Predictive models for identifying potentially valuable wetlands

Abstract

Several jurisdictions in North America have developed quantitative procedures for evaluating wetlands. Since evaluations can be costly and time-consuming because of the required field work and data collection, a simple method of identifying potentially valuable wetlands prior to conducting formal evaluations is useful. Using the wetland evaluation system developed by the Ontario Ministry of Natural Resources, this paper demonstrates that a small subset of easily measured wetland features can predict the number of points a wetland would score if it underwent a full evaluation. We developed a series of predictive models using multiple regression analysis of wetland evaluation data from both Southern and Northern Ontario. These models are of special relevance in Northern Ontario where a large number of unevaluated wetlands are located. This research may also provide a case study of a predictive procedure applicable in the assessment of wetlands for resource managers and planners in jurisdictions outside Ontario. © 1997 Taylor & Francis Group, LLC.