Spatial Aggregation, Precision, and Power in Surveys of Freshwater Mussel Populations

Abstract

We studied aggregation in 76 populations of freshwater mussels from relatively homogeneous surroundings in a wide range of habitats. Chi-square tests for spatial aggregation found only 53% of mussel populations significantly (p < 0.05) aggregated. The variance of replicate mussel samples (s2) varied with the mean number collected (m) as 1.49m1.17, but conformed to the general variance relation found for other aquatic taxa (m1.5) at m > 1. The number of replicate samples ([Formula: see text]) required to estimate mussel abundance with a given level of precision (D = SE/m) is approximately m−0.5D−2. Sampling mussels with large quadrats requires between 5 and 25 samples for 20% precision. Sampling designs to determine significant impacts (α = (β = 0.05) require 7–50 samples of each population to detect doubling or halving of the population density, or three to nine to detect order-of-magnitude changes. Large sampling units are recommended to ensure acceptable sampling precision and accurate chi-square analyses of spatial aggregation and to permit ecologists to detect significant impacts on freshwater mussel populations.