CORRESPONDENCE AND RELATIVE PRECISION OF STREAM HABITAT FEATURES ESTIMATED AT 2 SPATIAL SCALES

I compared visual estimates of stream bed substrate composition, stream bank land use, stream bank erosion, and instream shelter for fish made at two different spatial scales: the transect scale, where observations were made along transect lines; and the segment scale, where observations were made within relatively large areas centered on transects. Overall, estimates of habitat conditions based on observations at the two spatial scales were strongly correlated (r greater-than-or-equal-to 0.78), and mean differences between estimates based on the two techniques were relatively small (less-than-or-equal-to + 5.3%). Estimates of silt and gravel substrate differed significantly, but absolute differences between estimates at the two spatial scales were small. I recommend making observations at the transect scale because estimates at the transect scale were more precise and there were no consistent differences between estimates made at the two spatial scales. I recommend a hybrid approach for estimating substrate, where substrate composition, rather than only the dominant substrate, is recorded within quadrats at transect-points. This approach should reduce assumed bias associated with transect scale estimates of silt and gravel substrate.