Upland vegetation monitoring for the Blue River Landscape Study: Development of protocols and initial measurements

Steven A. Acker

Dept. of Forest Science

Oregon State University

Version of Jan. 22, 2000

Final edits May 1, 2000

Outline

1. Objectives and Design
   1. Overview
   2. Selection of patches
   3. Selection of plots
   4. Monumenting and documenting plot locations
   5. Plot layout
   6. Plot measurements
      1. Canopy closure
      2. Tree inventory
      3. Intensive tree measurements
      4. Understory measurements: herbs, shrubs, and tree regeneration
      5. Coarse woody debris: snags and logs
      6. Environmental variables
   7. Frequency of measurements
   8. Data management and quality control
2. Results of first year’s measurements
   1. Patches sampled in 1998
   2. State of the data
   3. Estimated effort for data collection and quality control
   4. Pre-harvest stand conditions
      1. Live trees
      2. Tree regeneration
      3. Vascular plant species composition and cover
      4. Coarse woody debris
3. Discussion
   1. Comments on monitoring design
   2. Problems in data collection
   3. Produce regular reports

I. Objectives and Design

Overview

The overall objective of this monitoring effort is to document changes of upland vegetation following timber harvest designed to approximate natural disturbance regimes in the upper Blue River watershed. There are a variety of attributes of the vegetation that are of interest, either for their contributions to biological diversity or for their commodity values. These attributes include tree regeneration, growth and mortality of residual trees, log and snag amounts and persistence, vascular plant dominance and diversity, and plant biomass and production.

The Blue River Landscape Study divides the portion of the watershed available for timber harvest into three "Landscape Areas" with contrasting harvest and reforestation prescriptions. The Landscape Areas differ with respect to rotation age and retention level, among other prescription elements. The basic prescription for Landscape Area 1 is a 100-year rotation with 50% retention; the basic prescription for Landscape Area 2 is a 180-year rotation and 30% retention; the basic prescription for Landscape Area 3 is a 260-year rotation and 15% retention.

Selection of patches

The spatial distribution of the Landscape Areas reflects the inferred historical pattern of natural disturbance regimes, which varies across the watershed. Variation across the watershed in potential productivity and natural vegetation (i.e., plant associations) is correlated with variation in natural disturbance regime (e.g., both vary with elevation). Thus it will not be possible to observe all the major plant associations of the watershed on all three Landscape Areas. In addition, stands to be harvested vary in seral stage, including mature stands, old-growth stands, and stands which are difficult to classify as either due to salvage logging or other factors. It is unlikely that all seral stages will be well represented even within a particular combination of plant association and Landscape Area. Thus this monitoring effort will concentrate on those combinations of harvest prescription, plant association, and seral stage which are common in any particular set of timber sales. To have confidence in conclusions about the various treatments, it is critical to replicate observations. The unit of replication will be the patch. A patch will be defined as a contiguous area within a sale unit, or in some cases, a contiguous area spanning adjacent sale units. At least three separate patches per combination of harvest prescription, plant association, and seral stage will be monitored. In cases where there are more potential patches than necessary for a combination of Landscape Area, plant association, and seral stage, patches to monitor will be selected randomly from the potential patches.

Selection of plots

Upland vegetation within patches will be measured using fixed-radius plots of 0.1 ha area (slope-corrected). Plots will be placed so that perimeters are at least 50 m from the nearest road or boundary with a stand of different age, structure, or prescription. The centers of plots will be separated by at least 100 m. Plots will not be located within obvious riparian-influenced areas. Within these constraints, plot centers will be located in a random manner. Proceeding from one end of the patch, the location of the first plot is determined by choosing a random point in the 1^st 100 m of the major axis of the patch, and a random point on the minor axis of the patch at that point on the major axis. The second plot is located by selecting a point at a random distance between 100 m and 200 m from the first plot along the major axis of the patch, and a random point along the minor axis of the patch at that point. Additional plots in a patch are located similarly.

