Freshwater fish predictive modelling for bio-assessment; a scoping study into fish bio-assessment models as national indicators
Loading...
Date
2013-01-01
DOI
Authors
Journal Title
Journal ISSN
Volume Title
Publisher
Wairesearch Limited
Rights
Abstract
The match between the biota expected at a site in the absence of impacts and what is found there when testing is a robust and popular bioassessment method in many countries worldwide.
The difference between the assemblage found and that expected is measured as the observed/expected ratio and is the basis of the RIVPACS approach initially developed in the United Kingdom using invertebrates.
When the observed and expected assemblages match the O/E score is 1. An O/E score less than 1 means some impact and more than 1 suggests better than expected biota.
This scoping trial of the feasibility of a fish predictive RIVPACS type bioassessment using the New Zealand Freshwater Fish Database (NZFFDB) and predictive models of fish distribution from the Freshwater Ecosystems of New Zealand (FENZ) was not successful.
This study revealed that the predictive bioassessment approach in this case failed mainly due to the lack of suitable predictive models for this not because of problems with the predictive bioassessment approach.
The problem with the available FENZ fish predictions used in this study is that they were developed to predict how the fish assemblages are today allowing for many land-use impacts rather than the predictions of the assemblages that would be expected in the absence of impacts crucial to RIVPACS type models.
Regional O/E fish models have been successfully applied with fish in New Zealand by taking all the steps in the RIVPACS process but have generally not been taken up by resource managers.
To validate the data used in this study an Index of Biotic Integrity (IBI) was successfully applied to the observed (NZFFDB) and predicted fish assemblages (FENZ) and revealed their suitability for bioassessment.
However, an assessment of the observed/expected IBI results was, like the fish community O/E unsuccessful, again because the predictions are for actual rather than expected fish communities.
The conclusions from this study are that predictive bioassessment models have great potential for use in New Zealand but there are no shortcuts. Consequently, new predictive models must be produced based on reference sites and using habitat descriptors that are least influenced by human impacts.
In the meantime the IBI is a useful measure of the biotic integrity of freshwater in New Zealand, and improvements in sampling will mean that the IBI can be updated and improved.
Description
Keywords
Citation
Freshwater fish predictive modelling for bio-assessment; a scoping study into fish bio-assessment models as national indicators, 2013, pp. 1 - 2424