AN APPROACH TO ADVANCED ECOSYSTEM MODELING 
USING VECTOR AND RASTER DATA 
Stanley A. Morain 
Department of Geography and Earth Data Analysis Center 
and 
Amelia Budge 
Earth Data Analysis Center 
University of New Mexico 
Albuquerque, New Mexico, USA 87131-6031 
Commission VII, Working Group 5 
KEY WORDS: Key Words: Ecosystems, Relational Databases, Object Databases, Modeling, Self-adaptation, satellite 
measurements, genetic algorithms, fuzzy rules, virtual reality, visualization 
ABSTRACT 
Complexity science asserts that ecosystems organize into delicate relationships that teeter on the edge of Order and Chaos. 
This edge is maintained and governed by imprecise “rules” of nature (i.e., it is fuzzy), but it can be modeled. The challenge 
in the next few decades will be to learn how to apply spectral, spatial, and temporal technologies so as to mimic the self- 
adaptive behavior of natural systems. An even greater challenge will be learning how to integrate gene sequencing data, 
genetic network functions, cellular automata, and advance computational strategies into cohesive data structures for model 
development. Advanced ecosystem modeling should move in directions that develop genetic algorithms that: (a) emulate 
self-adaptive behavior through rule-based, stimulus-response mechanisms; (b) are capable of "learning" through artificial 
neural networks; and, (c) permit evolutionary changes to be animated and visualized in a virtual reality framework. This paper 
conceptualizes a genetic algorithm that sandwiches a relational database of ecosystem and environmental parameters between 
two object databases that integrate plant genetics and atmospheric conditions with ecosystem responses. 
INTRODUCTION example, the strategy that aims toward desired future 
conditions), modelers should provide managers with tools 
that simulate evolution in natural systems. 
One of the biggest challenges in biogeography is to link 
temporal and spatial scales that range from cellular and This paper conceptualizes a genetic algorithm for 
molecular on the one hand, to whole landscapes on the monitoring ecosystem change (specifically changes that 
other. Nature progresses in complexity from bottom to can be credited to human impacts and climate warming), 
top, but has evolved from stimuli operating from top to and that, in turn, affect the speed and directions of plant 
bottom. The first level of complexity begins with DNA species migration and distribution. The concept employs 
chemistry and spreads through many functions that satellite measurements for three purposes: (a) as a source 
characterize individuals and species. The second level of "continuous" data that may contain fractal dimensions 
involves interactions that lead to communities that then and n" order periodicities like the type recently described 
interact with the environment and feedback to level one by Stolum (1996); (b) as a means for change detection at 
processes. Finally, the third level of complexity involves the Earth's surface; and 9 as part of the input for 
atmospheric, geophysical, and hydrospheric processes that articulating fuzzy rules directing self-adaptive behavior. 
trigger a species' adaptive responses to these external Other in-situ measurements are needed to develop eco- 
stimuli. physiological rules; and, both types of rules need to be 
; combined into object and relational databases that can 
The need to develop self-adaptive ecosystem models will interact and adapt on their own to stimuli. 
increase as human impacts on natural environments 
intensify. Human activity sets in motion accelerated ENABLING TECHNOLOGIES 
feedback loops that govern ecosystem reactions and 
adaptations that may never be fully understood. Homo The construct adopted here is referred to as a genetic 
sapiens cannot interpret or predict the importance of their algorithm to emphasize that it integrates genetic 
own economic or political imprints on landscapes because information with an approach for monitoring observable 
they cannot separate their perceptions of landscapes from landscape changes. It provides a framework for 
the biogeochemical forces that drive those systems. integrating retrospective data ...and future data...into an 
Humans are aware of their place in nature, but do not archive for more realistic resource management. Object 
understand how systems evolve in response to their and relational databases would be assembled from 
presence. Instead of resource management schemes distributed, electronic sources, and integrated Int 
whose outcomes are biased by human expectations (as for functional, interactive data sets to which dynamic rules are 
504 
International Archives of Photogrammetry and Remote Sensing. Vol. XXXI, Part B7. Vienna 1996