Given the 50 m buffer width, 100 m between plot centers, and slope-corrected plot radius of approximately 18 m, the minimum size of patch to accommodate 3 plots is about 5-6 ha. For patches that are larger than 5-6 ha, proportionately more plots may be established. On the other hand, patches smaller than 5-6 ha may be included if the patch represents a common combination of prescription, plant association, and stand type that lacks an adequate number of larger patches. In this case, fewer than 3 plots per patch will be established.

Monumenting and documenting plot locations

The centers of plots will be marked with rebar, pounded into the ground so that only a few inches remain exposed, with a pvc pipe over the rebar. The pvc will be labeled with the plot identifier and the date. Coordinates of each plot center will be recorded with GPS equipment. In addition, three trees marked for retention that are near to the plot center will be used as witness trees. If the trees closest to the center of the plot are small and/or apparently in poor health, larger, vigorous trees farther from the center may be used as witness trees. Witness trees will be tagged at breast height (see "Plot measurements" below), and tag number, diameter at breast height, and species will be recorded. The bearing, slope distance, and slope from the center of the plot to the nearest face of each witness tree will be recorded.

Plot layout

As mentioned above, fixed-radius plots of 0.1 ha area will be used. The radius of a 0.1 ha plot is 17.84 m. Plot boundaries will be slope-corrected so that plots will be 0.1 ha in horizontal area. To achieve this, plot radii will be corrected for the effects of slope (i.e., 17.84 m horizontal distance corresponds to a longer distance along steep slopes).

Plot measurements

Plot measurements will encompass vascular plants (trees, saplings, seedlings, shrubs, and herbs), coarse woody debris (snags and downed logs), and some site characteristics.

Canopy closure

The first measurement will be of canopy closure. Twelve readings per plot will be taken with a moosehorn densiometer. Readings will be taken at 5.0, 10.0, and 15.0 m (slope-corrected) away from the center of the plot on the 4 subcardinal directions (NW, NE, SE, SW). Measurements are taken facing the center of the plot.

Tree inventory

Within the entire 0.1 ha plot, for all trees >= 5 cm diameter at breast height (DBH), species, DBH, canopy class, overall vigor, and crown ratio will be recorded. Trees that are marked for retention will be tagged at breast height with aluminum tags and nails. All trees between 5 and 18 cm DBH should be tagged, since 18 cm DBH is the lower size limit for trees to harvest. Tags will be placed facing the center of the plot.

Breast height will be defined as 137 cm, measured along the bole, from the uphill location. The possible values for canopy class are:

D = dominant               Crown emerges from the general canopy layer, and so receives light from the top and the sides

C = co-dominant          Crown extends to the top of the general canopy layer, and so receives light from the top, but not much from the sides

I = intermediate            Crown extends into the lower portion of the general canopy layer, and so receives mostly filtered light from the top and the sides

S = suppressed             Crown completely beneath the general canopy layer

Overall vigor is one of the following:

1 = good vigor              No apparent signs of distress (e.g. discolored foliage, paucity of leaves)

2 = fair vigor                 Some signs of distress apparent

3 = poor vigor              Extreme distress apparent (i.e. death imminent)

Crown ratio is defined as the proportion of a tree's total length for which at least 1/3 of the bole's circumference is covered by live crown. Widely scattered branches are not included. Gaps along the bole within an otherwise continuous crown are subtracted from the crown ratio. Crown ratio will be estimated to the nearest 5 percent. If epicormic branches are present, (short branches arising directly from the bole), these should be noted in a comment.

Intensive tree measurements

A subsample of trees >= 10 cm DBH will be selected for height measurement, increment coring, and measurement of crown width. To determine which trees to include in the subsample, first tally the number of trees by species from the tree inventory (excluding those with broken or dead tops, or other obvious, serious damage or disease). For species with fewer than 5 individuals, no intensive measurements will be taken. For species with 5 to 10 individuals, all trees will be measured. For species with more than 10 individuals, determine the range of diameters by subtracting the smallest diameter from the largest diameter. Then divide the trees into three groups, each group including 1/3 of the range of diameters. The first group will include all trees from the smallest to those with diameter equal to the minimum diameter plus 1/3 of the range. The second group will include all trees with diameters between the minimum plus 1/3 of the range and the maximum minus 1/3 of the range. The third group will include all trees larger than the maximum diameter minus 1/3 of the range. Use a random number generator on a pocket calculator, or other randomization device, to pick 3 trees from the 1^st group, 4 trees from the 2^nd group, and 3 trees from the 3^rd group. If any group lacks the specified number of individuals, randomly choose additional trees from the other groups.

