Our objectives were five-fold:
Objectives 1-3: Classification and environmental relationships of forest and meadow communities; fire history of forest plots.
Forest communities were sampled in 1981 and meadow communities, in 1982, using the reconnaissance method of Franklin et al. (1970) (Franklin, J. F., C. T. Dyrness, and W. H. Moir. 1970. A reconnaissance method for forest site classification. Shinrin Richi 12:1-14). We apply the term ‘meadow’ fairly broadly to encompass a diversity of non-forested types, including mires, fens, grasslands, lithosolic (rock-garden) types, and the sparsely vegetated ‘pumice plains’ of the High Cascades. Sampling of forest communities was restricted to the Separation Creek drainage, located centrally on the west side of the Cascade crest. Sampling of meadows was more extensive, covering most regions of the wilderness. It also included a number of sites adjacent to the Wilderness boundary. Potential sampling sites were identified from topographic maps and aerial photographs. In total 162 forest plots and 152 meadow plots were sampled. Plot locations are marked on USGS topographic maps (see the PDF file, TSBR topo maps with reconnaissance plot and ecotone transect locations.pdf, under Related Materials).
Objectives 4 and 5: Characterize spatio-temporal patterns of tree invasion into meadows and their abiotic and biotic controls; establish a system of permanent transects to monitor future changes in vegetation
In 1983, we established and sampled a series of 21 permanently marked, 2-m wide transects spanning forest-meadow boundaries at 17 locations (four locations were sampled with two transects each). Each transect began within the forest and extended into adjacent meadow, well beyond the last invading tree. Transects vary in length from 58 to 220 m (data entity 11) and encompass a range of physical environments (e.g., montane to subalpine, xeric to hydric) and vegetation types (data entity 12). Steel or aluminum reinforcing bars (rebar) mark the ends and intermediate points of each transect. Details on rebar locations in 2009, the final sampling date, can be found in the document, TSBR_ecotone transect directions and details.pdf (see Related Materials). Start- and end-point UTMs and elevations are listed in data entity 11. Aerial photographs showing transect start and end points are contained in the document, TSBR_aerial photos of ecotone transects.pdf (see Related Materials).
At each location, a circular 500 m2 plot was subjectively placed in visually homogenous vegetation. We recorded the environmental characteristics of each plot, including elevation, slope, aspect, landform, and topography (data entities 1 and 2). Cover (%) of ground-surface conditions (data entities 3 and 4) and plant species (data entities 5 and 6) were estimated. Additional data were taken on forest structural attributes, including numbers and diameters of live and dead trees, ages of one or more dominant trees (obtained from increment cores), and tree heights (data entities 7-8). Fire history was inferred from the age structure of trees. Finally, a soil pit was dug, particle size distributions were determined, and soil profiles were described (data entities 9 and 10); in wet meadows, depth to water table was recorded. Detailed descriptions of field methods, analytical approaches, results, and the final classification of forest and meadow types can be found in Halpern et al. (1984) and Halpern (2000).
Objectives 4 and 5: Characterize spatio-temporal patterns of tree invasion into meadows and their abiotic and biotic controls; establish a system of permanent transects to monitor future changes in vegetation.
Transect establishment and dates of sampling.- In 1993, 15 of the original transects were resampled. At this time each transect was widened to increase the sample of invading trees. In addition, one transect was lengthened (3A, and two transects (5A and 20A) were added to increase the sample of trees (5A) or to provide a contrast in aspect at the same site (20A). In 2009, all of the original transects were resampled for ground vegetation and all but one (Cow Swamp, transect 9) was resampled for trees. Neither of the temporary transects (5A, 20A) nor transect extensions (3A) was resampled for trees in 2009 (for details see TP064_tree sampling history.xlsx under Related Materials).
Characterizing physical environments, vegetation types, and soils (entity 12).- In the establishment year (1983), a series of soil pits was dug adjacent to each transect, subjectively located to represent the range of plant communities present. At each pit we recorded the dominant ground-layer species, elevation, slope, aspect, landform, and topography. We also determined litter depth and soil particle size distribution; described soil horizons and depth of rooting; and recorded any unique characteristics (e.g., presence of charcoal or a water table).
