Long-term log decay experiments at the Andrews Experimental Forest, 1985 to 2185

CREATOR(S): Jerry F. Franklin, Mark E. Harmon
ORIGINATOR(S): Jerry F. Franklin
OTHER RESEARCHER(S): Jay M. Sexton, Becky Fasth, Jack Booth, John Moreau, Timothy D. Schowalter, John D. Lattin, Rick G. Kelsey, Elaine R. Ingham
ABSTRACTOR: Mark E. Harmon
25 Apr 2005
20 Sep 2018
Inorganic nutrients, Organic matter, decay rates, decomposition, nutrient cycling, inorganic nutrients, nutrients, woody debris, coarse woody debris, organic matter, invertebrates, logs
The experiment that generated these data is designed to test the effect of species of log, tissue type, soil contact, insects, and diameter on the rate of log decomposition and release of nutrients of logs (large, downed and dead wood and bark).
Experimental Design - TD014:
Description: The experiments are being conducted at six sites located within intact old-growth Douglas-fir-western hemlock forests. The experimental design is a split-split plot in time. Each site represents a block, and log species is the main plot effect. Substrate layer (that is, inner bark, outer bark, sapwood, and heartwood) is the subplot effect. Logs were sampled (or in some cases) resampled annually for first 5 years, and then sampled at 2, 4 and 10 year intervals. This allows for a greater temporal resolution initially, but decreases the resolution as time passes and decomposition slows allowing for long-term patterns to be seen.
Field Methods - TD014:

Logs of the four species used in the experiments were removed from four locations during September 1985. Low- standard access roads were constructed at each site during July and August 1985 to place the logs and then the roads were closed. Logs were placed on either side of the access roads after the 50 m point to reduce the microclimatic effects of stand edge on the experiment. To reduce the effects of road work on the stand, trees were not felled unless the road could not avoid them. Saving old-growth trees and large snags was given priority over saving small trees. Wherever possible, preexisting skidding roads were used. Logs crossing the road were removed so that a minimum of damage to the logs and soil occurred. The logs used in the experiment met specifications for diameter, length, amount of bark cover, and degree of decay. Logs that ranged in diameter between 45 and 60 cm over their length were deemed suitable. The final log length at the experimental sites was 5.5 m. Damage to the bark during the yarding, transport, and placement was minimized; however, logs with more than 10 percent of their bark missing were considered unsuitable. Logs with large decay columns, conks, or both were rejected. Logs were transported to the experimental sites on short-bed logging trucks and placed along the access road by mobile loader. Enough logs of each species were placed so that destructive samples could be made at least 18 times after placement--a total of 530 logs.

Invertebrates were excluded in a subsample of logs by placing them onto polyester mesh and then building an A- frame over the log. The frame was then covered with 0.8-mm polyester mesh. One-half of the exclosure length was also covered with a 0.1-mm mesh to exclude small insects, such as ambrosia beetles.

The logs were sampled in 1985 to determine the initial characteristics of each log. After that point subsets of logs were sampled to determine how their characteristics change over time. In some cases logs were resampled so as to preserve intact logs for future study. While methods have been standardized over time to the degree possible, the fact that logs are decomposing and changing also means that some of the methods have to change so that comparable data can be collected.

Bark coverage, length, and diameters were measured to characterize the initial condition of each log. The total area of bark missing was measured to the nearest 0.01 m2 using a gridded 20- by 50-cm metal frame. Diameter was measured at both ends and the middle of the log by using calipers. The maximum and minimum diameters were measured at each point to calculate the mean diameter. Log length was measured to the nearest 1 cm and represented a mean length when the ends were not cut parallel. The total length of the log suspended off the ground was measured to the nearest 0.1 m. The number of freshly cut branches was counted on each log to indicate if the log was from the bole or the crown of a tree. Log volume was calculated using Newton's formula:

V=L (Ab+4Am+At)/6

where V is the volume, L is the length, and Ab, Am, and At are the area of the base, midpoint, and top of the log. Surface area was calculated assuming the logs were frustrums of cones.

After 1985 these measurements were repeated with the exception that diameter measurements were also taken at the point where cross-sections were removed. In some cases diameter was estimated from digitized photographs of the cross-sections. Because the missing bark area increased as the logs decomposed, the length and width of missing bark sections was hand measured because using the original metal frame was not practical.

To characterize initial density, moisture content, and volume of tissue in each log, a cross-section 8 to 10-cm thick was removed from each end. All cross-sections were wrapped in black plastic bags and stored at 2°C until they were processed. The cross-sections were photographed and these were digitized to estimate the volume of outer bark, inner bark, sapwood, and heartwood.

The logs that were sampled after 1985 were determined randomly by drawing from the list of available logs.

