Massey Documents by Type
Permanent URI for this communityhttps://mro.massey.ac.nz/handle/10179/294
Browse
1 results
Search Results
Item The development of predictive models to enhance biological assessment of riverine systems in New Zealand : a thesis presented in partial fulfillment of the requirements for the degree of Doctor of Philosophy in Ecology at Massey University, Palmerston North, New Zealand(Massey University, 2003) Joy, Michael KevinA suite of new regional and national lotic freshwater bioassessment tools were developed for New Zealand. This work permits the inclusion of freshwater fish in bioassessment, a component of the fauna previously largely ignored. The multivariate predictive models developed gave a number of advantages over the existing albeit overextended single-index approach (the macroinvertebrate community index) used by regional authorities. To acquire the data for constructing the models more than 500 sites were sampled over three North Island regions. The sites were selected to represent least impacted conditions known as reference sites so that the biotic communities sampled would representing the best attainable or the goal for resource managers. Models were constructed to predict the biota representing best available conditions based on the non human influenced physicochemical variables defining the sites. The predicted and observed assemblages were then compared using an observed over expected ratio (O/E) so that scores less than 1 represent less [i.e. fewer] species observed than expected. This (O/E) ratio is more than simply the assessment of species richness, as only those species predicted are included in the ratio. Reference site multivariate predictive models using fish and macroinvertebrate assemblage groups were developed for bioassessment in the Manawatu-Wanganui Region. Two reference site multivariate predictive models using individual fish and decapod species were developed for the Auckland region. The first used traditional linear discriminant function analysis and the second used artificial neural networks (ANNs). A model to predict the spatial occurrence of fish and decapods was developed for fish in the Wellington Region using Geographic Information Systems (GIS) and ANNs. The remotely sensed data was available for all rivers in the region so the predictions could be extended over the entire stream network to produce a fish map. Finally an index of biotic integrity (IBI) using fish was developed for the entire country and evaluated using remotely assessed environmental data. Exhaustive evaluations of predictions from all the models confirmed their credibility as a biomonitoring.