|Riparian forests supply wood to adjacent stream channels (McDade et al. 1990, Benda et al. 2002). Such riparian-derived wood may interact with stream flow to alter the hydraulics, sediment dynamics, geomorphology (Gurnell et al. 2002), and biological productivity (Smock et al. 1989, Wallace et al. 1995) of the receiving stream reach. Where inputs are greatly reduced for long periods, instream functional LWD will decline (Murphy and Koski 1989, Bilby and Ward 1991) and stream channels may evolve towards a simpler geomorphic structure (Bilby and Ward 1991) with a reduced capacity to produce fish (Fausch and Northcote 1992) or fish-food organisms. The effects of reduced wood supply on channel geomorphology and biological production may persist for many decades (Murphy and Koski 1989). To reduce these undesirable effects, management agencies commonly require undisturbed or minimally-disturbed riparian buffers adjacent to stream channels (Young 2000, Lee et al. 2004) to maintain normal LWD inputs and other processes. Wood enters a stream reach from the adjacent riparian forest via several mechanisms, including: bank erosion, tree fall from natural mortality or windthrow, and landslides (Swanson and Lienkaemper 1978, Keller and Swanson 1979, Murphy and Koski 1989, Benda et al. 2002). The processes that deliver LWD from the riparian forest to a stream channel operate over different lateral distances from the channel (Benda et al. 2002, May and Gresswell 2003). Bank erosion will normally only deliver trees in close proximity to the channel edge (Murphy and Koski 1989, Benda et al. 2002). Tree mortality will deliver LWD from distances up to about one tree height from the channel (Van Sickle and Gregory 1990). Similarly, windthrow will normally deliver wood only within one tree height of the channel (Grizzel et al. 2000, May and Gresswell 2003). Landslides and debris torrents may introduce wood from longer distances from the channel (Benda et al. 2002, May and Gresswell 2003). The location of the wood source will also influence the piece size entering the stream. The buffer width needed to maintain normal inputs of LWD to a stream reach should depend on the dominant delivery mechanism(s) at the site. Recent modelling (Hogan and Bird, FSP# Y062170, unpublished) suggests that LWD source distances can be strongly influenced by the nature of the delivery process. However, riparian buffers to maintain LWD inputs are largely based on tree-fall models (Van Sickle and Gregory 1990) and limited empirical data from sites where tree-fall was the dominant LWD delivery mechanism to the stream (e.g., McDade et al. 1990). Certain LWD delivery mechanisms are strongly connected to the physical processes that determine channel type and size; thus, characteristics of a stream reach such as channel type, channel size, confinement, and the nature of the riparian vegetation may provide information about the relative influence of different LWD delivery mechanisms and the buffer width needed to maintain normal inputs of functional LWD. These channel characteristics are also the factors used in channel typologies such as the Channel Assessment Procedure (Church 1992, Anonymous 1996). By considering such factors, managers may be able to determine more precisely the buffer width required to maintain a specified proportion of expected LWD inputs with a given probability. |
We propose to determine LWD source distance curves (e.g., McDade et al. 1990) as functions of channel type (riffle-pool, cascade-pool, step-pool) and size (bankfull width) at sites with undisturbed old-growth or mature riparian forests in physically similar watersheds over a wide range of biogeoclimatic zones. Watershed and reach selection criteria will follow those used by Hogan and Bird (FSP# Y062170), as will the measurement of stream reach characteristics (channel width, depth, gradient, substrate, relative roughness, depth of incision, floodplain width, confinement, valley side-slope angle). We will determine LWD source distance curves using the direct survey methods of McDade et al. (1990). Where possible, we will assign a delivery mechanism (bank erosion, tree mortality, windfall, land sliding) to individual LWD pieces (e.g., Benda et al. 2002); we will also assign a geomorphic function (lateral scour, under scour, plunge pool, log step, log jam) wherever possible. We will determine riparian forest characteristics (species composition, height distribution, stem density) using standard forest mensuration methods. We will also survey coarse woody debris within one mean tree height of the channel to obtain information on tree-fall directionality for modelling purposes. LWD source distance curves for the sites will be summarized as quantile plots of the distribution of lateral distances at which specified proportions of the cumulative LWD input (by volume and by number) are encompassed for different channel types; quantile plots of the distributions of source distances at which specified proportions of cumulative LWD inputs were attained allow resource managers to better link management actions (buffer width) to the probability of attaining management goals (maintenance of specified LWD inputs). Distance to attain a specified proportion of expected inputs (e.g., the 90% of cumulative inputs) will be used as a dependent variable in general linear models to examine the effects of channel type, channel size, riparian forest height, and dominant delivery mechanism. Riparian forest characteristics will be used to model the LWD source distance curves expected from tree-fall models (Van Sickle and Gregory 1990, Bragg 2000) and compared to observed source distance curves.