Development of a Shoreline Degradation Index (SDI): monitoring environmental change using landscape ecology metrics and high resolution remotely sensed imagery

Thirty-five common landscape level metrics were analyzed and combined to create seventeen separate Shoreline Degradation Indices (SDIs). These SDIs were generated using multiple linear regression, multiple discriminant analysis, genetic neural networks, standard feed forward back propagation neural networks and manual combination. 400 shoreline units were then rated by each of the seventeen SDIs and compared to ratings of the same 400 shoreline landscape units produced by expert image analysis. There was very little difference in the performance of several SDIs despite differences in the included metrics and weights produced by the different methods. The ratings from all methods were able to reflect the expert ratings in general. Conclusions indicate that the ability of the landscape metrics to discriminate between levels of shoreline degradation is contained within a relatively few metrics. In any of the current forms the SDI is suitable only for general indication of shoreline health.