Predicting patterns of richness, occurrence and abundance of small fish in New Zealand estuaries

Abstract

Estuarine fish habitats are vulnerable to human impacts and are poorly studied. We surveyed 69 of New Zealand’s 443 estuaries across 1500 km to: determine species composition of small fishes; model and predict their richness, occurrence and abundance; test marine classification schemes as a basis for Marine Protected Areas; and inform impact mitigation measures. Boosted regression tree models produced acceptable fits for richness and occurrence at estuary and site scales and abundance at the site scale. Richness was greatest in northern North Island; the best predictors were estuary area and area of intertidal habitat. Within estuaries, richness increased towards the head, as water clarity declined and the substratum became muddier. Air temperature, estuary and intertidal area, tidal range and freshwater and seawater influx were the best predictors of occurrence at the estuary scale; water temperature and salinity were important at the site scale. Biological classification schemes seldom improved model fits and have little predictive utility. Richness predictions were made for 380 estuaries and occurrence predictions for 16 species. These predictions inform resource managers about estuarine fishes within their jurisdiction, bypassing the need to undertake expensive field surveys. However, sampling of environmental predictors is still required to drive some models.