sophisticated models incorporate data from multiple weather stations, multiple locations, and multiple time periods. Many of these models can use, or benefit from, predictions of three-dimensional temperature, wind, and precipitation from complicated computational fluid dynamics (CFD) weather prediction models. In the mid-1990s, these CFD models required mainframes or powerful workstations and simulations took days to complete. Now, these simulations can run on powerful PCs, including laptops, sometimes within several hours. This allows very complex behavior of atmospheric pollutants to be simulated in a relatively short amount of time. These models are especially useful in areas with very rough terrain or in situations where wind shifts and changing weather conditions are a factor. One model in this class that is becoming increasingly popular is the CALPUFF model, which can simulate the effects of time and space-varying meteorological conditions on pollutant transport, transformation, and removal.

The Next Step: Harnessing the Power of Multiple Computers
CALPUFF and other complex models require significant computational resources that challenge even high-performance PCs. To address this challenge, computers can be connected and coordinated to run in tandem. Computer clustering helps to reduce the execution time of large computing tasks through parallel processing. Clusters of similar computers (e.g., Linux clusters) can be synchronized to achieve supercomputer-level performance at a low cost. Clusters are frequently used in mesoscale meteorology modeling applications. Until recently, these capabilities were only available to universities and other public, research-based organizations. Now clustering techniques are becoming available in the private sector. The application of clustering techniques to currently available dispersion models greatly reduces execution time for large models with minimal additional capital investment. CALPUFF's unique computer code design lends itself particularly well to clustered computing approaches.

Often, analysts must run CALPUFF for long periods (e.g., five model-years), using a number of different scenarios that involve changes to stack height, assumed emission controls, etc. A single CALPUFF simulation can take several days to run, causing a significant bottleneck in the air-quality analysis, especially when many separate simulations are needed. Running CALPUFF on a cluster greatly reduces the bottleneck. In a recent benchmarking test that involved running CALPUFF on clustered computers, a single CALPUFF simulation that took 6 days to run on a single PC ran in approximately 13 hours on a cluster of 12 computers. Adding 12 more computers to the cluster reduced the run time to approximately seven hours. Cluster computing would be of interest to private businesses and regulatory agencies that have a sufficient number of computers to form a cluster.

Conclusion
Environmental managers face a growing number of responsibilities. Increased concerns about terrorism have brought additional responsibilities. Computer-based air dispersion, explosion, and accidental release models have evolved rapidly during the past few years. Recent innovations in more effectively managing computing resources (such as parallel computing and clustered computing) will allow environmental managers to tackle even more complicated problems efficiently now and in the coming years.

This article originally appeared in the 04/01/2005 issue of Environmental Protection.