| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
a Dep. of Forestry, 228 Cheatham Hall, Virginia Polytechnic Inst. and State Univ., Blacksburg, VA 24060
b MeadWestvaco Corp., P.O. Box 1950, Summerville, SC 29484
* Corresponding author (meisenbi{at}vt.edu)
 |
ABSTRACT |
|---|
 TOPNOTES ABSTRACT Review of Harvesting Disturbance...Evaluating Changes in Forest...Evaluating Changes in Soil-Site...MATERIALS AND METHODSRESULTS AND DISCUSSIONCONCLUSIONSREFERENCES |
|---|
Abbreviations: ANCOVA, analysis of covariance • BMPs, best management practices • DB, dry harvested and bedded treatment • DEM, digital elevation model • DF, dry harvested and nonbedded treatment • HRI, harvest residue index • NRCS, Natural Resources Conservation Service • PDI, physical disturbance index • RCSB, rank change based on stand biomass • RCSI, rank change based on site index • RCTB, rank change based on individual tree biomass • SCFC, South Carolina Forestry Commission • SSURGO, NRCS soil survey geographic database • USDA, United States Department of Agriculture • USGS, United States Geologic Survey • USLE, universal soil loss equation • WB, wet harvested and bedded treatment • WF, wet harvested and nonbedded treatment • WMB, wet harvested, mole-plowed, and bedded treatment
|
Review of Harvesting Disturbance Concerns |
|---|
|
TOPNOTESABSTRACTReview of Harvesting Disturbance...Evaluating Changes in Forest...Evaluating Changes in Soil-Site...MATERIALS AND METHODSRESULTS AND DISCUSSIONCONCLUSIONSREFERENCES |
|---|
Studies on the effects of trafficking associated with wet-weather harvesting in the United States have shown that disturbances, especially those associated with skid trails, can reduce the height and diameter growth and survival of young trees (Youngberg, 1959; Perry, 1964; Moehring and Rawls, 1970; Hatchell et al., 1970; Shoulders and Terry, 1978; Lockaby and Vidrine, 1984; Scheerer, 1994; Aust et al., 1995; Heninger et al., 2002). In general, there have been mixed results with regard to the effects harvesting disturbances and site preparation have on site productivity in the southeast, but they appear to be very site specific (Miller et al., 2004; Miwa et al., 2004). Declines in productivity are attributed to repeated trafficking that causes reduced soil physical quality, erosion, nutrient loss, and organic matter loss (Greacen and Sands, 1980; Powers et al., 1990; Sheriff and Nambiar, 1995; Kozlowski, 1999; Miwa et al., 2004; Miller et al., 2004).
Many states, such as South Carolina, have responded to long-term productivity concerns about wet-weather harvesting by incorporating harvesting best management practices (BMPs) as a means of protecting site quality by limiting rutting and compaction (Aust and Blinn, 2004). An issue generally not considered by BMP requirements is the resilience of some sites to certain types of disturbance or the positive role site preparation can play in the recovery of the site (Miwa et al., 2004; Miller et al., 2004).
|
Evaluating Changes in Forest and Site Productivity |
|---|
|
TOPNOTESABSTRACTReview of Harvesting Disturbance...Evaluating Changes in Forest...Evaluating Changes in Soil-Site...MATERIALS AND METHODSRESULTS AND DISCUSSIONCONCLUSIONSREFERENCES |
|---|
Normal distributions of net primary productivity, volume, biomass, and site index are all stretched or shifted, relative to prior rotations, due to these confounding factors such that production at the end of the second rotation may even be higher than the prior rotation only because technological improvements might mask site degradation (Burger, 1994). Computer modeling has been used to account for these factors, but models are imperfect and sometimes unsatisfying because of scaling issues (Proe et al., 1994). The ability to make a compelling field evaluation of treatment and disturbance effects on changes in soil-site quality would be an important achievement (Comerford et al., 1994).
