HJA newt logoHJ Andrews
Login Donate

Comparison of terrestrial versus aquatic decomposition rates of logs at the Andrews Experimental Forest, 1985 to 2050

Mark E. Harmon

Mark E. Harmon

Mark E. Harmon

Becky Fasth

Mark E. Harmon

3 Jun 2005

22 Jan 2014

Coarse woody debris, Decay, Decomposition, Geomorphology, Logs, Woody debris, Inorganic nutrients, Organic matter

This study is contrasting decomposition of logs within a stream versus on an adjacent terrestrial site at the H. J. andrews Experimental Forest. Changes in log volume associated with fragmentation and changes in wood density associated with leaching and respiration are being followed.


Experimental Design - TD017 :


This experiment was designed so that the decline in log density, volume and dry mass over time could be compared between log species and stream and upland sites using analysis of covariance. A total of 180 logs were used in the experiment. Thirty logs of each of three species (i.e., red alder, Douglas-fir, and western hemlock) were added to each site (i.e., stream versus upland). Periodically a subsample of logs will be removed from each site and log volume, mass, and bark cover will be measured. In addition, each removed log will be thoroughly dissected to determine the volume of decay and insect galleries and the fractional volume of the substrates remaining. Samples of each substrate will be taken in order to estimate the substrate density, moisture and nutrient content. Three logs of each species will be removed from each site at each sample time. The first sample will be removed after 2 yrs of decay and then at 4 yr intervals until 32 yrs after the initiation of the experiment.

In addition to examining the decay of logs, their movement downstream will be followed. The exact nature of the statistical analysis used for this aspect of the study has not been determined.

Field Methods - TD017:


Source Areas:

The logs used in the establishment of this experiment were taken from four locations in the Blue River Ranger District (Figure 1). Western hemlock logs were taken from subcanopy trees felled between June 5-7 near or within the Upper Lookout Creek study site. Age of the trees felled were in greater than 100 yrs and all trees were in suppressed and intermediate canopy classes. The forest at this site old-growth Douglas- fir, although salvage operations had occured in the past.

Douglas-fir logs were removed from a mature forest growing adjacent to the rock quarry on Forest Service Road No. 15-130. Trees were felled on June 10 and June 18. Most of the felled trees were in the suppressed and intermediate crown classes, although two trees were codominant. One tree was being attacked by bark beetles, but the foliage was green (logs 71, 72, 73). The age of trees felled was 130 yrs.

Red alder was removed from two sites off of the H.J. Andrews Experimental Forest. Thirty-eight logs were removed from the North Quartz Creek site which was adjacent to Forest Service Road No. 15-126. Felling of these trees started June 9 and was completed on June 13. Tree age was generally less than 30 yrs old and canopy class was dominant to codominant. The forest was a successional riparian type; dominated by red alder. The second source area for red alder was from Cougar Creek near the maintenance storage area along Forest Service Road No. 19-411. Felling at this site occurred on June 19. Trees in dominant to codominant canopy classes were used and tree age was less than 30 yrs. As at Quartz Creek the forest was a successional riparian type dominated by red alder.

Transport of Logs to Study Site:

Logs were moved from source areas to the study site by a number of processes. Western hemlock logs were carried to the study site by a crew of 6-8 using rope timber carriers. Red alder and Douglas-fir logs were transported to the study site in pick-up trucks. At the Quartz Creek site, red alder logs were moved From the stream to the road using a slack-line system.

Placement of Logs:

Logs on the upland site were placed using a crew of 6-8 workers with rope timber carriers on June 17 and June 20. All but 8 of the logs were placed on June 17. Thirty positions were flagged at 10 m intervals along a "cat-trail" that ran through the was placed in a line perpendicular to the road. The location of each log with respect to the road (i.e., nearest road, middle, farthest from road) was randomly selected. Logs were placed as they "fell" so that logs may have been suspended by underlying logs.

Logs were placed in the stream site by a combination of man-power and a slack-line system driven by an electric winch. Stream placement ( occurred on June 21, 24, and 25. Ninety logs were added to the stream over a 150 m mapped reach Prior to placement the location and orientation of logs was marked with colored flags in the stream channel. Logs were moved to the flag and placed so that the center of each log lay over the flag. In order to speed log location and placement and to spread the logs out over the stream reach, they were systematically spaced. Six logs, two of each species, were added to each 10 m segment. The order of species within any 5 m segment was random. Within each 10 m segment the position of the log with respect to the stream channel was randomly chosen. Two of the logs in each 10 m reach were placed in the zone of current flow, two were placed half-in the water and two were placed between the current stream level and that of winter flow. Orientation with respect to stream flow was chosen randomly for each log and four orientations were used. Angle to the current was ( either 0, 45, 90, or 135 degrees. After placement, log location was noted on the map of the stream reach.

