Premium
Accuracy and Precision of Selected Stream Habitat Estimates
Author(s) -
Wang Lizhu,
Simonson Timothy D.,
Lyons John
Publication year - 1996
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(1996)016<0340:aaposs>2.3.co;2
Subject(s) - habitat , accuracy and precision , substrate (aquarium) , streams , environmental science , statistics , riparian zone , vegetation (pathology) , fish <actinopterygii> , mathematics , hydrology (agriculture) , soil science , ecology , geology , computer science , fishery , biology , geotechnical engineering , medicine , computer network , pathology
The precision and accuracy of estimates of stream habitat variables were evaluated in three southern Wisconsin streams. Among‐observer precision was estimated from six observers with three different levels of experience. The precision of estimates for 27 stream habitat variables was generally good; 46% of the confidence intervals of the overall means for six observers were less than or equal to the field measurement precision. Stream width and water depth were estimated most precisely; these were followed by substrate composition, cover for fish, and bank susceptibility to erosion. Estimates of bank vegetation or land use and gravel embeddedness were the least precise. We determined accuracy of quantifying substrate composition by comparing direct visual observations from the same six observers with values obtained from digitized photographs of the same area of substrate. The results showed that 73% of the differences between visually estimated substrate values and digitized values were less than 5 percentage points, and none were greater than 12 percentage points. Accuracy and precision were lowest in the streams with the most heterogeneous habitat. Overall, our results indicate that visual estimates of substrate composition are sufficiently accurate for many fisheries applications. No significant differences in accuracy were found for estimates of substrate compositions obtained by the least and the most experienced observers. Variation among trained observers was relatively low for most habitat variables but fairly high for bank vegetation or land use and for gravel embeddedness. If initial training is adequate, experience of observers has little effect on the accuracy and precision of estimates.