Both total height and height to the base of the live crown will be measured. Whether using a survey laser or clinometer and distance measuring device, the angle to the top of the tree should be no more than 50 degrees. The base of the live crown is the lower point on the continuous crown (i.e., ignore foliage below gaps along the bole). Two relatively short cores will be obtained from each of these trees at breast height, with the cores separated by at least 90 degrees. For trees <= 18 cm DBH, only 1 core will be taken. For each core, bark thickness, sapwood thickness, and 5- and 10-year radial growth will be recorded. No attempt will be made to measure sapwood for hardwood species. For species with furrowed bark, cores are taken on ridges, not crevices. Four radii of the crowns of the trees selected for height measurement and coring will be also be measured along cardinal directions.

Understory measurements: herbs, shrubs, and tree regeneration

Quantitative measurements of saplings, seedlings, shrubs, and herbs will be structured around two parallel line transects. The endpoints of these transects will be located 10.00 m, slope-corrected, from the center of the plot, on the subcardinal compass bearings. A declination setting of 19.5 degrees east will be used throughout. These four points will also be marked with rebar, which will be brightly painted. The two line transects will run in a clockwise fashion from adjacent points: line 1 will run from the NW point to the NE point; line 2 will run from the SE point to the SW point.

These line transects will be used for line-intercept measurements of cover of shrubs and tree regeneration, and for location of 1 x 1 m quadrats for other measurements. The perimeters of the 1 x 1 m quadrats will be slope-corrected. In the quadrats, estimated cover of herbs and low shrubs (e.g., salal, Oregon grape, common snowberry) by species will be recorded. The minimum cover value recorded is 0.1%; vascular plant species in a quadrat but with cover <0.1% are recorded as 0.1% cover. In addition, tree regeneration will be tallied by species and size class. Seedlings will be defined as trees < breast height and will be recorded in four height classes: 1 (0 to 10 cm tall); 2 (10 to 25 cm tall); 3 (25 to 75 cm tall); and 4 (75 to 136 cm tall). Saplings will be defined as trees >= 137 cm tall and < 5.0 cm DBH. Saplings will be recorded in 1-cm DBH classes from 0 (i.e., up to 0.9 cm DBH) to 4 (i.e., 4.0 to 4.9 cm DBH). Also in the 1 x 1 m quadrats, basal diameter of upright shrub species will be recorded. Basal diameter is measured at the root collar, or slightly above if swollen. The minimum diameter recorded is 0.1 cm. Upright shrubs include species such as rhododendron and vine maple, but not salal or dwarf Oregon grape. Definitions of tall shrubs vs. low shrubs and herbs follows Dyrness (1973). Finally, litter depth and degree of soil disturbance will be recorded in or near the 1 x 1 m quadrats (see below for details).

Three 1 x 1 m quadrats will be measured along each of the 2 lines, with the upper left-hand corner of the quadrats (if looking from the center of the plot) on the lines at 1.00, 6.00, and 11.00 m. Quadrats will be placed on the inside of the line transects as viewed from the center of the plot.

After the quantitative measurements of live vegetation have been taken, a complete list of vascular plant species on the plot will be recorded. The entire 0.1 ha plot will be inspected briefly to identify any species not included in the quantitative measurements.

