|
Using Regional Atmospheric Models for
Commercial Applications
Roni Avissar and Craig Tremback
Regional atmospheric models are used to predict
or diagnose atmospheric variables relevant to weather, climate,
and air pollution. Regional models provide much more detailed
simulation of atmospheric variables than do global models. For
example, the basic computational unit in global models typically
represents an area on the order of 10,000-100,000 km2. By comparison,
the basic computational unit in regional models typically represents
an area of 100-1,000 km2. As the price-performance ratios of
computer systems have continued to improve, the basic computational
unit in regional models is routinely approaching 1-10 km2, even
for operational applications. So while a state the size of Colorado
would be represented by one or two points in global atmospheric
models, Colorado would be represented by a grid of 1,000 to
10,000 or more points in regional models, offering much more
detail on the variables that it simulates. This capability is
important in regions characterized by a large variability, as
is the case for Colorado where the topography varies significantly
within small distances.
Regional Atmospheric Modeling System (RAMS)
The Regional Atmospheric Modeling System (RAMS), available from
ATMET, LLC (www.atmet.com), is a multipurpose, numerical prediction
model designed to simulate atmospheric processes spanning in
scale from the hemisphere to turbulence around buildings and
in tree canopies. RAMS is a three-dimensional model based on
the fundamental equations describing the conservation of mass,
energy, motion, water in its three phases, and any gaseous or
aerosol material. It integrates these equations in time starting
from the three-dimensional fields of temperature, humidity,
winds, and pressure, which are derived from the global network
of atmospheric observations and large-scale model information
available from the National Centers for Environmental Prediction
(NCEP).
The ability of RAMS to produce high-resolution weather forecasts
and simulations can be attractive for many commercial applications,
as explained in the following two examples.
Electrical Power Demand and Production
Electrical power needs are strongly dependent on the amount
of energy needed to control the temperature and humidity in
public and private buildings. Consequently, to anticipate demand
better, power companies need accurate, high-resolution forecasts
of various meteorological variables, such as air temperature,
humidity, cloud cover, and precipitation. To manage power production
and planning activities better, power companies want short-term
(i.e., 24 hours) through seasonal forecasts.
RAMS has successfully provided a 1- to 10-day weather forecast
for GPU, Inc.1 The simulated domain for this project used three
levels of numerical nested grid, with the highest resolution
at 4 km, to provide high-resolution forecasts of temperature
and humidity to load forecasters. These forecasts were compared
to actual temperature and humidity observations each day to
verify the accuracy of RAMS predictions, and the verification
products were automatically posted to a Web site for project
researchers’ use (Figures 1a, b and Figure 2).
Power utility companies have additional needs for high-resolution
weather forecasts. For example, anticipating the need to dispatch
emergency maintenance teams to weather-damaged areas of an electrical
network is a valuable management tool. However, prediction of
thunderstorms is challenging because their formation is driven
by chaotic motions that can be affected by many atmospheric
and land characteristics. For example, atmospheric humidity
and stability, winds, topography, and abrupt change of landscape
(e.g., the land-to-sea change) affect the timing and location
of such storms. RAMS has demonstrated its ability to simulate
such complex weather events by correctly predicting the timing
and location of storm cells as squall lines moved through topographically
varied areas.
Emergency Response to Catastrophic Pollution
Accidents or Terrorism
The terrorist attacks on September 11, 2001, the bombing at
the Atlanta Olympic Games, the anthrax releases, and the continuing
concern with terrorist activities around the world have caused
the U.S. military and government agencies to increase funding
substantially for the development of an emergency response capability
that will provide decision makers with accurate data during
such emergencies. However, the possibility of accidental release
of toxic materials has existed for many years. Scenarios of
concern include radioactive releases from nuclear power plants
and toxic releases from somewhat specialized activities, such
as the launch activities at Cape Canaveral Air Force Station
(CCAFS) and Kennedy Space Center (KSC). Regional weather forecasting
models, in conjunction with pollution dispersion models, can
provide important information for decision makers during such
events.
The Eastern Range Dispersion Assessment System (ERDAS) RAMS
(Tremback et al., 1994) was developed under funding by the U.S.
