CHAPTER I
RESULTS AND DISCUSSION

Bulk Density

Concern about bulk density increases from stumping was the major focus of this study, but the results after twenty years show that there is little difference in bulk density between the stumping and no stumping treatments, among fertilizer treatments, and among treatments and controls (Fig. 1.2). These results are consistant at all sites studied (Fig. 1.3). Bulk densities were higher for the stumping treatment, after more than twenty years, at all sites except Gates (Fig. 1.4). The average stumping treatment bulk density was 3% higher than no stumping for the combined data. When the sites are considered separately, only Apiary shows a statistically significant difference in bulk density for treatments (stumping and nitrogen application) with the two-way ANOVA model. When the main effect of stumping is considered separately, Apiary and LaGrande show a statistically significant increase caused by stumping alone (Table 1.2). However, the average stumping treatment bulk density was only 3% higher than no stumping for the combined data. When the sites are considered separately, only Apiary shows a statistically significant difference in bulk density for treatments (stumping and nitrogen application) with the two-way ANOVA. When the main effect of stumping is considered separately, Apiary and LaGrande show a statistically significant increase in bulk density caused by stumping alone (Table 2.2). When all sites are considered together (n=40, p=0.05), there are no statistically significant differences in bulk densities between the treatments.

The bulldozing activity shows no effect on soil bulk density variability after 20 years. For the combined data the variation in bulk density measures was not different between stumped and no-stumping plots. Stumping treatments reduced within plot bulk density variability (subplots) at all sites except Apiary. Among-plot variability was reduced by stumping at Gates, Sweethome and LaGrande (Table 1.3).

The higher bulk densities measured on the no-stumping plots at Gates may be the result of the chance grouping of these randomly assigned plots into one area that has higher density or that was the center of activity during stumping or logging. The soils at Gates also contained many large decomposing logs. The variability in bulk density in the no-stumping plots at Gates was the highest in the study. The statistically significant difference at Lagrande and Apiary caused by the stumping may be due to the finer texture of these soils. Apiary also has a high clay percentage and there were heavy rains the month before the treaments were applied.

The bulk density results from Sweethome show recovery from stumping in the second decade after treaments (Fig. 1.5). The data taken in 1991 (Thies et al. 1994) and again in 2003 show a decrease in the bulk density of the stumped plots of 7%, while the results from the no-stumping plots are almost the same (-1%). The 2003 results show that the stumping and no-stumping plots are now very similar (0.97 and 0.96 g cm-3, respectively). The no-stumping plots have already returned to the bulk densities measured in the forest plots in 1991 (0.96 g cm-3). At the rate of recovery seen in the stumped plots, they will recover before 2010. These recovery rates for the surface soils are similar to those found in other studies (Froelich et al. 1985).

This study sampled bulk density only to 20 cm. While this is the main area for root growth, other stumping studies have shown that compaction often occurs at depth, while surface bulk density can sometimes decrease (Page-Dumroese et al.1998). Future studies of the bulk density at deeper levels on these sites may be warranted if growth data indicate lower growth on these plots. There is little evidence of fertilizer effect on bulk density.

While it is clear that the stumping treatments have increased the soil bulk density at all sites except Gates, and that recovery is slow, the increase is not likely to affect tree growth (Wass and Smith 1997; Heninger et al. 2002), although there may be some detrimental effect on early root development (Page-Dumroese et al. 1998). It should also be noted that stump removal on these sites was done in 1978-80 by pushing with a bulldozer. Today stump removal would most likely be done with an excavator, which should cause less soil disturbance (Bloomberg and Reynolds 1988; Wass and Smith 1997). Stumping not only causes increases in bulk density through compaction, it can also lower bulk density by mixing forest floor material into mineral soil and loosening soil around extracted roots. Keeping in mind that these sites were chosen for lighter soils and that stumping was intended to be done when soils were dry, it could be concluded that the increased soil density caused by the stumping treatments is not severe enough to rule out the practice if it successfully reduces P. weirii mortality.

Nitrogen

After more than twenty years the average mineral soil nitrogen concentration was lower in the stump removal treatments compared to the no-stumping treatments and the adjacent forest for the combined data (Figure 1.6). This reduction occurred at all sites (Fig. 1.7). The average reduction in mineral soil nitrogen in the stumped plots compared to the no-stumping plots was 20%. This difference is statistically significant individually at Gates and Sweethome with the two-way ANOVA (p = 0.05). Comparing main effects with statistical analysis shows that the reduced mineral soil nitrogen was caused by the stumping treatment and is not related to the fertilizer application. The lower nitrogen concentration is a statistically significant result of stumping at all sites except Hoodsport (Table 1.2). At all sites except Hoodsport, the reductions of nitrogen concentration are incremental from controls (undisturbed forest) to no stumping (clearcut) to stumping (clearcut and stumped) (Fig. 1.8).