Coarse woody debris: snags and logs

All snags and logs within the 0.1 ha plot will be measured. The minimum size for coarse woody debris is 10 cm (DBH for snags, diameter of the larger end for logs). An upright piece of CWD that is not self-supporting (leaning on a snag or tree) is a log, not a snag. Snags marked for retention (or between 10 and 18 cm DBH) will be tagged. For snags with intact tops DBH will be measured. For broken snags, DBH and total height will be measured, and top diameter will be estimated ocularly, or from measurement of the end of the fallen piece, if found. For downed logs, the portion within the plot will be measured. The length and the two end diameters of all pieces will be recorded (note that the diameter of the smaller end can be less than 10 cm). For ends of logs that are not round, two perpendicular diameter measurements are averaged. For broken pieces with jagged, pointed ends, the pointed portions are excluded from diameter measurements. Only the segments of logs within plots will be measured, so that if one or both ends of a log lie outside of a plot, the diameter(s) where the log crosses the perimeter of the plot would be recorded. For all snags and logs the decay class (from 1 to 5) will also be recorded. Definitions of the decay classes of snags and logs follow Maser et al. (1988).

Environmental variables

Environmental variables recorded will include slope, aspect, topographic position, degree of soil disturbance, and litter depth. Slope, aspect, and topographic position will be recorded from the center of plot. Slope in degrees will be recorded both up the slope and down the slope (in the direction of the aspect). Slope is measured from the center to the edge of the plot. Topographic position will be recorded as either top 1/3 of slope, middle 1/3 of slope, bottom 1/3 of slope, or flat. Topographic position is relative to the entire hillslope on which the plot is located. Soil disturbance and litter depth will be recorded for each of the 1 x 1 m quadrats within which cover of herb species, numbers of tree saplings and seedlings, and basal diameters of upright shrubs were recorded. Soil disturbance will be recorded using the four categories developed for Watersheds 1 and 3 on the Andrews Experimental Forest (Halpern 1988):

1) undisturbed (soil surface similar in appearance to areas not logged and burned, with minimal mixing of soil and litter and no evidence of fire);

2) disturbed-unburned (disturbance from logging evident; litter removed or mixed with mineral soil but minimal evidence of fire);

3) lightly burned (surface litter charred by fire but not completely removed);

4) heavily burned (surface litter completely removed by intense fire).

Litter depth will be measured approximately 1 m farther along the line transect from the lower right-hand corner of each quadrat (e.g. for quadrat 1 at 3.0 m along the line and 1.0 m off the line towards the center of the plot). Litter depth will be measured by making a small, temporary cut into the forest floor with a trowel.

Frequency of measurements

Measurements will be made prior to harvest and in the first year after harvest. Subsequent measurements will be made every 5 years, unless resources permit more frequent measurement.

Data management and quality control

Following the field season, data will be checked for completeness at Oregon State University and any resolvable problems will be attended to. Data will be archived in the Forest Science Data Bank at Oregon State University.

II. Results of first year’s measurements

Patches sampled in 1998

Only Landscape Areas 1 and 2 were represented in the two timber sales included in the Blue River Landscape Study that had been offered for sale by 1998. Both sales were located in the western hemlock zone; based on the judgement of Blue River Ranger District personnel, it appeared that most of the upland area in the sale units was covered by dwarf Oregon grape, salal, or closely related rhododendron plant associations. Thus it was not deemed practical to stratify sampling within these two timber sales on the basis of plant association. Measured patches included both mature and old-growth stands. The distribution of measured patches by Landscape Area and seral stage is summarized in Table 1.

State of the data

Data from 1998 have been typed into computer database files. The computerized data have been checked for completeness and logical errors. Items requiring field checks were attended to during the 1999 field season. The data are archived in the Forest Science Data Bank at Oregon State University (study code = TV048).

Estimated effort for data collection and quality control

Field data were collected by a seasonal crew hired and supervised by personnel from the Blue River Ranger District. The work was performed between mid-July and late November, with the bulk of the work in July, August, and September. Although all field sheets included space for recording personnel, individual names were not recorded in all cases. Thus, some assumptions are necessary to estimate the total field effort. Our best estimate is that 2064 hours was expended on the fieldwork (43 days, 8 hours/day, an average of 6 people on the crew).

