Johnston, N. T.; Calla, K.; Down, N. E.; Macdonald, J. S.; MacIsaac, E. A.; Witt, A. N.; Woo, E. 2007. A review of empirical source distance data for the recruitment of large woody debris to forested streams. Fisheries Project Rep. RD119. Victoria, BC: British Columbia Ministry of Environment. 41 p.
We have assembled published or Internet-accessible unpublished measurements of cumulative inputs of large woody debris (LWD) to stream channels as a function of lateral distance from the stream channel (i.e., "LWD source distance curves") in order to examine relationships between source distances and stream characteristics that are expected to be associated with dominant input processes. We obtained 137 source distance curves from 13 separate studies, most from coniferous forests along the Pacific coast of the U.S.A. From each LWD source distance curve, we determined graphically the distances to provide specified proportions of the cumulative LWD input (i.e., the distance for 50%, 75%, 80%, 90%, or 95% of the cumulative number or volume of LWD pieces at a reach). We used these data to examine the effects of input process (bank erosion, tree fall, land slides), stream channel type (riffle-pool, cascade-pool, step-pool), stream width, and riparian vegetation height on the source distances at which specific proportions (usually 90%) of the cumulative LWD inputs were attained. Few data sets provided complete coverage of the suite of variables that we wished to examine, so different analyses are based on different subsets of the data.
LWD source distance curves were variable within geographical areas with similar vegetation types and topography. Some of this variation is attributable to site characteristics which influence the mechanisms through which LWD enters the channel. Source distances which accounted for 90% of the cumulative numbers or volumes of LWD pieces increased with increasing tree height. The source distance which supplied 90% of the cumulative volume of LWD inputs differed among LWD delivery processes, being greater where tree fall was identified as the dominant input process than for bank erosion and land slides. Channel type and size influenced source distances for volume inputs, but the effect of channel type varied with channel width. Volume source distances at riffle-pool channels varied little with channel width, but source distances generally increased with channel size at cascade-pool and step-pool channels. We provide quantile plots of the distribution of lateral distances within which specified proportions of cumulative LWD inputs originated, which allow managers to determine riparian buffer widths that will attain specific management objectives for LWD recruitment with a specified probability.