Premium
Optimizing Sampling Effort for Sampling Warmwater Stream Fish Communities
Author(s) -
Peterson James T.,
Rabeni Charles F.
Publication year - 1995
Publication title -
north american journal of fisheries management
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 0.587
H-Index - 72
eISSN - 1548-8675
pISSN - 0275-5947
DOI - 10.1577/1548-8675(1995)015<0528:osefsw>2.3.co;2
Subject(s) - sampling (signal processing) , fish <actinopterygii> , streams , fishery , environmental science , computer science , biology , computer network , filter (signal processing) , computer vision
We measured the variation of some commonly used fish community attributes for two warmwater stream reaches during June–October 1992 and 1993. Habitat‐specific variation among samples, expressed as coefficients of variation (SD/mean) collected throughout the study ranged from 0.13 to 1.58 and were lower for community‐level attributes such as species richness (total number of species) and total fish biomass than for biomass estimates of bleeding shiner Luxilus zonatus , longear sunfish Lepomis megalotis , and rainbow darter Etheostoma caeruleum . Corresponding estimates of the number of samples needed to ensure 20% precision at the 95% confidence level ranged from 2 to 245 and indicated that fewer samples are needed to precisely estimate community level attributes than to estimate individual species biomass. A significant negative relationship ( P < 0.05) between coefficients of variation and predicted sampling efficiency suggested that low sampling efficiencies may increase sample variance. Significant heterogeneity of variance ( P < 0.05) among habitat types suggested that physical habitat characteristics also influenced sample variance. Mixed‐model analysis of variance was used to examine spatial variance (between sampling locations, across time) and temporal variance (among sampling periods, across locations) for species richness and fish biomass. Eighteen variance components were significant, ( P < 0.01) and in 12 of these, spatial variation exceeded temporal variation. When age‐0 fish were excluded from analysis, spatial variation exceeded temporal variation in 13 of the 14 significant components. Our results indicate that the optimum sampling strategy for warmwater streams during June–October includes the collection of many samples from all habitat types during one sampling period in September–October.