The computerized data structure for the field data was created by Howard Bruner, a faculty research assistant in the Dept. of Forest Science. Howard also carried out quality control and archiving after data were entered. He was aided by Gody Spycher of the Dept. of Forest Science and myself. Howard spent 96 hours setting up the data structure and subsequently cleaning and archiving data.

Data entry was carried out by student hourly workers under the supervision of Gody Spycher. Data entry required approximately 55 hours.

Pre-harvest stand conditions

Live trees

Stand structure, tree species composition, and bole volume varied between groups of patches representing different combinations of Landscape Area and seral stage, and to a lesser extent, between patches with groups (Tables 2-5). Stand structure is summarized in Table 2. The 95% confidence intervals (95%CI) for different seral stages computed by Spies and Franklin (1991) from an extensive set of plots in natural mature and old-growth stands provides a basis for comparison. Among mature patches representing Landscape Area 1, 3 of the 4 patches fell outside of the 95%CI for mature stands for either trees per ha (tph), basal area, or both. Patch 8 had lower than expected tph. Patch 9 was higher than expectations for both tph and basal area. Patch 2A had lower than expected basal area. All of the mature patches in Landscape Area 2 had lower than expected tph and basal area. Five of the seven old-growth patches in Landscape Area 2 fell outside of the 95%CI for old-growth stands for either tph, basal area, or both. Patches 3X, 3Y, and 5E all had higher than expected basal. Patch 3Y also had lower than expected tph. Patches 5B and 5C had higher than expected tph.

The lack of valid tree height information for patch 5D represents a shortcoming of the data collection on several of the patches. Many height measurements were recorded with excessively large angles to the top of the tree. This renders height measurements very inaccurate, and in some cases produces nonsensical data. The field sheet indicated that top angles should not exceed 45 degrees; however, top angles up to 89 degrees were recorded. Based on analysis of other datasets, and careful examination of all height measurements from 1998, I decided to consider as invalid all tree heights with a top angle greater than 55 degrees. This resulted in discarding 104 of the 695 heights measured in 1998. This had the greatest effect on patch 5D where about two-thirds of the heights were discarded. About one-quarter of the heights were discarded for patches 4A and 5C.

Tree species composition varied between the three groups of patches, especially from the perspective of relative density (Tables 3 and 4). The mature patches in Landscape Area 1 had more stems of Douglas-fir than any other species; Douglas-fir accounted for more than 90% of basal area (except for patch 2A, the patch added to assess an area of unsuited soils). Shade-tolerant species accounted for less than 10% of basal area. The mature patches in Landscape Area 2 had more stems of western hemlock than any other species; Douglas-fir accounted for between 80% and 90% of basal area. Shade-tolerant species accounted for at least 10% of basal area. In most of the old-growth patches in Landscape Area 2, there were more stems of western redcedar than any other species (the exception was a patch where stems of western hemlock were most abundant); Douglas-fir accounted for between about 50% and 80% of basal area. Shade-tolerant species accounted for about 20% to 50% of basal area.

Information from the Willamette National Forest Plant Association and Management Guide (Hemstrom et al. 1987) provides a standard for comparison of bole volumes on the sampled patches (Table 5). Hemstrom et al. (1987) computed total stand volume, to a 4-inch top, for the mature and old-growth stands they measured in the various plant associations. Table 5 includes total stand volume including top and stump, so values might be expected to be somewhat higher than those reported by Hemstrom et al. Among the plant associations similar to the sampled patches (i.e., western hemlock types with dwarf Oregon grape and/or salal, with or without rhododendron), the lowest value was for western hemlock/salal (844 m³/ha, standard error 79) and the highest value was for western hemlock/rhododendron-dwarf Oregon grape (1146 m³/ha, s.e. 97). Values for the sampled patches were compared to the lower value minus 2 standard errors, and to the higher value plus 2 standard errors.

Among the mature patches in Landscape Area 1, only patch 2A on unsuited soils fell outside (below) the expected range. All of the mature patches in Landscape Area 2 were lower than the expected range of bole volume. Volume on 3 of the 7 old-growth patches in Landscape Area 2 were higher than the expected range (patches 3X, 3Y, and 5E); volume on the other old-growth patches was within the expected range.