Map Preparation:

Maps of both the stream and upland study sites were prepared. In the case of the upland site a tape and compass were used to measure the location of the tributary stream, cat road, and log locations relative to Forest Service Road No. 1506. Louis Stubecki and James Westman surveyed the upland map.

Preparation of the stream map was more complex than the upland map. Randy Wildman supervised the mapping of a 300 m stream reach. The upper 150 m section was used to place the logs, while the lower 150 m was mapped in order to track log movement. A compass line was surveyed down the study stream reach and white PVC posts were placed on the stream banks at 10 m intervals. At 1 m intervals the present flow and winter flow was measured with a staff rod held perpendicular to the compass line. Location of rocks larger than 1 m in diameter, logs larger than 15 cm diameter and 1.5 m long, trees and snags in or adjacent to the stream channel and bedrock exposures was measured with the staff rod and sketched onto the map. Finally, the location and orientation of the experimental logs was added to the map immediately following log placement.

Log Descriptions:

Before placement, each log was tagged and measured to determine exterior dimensions (see format 1). Each log was tagged with four tags; two tags on each end of the logs. The diameter at each end of each log was measured with a diameter tape to the nearest 0.1 cm. Mean log length was measured to the nearest 1 cm after end cross-sections had been removed. After placement bark cover was estimated to the nearest percent. Only zones with sapwood showing were considered bark free.

In subsequent samplings of decayed logs the bark cover was estimated and the diameter and distance from the large end of the cross-section that were removed was recorded.

Sampling of Cross-sections:

For the undecayed logs cross-sections 3 cm thick were removed from both ends of each log with a chainsaw. For red alder and Douglas-fir logs, cross-sections were removed within a few hours of tree felling. Logs of western hemlock were sampled from a few hours after felling to three days after felling. This was a period of wet weather so drying during this period of storage was probably minimal. Before the cross-sections were removed, the outer 5-10 cm of each end of the log was trimmed away so a clean, undried section of log could be sampled. All cross-sections were sealed in plastic bags and those not processed within 24 hours were stored under refrigeration.

For the decayed logs 6 cross-sections were removed with a chainsaw systematically along the length of each log. The position and diameter of each of these cross-sections was recorded. In addition the end diameters were remeasured. In 1987 the large end diameter was measured, and 1989 & 1991 the small end diameter was measured. After this time both ends were measured in the field. For the 1987, 1989, and 1991 years, the missing diameter was estimated from the original diameter and the degree of bark loss. Tha is if all the bark was lost this was subtracted from the original diameter.

For undecayed logs , cross-sections were sampled to determine the fractional volume, moisture content, and density of each substrate. Fractional volume of each of four substrates (i.e., heartwood, sapwood, inner bark and outer bark) was calculated from the mean radial thickness of each substrate. Thickness was measured to the nearest mm. Heartwood and sapwood radial thickness was determined by measuring along the long and short axis of the cross-sections. Bark thickness was measured for 1-4 locations on each cross-section depending upon the variation in thickness observed. The methods used to distinguish substrates for the three species differed. In the case of Douglas-fir all four substrates were quite distinctive and easily separated. The pH indicator alazarin-red S was used to distinguish heartwood from sapwood. When applied to Douglas-fir sapwood the indicator turns red, whereas on heartwood the indicator turns yellow. The heartwood-sapwood boundary of western hemlock ws more difficult to determine than for Douglas-fir. After examining many cross-sections, we determined that the sapwood of western hemlock was slightly darker than the heartwood. This was especially apparent when the cross-section was "split open". The darker coloration was caused by the higher moisture content of the sapwood. In the case of red alder there was no way to reliably separate heartwood from sapwood. Therefore, all wood was considered sapwood for this species.

The volume of each substrate was calculated from the radial measurements from each end and log lengths by assuming logs were frustrums of cones. Inner and outer bark volumes were adjusted to reflect losses prior to placement. Fractional volume was calculated using the total volume based on radial measurements. This volume was slightly lower than that estimated from taped diameter measurements.

Photographs were taken to document changes in the cross-sections as they began to decay. These have not been digitized as of 1991.