In each year, cover of all vascular plant species was visually estimated in a series of 1 x 1 m plots spaced 1-2 m apart on alternating sides of each transect (data entity 15). Trees were sampled in contiguous plots along both sides of each transect, to varying distances from the center line (data entity 14). Species, height, diameter (basal or breast-height), position along the transect, and age (1983 and 1993) were recorded for each individual (sampling details follow).
Sampling ground-layer vegetation and tree cover (data entity 15).- Vegetation data were collected in 1 x 1 microplots on alternating sides of each transect (0-1 m on the left, 1-2 m on the right, etc.). Occasionally, on longer transects, microplots were spaced at 2-m intervals for portions of the transect. In some instances, microplot spacing deviated from systematic for portions of the transect. Site-specific details on the spacing of microplots is provided in the document, TSBR_ecotone transect directions and details.pdf (see Related Materials). Cover was estimated for ground-surface conditions (mineral soil, fine litter, coarse litter, etc.) and for each vascular plant species. For tree species, cover was estimated by size class (class 5 = total tree cover; class 6 = trees < 1 m tall; class 7 = trees > 1 m tall and < 10 cm dbh; class 8 = trees 10-30 cm dbh; class 9 = trees 30-100 cm dbh; and class 0 = trees > 100 cm dbh).
Sampling tree sizes and positions along the transect (data entities 13 and 14).- Data on tree size (basal or breast-height diameter and height) and position (distance along the transect) were collected in a continuous "belt" on both sides of the transect. Initially (1983), all trees including first-year seedlings, were sampled in a continuous 2-m wide belt (1 m on each side of the transect). In 1993 and 2009, the sampling belt was increased to varying widths for varying distances down the right or left side of the transect. Estimates of tree density were adjusted accordingly. In 1993 all trees > 2 yr old were assigned uniquely numbered tags (nailed to the bole or loosely wired around the base). Tags were added in 2009 as needed, for trees meeting particular age (e.g., > 2 yr) or height thresholds (e.g., 10 cm). Site-specific details on belt widths at each sampling date are provided in an excel spreadsheet, TP064_tree sampling history.xlsx (see Related Materials). Site-specific details on tree tagging in 2009 are contained in the document, TSBR_ecotone transect directions and details.pdf (see Related Materials).
Estimating tree ages/establishment dates (data entities 13 and 14).- In 1983 and 1993, age at the time of sampling was determined for all trees measured for diameter and height. One of three non-destructive methods was used, depending on diameter and species. For trees with sufficiently large basal diameters, increment cores were extracted. Cores were taken as low to the ground as possible and coring height was recorded. For samples that did not include the pith, the number of missing rings was derived from an estimate of the distance to pith divided by the average width of the inner 5-20 rings present in the core (see Miller 1995). To obtain a final estimate of tree age, an estimate of age-to-coring-height was added based on regression equations developed from destructive sampling of small trees off the transects (see Miller 1995). For trees that were too small to core, terminal bud-scale scars were counted. Scars were counted only to the point where they were obscured by basal wounds, bark expansion, or ground-layer mosses. As with cored trees, an estimate of age to the height of the last bud-scale scar counted was added to the bud scar tally. Because this method of aging typically underestimates true age, species-specific regression equations comparing bud-scar tallies to ring counts were used to adjust final age estimates (Miller 1995). Finally, for small individuals of mountain hemlock (a species that does not produce annual bud-scale scars), ages were predicted from height or diameter. Regression equations of age vs. size were developed from destructive sampling of trees off the transects (Miller 1995). Tree ages were not determined in 2009, the last sampling date. For additional details on aging see Halpern et al. (1984), Miller (1995), and Miller and Halpern (1998).
Permanent photo points.- Photo points were established at the ends of, and at various points adjacent to, each transect. Photos were taken of most sites in each sampling year. A subset of matched photos is provided for each site in the zip file, TSBR_repeat photos of ecotone transects.zip (see Related Materials).