Tree regeneration

Density of seedlings and saplings was generally low (Table 6): 3 of the sampled patches had no tree regeneration. Only 2 of the sampled patches had more than 1 seedling and sapling per m². In all of the mature patches with tree regeneration, the most abundant species were shade-intolerants (except in patch 8, where western hemlock was as abundant as Douglas-fir). In all of the old-growth patches with tree regeneration, the most abundant species was a shade-tolerant.

Vascular plant species composition and cover

A total of 94 species were recorded on the 14 patches sampled in 1998 (Table 7). These included 14 tree species, 24 shrub species, and 54 forb species. No identifications were attempted within the grass family or the sedge genus. There were 3 exotic species on the list. Twelve species occurred in all 14 patches sampled in 1998; 28 species occurred in only 1 patch.

Average richness of vascular plant species varied by more than a factor of 2 between the various patches, from 15 species per 0.1 ha plot to 36 species (Table 8). The mature patches in Landscape Area 2 tended to have higher species richness than patches in the other 2 groups.

There were some differences between the groups of patches with respect to understory cover. Cover of herbs and low shrubs was about 40 to 50% on the mature patches in Landscape Area 1 (except for patch 2A, on unsuited ground), over 50% on the mature patches in Landscape Area 2, and 36% or less on the old-growth patches in Landscape Area 2 (Table 8). In all 3 groups of patches, the herb or low shrub species with highest cover was usually salal or dwarf Oregon grape. Although there was considerable variation in the cover of tall shrubs and tree regeneration within groups of patches, the mature patches in Landscape Area 1 generally had the highest values and the old-growth patches in Landscape Area 2 generally had the lowest values (Table 8). In both of the groups of mature patches, the species with highest cover was always either western rhododendron or vine maple; in the old-growth patches in Landscape Area 2 western redcedar was most commonly the species with highest cover.

Coarse woody debris

Downed logs in the sampled patches represented the full range of decay states, with 7% in the least decayed state, class 1, 24% in decay class 2, 31% in decay class 3, 21% in decay class 4, and the remaining 13% in decay class 5. Nearly half the logs were of unknown species, 33% were Douglas-fir, 11% were western hemlock, 9% were western redcedar, and 1% were other species. The majority of snags were of decay class 2 (54%), with 25% in decay class 3, 14% in decay class 1, and 6% in decay class 4. No snags of the most decayed state, class 5, were recorded. The majority of snags were Douglas-fir (60%), 20% were western redcedar, 9% were western hemlock, and 10% were other species or unknown.

The 3 groups of patches varied both with respect to the number and mass of logs, and with respect to the numbers of snags (Table 9). Mass or volume of snags have not yet been computed due to complications imposed by missing or erroneous data. The mature patches in Landscape Area 1 generally had the fewest logs (under 200 per ha, except for patch 2A on unsuited soils), and the lowest mass of logs (between about 10 and 30 Mg/ha). The mature patches in Landscape Area 2 had more logs (between 200 and 300 per ha), but mass of logs was quite variable. The old-growth patches in Landscape Area 2 had the highest number of logs between 200 and almost 500 per ha), and the greatest mass of logs (between about 40 and 140 Mg/ha). The mature patches in Landscape Area 2 had the lowest number of snags (between about 20 and 50 per ha), while the mature patches in Landscape Area 1 had between about 40 and 110 snags per ha, and the old-growth patches in Landscape Area 2 had between 20 and 170 snags per ha.

Compared to the extensive survey of natural stands reported by Spies et al. (1988), the mature patches all had low numbers of logs, while about half the mature patches had at least the expected mass of logs and number of snags. For mature stands in the Oregon Cascades, Spies et al. (1988) reported means of 413 logs per ha (standard error 45), 23 Mg/ha mean log mass (s.e. 3), and 109 snags per per ha (s.e. 17). Fewer than half of the old-growth patches in Landscape Area 2 had as many logs as expected (mean 417 per ha, s.e. 30). Most of the old-growth patches had at least the expected mass of logs, however (mean of 73 Mg/ha, s.e. 8). Most of the old-growth patches also had at least the expected number of snags (mean of 60 per ha, s.e. 5).

