|
“A complex field such as oceanography tends to be subject to
two opposite approaches. The first
is the descriptive, in which several quantities are measured simultaneously
and their inter-relationships derived by some sort of statistical method. The other approach is the synthetic one,
in which a few reasonable although perhaps oversimplified assumptions are
laid down, these serving as a basis for mathematical derivation of
relationships.”
-
Gordon A. Riley (1946)
With these words, Gordon Riley began his revolutionary development of
the first numerical, dynamic, mechanistic model of a marine ecosystem
– in his case, the annual phytoplankton dynamics on Georges
Bank. Since that time,
the field of marine ecosystem modeling grew with a particular flowering
during the 1970s with the development of systems ecology by HT Odum and colleagues, resulting in many still classic
models designed to understand system structure and function. The field exploded in the mid-1990s with
the development of powerful desktop computers and has come into its own as
a widely-used tool in both research and management of marine ecosystems.
Modeling is the process of building a mathematical abstraction of an
actual system. The purpose in ecological modeling is often systems
analysis, an approach that attempts to understand fundamental attributes of
ecosystems with the goal of predicting behavior. Or a model could be purely
descriptive, a way of discovering the mechanisms that explain the structure
or behavior of an ecosystem.
Some advantages include:
·
Models provide a conceptual framework within
which specific studies and experimental designs may be evaluated.
·
The process of modeling often identifies
areas where information or data is inadequate or entirely lacking.
·
Simulation analysis can determine sensitive
or controlling parameters that might otherwise be overlooked, thought
unimportant, or completely ignored.
·
Models provide a tool to aid both current
and future research by directing the effort to fill data gaps and providing
better understanding of controlling parameters and interactions.
·
Modeling experiments (often referred to as
model scenarios) allow "what-if" analyses of system behavior.
·
Models serve as an important
hypothesis-generation tool.
|
