Although models are now routinely used for addressing environmental problem
s, both in research and management applications, the problem of obtaining t
he required parameters remains a major challenge. An attractive procedure f
or obtaining model parameters in recent years has been through inverse mode
ling. This approach involves obtaining easily measurable variables (model o
utput), and using this information to estimate a set of unknown model param
eters. Inverse procedures usually require optimization of an objective func
tion. In this study we emulate the behavior of a colony of ants to achieve
this optimization. The method uses the fact that ants are capable of findin
g the shortest path from a food source to their nest by depositing a trail
of pheromone during their walk. Results obtained with the ant colony parame
ter optimization method are very promising; in eight different applications
we were able to estimate the 'true' parameters to within a few percent. On
e such study is reported in this paper plus an application to estimating hy
draulic parameters in a lysimeter experiment. Despite the encouraging resul
ts obtained thus far, further improvements could still be made in the param
eterization of the ant colony optimization for application to estimation of
unsaturated flow and transport parameters. (C) 2001 Elsevier Science Ltd.
All rights reserved.