ecovis

ecovis.org Virtual Ecosystems

One of the great habits taught in Engineering School is to first draw a picture of the problem -- a visual model. Take the provided input data, hang it onto a picture, and visualize what is needed to solve the problem at hand. The result of this process is almost always a clear and rapid solution. And, once you solve the problem, you draw a picture of the solution.

Years later, I ended up in Marine Ecology where visual models of ecosystems are rarely used. Instead, differential equations, statistical analyses, charts, and tables attempt to summarize ideas and/or field data. Since only a limited number of people are trained to understand this kind of presentation, the ecological decision makers are usually unable to comprehend the analysis. Furthermore, the use of numbers alone leaves the false impression that something is accurate. And, I find it impossible to visualize a complex ecosystem without a picture.

Take a minute to study the the little crabs as they follow the mouse around the screen (Internet Explorer only). Then try to picture a chart, table, or equation describing their dynamics. I think you'll see what I mean.

Scientific visualization is the science and art of rendering complex data sets and their contained interrelationships in graphical forms that are comprehensible to scientists and lay people alike. Virtual ecosystems are computer-generated visualizations of all or part of an ecosystem under study. Data used to generate the picture can be information brought in from the field, estimated numbers, or a combination of both. The resulting pictures can be accurate physical representations or highly abstract symbols. Usually, the visualization is a picture of a simulation.

Some of the ways in which computer-based visualization tools can impact ecological studies are:

	Improved understanding of the dynamic interactions between species and their environment.
	Clear communication of alternatives to the scientist and non-scientist.
	While a map alone may clearly indicate the extent of an environmental change, a virtual ecosystem can communicate the experience of being in the environment.

With virtual ecosystems, the computer is being used in essentially the same way as an experimental physicist uses laboratory apparatus to explore the structure of the physical world. As input parameters are changed, the scientist observes any changes in the ecosystem under study. One of the big advantages of virtual ecosystems is that they appeal to a strong aspect of human cognition - pictures and pattern recognition. As a result, virtual ecosystems help bring understanding to complex relationships. They help make big and complex problems comprehensible. Computer simulations of nature are powerful tools because one can observe the order that emerges from seemingly random or unrelated input data and iterative experimentation. The result is a pattern visualization that is easy to see.

Virtual studies in artificial life, complex systems, fractal geometry and other related fields have brought us a lot closer to understanding the fundamental order of nature. But, virtual ecosystems can easily lure one into a false sense of reality. Like statistical analysis, virtual ecosystems can rarely be validated against real data. The reason is that the system under study is so large and complex that meaningful validation is impossible. So, one relies on human logic or cognition -- a perilous substitute for reality.

Consider schooling fish. When a computer simulation and a real life system both show identical schooling behavior, it doesn't mean that both systems are driven by the same internal schooling mechanisms or instincts. An observation of general behavioral similarities does not validate the model. The fallacy that similarity equates to truth and reality is the trap that has ensnared many observers.

Nonetheless, computer simulations have great value as pattern generators. For it is in these experiments that striking common similarities between seemingly unrelated systems have emerged. It is these similar patterns that have the potential for generating theories (and perhaps fact) regarding the fundamental order of both living and non-living systems. And, computer models provide powerful precursors to actual field studies.

It is with this sense of both caution and hope that I wish to simulate real marine ecosystems in their 2D and 3D environments. In doing so, I hope to achieve a greater general understanding of these ecosystems. Exactly where I'm heading with my studies is still an open question -- which is the way I want it for the moment. But I do want to lay the groundwork by producing some interactive Web-based visualizations (Java applets) that demonstrate some key ideas that have been already established as well as some actual applications.

Much of this material is derived and paraphrased from books, lecture notes from courses I've taught, and many fine web sites. You can access the various subjects by using the menu pop-out at the left of the screen. This virtual ecosystems section is broken down into two areas:

Key concepts - Here I summarize ideas that are applicable to the computer simulation of marine ecosystems. The emphasis is on concepts applicable to the study of Marine Ecology and not the mathematics. You can get more detail for each subject (including the mathematics) by accessing the sites listed in the Additional Resources section at the bottom of each page.The list of these concept pages is contained in the pop-out menu to the left of the screen. Most pages have (or will have) java applets to demonstrate the ideas.

Applications - Here I provide a growing number of pages that demonstrate real models and applications developed by others. Where possible, each application will be visualized with a Java applet.

I hope that you will experiment with or add to the ideas in these pages. If you're interested in playing right now, take a look at the fish schooling Java applet.