|
Evaluating Changes in Soil-Site Productivity Using Rank |
|---|
|
TOPNOTESABSTRACTReview of Harvesting Disturbance...Evaluating Changes in Forest...Evaluating Changes in Soil-Site...MATERIALS AND METHODSRESULTS AND DISCUSSIONCONCLUSIONSREFERENCES |
|---|
Many conceptual models exist for defining the biotic, abiotic, and cultural practices that influence forest productivity (Switzer, 1978; Burger, 1994; Morris and Miller, 1994). Morris and Miller (1994) described forest productivity as a function of plant potential, climate, soil-site quality, and catastrophe. This definition is useful because it separates soil-site quality from the major confounding factors associated with comparing productivity between rotations. We can further hypothesize that the change in site-soil quality will be a function of silvicultural treatments, inherent site factors, and harvesting disturbance.
The objectives of this paper are to address four questions. (1) Does wet-weather harvesting reduce soil-site productivity? (2) If so, does conventional bedding restore productivity? (3) If conventional bedding does not restore the reduced productivity of wet harvested sites, will an experimental mole-plow treatment succeed? (4) Can a model be developed that indicates which soil-site factors maintain or diminish site productivity after harvesting disturbance?
|
MATERIALS AND METHODS |
|---|
|
TOPNOTESABSTRACTReview of Harvesting Disturbance...Evaluating Changes in Forest...Evaluating Changes in Soil-Site...MATERIALS AND METHODSRESULTS AND DISCUSSIONCONCLUSIONSREFERENCES |
|---|
General Experimental Design and Treatment Installation
In 1992, three 20-ha, bedded, loblolly pine plantations were selected based on similar age (ages 20, 23, and 25), soil, and hydrologic conditions. Each plantation (block), located approximately 2.5 km apart, was subsequently divided into six 3.3-ha "operational-scale" treatment plots. Treatments were randomly assigned and included dry harvests with conventional bedding (DB) and without bedding (DF) and wet harvests with (WB) and without bedding (WF). The fifth treatment was an experimental mole-plow and bedding combination (WMB) designed to equilibrate the water table with subsurface channels after surface drainage had been disrupted. The sixth plot, an unharvested control, was not utilized in this study.
Harvesting was conducted by conventional commercial logging operations using feller buncher–grapple skidder systems. Despite being contiguous within each block, the treatment plots were laid out as individual harvest units with separate decks and skid trails to assure the treatments were fully independent. In the fall of 1993, two plots on each block received a dry-weather harvesting treatment. In the spring of 1994, the remaining three plots on each block were harvested during wet conditions to maximize soil disturbance. Bedded sites were sheared and drum chopped before bed installation. Mole plowing was done in October 1995, and bedding in November 1995. In July 1995, chemical weed control in the form of imazapyr (1.2 L ha–1) and glyphosate (5.6 L ha–1) was applied to each harvested unit. The sites were hand planted in February 1996 with best first generation, open-pollinated family, loblolly pine seedlings provided by the MeadWestvaco Corp. (formerly Westvaco Corp.) nursery. Plots were double planted to emphasize productivity as a result of site quality over stocking and survival effects. Extra seedlings were culled from double plantings that remained after the first year of growth.
Before harvest each treatment plot was overlain with a 20 by 20 m grid of permanent, 0.008-ha measurement subplots. A total of 1170 subplots were installed and all subsequent stand measurements were collected at these "polypedon-scale" subplots. Heights and diameters of all trees within each subplot were measured before treatment installation. The sampling intensity was 20% of the entire study area.
After harvest, site disturbances associated with each treatment were characterized by the percentage of coverage of five harvesting residue classes and five physical disturbance classes preceding site preparation (Eisenbies et al., 2004). An index of soil physical disturbance was determined by assigning an integer value to successively increased levels of soil disturbance (undisturbed [1], compacted [2], shallow rutted [3], deep rutted [4], churned [5]) and calculating a weighted average based on percentage of coverage. Harvesting residue biomass was calculated from percentage of coverage by regression (Eisenbies et al., 2002). A second inventory of height and diameter was conducted at age 5 at the 0.008-ha subplots across the study.
