The study site is located in the Blue River watershed of the Central Cascades Adaptive Management Area (AMA) in the Willamette National Forest, Oregon. Stands were sampled between 44.0 and 44.5 degrees N, and 122.0 and 123.0 degrees W. The Blue River watershed consists of 23,900 ha of conifer-dominated forest on steep volcanic terrain of the Cascade Mountain Range, ranging from 317 – 1,639 m in elevation (Cissel et al. 1999; Figure 1). Annual precipitation averages 220 cm (ranging from 55 to 361 cm), deposited as rain or snow in higher elevations, mainly between October and April. The winters are mild and wet with average temperature of 2 degrees C in January (ranging from -1.5 to 7.3 degrees C), and the summers are warm and dry with average temperature of 19 degrees C in July (ranging from 15 to 22 degrees C). The northern section of the watershed consists of a narrow band of high elevation, Abies amabilis (Dougl.) Forbes (Pacific silver fir) and Abies procera Rheder (Noble fir) dominated forest (hereafter, “true fir series;” Logan et al. 1987). Most of the watershed is lower elevation forest dominated by Pseudotsuga menziesii (Mirb.) Franco. (Douglas fir) and Tsuga heterophylla (Raf.) Sarg. (western hemlock; hereafter, “western hemlock series;” Logan et al. 1987).
The Blue River watershed has had decades of fire suppression and timber extraction. Historical fire regimes varied in frequency and severity within the watershed (Weisberg 1998). Forests in true fir series burned infrequently (mean fire interval of 260 yr), but fires were severe with high mortality (> 80%). The fire return interval for the western hemlock series ranged from 100 to 180 yr with less severe burns (40-80% mortality). Consequently, forests in the two plant series will be managed differently in the Blue River watershed, as part of the Landscape Plan (LP) which is an adaptive management approach using historical fire history as reference for management (Cissel et al. 1999).
Lichen communities were sampled in forest stands according to a stratified random design. Forest stands were stratified by four attributes, modified from Cissel et al. (1999; Figure 2):
Upland stands were at least two tree-heights (about 105 m) from perennial streams (hereafter referred to as “perennial fish-bearing”) and one tree-height (about 52 m) from all other streams (USDA and USDI 2001). Riparian stands were defined as having some part of the stream within or immediately bordering the plot boundary. Intermittent streams formed narrow channels and the stream-bank vegetation was similar to that of upland slopes. Perennial streams formed wider channels and vegetation along the stream banks was characteristic of riparian areas, including hardwood trees and shrubs. We were unable to identify stands along perennial streams with a definite remnant cohort; therefore, these stands were stratified by the age-class of the co-dominant tree cohort, ignoring the remnant stratum (Figure 2).
The design yielded 34 possible stand types for the western hemlock series and 22 for the true fir series, of which we sought to sample three stands each. However, some stand types were sampled with fewer stands or were not sampled due to their scarcity or absence in the landscape (such as stands with remnant retention greater than or equal to 30%). The 50% remnant retention class was uncommon in the landscape at the time of sampling and was therefore under-represented compared to other retention classes. The 50% retention class will become more prominent in the future landscape as managed under the LP (Cissel et al. 1999). We sampled 27 stand types in the western hemlock series, 6 of which were sampled with less than 3 stands. In the true fir series we sampled 18 stand types, 10 of which were sampled with less than 3 stands.
Stand Selection and Plot Installation
We located stand types described above from aerial photos. Most stands were within the Blue River watershed, however some stands were located outside of the watershed, but still on the Willamette National Forest and within the AMA boundaries (Berryman 2002). Stands sampled outside of the Blue River watershed represented stand types that were scarce or absent in the watershed. Stands were sampled in the summers of 1997-1999 using one permanent plot (34.7 m radius, 0.4 ha) per stand following the FHM plot methodology (McCune et al. 1997b). A total of 117 plots were permanently installed for this study.
Once the stand was located on the ground, a reference point (RP) was established along the road to assist in future plot relocation. From the RP (typically a tree) we chose an approximate azimuth into the stand. The RP was labeled with metal tags indicating the azimuth and distance to plot center. This azimuth was followed for 46.0 m (not slope-corrected) plus a two-digit random integer. Plot center was located no less than 46.0 m from: designated reserve areas in timber sales (other than stream buffers); the stand edge; roads; campgrounds; and power lines. Failing this, another random number was chosen and the same azimuth was followed until plot center was located outside of these exclusive areas and at least within 46.0 m of the stand edge. Riparian plots were installed to border the stream edge.