To characterize how density, moisture content, and volume of tissue in each log changed after 1985, five 8 to 10-cm thick cross-sections, spaced out at roughly equal intervals were removed from each sampled log. The diameter of each cross-section and its location relative to the large end was measured. For some year logs that had already been sampled were resampled to increase the temporal resolution of changes. In this case two of the cut segments that remained after the initial sampling had cross-sections removed from their center. The time between the initial sampling and the resampling was kept to a minimum to reduce the effects of decomposer colonization from the cut ends. As the logs decomposed some became weak enough that cross-sections could not be removed without additional support. Bark was kept on the cross-section by initially using canvas strips, then duct tape, and finally cardboard wrapped with saran wrap. In addition cross-sections that were too weak to be carried without breaking were placed on octagonal plywood backer boards. To reduce movement the cross-sections were saran wrapped onto the backer boards. All cross-sections were placed in a black plastic trash bag and stored in refrigeration until they were processed.

Laboratory Methods - TD014:

Processing of the cross-sections included many measurements to characterize density, moisture content, and volume of tissue and extent of decay.

The cross-sections were photographed and these images were digitized to estimate the volume of outer bark, inner bark, sapwood, and heartwood. To aid identification of tissues and zones of decay, the different zones on cross-sections were marked with different colored pens. The initial digitizing system used slides that were back projected onto an opaque digitizing pad. The programs used to determine the areas was vector-based and while it could determine the areas of the various tissues and decay zones for the first three years, as the decay patterns became more and more complex, it became impossible to use and accurately determine areas. In 2016-2017 an alternative system based on pixels was used instead. In this case the slides were scanned and analyzed using the Image-J software. The first step in the Image-J process was to calibrate the linear horizontal distance on the image using the scale on the photo to determine how many pixels were in 70 cm. The total area of the cross-section was determined by digitizing the outermost boundary and selecting the area inside the boundary to determine its area using the threshold function. When bark was missing due to cross-section handling, the distance between the outmost bark and the wood was estimated. When bark had fallen off due to natural processes, the outermost boundary used was that of the wood. The next area to be determined was that of total wood by digitizing along the bark-wood boundary. When sapwood was missing due to handling the outer wood area was also estimated using the remaining sapwood. The heartwood area was then determined by digitizing along the sapwood-heartwood boundaries indicated on the images that had been marked with magic markers prior to the photography. In cases in which the sapwood-heartwood boundary was not marked on the image (typically ABAM and TSHE in advanced stages of decomposition) the average thickness of the sapwood on the undecayed initial samples was used to approximate the boundary. The area blue-stained by fungi was determined using the threshold function to identify darker areas of sapwood. This threshold was determined by comparing areas that clearly had blue-stain with the areas meeting the threshold; however soil contamination, bark dust, shadows, and other non-blue-stain processes could make the sapwood appear dark. It is therefore not a very reliable indicator of the area with blue-stain and might be considered more of a qualitative index as to whether blue-stain was present or not. The boundary of all the wood with decay was digitized and the area of each decayed zone was summed when there was more than one zone. This process was repeated for the heartwood to determine the area of heartwood that was decayed. These are all general methods, the exact method of digitizing, order of digitizing, creation of intermediate images (e.g., wood area or heartwood area) for further processing depended on the arrangement of the decay patterns found in a cross-section.

In addition, the following areas could be determined from these primary measurements: The bark area was determined by subtracting the wood area from the total area. The sapwood area was the difference between the total wood area and the heartwood area. The area of decayed sapwood was determined as the difference between the total area of decayed wood minus the area of decayed heartwood. The diameter was estimated from the total area assuming that the area was a circle (often it was elliptical).

The area of outer versus inner bark was initially calculated using the inner bark thickness measurements and the wood diameter based on the outermost wood surface. As it became more difficult to measure the inner bark thickness due to decay, the area of each bark tissue was estimated using the initial proportions of outer versus inner bark and applying that to the total bark area. This assumes that when outer bark is lost from the circumference, that inner bark is lost as well. Based on observations, this assumption conforms to how bark is typically lost during the decomposition process.

For the 1985-2005 period wood density was sampled radially along a series of transects along a cross-section that went from the top to the bottom through the center, and the two sides to the center. Rectangular blocks of wood were rough-cut to create these transects by chainsaw forming an X-pattern. One of the blocks sampled in the vertical direction and two shorter blocks were used to sample in the horizontal direction. Portions of the cross-section with large knots were avoided when possible. A table saw was used to trim the blocks to precise cross-sectional area (75 by 50 mm), and then smaller blocks were cut--either 25 mm or 38 mm wide in the radial dimension. The prepared blocks were categorized by sapwood, mixed, or heartwood types. Samples were identified as to type, log number, end, and piece by bar- code labels. All subsequent measurements (weight and dimensions) were referenced to these labels by microcomputer and a bar code reader. The volume of knots within each block was visually estimated by checking against prepared standards.