Air Force and NASA. The primary mission of ERDAS is to provide
improved forecasts of dispersion from sources at KSC/CCAFS for
up to 24 hours in an emergency response mode. A variety of pre-determined
release scenarios will allow initial estimates within several
minutes of event notification and specification of strength
of the release. Examples of the potential release scenarios
include the following:
-- Catastrophic launch accidents;
-- Spills of toxic chemicals/fuel at launch pads and storage
facilities;
-- Plumes from solid rocket motor tests at launch pads;
-- Toxic chemical venting from storage facilities;
-- Exhaust ground clouds from normal launches of Titan, Atlas,
and Delta rockets and space shuttle vehicles.
The secondary missions for ERDAS include planning
support for proposed missions (launches several hours into the
future) and scheduling operations in which the potential exists
for an accidental release.
ERDAS is configured to use the forecast weather information
from the RAMS model available within the ERDAS production cycle.
The system can also go offline to calculate meteorology for
hypothetical situations or alternately can use previously calculated
RAMS fields available from archives. These simulated meteorological
fields, in turn, could be used to conduct research on the impact
of a variety of activities, such as new system test firings,
launch vehicle abort clouds, controlled vegetative burning,
pesticide applications, evaluation of existing regulatory and
emergency response models, defense against legal claims from
air contaminants, and accident reconstruction. ERDAS could also
have applications as part of a training and education system
for KSC/CCAFS personnel.
The first version of ERDAS was installed at CCAFS in 1994. Following
a successful, two-year evaluation period and with the growing
capabilities of computer systems through the 1990s, additional
projects enhanced the overall system and provided additional
meteorological features, including the direct simulation and
forecasting of thunderstorms.
The RAMS’ grid configuration geography around KSC and the immediate
KSC/CCAS region encompasses a rugged area of coastlines near
the launch site. The grid configuration becomes even more complex
when considering the land use characteristics of forest, wetlands,
agriculture, and urban areas. Local wind circulations developing
in this landscape, which can trigger the formation of thunderstorms
under appropriate conditions, can be properly predicted only
if a very high resolution model is applied to this region. Wind
speeds and directions can change dramatically over a very small
area. An initial meteorological forecast must be accurate if
the dispersion of toxic materials is to be predicted correctly
(Figures 3a and 3b, page 22).
While ERDAS was developed specifically for the activities at
KSC/CCAFS, it was designed for easy customization to other scenarios
where the release of toxic airborne materials is a concern.
Continued improvement of computer price/performance ratios is
allowing this type of software system to be hosted on hardware
that costs a fraction of comparable systems from a decade ago.
Continued improvement in regional meteorological models has
dramatically increased the resolutions at which reliable weather
forecasts can be made.
Summary
Regional atmospheric models have many applications in predicting
or diagnosing high-resolution weather, climate, and air pollution
events. The two examples given here illustrate the potential
of this powerful tool. They also emphasize the need to use spatial
distributions of land characteristics, which can best be estimated
from satellite imagery. Furthermore, new techniques permitting
the assimilation of meteorological fields, such as cloud cover
derived from satellite imagery, are promising to improve the
forecasting capability of such models and are currently under
development.
1GPU, Inc., was a power company that operated in eastern Pennsylvania
and merged in 2001 with First Energy, which operates in part
of Ohio, Pennsylvania, and New Jersey (see www.gpu.com and www.firstenergycorp.com).
References
Tremback, C.J., W.A. Lyons, W.P. Thorson, and R.L. Walko, 1994.
An emergency response and local weather forecasting software
system. Preprints of the Eighth Joint Conference on the Applications
of Air Pollution Meteorology, Nashville, TN, 23-28 Jan 1994.
Acknowledgments
Funding for this research has been partly provided by NASA
Grant Number NAG5-11402.
About the Authors
Roni Avissar is the W.H. Gardner Jr. Professor and Chair
of the Department of Civil and Environmental Engineering at
Duke University. He has used numerical models of the atmosphere
and the ocean for various research projects at the different
scales (micro- to global scale), with a particular interest
in convective precipitation triggered by land-atmosphere interactions.
Craig J. Tremback is a Senior Scientist and the President of
ATMET, LLC (Atmospheric, Meteorological and Environmental Technologies)
in Boulder, Colorado. He is one of the main architects and developers
of RAMS and he has used this model for many applications, including
regional climate simulations, providing meteorological simulation
results to photochemical models, development of an emergency
response software system, and several operational forecasting
applications.
Back
|
