Ecological response differentials: an alternative benchmark to inform stream and river bioassessment

Abstract

Summary Contemporary patterns of land development mean that suitable reference sites for low‐gradient stream and non‐wadeable river biomonitoring are difficult to locate, requiring alternative approaches for benchmarking human impacts. We use two strongly harmonised macroinvertebrate data sets for wadeable streams (180 sites) and non‐wadeable rivers (30 sites) to (i) quantify metric response differentials across a regional landcover pressure gradient in northern New Zealand; (ii) interpolate reference condition for waterway types lacking comparable reference sites and (iii) inform development of narrative condition bands for data‐poor waterway types. Cumulative distribution functions of per cent waterway length indicated that near‐shore littoral samples from non‐wadeable rivers had fewer total and EPT (Ephemeroptera, Plecoptera and Trichoptera) taxa than wadeable streams, which had higher % EPT abundance and lower % Crustacea abundance. Community composition differed significantly among non‐wadeable river and wadeable hard‐bottom and soft‐bottom stream samples, reflecting factors associated with catchment altitude, slope, rainfall and landuse intensity. Ephemeroptera, Plecoptera and Trichoptera richness was the only metric evaluated that significantly reflected landcover across all waterway types. The intercept of the 60th percentile quantile regression line at 100% native vegetation cover provided the best approximation of median EPT taxa richness in an independent reference site data set for hard‐bottom streams. Applying the same interpolation to samples from soft‐bottom streams and non‐wadable rivers that lacked regional reference sites indicated median reference expectations of 9 and 11 EPT taxa, respectively, and a consistent differential across the pressure gradient relative to hard‐bottom streams of seven taxa and five taxa, respectively. The derived response differentials enabled the development of narrative bands for directly comparing ecological condition of data‐poor stream and river types. Quantification of ecological response differentials can provide an alternative means of benchmarking macroinvertebrate metrics in the absence of suitable reference sites.