The problems that have delayed computation of volume and mass for snags concern snags with broken tops. Such snags require measurements of height and top diameter. The problem cases include illogical height measurements (angle to top of tree less than angle to base), height measurements with an excessively large top angle (over 55 degrees, see comments under "Live trees," above), and no top diameter recorded. It most cases it will probably be possible to estimate missing or erroneous data using taper equations and other assumptions. However, accurate measurements would be preferable.

III. Discussion

Comments on monitoring design

Two points made in a recent review of monitoring on National Forests (Morrison and Marcot 1995) are worth considering. Morrison and Marcot state that "monitoring ... of ecosystem elements and conditions should be designed to test clearly stated hypotheses." This is a key step in evaluating the adequacy of any monitoring effort, and something that should be considered for monitoring associated with the Blue River Landscape Study.

Morrison and Marcot also stress that for optimal design of monitoring studies, it is necessary to have measurements both before treatments, and in locations which are left untreated. These represent control observations in time and space, respectively. Without controls in both time and space, it is not possible to know whether any observed changes are due to the treatment, or would have occurred in any case.

It is important to acknowledge that the current design for monitoring of upland vegetation in areas harvested under the Blue River Landscape Study does not include controls in space. Within limitations due to sample sizes and measurement error, the design should be adequate to identify change in the sampled patches. However, strictly speaking it will not be possible to state that any changes observed are due to management activities. If identifying cause and effect is an objective of the monitoring effort, establishment of control plots or identification of other means of obtaining control data will be necessary. Using the definitions of Morrison and Marcot (1995), it may be reasonable to consider the current design as "effectiveness monitoring" ("is the management activity having the desired effect?"), but it is not "validation monitoring" ("are the basic assumptions and foundations used for developing forest management guides sound and correct?"). I would speculate that validation monitoring is more likely to elicit participation of scientists than is effectiveness monitoring.

Problems in data collection

In the process of performing quality checks on the data and preparing this report, several systematic problems in data collection came to light. Communication with field crews in the future should minimize the occurrence of these problems.

It was noted that many height measurements of live trees and snags were recorded with such large angles to tree tops that the data were discarded. Field crews should be told repeatedly that top angles of 45 degrees or less are preferable, and top angles greater than 55 degrees are unacceptable. For snags, there were also numerous cases of angles to tree tops that were less than angles to tree bases. It would be worthwhile to ask field crews to remember to check that this does not occur. It also apparently would be worthwhile to remind crews that for all snags with broken boles, the top diameter must be measured or estimated.

The line intercept measurements also contained some repeated errors. These included cases in which the end value of an intercept was less than the start value, and also cases in which no end value was recorded for an intercept. The importance of checking these data in the field before leaving a plot should be emphasized.

Finally, the use of 4- or 5-character acronyms for plant species ("Garrison codes"), though convenient, almost invariably generates some confusion when data are analyzed. Two problems seem common. The first is neglecting to include the integer at the end of the acronym when necessary (for example, recording "ARCA" for Artemisia cana, silver sagebrush, when "ARCA3" for Aralia californica, elk clover, is meant). The second is recording an acronym that has the same sound, but a different spelling than the intended species (for example, recording "ANDI" for Antennaria dimorpha, an east-side species, when "ANDE" for Anemone deltoidea is meant). Crews should be reminded to write the full scientific name whenever there is doubt. They should also be encouraged to keep in mind the drawbacks of the acronym system.

Produce regular reports

I encourage whoever is involved in the monitoring effort in the future to prepare a report on each year’s field data. When done conscientiously, summarizing data requires a degree of quality control that is difficult to achieve in any other way. Summarizing and interpreting the data also is necessary to assess whether monitoring is likely to provide the answers it was intended to produce, and may also suggest additional uses of the monitoring data for management or research questions.

