Chris Andrews
Models can represent the
physical state of an object or they can describe the
behavior and interrelationship of objects in a system. In
the geospatial technical world, a model is most commonly
thought of as a Digital Elevation Model (DEM) or a map
projection. DEMs are digital depictions of topography on a
part of the globe. Map projections allow three-dimensional
features on the Earth to be described, or modeled, on a
two-dimensional surface. Fields that use geospatial data
routinely employ more behaviorally descriptive models.
Statistical models describe the distribution of occurrence
of a set of events or observations. Algorithmic models use
tailored equations to represent the inputs, behavior, and
outputs of a system. Stochastic models use algorithms or
statistical formulae with random number generators to
create a “population” of results that may be statistically
analyzed and compared with real systems. The variety of
descriptive system models reflects the many different
applications to which they may be applied.
The Long Term Ecological
Research (LTER) project, funded by the National Science
Foundation, examines chemical, mineralogical, and
biological changes at 26 sites around the world. LTER
scientists and students collect data across broad spatial
areas and extended time frames. LTER studies rely on both
geographic models that represent the topography of part of
the surface of the Earth as well as behavioral models that
allow researchers to test hypotheses about how the
organisms and nutrients in study areas behave. A wide
variety of aerial and satellite data are used, for example
the Niwot Ridge LTER site contracted with a private vendor
to obtain high resolution DEM’s and digital orthophotos of
its Colorado research site. The topographical information
helps researchers understand variations in biological
diversity that may be caused by sun and wind shade due to
terrain variation in an alpine tundra ecosystem. The H.J.
Andrews Experimental Forest LTER team employed elevation
data as well as more detailed species data in a grid-based
simulation model to explore forest succession after clear
cutting and harvesting. (More information about LTER may
be found at http://lternet.edu/.)
Forest fire management also
presents a diverse research area that uses a variety of
spatial modeling tools. By the late 1990s enough satellite
imagery data had been collected so that scientists could
survey the historical record of imagery in places as
diverse as Spain and California to statistically model the
spatial distribution of forest fires in remote locales.
Researchers used orthorectified imagery to accurately map
the boundaries of burned out areas. Elevation and fuel
types are two of several remotely sensed parameters that
factor into the FARSITE simulation model, a deterministic
simulation model created by the United States Department
of Agriculture Forestry Service (USDA-FS) to study
wildfire propagation. FARSITE is an example of a single
wildfire model that outputs smoke production, scorch
height, fire surface area, and other fire characteristics
for each digitally simulated event.
The BlueSkyRains project is a
collaborative project developed by the USDA-FS to provide
real-time models and forecasts of forest fire smoke
dispersion for Northwestern U.S. regional wildfires.
BlueSkyRains combines weather, smoke dispersion, and fuel
consumption with meteorological information and a Web
interface to provide interactive forecast maps (http://www.blueskyrains.org).
The part of the geospatial
community that is focused on data collection and accurate
data representation provides a critical foundation for a
wide variety of projects that use the community’s data.
The static representational models of the Earth developed
in the technical world help to improve the accuracy of
behavioral models that are used to understand and forecast
events in the natural system. As data collection and
representation improve and become more fine-grained, the
potential opens up to allow behavioral models to have
greater detail and broader scope.
About the Author
Chris Andrews has been an
advocate for standardizing and extending GIS technology in
the past eight years, programming and listening to
customers in a variety of environments from private
industry to the Kennedy Space Center. Chris is currently
employed as a GIS Solution Architect at Idea Integration
in Denver, Colorado, and may be contacted at
[email protected].
Back