Cross-sections were dissected to determine moisture and density. The mean exterior dimensions of these samples were measured to the nearest mm in order to calculate volume. Pieces of sapwood of Douglas-fir and western hemlock, and inner bark and outer bark of all three species were assumed to be rectangular parallelepipeds and volume was calculated from the Comparison to volumes measured using water displacement indicate these volumes estimates were accurate and unbiased. After measurement, the outer bark and inner bark were separated using a chisel. The heartwood of Douglas-fir and western hemlock, and the sapwood of red alder was cut in triangular pieces. The volume of triangular shaped pieces was calculated using Hero's formula:

A = [s(s-a)(s-b)(s-c)^0.5

where A is the area, a, b, and c are the lengths of the three sides and


The volume, V, was therefore:

V = Ad

where d is the thickness.

Based on susequent studies, it appears that Hero's formula underestimated the volume of triangular pieces by 24.7 %. The calculated volume of triangular pieces was therefore multiplied by 1.247 to correct for this bias. Starting in 1987, the volume of triangular pieces was estimated using a formula for a sector of a circle. This formula did not have the same bias as the Hero formula. The exact formula used was: A=S/360*pi*R^2 where S is the angle of the sector in degrees, and R is the mean radial lenght of the sector.

In 1985 wet weight was measured to the nearest 1 g on an electronic, digital scale. Starting in 1987, wet weight was recorded to the nearest 0.01 g. The 1985 samples were air dried for 3-5 weeks and then oven dried for 4 days at 55 degrees C. Samples were allowed to cool to room temperature and weighed to the nearest 0.01 g on an electronic digital scale. The based on a subsampling of 95 samples that were redried and weighed just after removal from the oven, the moisture content at this times was 2.5%. The initial density and moisture content was adjusted in the TD173.dat file to remove the effect of this slight increase in moisture content.

Samples measured in 1987 and after were dried at 55 degrees C for 5-7 days immediately following wet weight determination. Oven dry weights were measured immediately following removal from the oven.

Total mass and fractional mass were calculated using substrate volumes and the mean density of substrates for each log. In the case of red alder, outer and inner bark were combined because the thickness of inner bark was not measured accurately enough (it was often less than 1 mm thick). Subsequent analysis of the Douglas-fir and hemlock bark densities indicates that thickness was not measured accurately. In future analysis it may be best to combine outer and inner bark results.



The elevation of the site is approximately 2500 ft 762 m). The upland site is situated on a gently sloping (<10% slope) bench that is 20 m above Lookout Creek. The stream site is within the "active" channel or zone exposed to annual high flows. The channel gradient at the point of log placement is 7 to 10% and the general downstream direction is west. Much of the reach 50 m or more below the easternmost portion of the reach is bordered by terraces formed by the 1964 flood. Vegetation on these areas consists of tree seedlings <2 m tall, low shrubs and herbs. Approximately 50 m from the easternmost portion of the reach there is a large log jam that completely crosses the channel. Another log jam that partially crosses the channel occurs 140 m from the easternmost portion of the reach.

Both upland site and the stream reach have old- growth Douglas-fir/western hemlock forest, although the stream reach has a larger component of western red cedar than the upland site. The habitat type of the upland site is Silver fir/ Alaskan ( huckleberry/ bunchberry (Abies amabilis/ Vaccinium alaskense/ Cornus canadensis) (Dyrness et al. 1974, Hemstrom et al. 1982). Both sites have had some salvage operations although the extent of salvage in the stream channel appears to be minor. Salvage 1967-68 and 1979-80.

Climatic conditions have not been determined at the site. However, precipitation at the nearby Earthflow site for 1979-86 was 205 cm. Thermograph data from Upper Lookout Creek had a mean annual air temperature of 5.5 C, with minimum and maximum values of -7.1 and 28.9 C, respectively.

Alnus rubra, Pseudotsuga mensiezii, Tsuga heterophylla

Garrison et al., 1976
Garrison, G.A., J.M. Skovlin, C.E. Poulton and A.H.Winward. 1976. Northwest Plant Names and Symbols for Ecosystem Inventory and Analysis (4th ed.). Gen. Tech. Rep. PNW-46. U.S. Department of Agriculture, Forest Service, Pacific Northwest Forest and Range Experiment Station. 267 p.. (http://www.fs.fed.us/pnw/publications/pnw_gtr046/)

H.J. Andrews Experimental Forest along the upper section of Lookout Creek.

4 years

Study continues and further data collection is planned

Data is updated as needed

"Ground condition" is the range of dates during which the site was visited and data collected.