This trend is repeated in the mineral soil concentrations of total carbon (Fig. 1.9). average decrease is 22%, and again, the result of the stumping, not fertilization. This is a statistcally significant difference for all sites (n=40). The difference in total mineral soil carbon is a statistically significant result of treatments individually at Gates andSweethome, and a significant result of stumping at all sites except Hoodsport (p = 0.05). Forest floor measurements showed a reduction in depth and total carbon at all sites in the stumped plots compared to the no stumping plots (Fig. 1.10). The average reduction in forest floor depth over all sites was 24%, and the average reduction in forest floor total carbon was 6%. Both of these measures were a statistically significant stumping effect overall, and individually at Gates and Sweethome. Despite forest floor total N concentrations that are similar in the stumping and no stumping plots, forest floor N in kilograms per hectare declines at all sites from no stumping to stumping (Fig. 1.11). The average decline is 28%. This is a statistically significant result of stumping for overall means (n=40) and individually at Hoodsport, LaGrande and Sweethome.

Combining the mineral soil and forest floor nitrogen data shows that the overall losses of kg N ha-1 are again a statistically significant result of stump removal overall and individually at all sites except Hoodsport (Table 1.2). This decrease in the N pool of the forest floor and top 20 cm of the mineral soil caused by stump removal averages 19%. Combining mineral soil and forest floor carbon data shows an average decrease of kg C ha-1 of 21% as a result of stump removal.

The loss of mineral soil N and C, and forest floor depth and mass caused by the added disturbance of stumping indicates a reduction of organic matter. Soil organic matter has been shown to have an important role in forest ecosystems, acting as a reservoir for nutrients (Weber et al 1984), as well as enhancing moisture holding capability and contributing to soil structure. Soil biota necessary for nutrient cycling are dependent on soil organic matter, and alterations in organic matter can have large effects on soil and site properties (Harvey et al. 1994; Jurgensen et al. 1997). Nitrogen is generally the most limiting nutrient in Pacific northwest forests. The largest reservoir of nitrogen in these temperate conifer forest ecosystems is in soil organic matter. The portion of the N pool in the upper 30cm of mineral soil may be 50% or greater with another 5 or10% in the forest floor (Dyck and Beets 1987, Page-Dumroese et al., 2000, Klopatek, 2001, Finer et al. 2003). Ross et al. (1995) reported 5-7% of forest ecosystem N was found in soil microbes. The 20% reduction in total N in the top 20cm of soil, and the 28% reduction in forest floor kg N ha-1 seen to result from the stumping treatments should be a concern to managers for long and short term productivity. Studies of N levels in soil depths below 20 cm would be necessary to see if the N has merely been leached to a deeper position in the profile (Carter et al. 2002).

Many studies have shown losses of carbon and nitrogen after clear-cutting (Clayton 1985, Pennock 1997), as the normally tight nutrient cycle becomes "leaky" from disturbance. Increased N mineralization can lead to leaching of nitrate. These reductions are related to changes in organic matter and forest floor material (Aber et al. 1978). Clearcutting also changes the soil biota communtities (Fons and Klinka 1990). Of interest in this study is the added incremental reduction caused by the stumping treatment (Figure 1.8), and the substantial losses of N in the A horizon. The clear-cutting alone has caused a reduction in organic matter by physical removal, increased decomposition, lack of litter fall, mineralization and leaching (Holmes and Zak 1999). The added activity of the stumping operation has augmented the removal of these organic components. Although generally inputs of N can make up for losses from harvest, the increased loss of N in these stumped plots may be large enough to affect site productivity. In studies in Saskatchewan, Pennock and van Kessel (1997) found losses of exchangeable calcium and magnesium, soluble organic phosphorus, cation exchange capacity (CEC) and base saturation in soils associated with the lower nitrogen and carbon 6-20 years after clear-cutting. It is possible these other nutrient losses are present at these five stumping sites, and probable that losses of the organic component have decreased CEC.

The additional loss of C and N from stump removal is similar to losses from soil disturbance noted in studies of mechanical site preparation, a standard silvicultural practice widely used to improve seedling establishment and reduce competition after harvest. Vitousek and Matson (1985a) showed the importance of organic matter in immobilizing mineralized N on coniferous sites undergoing site preparation. Burger and Pritchett (1984) showed total N and organic matter reductions and increased mineralization. Reductions of total nitrogen and carbon, along with reduced available phosphorus were found to result from site preparation in Alberta, Canada, (Schmidt et al 1996). Simard et al. (2003) found that removing the forest floor during site preparation in eastern British Columbia could reduce the N pool by 28%. Piatek and Allen (1999) showed a 13% reduction in the N pool with heavy site prep and disking. Merino et al. (1998) shows similar reductions of total N in mineral soil caused by harvest and site prep consisting of disking, along with significant reductions of soil organic matter and some decline in CEC.

