Performance of the Macroinvertebrate Community Index: Effects of sampling method, sample replication, water depth, current velocity, and substratum on index values

The influences of sampling method, water depth, current velocity, and substratum on two macroinvertebrate‐based biotic indices were investigated based upon 523 samples from 55 stony riffle sites on 23 New Zealand streams. A single hand‐net sample estimated the site Macroinvertebrate Community Index (MCI) within ± 15% and four replicates yielded ± 10%. Between 8 and 10 replicate Surber samples produced ± 10% precision. Quantitative MCI (QMCI) values were more variable, with 10 or 11 replicate Surber samples required for ± 10% precision. Two procedures for detection of statistically significant differences between paired MCI or QMCI values are described. The detectable difference method (equal sample sizes) is preferred for statistical reasons but a f‐test method can be used for unequal sample sizes. MCI and QMCI were relatively independent of depth, velocity, and substratum within the sampled ranges of these variables. This is a major advantage for the assessment of water pollution or enrichment using these indices. However, to avoid possible complications brought about by extreme values, sampling within the following ranges of these variables is suggested: depth 0.1–0.4 m, velocity 0.2–1.2 m s −1 , and substrate 60—140 mm median diameter.