Plot center was marked with steel rebar and PVC pipe to increase the possibility of plot relocation after major disturbances, such as a fire or timber harvest. Three RP trees near plot center were tagged to reference the plot center with an azimuth and distance to facilitate relocation
Lichen Community Survey
The Forest Health Monitoring (FHM) lichen method was used to sample lichen communities in each stand. These data will be used as a baseline for long-term monitoring of lichen communities in the managed landscape. In each FHM plot, the surveyor completed a maximum two-hour ocular lichen community survey. The survey method consisted of two parts performed simultaneously (McCune et al. 1997): 1. The field surveyor used a field check list of easily identifiable lichens in the field and sometimes collected voucher specimens of these species. Voucher specimens were collected for all other species that could not be easily identified in the field. The collection represented the species diversity and composition of epiphytic macrolichens in the plot as fully as possible. The population sampled consisted of all macrolichens occurring on woody plants, excluding the 0.5 m basal portions of trees and shrubs below 0.5 m. Given the large plot area, lichen litter and fallen branches provided a sample of the canopy lichens. 2. The abundance of each species was estimated using a five-step scale (modified from McCune et al. 1997): 0 = absent; 1 = rare (less than 3 individuals per plot); 2 = uncommon (4-10 individuals per plot); 3 = common (greater than10, but less than 40 individuals per plot); 4 = very common (greater than 40 individuals per plot, but less than half of the available substrate was covered by the species); 5 = abundant (more than half of the available substrate had the species present).
The FHM method has been used to sample lichen communities in over 1,000 plots for the FHM program nationwide (McCune 2000) and has been used by the Pacific Northwest Forest Service Air Quality Biomonitoring Project for nearly 1,000 lichen community plots in Oregon and Washington (L. Geiser, unpublished data). Field methods have been documented for repeatability and quality assurance and are described in McCune et al. (1997).
Lichen Biomass Sampling
We estimated epiphytic lichen biomass from lichen litterfall on the forest floor in litter plots, sub-sampled in each stand. Epiphytic lichen biomass includes all epiphytic lichens growing on boles and branches of trees and tall shrubs. Collecting lichen litter in late summer (late August through October) avoids the large and variable amounts of litter that can occur in winter months due to large storm events (Stevenson and Rochelle 1984; Esseen 1985). Late summer litter does not represent annual litterfall because lichen litter in the forests of the western Cascades is eaten and decomposes rapidly (McCune and Daly 1994). However, such samples can be used to estimate epiphytic lichen biomass at the stand level (Neitlich 1993; McCune 1994; Peck and McCune 1997; Sillett and Goslin 1999). Annual variation in litterfall is one source of error in such estimates. Hence, this method should be based on samples collected during one late-summer period and is best used for estimating large relative differences in epiphytic lichen biomass among stands over a large area (McCune 1994).
Lichen litter was collected in a minimum of one stand per stand type. Of the 117 stands in which we collected lichen community data, we sampled 63 stands for lichen litter biomass. The 63 stands were chosen to include the full range of stand types included in the 117 stands. In each stand, epiphytic macrolichen litter was sampled in 2 m radius circular plots ("litter plots"). Depending on the stand age and complexity of canopy structure, ten to fifteen litter plots were sampled for each stand (McCune 1994). Stands with obviously low lichen biomass (e.g., even-aged young stands, less than 20 yr) were sampled with ten litter plots. Old growth (greater than 200 yr), mature stands (81-200 yr), and most stands with remnants were sampled with 15 litter plots.
Litter plots were placed along three transects per stand at randomly selected intervals, but constrained to 12-30 m between plots (two transects if sampling only ten litter plots). Transects were established parallel to the contour, intersecting the FHM plot center for the first transect. The other two transects were parallel to the first, separated by 12 m. This achieved interspersion throughout the stand. Some litter plots were placed outside of the FHM plot boundaries, though still within the stand. Five litter plots were sampled per transect.
