Towards Monitoring of Long-term Trends in Terrestrial Ecosystems

Ecological monitoring is the purposeful observation, over time, of ecological processes in relation to stress. It differs from biological monitoring in that ecological monitoring does not consider the biota to be a surrogate filter to be analysed for contaminants, but rather has changes in the biotic processes as its focal point for observation of response to stress. Ecological monitoring methods aimed at detecting subtle or slow changes in ecological structure or function usually cannot be based on simple repetition of an arbitrarily chosen field measurement. An optimum method should be deliberately designed to be ecologically appropriate, statistically credible, and cost-efficient. Ecologically appropriate methods should consider the ecological processes that are most likely to respond to the stress of concern, so that relatively simple and well-defined measurements can be used. Statistical credibility requires that both Type I and Type II errors be addressed; Type I error (a false declaration of impact when none exists) and Type II error (a false declaration that no change has taken place or that an observed change is random) are about equally important in a monitoring context. Therefore, these error rates should probably be equal. Furthermore, the error rates should reflect the large inherent variability in undomesticated situations; the optimum may be 10%, rather than the traditional 5% or 1% for controlled experiments and observations.