As in the site preparation studies, the added reduction in organic matter from the stumping operation may come from a number of causes. Besides the dilution of the A horizon with mineral soil brought up with the roots, mixing, aeration and increased temperatures and moisture can lead to increased decomposition, mineralization and leaching. The destruction of soil stucture can release protected, low C/N ratio organic matter (Caldero et al. 2001). The lack of N immobilization due to reduction of soil microorganisms and vegetation and removal of litter and the associated carbon to enable microbial immobilization (Vitousek and Matson 1985b) can contribute to nutrient losses. Removal of large woody roots, while leaving fine roots with their higher N concentration, will likely increase decomposition rates and limit immobilization. The timing of these stumping operations in the fall, at the end of the dry season, may encourage increased leaching with winter rains and spring run-off, with less plant and microbial activity and immobilization. Wet, anaerobic conditions along with increased mineralization may increase denitrification.

As in site preparation, the degree of loss of N and organic matter is related to the degree of disturbance. Nohrstedt (1994) notes that lysimeter studies have shown increasing inorganic N in soil solution with increasing disturbance of the humus layer. Site preparation techniques that involved disking showed the most losses. Distubances caused by stumping operations will vary. Displacement caused by uprooting the stumps can bring B horizon material to the surface that is low in organic matter. Other areas may only have the forest floor displaced. These stumped sites were never systematically evaluated for degree and amount of distubance, but due to their size, they could be considered highly disturbed (Thies and Nelson 1988). Wass and Smith (1997) measured 58% of a harvested area disturbed by stumping operations. An additional 16% was disturbed by windrowing extracted stumps. Quesnel and Curran (2000) found 16% of their push-over treatments disturbed the soil enough to fail to meet forest practice guidelines. Other studies (Smith and Wass 1989; Davis and Wells 1994) have found up to 95% of an area where stump removal was done by bulldozing to have soil displacement. It is reasonable to conclude that keeping soil distubance to a minimum will mitigate losses of N and organic matter.

These losses will take place in the first years after treatment. Tuttle (1985) found increases in N three years after site prep treatments. The results from Sweethome from 1991 and 2003 indicate that the leakage has stopped, but the deficit remains (Figure 1.12). While the 2003 mean for total N in the no stumping plots is 6% lower than the 1991 mean (Thies et al. 1994) , the stumped plot total N has increased 8%. This shows that some recovery is underway, but the difference is still statistically significant. One factor that may lead to the small recovery of the stumped plots is the young age of the stand. The trees and other vegetation are accumulating nutrients during this period of recovery and fast growth. Inputs into this system may be accumulating in above ground vegetation at the expense of the soil pool. Additions from litter fall will be small. Studies by Aber (1978) and Parker et al. (2001) indicate that recovery of forest floor may take 30 years in eastern hardwoods.

The one-time addition of nitrogen has not mitigated the low nitrogen levels found after 20 years on these sites. Forge and Simard (2000) found that fertilization can increase nitrogen losses from clearcuts in the first year after application and may be due to changes in soil microorganisms. Although total microbial biomass was not increased with fertilization, reduction of fungal species and increases in bacterial species may lead to faster N turnover and more mineralization and leaching. Increases in populations of denitrifying bacteria in fertilized areas may increase N losses. Little immobilization would be likely with fall application on these disturbed, unplanted sites.

Not only has the initial clear-cutting reduced the biomass and altered the species composition of the soil microbes, and thus its capability to immobilize nitrogen, the added disturbance and root removal of the stumping operation seems to have further reduced this capability for nitrogen retention. Stumping has also been shown to adversely affect Douglas-fir mycorrhizae (Page-Dumroese et al. 1998). It may be that the positive effect of stumping in controlling Phellinus is not only from removing the inoculum, but also by making the habitat generally less favorable to fungal species.

Site Differences

The studies were carried out on five different sites and five different soils across the Pacific Northwest. Although the finer grained soils of Apiary and LaGrande showed the most compaction, there is little evidence that the general effects caused by the treatments differ from site to site or soil to soil. The soils varied from the very rocky at Hoodsport to the fine loess and ash of LaGrande, and the results were similar. There was a small increase in bulk density, and a significant decrease in total nitrogen, carbon and forest floor depth. Stumping increased variability in soil bulk density at Apiary, but decreased the variability at the other sites.

Nitrogen concentrations measured for this study allow comparisons of N pools of all sites in the forest floor and top 20 cm of the mineral soil (Figure 1.13). Nitrogen levels at Hoodsport were low, and showed a similar percentage reduction in the mineral soil (20%), but twice the reduction in forest floor kg N ha-1(54%). These low nitrogen sites may be most susceptible to damage from stumping.

It seems clear from this study that when deciding whether a site is suitable for stumping treatment, in addition to considering the physical properties of soils and operational procedures that minimize forest floor and mineral soil disturbance (Quesnel and Curran 2000), the productivity of the site, nitrogen pools and nutrient cycling must be considered. On a properly selected site, increases in bulk density do not seem to be a long term concern. A low productivity site may not be suitable or it will need fertilization or organic matter inputs after stumping treatments. The continuing evaluation of growth and foliar data on these sites will give us a better idea of long term effects of the N losses. The long recovery time needed to restore soil organic matter and dependence of functioning soil biota on soil organic matter must be taken into account in order to determine whether the loss of productivity caused by the additional damage to the organic component of the soil by stump removal for P. wierii control can be offset by the reductions in tree mortality.