Epiphytic macrolichens were divided into three functional groups based on their roles in the forest ecosystem (McCune 1993). These groups include "cyanolichens," which consist of all nitrogen-fixing lichens with cyanobacteria present as either the primary or secondary photobiont; the major contributors to this group included primarily Lobaria oregana and to a lesser degree L. pulmonaria. "Forage lichens" consist of all pendulous fruticose lichens. These are used for forage by wildlife, primarily the genera Usnea, Alectoria, and Bryoria. "Matrix lichens" account for all remaining green-algal macrolichens, primarily foliose in growth form, most commonly Platismatia and Hypogymnia.
We modified McCune’s (1994) litter-pickup method for estimating stand-level lichen biomass to expedite sampling across many stands at the landscape scale. The "reference method" was developed based on visual biomass estimates of thalli to sample lichen litter from the forest floor more rapidly, while maintaining a similar level of accuracy to that obtained with the litter-pickup method. The visual estimates of lichen biomass were made using reference lichen samples for calibration. This method was adapted from Rosso et al. (2000) and Campbell et al. (1999), in which they visually estimated biomass of lichens and bryophytes in the forest canopy using air-dried reference samples for calibration. The reference method is also a modification of the abundance classes (defined by grams of lichen) used by Stevenson et al. (1998) to estimate arboreal forage lichen biomass. We modified these approaches to visually estimate lichen litter biomass by functional group on the forest floor.
We estimated lichen litter biomass during the late summers of 1997–1999, during which each of the 63 stands was sampled once for lichen litter. Within each litter plot, oven-dried samples from each functional group (0.1, 1.0, 5.0, 10.0 g) were used as references for calibrating estimates of lichen litter biomass in the field. To assess reliability of the method, estimates from the reference method were calibrated against true litter-pickup masses for one litter plot in each of 16 different stands. Two field collectors calibrated their biomass estimates from the lichen litter plot to true lichen masses (16 litter plots per field collector). The “picked-up” specimens were air-dried, then oven-dried at 60°C for 24 hours, and then weighed to the nearest milligram in the lab. Daily calibrations were also made between estimates of biomass for individual clumps of lichen litter and true lichen masses for each functional group. These calibrations allowed field collectors to gauge the accuracy of their litter estimates. We also calibrated litter estimates between field collectors to improve precision of the estimates.
Lichen biomass data was entered at the sub-plot level for each stand. Data were cross-checked once for data entry errors. Lichen biomass data were averaged across all subplots by functional group for each stand, resulting in one mean biomass value at the stand level for each functional group: cyanolichen, matrix lichen and forage lichen.
Stand Measurements and Derived Stand Variables
At each stand, basic plot level information was recorded and the “stand type” for the plot was classified. Plot level information included: the date of sampling (note that all measurements for a plot were taken in one day, including lichen communities, lichen biomass, overstory trees and plot data); latitude and longitude coordinates and elevation (m) were taken from the GPS unit at plot center; percent slope was taken at plot center (averaging both up-hill and down-hill measurements using a clinometer); aspect was taken at plot center (degrees east of true north using a compass); topographic position of the stand in relation to the surrounding landscape was classified; and a code of 0 or 1 was assigned if lichen biomass was collected (0=no biomass, 1=lichen biomass was collected).
To classify the “stand type” we verified and recorded information for four main attributes. The overall vascular plant series for the FHM plot was determined using the Willamette National Forest Plant Association and Management Guide (Logan et. al. 1987), and the topographic position was recorded (upland or riparian intermittent stream, riparian perennial fish-bearing stream or riparian perennial non fish-bearing stream). The age class of the younger cohort was estimated for the stand, or, if the age class was difficult to determine, we cored representative trees to better estimate stand age. Old growth was defined as stands > 200 yr with highly variable canopy layers. Total percent canopy cover of remnant trees was used as an estimate of total percent retention of remnants for a stand. We estimated canopy cover of remnants in the field from dbh and crown width. J. Mayo (unpublished data) developed a table for estimating canopy cover of trees from dbh. This table is based on the relationship of dbh to crown width, from which percent canopy cover by each remnant tree was calculated. Remnants were typically Pseudotsuga menziesii, and in some cases Tsuga heterophylla, Thuja plicata, or Abies procera. Remnant tree age was not measured.
We calculated the heat load index and potential direct incident radiation for each stand. The heat load index represents the amount of heat a site potentially receives and is derived from models based on latitude, slope, and aspect (McCune and Keon 2002). Potential direct incident radiation (MJ/cm2/yr) represents the amount of light a site potentially receives, and is also derived from latitude, slope, and aspect.
