The Relationship Between Productivities of Salmonids and Forest Stands in Northern California Watersheds

species and rock bass (Ambloplites rupestris) in the Ozark Mountains. These growth increments, in turn, were correlated with climatic variables, including July rainfall and stream discharge. Although many studies have evaluated salmonid response to logging (e.g., Connolly and Hall 1999, Tschaplinski 1999, Young et al. 1999), the relationship between upland productivity and salmonid productivity in the Pacific Northwest has not been examined. For forest managers, it is important to know whether the productivity of potential upslope harvest areas is directly related to, or unrelated to, the capacity of associated stream reaches to support salmonid fishes. How upslope productivity, which likely influences timber harvest objectives, may be manifested in salmonid growth has the potential to significantly impact timber harvest plans. Because of the key role that riparian vegetation plays in providing an inflow of organic matter and in influencing trophic pathways in streams (Cummins et al. 1989), riparian condition may be a better predictor of salmonid productivity than upland stand productivity. Upland stand productivity may not be a good predictor of riparian condition, as the structure and dynamics of riparian vegetation is strongly affected by complex interactions among hydrology and local geology (Naiman and Decamps 1997). Lack of correspondence between upland and riparian productivity may be especially marked where the riparian zone is dominated by nitrogenfixing red alder (Alnus rubra), as its growth may be relatively independent of soil nutrients (Binkley et al. 1994). The objective of this study was to quantify relationships between indices of salmonid productivity and upland and riparian forest productivity in small coastal watersheds of northern California. We asked whether: 1) upland site productivity predicted riparian site productivity; 2) either upland or riparian site productivity predicted salmonid productivity; and 3) other parameters explained more of the variance in salmonid productivity than upland or riparian site productivity. Methods Study Area The relationship of upland productivity, riparian condition, and salmonid productivity was evaluated by analyzing data collected from 25 forest stands and their adjoining streams in coastal watersheds of northern California. Sites were located in small watersheds in the Klamath River, Maple Creek, Little River, and Mad River drainages in Humboldt and Del Norte counties (Figure 1). The regional climate is characterized as marine west coast, with an annual average precipitation of 170–200 cm, approximately 75% of which falls as rain between November and March. Discharge in these basins during the period of sampling (October 2001 to September 2002) was slightly below average, based on 39 years of record (Freeman et al. 2003). Geology of the sites includes Mesozoic and Paleozoic metamorphic rocks, Mesozoic granitic rock, Mesozoic sedimentary rocks, and serpentinized ultramafic rocks (Harden 2003). The study was conducted on lands of Green Diamond Resource Company, in areas of second growth redwood (Sequoia sempervirens) and Douglas fir (Pseudotsuga menziesii). Sites were selected to provide variation in Site Index values of the forest stands, with adjoining stream reaches in which the fish community was dominated by resident coastal cutthroat trout (Oncorhynchus clarki clarki) and rainbow/steelhead trout (Oncorhynchus mykiss irideus). Cutthroat and rainbow trout were chosen as the target species for analysis because they are present in streams year-round. Small numbers of juvenile coho salmon (Oncorhynchus kisutch) were observed in one stream. Other vertebrates present in the study sites included the coastal giant salamander (Dicamptodon tenebrosus), tailed frog (Ascaphus truei), and ammocoetes of Pacific lamprey (Lampetra tridentata). Stream reaches were generally located in secondand third-order tributaries, with moderate gradients ranging from 2 to 9%, and catchments varying in size from 173 to 2,233 ha (Table 1). Study reach lengths were established using the proportional-distance designation (40 average stream width), a method adopted by the US Environmental Protection Agency Environmental Monitoring and Assessment Program (EMAP) program and described in Barbour et al. (1999). Reach lengths averaged 125 m and ranged from 60 to 200 m. Red alder was the dominant hardwood species found in the riparian areas along the reaches. Hardwood species present included red alder (42% of the total number of riparian trees), tanoak (Lithocarpus densiflorus) (11%), Pacific myrtle (Umbellularia californica) (3%), and bigleaf maple (Acer macrophyllum) (2%). Upland Forest Productivity Upland productivity was estimated using Site Index, which is the primary tool used worldwide in estimation of forest stand productivity (Stearns-Smith 2001). For a given tree species, Site Index is defined as the average tree height at some fixed age (commonly tree height at 50 years) attained by dominant and codominant site trees that are selected to reflect site potential. The height of dominant trees is used as the measure of productivity because, unlike timber volume, it is relatively unaffected by density except in extremely dense or sparse stands (Stearns-Smith 2001). A different age-height model is required for each species because height growth patterns are not the same across species (Nigh 1997). Site Index values are often grouped into five classes from the Site Index curves, with I being the highest and V being the lowest. Site Index values of redwoods in upland areas of each watershed were provided by Green Diamond Resource Company, where data were obtained by cruise estimates, extrapolation from cruise estimates, field estimation, and photo interpretation. Values were fieldverified by measuring tree height and coring dominant trees from five (20%) of the study sites. At each stream reach, four evenly spaced transects were established perpendicular to the channel. The dominant upland tree along each of the four transects, on each side of the channel (n 8 trees per reach), was located by visual inspection. Because some transects did not contain a dominant or codominant tree due to crown damage or rocky soils, the number of trees cored per site for upland Site Index averaged 5.6. Upland was delimited as the region between the ridge top and the upslope edge of the riparian zone. Upslope edge of the riparian zone was delimited by abrupt changes in gradient and changes in vegetation composition. Tree height was measured with a laser rangefinder, and age was determined by coring trees with an increment borer. Averaged values were used to determine Site Index for the stand by using a lookup table from Hanson et al. (2003), modified from Krumland and Wensel (1977). Site Index lookup tables link tree height at present age with tree height at a common base age. Field measurements of Site Index were only weakly correlated with Site Index values provided by Green Diamond Resource Company (r 0.13). Because the field verification sample size was small and the Green Diamond Resource Company data averaged values for the entire watershed, no adjustments were made to their data set. 74 WEST. J. APPL. FOR. 22(2) 2007 A small proportion of the catchment areas examined (average 12%) extended beyond lands within Green Diamond Resource Company ownership. These areas had land use similar to those on Green Diamond Resource Company property. Thus, Site Index values supplied by Green Diamond Resource Company were assumed to apply to all uplands within the watershed boundaries. Drainage area was estimated by digitizing the watershed boundaries associated with corresponding stream reaches. Riparian Condition Riparian attributes measured at each site included the Site Index of red alder, riparian canopy composition (percentage of hardwoods), and percentage of canopy coverage. In five sites, Site Index of redwoods within the riparian area was measured to evaluate the correspondence of redwood Site Index between riparian and upland areas. In March 2003, following leaf-out of riparian hardwoods, measurements of each parameter were made along four evenly spaced transects perpendicular to the stream reach. Percentage of canopy cover was estimated using a spherical densiometer located 1.2 m above the water surface at the center of the channel width. Four readings (upstream, downstream, right bank, and left bank) were taken at each of the four transects and averaged. Canopy composition of riparian vegetation was quantified along both sides of the stream channel using a line transect and point intercept method (Bonham 1989). Points were established every 2 m along transects extending to the edge of the riparian area. Riparian width was determined for each transect by a change in uniform slope and vegetation type. Percentage of hardwoods within the riparian zone was estimated as the number of point intercepts of hardwoods divided by total point intercept of all trees within the