Assessing the relations between aquatic habitat indicators and forest harvesting at watershed scale in the interior of British Columbia

Selection of appropriate aquatic habitat indicators is important for supporting sustainable forest and watershed management. To date, many identified watershed indicators have not been well tested. To address this issue, we sampled 28 stream reaches in the central and south interior of British Columbia, Canada. The selected reaches were similar with respect to gradient, watershed and topographic characteristics but with various levels of forest harvesting. Aquatic habitat indicators were derived based on a field survey, involving substrates, channel dimensions, pools, and large woody debris (LWD), together with some ratios and derivatives generated from them. Timber harvest disturbance at the watershed scale was quantified by calculating historical forest clear-cut area based on the BC vegetation resources inventory database. To account for the subsequent forest recovery, equivalent clear-cut area (ECA) was calculated based on the dominant species of each stream watershed and its regrowth since being logged. Pearson correlation coefficients showed that percent ECA and percent historical clear-cut area could effectively quantify the cumulative logging disturbance by limiting the effect of factors not related to logging activity (i.e., watershed area and elevations). These two variables were then related to the aquatic habitat indicators. Pearson correlation analyses found that relative width (ratio of D to bankfull width) (D is the b axis diameter of the largest substrate particle found in the reach), relative roughness (ratio of D to bankfull depth), pool frequency and per piece LWD volume are sensitive to percent ECA, i.e., significant (p < 0.01) while independent of influence from non-forest logging activity related factors and spatial clustering of samples. This relationship also holds true for the last two indicators when related to percent historical clear-cut area, though with a reduced degree of correlation. In addition, mean LWD diameter also showed significant relationship with percent historical clear-cut area, but not with percent ECA. Overall, per piece LWD volume was considered as the best aquatic habitat indicator. The performance of mean LWD diameter was comparable to per piece LWD volume if percent historical clear-cut area was used to quantify forest disturbance. Implications of these findings for forest management strategies are discussed. (C) 2008 Elsevier B.V. All rights reserved.