Basal Area and Overstory Tree Measurements
Overstory basal area (BA) was measured with an angle gauge (note: a relescope was used in 1997 only) in five variable-radius subplots within each stand: one at plot center and the other four at 23.0 m in each cardinal direction from plot center (subplot labels were: plot center=1, N=2, E=3, S=4, W=5). A consistent BA factor (BAF) was used for all subplots within one stand, though the factor varied across all stands, depending on tree size and density. The goal for repeatability of BA was a standard error of less than 15% of the mean. Stand basal area (BA, m2/ha) was measured for all live and dead trees, separating hardwoods and conifers. However, at two plots (SB14 and SB17) we were unable to perform BA measurements because the stand was very young and dense (i.e., dog-hair stand). Instead, we did a count of trees per 1/10 acre for each subplot and no BAF was used. BA calculations for the plot level were made by averaging BA data across the 5 subplots to generate six BA variables as the plot level: total BA of live trees, basal area of conifers, BA of conifers, BA of hardwoods, BA of dead trees, and BA of remnants. In addition, we calculated the percent total BA that was comprised of remnant trees in the stand.
All overstory trees that were accounted for in the BA variable-radius subplots (or those trees that were “in”) were measured. Overstory tree measurements included: identifying species (species acronym recorded); forestry assessment of tree condition class; diameter at breast height (dbh); percent ratio of live crown class; and assignment of crown class (note: see class code descriptions in metadata file). Crown width was measured only for remnant trees, measured from the center of the tree bole, out to the tip of the crown edge (estimating by looking up). The longest branch was the starting point for measuring the crown width.
Laboratory Species Identification Methods and Data Entry
All lichen vouchers collected in the field were identified in the laboratory by Shanti Berryman, with the assistance of Dr. Bruce McCune. Lichen nomenclature followed McCune and Geiser (1997), and McCune’s key (2002, http://oregonstate.edu/~mccuneb/Usnea.PDF) to the genus Usnea in the Pacific Northwest. Thin layer chromatography was performed when necessary for identification. In addition, some specimens were sent to lichen specialists for verification. Representative voucher specimens were curated and deposited in the Oregon State University Herbarium (OSC).
All lichen community data were recorded on FHM lichen identification data sheets during identification in the laboratory and data were entered using compact format (text file; McCune and Mefford 1999) for PCORD using lichen species codes from the FHM lichen program species list. Duplicate collections of species at a given plot were combined following the FHM lichen program protocol (http://www.fia.fs.fed.us/library/field-guides-methods-proc/). All lichen community data entries were cross-checked once for data entry errors.
OTHER RELATED REFERENCES:
Campbell, J., S.K Stevenson, and D.S. Coxson. 1999. Estimating epiphyte abundance in high-elevation forests of northern British Columbia. Selbeyana 20: 261-267.
Logan, S.A., M.A. Hemstrom and W. Paulat. 1987. Plant association and management guide: Willamette National Forest. USDA Forest Service, R6-ECOL-TP. Portland, Oregon, USA.
McCune, B., J. Dey, J. Peck, D. Cassell, K. Heiman, S. Will-Wolf, and P. Neitlich. 1997. Repeatability of community data: species richness versus gradient scores in large-scale lichen studies. The Bryologist 100: 40-46.
McCune, B. and M.J. Mefford. 2004. HyperNiche. Nonparametric Multiplicative Habitat Modeling. Version 1.0. MjM Software, Gleneden Beach, Oregon, U.S.A.
Sillett, S.C. and M.N. Goslin. 1999. Distribution of epiphytic macrolichens in relation to remnant trees in a multiple-age Douglas-fir forest. Canadian Journal of Forest Research 29: 1204-1215.
Stevenson, S.K., A.N. Lance and H.M. Armleder. 1998. Estimating the abundance of arboreal forage lichens: User’s guide. B.C. Ministry of Forests, Land Management Handbook Field Guide Insert 7. Victoria, BC, Canada.
U.S. Department of Agriculture and U.S. Department of the Interior. 2001 Record of Decision for amendments to the survey and manage, protection buffer, and other mitigation measures standards and guidelines. Regional Ecosystem Office, Portland, OR.