1.0.2 What are the pitfalls in simulation?

Before describing the use of simulation, we want to point out some of its potential problems. Like all computer programs, simulations are subject to the old (?) adage, "garbage in, garbage out." Before a computer can yield valid results, the program must be correct and the input data must be accurate. It is wise to test your simulation on easily verifiable data before running a simulation.

It is safe to say that program accuracy will never be achieved in any complete ecological simulation. Simply put, there are always gaps in our understanding of nature, therefore incompletness is built into any simulation. We should be dubious of simulation results without some empirical verification of these results.

The outcome of a simulation is dependent on the random numbers used in the simulation. The simulation will give only one possible result. To find trends it is necessary to run many simulation to determine the probable behavior of an ecosystem.

For a more complete description of simulation validation see [Bart 95].

[Bart 95] Bart, J. 1995, "Acceptance Criteria for Using Individual-Based Models to Make Management Decisions," Ecological Applications, 5:2