ADAPTATION OF STATISTICAL RESIDUAL ANALYSIS FOR USE WITH REMOTELY SENSED IMAGERY TO AID IN BIOLOGICAL IMPACT ASSESSMENT (SIMPLIFIED, HABITAT, EVALUATION, DIVERSITY, OKLAHOMA)

The current difficulty with conducting biological impact assessment is due in part to the complexity of commonly used habitat-based methods for assessment. This research was, therefore, an attempt to devise a simplified technique which could be used as an assessment tool. The technique revolves around an easily quantified diversity index which can be obtained with remotely sensed imagery.;The analysis of variance statistics indicated that the simplified technique was repeatable. Initial regression analysis indicated a correlation, however, it was too low for the simplified technique to be desirable. Further investigation revealed that many of the outliers had something in common: they were sites where wetlands or flood plain forests had been evaluated by HEP without considering the quality of surrounding cover types. When these 'isolated' wetlands and flood plain forests were removed from analysis, the correlation improved significantly (F Value = 127.96, p Value = 0.0001, R-square = 0.8152). The best correlation was found when all 'isolated' sites were removed from analysis (F Value = 109.72, p Value = 0.0001, R-square = 0.8658).;These statistics indicate that the simplified technique is a valid approach, and can be used with confidence as an assessment tool. The results have biological validity and the output is in units readily understood by decision-makers. The results of the simplified technique should be amenable to integration with data from other disciplines such as regional planning as well as all other forms of impact assessment (physical, chemical, and socio-economic).;The feasibility of using the simplified technique was determined by applying the method to five proposed water resource projects where the U.S. Fish and Wildlife Service had previously conducted the 'Habitat Evaluation Procedures' (HEP). The five project areas (in Oklahoma) had forty-three sampling sites which were available for comparing the simplified technique to HEP. Parametric and non-parametric regression analyses were used to determine the relationship between the simplified technique and HEP. Analysis of variance techniques were used to determine the repeatability of the simplified technique.