Evaluation of a pump method for unbiased sampling of stream hyporheos

Abstract

The applicability of a pump method for field surveys of hyporheos was investigated for taxon-specific collection bias by comparing a pump method with a core method in a shore area of the lower reach of Kizu River in Japan. A sampling bias index for the first 10 l of pump samples compared with 10 l hyporheic water (equivalent) core samples showed that the pump method underestimated the abundance of most taxa, except for the heliozoan Centrohelida. Animals were classified into three groups based on body size and sampling bias index; animals with small bodies and high sampling bias indices generally had flatness indices near 1.0. The animals were also classified into three groups based on patterns of reduction in serial fraction values in consecutive pumpings. Although Crustacea tended to be overrepresented in the pump samples, relative densities among them were comparable to those in the core samples, except for harpacticoid copepods. Animals with an intermediate reduction pattern had a similar relative abundance in the 50-l pump and 1,000-cm3 core samples, except for Crustacea and Centrohelida. The results suggest that by using an adequate taxon-dependent volume of water, the pump method is useful for examining the community composition of hyporheic fauna.