IV. Literature cited

Dyrness, C.T. 1973. Early stages of plant succession following logging and burning in the western Cascades of Oregon. Ecol. 54: 57-69

Halpern, C.B. 1988. Early successional pathways and the resistance and resilience of forest communities. Ecology 69: 1703-1715

Hemstrom, M.A., S.E. Logan, and W. Pavlat. 1987. Plant association and management guide: Willamette National Forest. USDA Forest Service, Pacific Northwest Region, R6-Ecol 257-B-86. 312 p.

Maser, C., S.P. Cline, K. Cromack, Jr., J.M. Trappe, and E. Hansen. 1988. What we know about large trees that fall to the forest floor. Pp. 25-45 in C. Maser, R.F. Tarrant, J.M. Trappe, and J.F. Franklin, tech. eds. From the forest to the sea: a story of fallen trees. Gen. Tech. Rep. PNW-GTR-229. USDA PNW For. & Range Exp. Sta.

Morrison, M.L., and B.G. Marcot. 1995. An evaluation of resource inventory and monitoring program used in National Forest planning. Environ. Manag. 19: 147-156.

Spies, T.A., and J.F. Franklin. 1991. The structure of natural young, mature, and old-growth Douglas-fir forests in Oregon and Washington. Pp. 91-110 in L. Ruggiero, ed. Wildlife and vegetation of unmanaged Douglas-fir forests. Gen. Tech. Rep.

PNW-GTR-285. USDA Pacific Northwest Research Station, Portland, OR.

Spies, T.A., J.F. Franklin, and T.B. Thomas. 1988. Coarse woody debris in Douglas-fir forests of western Oregon and Washington. Ecology 69: 1689-1702

Table 1. Summary of patches within which permanent plots were established in 1998 for monitoring upland vegetation for the Blue River Landscape Study.

^aExtra set of plots to monitor an area of unsuitable soils.

Table 2. Stand structure of patches within which permanent plots were established in 1998. Standard errors in parentheses.

Table 3. Relative density (percent of total trees per hectare) for patches within which permanent plots were established in 1998.

^aSpecies codes: ACMA=bigleaf maple; CONU=Pacific dogwood; PSME=Douglas-fir; THPL=western redcedar; TSHE=western hemlock; OCFR=other conifer (western yew, grand fir, noble fir, Pacific silver fir, western white pine); OHRD=other hardwood (golden chinquapin, red alder).

Table 4. Relative basal area (percent of total basal area) for patches within which permanent plots were established in 1998.

^aSpecies codes: ACMA=bigleaf maple; CONU=Pacific dogwood; PSME=Douglas-fir; THPL=western redcedar; TSHE=western hemlock; OCFR=other conifer (western yew, grand fir, noble fir, Pacific silver fir, western white pine); OHRD=other hardwood (golden chinquapin, red alder).

Table 5. Tree bole volume and biomass of patches within which permanent plots were established in 1998. Standard errors in parentheses.

Table 6. Tree regeneration on patches within which permanent plots were established in 1998.

^aSpecies codes: CACH=golden chinquapin; PSME=Douglas-fir; THPL=western redcedar; TSHE=western hemlock.

^bThere are 4 size classes for seedlings, based on height: 1 (0 to 10 cm); 2 (10 to 25 cm); 3 (25 to 75 cm); 4 (75 to 136 cm). There are 5 size classes for saplings: 0 (0 to 0.9 cm DBH); 1 (1 to 1.9 cm DBH); etc.

Table 7. List of vascular plant species recorded on patches within which permanent plots were established in 1998.

Table 8. Vascular plant species richness, cover of herbs and low shrubs from 1 m² quadrats, and cover of tall shrubs and tree regeneration from line intercepts. Standard errors in parentheses.

^aCover computed by summing covers of all species in the group. ^bSee Table 7 for species acronyms.

^cData missing for 1 of the 2 plots. ^dOnly 1 plot in the patch, so standard error cannot be computed.

Table 9. Coarse woody debris amounts on patches in which permanent plots were established in 1998. Standard errors in parentheses.

^aOnly 1 plot in the patch, so standard error cannot be computed.