For the 2015 sampling cross-sections were sampled using two different methods as the log cross-sections had become decayed enough that it was not safe to cut them on a table saw. The first method was similar to the radial sampling in that bark and wood samples were removed from the top, bottom, and sides of the cross-sections. However, only one sample was removed for bark, sapwood, and heartwood per “transect”. The bark samples were combined into one sample and inner and outer bark were separated. The wood samples were kept separate for each transect and represented a physical average. The sapwood sample was generally rectangular in shape; the heartwood sample was a triangle centered on the pith (if it was visible). The dimensions, wet and dry weights of these samples were determined as in earlier years of processing. The second method used anticipated the method that will likely have to be used when the logs become extremely decayed. In this method each layer of tissue (outer bark, inner bark, sapwood, and heartwood ) is sampled as a series of concentric rings that are separated and weighed. For each ring of tissue the diameter, longitudinal thickness, and total wet weight was determined. When fragmentation removed a portion of the ring (typically for bark), the circumferential length of the tissue was measured as well as the radial and longitudinal length, and the total wet weight of the ring of tissue. Dry weights for the entire cross-section were determined by adjusting the wet weight using the moisture content of the samples removed in the first sampling.

The radial, longitudinal, and tangential dimensions were measured to the nearest 0.1 mm with calipers for samples removed from each cross- section cut. For samples removed radially along a transect, a polynomial relationship was found that could be used to estimate the cross-sectional area at any point to within 1.8 percent of the actual value by using the cross- sectional area at both ends and the middle. The tangential and longitudinal dimensions at both ends and the middle were used to parameterize this relationship for each cross- section.

The weight of each block or sample before and after oven drying was determined to the nearest 0.01 g by using an electronic digital balance linked to a microcomputer. Oven drying was generally at 55°C for seven days or until the dry weight stabilized.

After recording the thickness at four points of each cross-section, the outer and inner bark were also sampled for density and moisture content for each cross-section typically by removing subsamples from the top, side, and bottom that were then combined into one sample. Each subsample of bark was 20- to 30-cm length of the circumference. The radial, longitudinal, and tangential dimensions of these pieces were recorded. For radial dimensions, at least six measurements were made to give a reasonable average.

As decomposition proceeded it became difficult to measure some tissue thicknesses directly. In the case of inner bark extensive decomposition lead to a collapse of this tissue. Therefore the initial thickness measurements from 1985 were used instead. For sapwood, either fragmentation, complete decomposition, or decay obscured the sapwood-heartwood boundary. The initial thickness of sapwood in 1985 for a log was used to determine the approximate thickness of sapwood samples. The latter method was used extensively in 2015 for both Pacific silver fir, western redcedar, and western hemlock. For the first two species sapwood was often missing. For western hemlock the sapwood-heartwood boundary was not always obvious.

Outer and inner bark were separated by chisel for all species, except Pacific silver fir. The weight of each bark sample before and after oven drying was determined to the nearest 0.01 g using an electronic digital balance linked to microcomputer. Oven drying was at 55°C for seven days. Previously oven dried outer bark samples were soaked in water 48 hours, and then volume was measured to the nearest 1 cm3 by water displacement. This method was the preferred one to determine volume for density calculations. It was found that the irregular shape of the outer bark surface made volume estimates from dimensions highly uncertain.

Moisture content, calculated from the weight before and after oven drying, was expressed as a percentage of oven dry weight. Density was calculated as oven dry weight over green volume. For outer bark samples, volume was based on water displacement; for other tissues it was based on dimensional analysis for both the individual sample and the ring-based sampling methods.

In 1985 each tissue type was sampled from a subset of 10 logs of each species for nutrient content and cell wall carbon chemistry (that is, lignin, cellulose, acid detergent fiber). After 1985 samples were taken from each log that was processed. In 1985 a separate set of bark subsamples was removed and stored frozen until processing. In 1985 sapwood and heartwood samples were removed from the oven dried pieces as were inner bark, sapwood, and heartwood after that period. Outer bark samples after 1985 were taken from samples that were dried, but not used in the water displacement determination of volume.

All tissues were first coarse ground, and then fine ground with a Wiley mill to pass a 40-mesh screen. Nitrogen content was measured using microKjeldahl digestion. Concentration of assorted elements--including calcium, copper, iron, potassium, magnesium, manganese, phosphorus, sulfur, and zinc--were measured by inductively coupled argon spectroscopy. Acid detergent fiber, permanganate lignin, and cellulose were measured using the procedures of Goering and Van Soest (1970). Unextracted material was used for these analyses.

Permanent Plots - TD014:
Description: A map indicating the position of logs at each site was prepared to aid relocation (maps on file in RWU- 4356, Forest Science Laboratory, Corvallis, OR). Each log was marked with an aluminum numbered tag nailed at the top of each end. The position of fiberglass t-posts along each access road were surveyed using by a jacobs staff, compass, fiberglass tape, and clinometer. Backsight measurements were made at each position. The distance, compass bearing, and slope of each log was surveyed from the nearest fiberglass t-post.
This is a very long-term study (200 years). Other data sets utilizing the same logs include TD18 (nitrogen fixation) and TD20 (respiration). Data also exists for the insects attacking logs in first 3 years (Jack Lattin), seasonal variations in moisture content (Mark Harmon) and temperature (Tim Schowalter), and leaching losses (Mark Harmon).
For details see
Garrison et al., 1976
H. J. Andrews Experimental Forest. Pacific Northwest/Oregon Cascades.
Ground condition