CHAPTER I
METHODS
Study Areas
This study was conducted on five clear-cut areas in the Pacific Northwest
provided by cooperating public and private landowners (Figure
2.1). The five sites are; near Hoodsport, WA on the southeast edge of
Washington's Olympic Mountains, near Apiary, OR at the west end of the Columbia
River Gorge; near Sweethome and Gates on the west slopes of Oregon's Cascade
Mountain Range; and LaGrande in eastern Oregon's Blue Mountains. These root
rot infected sites were chosen to assess the treatment effects with varying
soils and environmental conditions(Table 2.1).
The four western sites consist of typical second and third growth forests
of the Tsuga Heterophylla Zone (Franklin and Dyrness 1973). The climax species
in these forests are typically Tsuga heterophylla and Thuja plicata, with
Psuedotsuga menziesii ( Douglas-fir), Acer macrophylla, and Alnus rubra
as earlier successional species. Silvicultural methods used in these forests
have been almost exclusively clear cutting followed by planting Douglas-fir.
The one site in eastern Oregon is in the Abies Grandis Zone. The climax
species in this area will normally be Abies grandis with Psuedotsuga menziesii,
Larix occidentalis, Pinus ponderosa, and Pinus contorta as successional
species. Silvicultural methods used in these areas tend to be more varied,
and clear cuts can be planted with resistant species such as Larix and Pinus
to mitigate root rot problems.
All sites had documented mortality of Douglas-fir caused by P. weirii present in the stand at the time of harvest. All sites were clear-cut harvested before the stumping treatment and planted with Douglas-fir seedlings after the treatments. Hoodsport treatment plots were established in 1976, LaGrande and Apiary in 1978, and Gates and Sweethome in 1980.
The sites cover a range of soil types. Soils at Hoodsport and LaGrande are Andisols, but differ greatly. Hoodsport soil is a very gravelly sandy loam 50-100 cm deep formed in ablation till over a compacted till that restricts drainage. Lagrande soils are a fine textured silt loam 80-160 cm deep formed from loess and volcanic ash over buried soil and colluvium. Due to geographic location and elevation, Lagrande is the least productive site. Apiary and Sweethome are Ultisols, older soils in areas that were not glaciated during the last ice age, and have developed more clay content. Soil at Apiary is a fine textured, deep (160cm) soil formed from loess and sandstone and basalt colluvium. Soil at Sweethome has the highest productivity rating and is a fine textured deep (160 cm) soil formed from sandstone, basalt and siltstone. The soil at Gates is classified as an Entisol, a younger soil formed from igneous agglomerate. This soil is a deep (160 cm) cobbly loam.
Plots
Each study area was divided into 30X30 meter subunits that were systematically searched before and after harvest for phellinus infested trees, down trees and stumps (Thies and Hoopes 1979). Infection was rated using an index of relative inoculum potential (INOC) based on the presence of incipient or advanced decay typical of that caused by P. weirii or ectotrophic mycelium typical of P. weirii near the root collar.
On a map of infected stumps, a treatment plot was positioned to cover areas of infected stumps. Plot centers were then located in the field based on the map, and the presence of infected stumps verified. The plots are 0.04-ha non-overlapping circular treatment areas surrounding a 0.02-ha circular measurement area. This system created a 3 meter buffer around each measurement plot. The 0.02-ha plots were rated for inoculum potential using the INOC rating of stumps in the plot plus 25% of the INOC in the buffer area. Plots were then stratifed by INOC ratings into treatment blocks of eight plots each. Eight plots were positioned in areas without infected stumps. After stump removal, plot centers were reestablished and permanently marked with fiberglass poles and rebar (Thies and Nelson 1988).
Additional subplot samples were taken in 1991 and 2003 outside each study area to establish a reference data set of undisturbed areas of the same soil and timber type (Thies et al. 1994). These plots were randomly selected in the least disturbed areas adjacent to the study areas.
Treatments
Treatments were applied the season following harvest. The treatments consisted of a factorial design of two stumping levels and four nitrogen application levels.
The stumping treatments were: B0, no stumps removed, and B1, all stumps removed. Removal of stumps was done under dry conditions with a bulldozer equipped with a brush blade, and with a splitting wedge on the back. Large stumps were split at least once with the wedge before attempting to remove them with the blade. Normal bulldozing techniques were used to remove as much of the stump and root system as possible. No effort was made to remove severed roots. Upended stumps were left where they came to rest whenever possible.
Nitrogen treatments were made before planting using ammonium nitrate as small prills broadcast with a cyclone seeder, crossing the treatment area in perpendicular directions to assure even coverage. The four levels of treatment were: F0, none applied; F1, 336 kgN/ha; F2, 672 kgN/ha; F3, 1345 kgN/ha.
The eight possible treatment combinations (B0F0, B0F1, B0F2, B0F3, B1F0, B1F1, B1F2, B1F3) were randomly assigned within each block of eight plots (Thies and Nelson 1988).
All sites were planted with Douglas-fir seedlings in the next planting season, and in the following seasons were interplanted and thinned to control stocking levels.