Ranking Procedure
The ranking procedure has four basic steps. First, determine the growth metric for treatment comparisons. Second, rank measurement plots within the study area. Third, determine the change in rank between time periods. Fourth, conduct a standard parametric analysis on the rank change data. Each step is detailed below.
= 0.05 level using prior rank as a covariate (SAS Institute, 2001). Means separations were determined by Fisher's protected least significant difference. To interpret the change in site productivity we compared all treatment responses using the dry harvested–bedded (DB) treatment as a reference, which was most similar to the operational treatment in the previous rotation and is considered the most operationally desirable. If change in rank was equal to or greater than the DB reference treatment we concluded productivity was maintained or improved. If change in rank was less than the DB reference we concluded productivity was diminished.
Regression Analysis
To assess cause and effect relationships, change in rank was evaluated at the polypedon-scale (0.008 ha) as a function of silviculture, site factors, and harvesting disturbance using multiple linear regression (SAS Institute, 2001) on a random subset of 198 observations stratified by soil physical and harvest residue disturbance. The candidate regressors were limited to those that were available for all 1170 polypedon-scale subplots (Table 1). Three models were constructed: change in rank based on site index (RCSI), individual tree biomass (RCTB), and plot biomass (RCSB). Final model selection was based on the Mallows' Cp statistic (Mallows, 1973; Mallows, 1995). Outliers were evaluated using the studentized residuals, diagonal elements, and "DFITTS" deletion influence (Montgomory et al., 2001).
|
Relative landscape position was quantified by calculating relative elevation and flow accumulation layers. Relative elevations for the 0.008-ha subplots were determined using a combination of field-derived elevations within the plots and local minimum elevations from a United States Geological Survey 30-m grid digital elevation model (DEM). Sites were surveyed using a differential level. Relative elevations were calculated by subtracting the minimum elevation within each neighborhood from the elevation of the subplot using the neighborhood statistics function in ARCGIS for surrounding square neighborhoods ranging in size between 4 ha (7 by 7 cells) and 40 ha (21 by 21 cells). The hydrology module in ARCGIS was used to calculate a flow accumulation layer, which represents the total number of cells in an elevation model that hypothetically flow to a given cell. Mean flow accumulations were also calculated for square neighborhoods ranging between 4 and 40 ha using the neighborhood statistics function in ARCGIS. Data were assigned to the subplots using the spatial join function in ARCGIS.
|
RESULTS AND DISCUSSION |
|---|
|
TOPNOTESABSTRACTReview of Harvesting Disturbance...Evaluating Changes in Forest...Evaluating Changes in Soil-Site...MATERIALS AND METHODSRESULTS AND DISCUSSIONCONCLUSIONSREFERENCES |
|---|
Based on tree growth after 5 yr (tree height, site index, tree biomass, and stand biomass) there were significant differences between dry and wet harvested sites (Table 2). The WB and WMB sites were generally more productive than the DB sites, and the bedded sites were more productive than the flat-planted sites. The exception was that there was no significant difference between the WB and DB height (5.4 and 5.1 m respectively) or site index (25.0 and 24.0 m respectively) values. There were significant differences in stem density, but none attributable to specific treatment conditions. The WMB (45 Mg ha–1) and WB (41 Mg ha–1) treatments had significantly higher stand biomass in the new rotation than the DB (34 Mg ha–1) treatment at age 5 (Table 2). These results imply that wet-weather harvesting had no adverse effect on tree performance on flat-planted sites, and that wet-weather harvesting may be desirable over dry-weather harvesting when site preparation is used.
