Discussion 
Wind River Experimental Forest, WA
H. J. Andrews Experimental Forest, OR (photo by Jay
Sexton)
The
decomposition rates we observed for Abies logs was generally lower
than reported for Abies concolor in the past (0.05 year-1)
(Table 9). However, our
current estimates for Abies concolor at SQNP and Abies amabilis
at HJAF are similar (0.051 and 0.051 year-1, respectively).
The lower values observed at other sites appears to be related to
less favorable climatic conditions for decomposition that our study
included. The decomposition
rates for Pseudotsuga and Tsuga that we found are generally
higher than those reported in the literature.
For example, past estimates of decomposition rates of Pseudotsuga
have ranged from 0.007 to 0.014 year-1, whereas our range
was from 0.014 to 0.015 year-1.
It should be noted that the earliest estimates for this genus
tended to undervalue the loss of volume as decomposition proceeded (Means
et al. 1985), although it is also possible that larger boles were included
in earlier studies, a factor that would also lead to slower decomposition.
For Tsuga, the most noticeable difference was for CHEF, with
the earliest estimates of decomposition rates (0.007 year-1)
being about one-third of the value we found.
It is not clear what would have lead to such a difference, although
as noted the earliest estimates of decomposition rates tended to discount
volume losses and were solely based on residual density.
At HJAF, our estimate for Tsuga was generally higher (0.023
year-1), although near the upper end of the reported range of
0.016 to 0.024 year-1 reported by Sollins et al. (1987) and
Graham (1982), respectively. Reported
values of decomposition rates for Pinus contorta range from 0.012
to 0.027 year-1 (Busse 1994, Fahey 1982), and our range, for a
potentially similar breadth of temperature conditions was 0.023 to 0.042
year-1. It may be possible that the sites we studied were slightly
wetter, or it may be that other factors are at play. Finally, estimates of Thuja decomposition reported by
Sollins et al. (1987) are quite similar to those we found, with values of
0.009 and 0.007 year-1, respectively.
Several
future analyzes need to be conducted to improve our understanding of
decomposition rates of dead trees. First,
a more detailed comparison of the estimation methods needs to be
undertaken. When
decomposition undergoes a true negative exponential form, then all the
methods we employed are likely to give similar estimates if the sample
size is adequate. However,
this is unlikely to be the case if the decomposition rate changes over
time. The most likely pattern
is that decomposition at first accelerates and then slows down (Harmon et
al. 2000). While the
decomposition vector approach reveals this pattern, it may inadvertently
bias the average upwards. This
could be due to several factors, but one may be that by resampling logs
one tends to under sample the earliest phases of decomposition when the
rates tend to be low. Another
issue may be the tendency for chronosequence estimates using linear
regression to fit the data at the start and end phases of the
decomposition process well, but miss the middle phases.
This might tend to underestimate the decomposition rate.
Second, the
effects of climate are still not clear.
In part this is because we still do not have a wide range of
climatic conditions for many species or genera.
It is also caused by the complex interactions between precipitation
and temperature. Temperature
has an independent effect, but by influencing drying rates it also can
alter the effectiveness of a given amount of precipitation.
There should be an exploration of the use of an algorithm that
models the interaction of these two climate factors.
Hopefully this will explain why a consistent climatic signal is not
detected.
Third, the
estimates presented here are for logs lying on or near the forest floor.
It is unlikely that these rates will apply to standing dead trees
or even logs suspended off the forest floor (especially in dry sites).
As with the interaction with temperature and moisture, the
interaction between position (i.e., standing, suspended, in contact with
the soil) and precipitation is likely to be complex. In particular as
climates become wetter, the decomposition rate of standing wood is likely
to increase. Conversely as
climates become drier, the more likely it is that wood in contact with the
soil will decompose faster than standing or suspended wood.
Unfortunately there are few comparisons among the various positions
at a single site. For forests
west of the Cascades and northern Sierra crests, it is likely that
standing dead trees will decompose faster than those that are suspended or
in contact with the ground. Past
studies indicate that standing dead could decompose 12 to 92% faster than
downed dead trees (Harmon et al. 2004).
For forests east of the Cascades and Sierra crests, including those
in the Rocky Mountains, standing dead trees appear to decompose at very
slow rates relative to those in contact with the soil.
Thus assuming a standing dead decomposition rate that is <20% of
the rates reported here would be a reasonable assumption.
Fourth, it
is apparent that decomposer species may be influencing the difference in
rates observed among species. Decay
resistance ratings give a rough approximation of differences among tree
genera; however, particularly within the low decay resistance class there
is considerable variation that cannot be accounted for. In a practical sense this may not make a difference as long
as differences among genera are consistent.
However, if there is an interaction between decomposer species and
the site, then a fuller understanding will be required to make reliable
predictions.