Soil Measurements
To determine the affect of the treatments on the soil, soil bulk density was measured, and soil and forest floor samples were collected and analyzed for total nitrogen from each treatment plot at all sites. The Sweethome site had previously been sampled for bulk density and total nitrogen in the mineral soil (A horizon) in 1991 for all treatments except for the 672 kg.ha-1 nitrogen applications (Thies et al. 1994). All sites were remeasured similarly in 2003 except that the forest floor was also samped. The forest floor samples were measured for depth, volume and weight.
Bulk density readings and the soil and forest floor samples for each plot were obtained in three randomly selected subplots created by division of each circular plot into 360 subplots of equal area by 10 concentric rings and 36 10o sectors. The subplot locations were generated by selecting random pairs of numbers: 1-10 (ring) and 1-36 (sector). If an obstruction, such as a stump or downed log, prevented sampling of a subplot, a new plot was selected.
Soil bulk density was measured at each subplot to a depth of 20 cm with a Troxler 3411B moisture/density neutron probe. (The 1991 Sweethome bulk density readings were made with a Campbell Pacific Nuclear neutron probe, model MC-1.) (Thies et al. 1994) Subplot bulk density values were averaged for the mean for each plot. The neutron probe results were validated by collecting bulk density core samples directly from some of the probe sites with a bulk density corer or by displacement method. These samples were oven dried, and weighed for comparison with the probe readings. At least 8 of these check samples were taken at each site.
Forest floor samples were collected by using a knife to cut around a 208 cm2 rectangular template to the mineral soil. The forest floor was then separated from the mineral soil and placed in plastic bags. The depth of the forest floor to the mineral soil was measured. The samples were oven dried, and weighed to determine mass. The three subplot samples from each plot were then composited into one sample for each plot, ground and analyzed for total N at the University of Washington using a Perkins-Elmer PE 2400 Series II CHNS/O Analyzer.
Neutron Probe
Bulk density measurements taken with the neutron probe were consistently
higher than the checks taken with the displacement and core methods
(see appendix B). The average difference was 0.13 g cm-3 and this
relationship was consistent across all sites and treatments and between
methods. When two depths of cores were taken one below the other, the
lower core averaged 0.07 g cm-3 lower than the probe, and the upper core
0.17 g cm-3 lower. The neutron probe results are reading a larger volume
than the core samples. Rousseva et al. (1988) found that field calibration
of the neutron probe rather than factory calibration is necessary for
matching results from differing methods. Cassaro et al. (2000) state that
this is due to the factory calibrations using denser materials than the
normal bulk density range found in soils. Cassaro et al. found that factory
calibrations tend to overestimate bulk densities below about 1.4 g cm-3.
The moist soil at depth has higher bulk density because of water content
and is closer to the factory calibration. No field calibrations were conducted
for the purposes of this study. The results were considered consistent
enough to use the neutron probe readings for comparing treatments.
Forest Floor
Forest floor samples were collected by using a knife to cut around a 208 cm2 rectangular template to the mineral soil. The forest floor was then separated from the mineral soil and placed in plastic bags. The depth of the forest floor to the mineral soil was measured. The samples were oven dried, and weighed to determine mass. The three subplot samples from each plot were then composited into one sample for each plot, ground and analyzed for total nitrogen at the University of Washington using a Perkins-Elmer PE 2400 Series II CHNS/O Analyzer.
Soil samples were collected from the upper 15 cm of mineral soil directly beneath the forest floor sample. The soil samples were placed in plastic bags for transport then air dried in the lab. The three subplots samples from each plot were composited, ground, and analyzed for total nitrogen.
Forest Reference plots
The original study had no reference or control plots. The forest plots we sampled were from the closest, similar, unharvested stand available. Gates and Hoodsport had adjacent uncut stands. The adjacent stand at Hoodsport had a greater slope than the study area. Apiary was sampled from a 60 year old naturally regenerated fire stand about two kilometers away because the entire area had been harvested. LaGrande is in a cut-over area and samples were taken under groups of approximately 60-80 year old trees adjacent to the study area. Sweethome forest sites sampled in 1991 are now a regenerated clear-cut. These sites were sampled as well as six more subplots in uncut and thinned timber 100 meters away.
Data Analysis
Statistical analysis was done using SAS for Windows (version 8.2, SAS Institute Inc. Carey, N.C.). Analysis of the total data was done by two-way ANOVA (PROC GLM, SAS Institute, 1999-2001 ), utilizing treatment means from each site (n=40,). Two-way factorial analysis of variance (PROC GLM, SAS Institute, 1999-2001) was used to examine data from individual sites for significant effects caused by treatment, stumping, fertilization and interaction effects of stumping and fertilization. Variables analyzed were bulk density, total mineral soil nitrogen and carbon, forest floor total nitrogen and carbon, forest floor weight, forest floor depth, and forest floor nitrogen per hectare. Comparisons of of significant differences of means for stumping, fertilizer and treatment effects were done utilizing Tukey's HSD and least squares. The factors of stumping and fertilization are considered fixed. The application of the treatment blocks in the five separate sites introduced a random factor. Differences in p values were observed and are called significant when less than or equal to 0.05.
|