|
|
Indeed, there were significant differences in stand biomass, height, and initial rank in the prior stand (Table 3) indicating there was site heterogeneity that may have confounded the analysis of current tree growth. The first-rotation trees on WMB sites were taller, had more biomass, and had a lower stand biomass rank than the dry harvested sites; although not significant, the WB sites trended the same. On the flat-planted sites, stand biomass of the previous stand on the WF (181) sites had a better rank than the DF (223) sites. Despite these differences, prior biomass and site index were not significant covariates in the analysis of current tree growth, although stem density was significant as a covariate to stand biomass (P = 0.0769) (Table 2). In any case, it appears that the 5-yr tree performance (Table 2) showing that wet-weather harvested and bedded sites performed better than dry harvested and bedded sites could be due to the fact that the wet harvested sites were better sites to begin with, or that an extended drought caused the more poorly drained (due to rutting disturbance) wet harvested sites to perform better. Thus, the conclusion that wet harvesting enhances growth when bedding is used, is probably inaccurate, although in either analysis you would certainly conclude that wet-weather harvesting is not harmful at the operational scale.
|
In all three rank analyses, the WMB treatment had an extremely low variance compared to the other treatments, which violated the constant variance assumption of ANCOVA (Table 4). Repeating the entire procedure with WMB removed from the analysis sufficiently restored compliance with the constant variance assumption, but it did not change the basic relationship or conclusions among the other treatments. Although the mole plow treatment (WMB) did not improve tree growth, site quality, or productivity rank (Table 2, Fig. 1), it greatly reduced their variance (Table 4), which may or may not influence overall productivity as the stand ages.
|
Looking back to the flat-planting result in the RCSI analysis, this study shows that wet-weather harvesting may damage site index if sites are not bedded; however, this was not supported by the RCTB or RCSB results (Fig. 1). One explanation of the RCSI result is that competition is sometimes initially suppressed on wet harvested sites (Aust et al., 1997; Lister et al., 2004; Murphy and Firth, 2004). Pines growing on the DF sites among competing herbaceous species will tend to have more height growth, while pines growing on WF sites have sufficient diameter growth to close the gap in relative biomass. Therefore, it appears that there was little or no wet-weather disturbance effect on this young stand. Whether or not there is a negative height and site index effect due to wet-weather harvesting, bedding appears to mitigate any potential loss in productivity associated with wet-weather harvesting relative to the operational reference (DB) according to all of the rank diagnostics.
A number of studies over the past half century on pine forest productivity in the southern United States have shown that survival and height growth on wet harvested sites can be diminished by 88 to 91% and 39 to 59%, respectively, due to wet harvesting; however, few were conducted longer than 1 or 2 yr (Miwa et al., 2004). Decreased growth in these other studies was attributed to increases in bulk density and alterations of site hydrology that interfere with root growth and plant water relations (Miller et al., 2004).
Some long-term growth responses to harvesting disturbance are available and they show decreasing influence of disturbance with age. Scott and Tiarks (2005) found few significant differences between wet and dry harvested slash pine forests on poorly drained flatwoods in Louisiana after 18 yr, and that site preparation and fertilization were far more influential on pine production. Another long-term study on 14- to 16-yr-old radiata pine (Pinus radiata D. Don) plantations in New Zealand found no significant differences in production between heavily disturbed and undisturbed sites on Andisols and moderate topography (50-m relief and 35% slopes) (Murphy and Firth, 2004). Our study corroborates the findings of these older studies. It appears that many forest sites are both resistant and resilient to harvest disturbance.
Polypedon-Scale Effects
The cause and effect relationships among harvesting and site preparation treatments and tree response were studied using regression analysis. We assumed that change in site-soil quality represented by RCSI, RCTB, and RCSB would be a function of inherent site factors, silvicultural treatments, and harvesting disturbance (Table 1). We tested the hypothesis that harvesting disturbance effects would be insignificant compared to site and silvicultural factors. Therefore, if the process of harvesting and stand replacement had no effect on site index or tree growth, theoretically, there would be no change in rank from the first and second rotations; that is, the highest ranked sites in the first rotation would still be highest in the second, and lowest ranked sites would be lowest.
Observations and measurements were made at the polypedon-scale (0.008 ha) (Table 1), which represents a fairly small unit of forest where discrete levels of disturbance and other site factors occur within the operational unit. Site index, height, and biomass were highly variable at this scale for both the pre- and postharvest stands (Table 5). The total ranges and interquartile ranges for the post harvest estimates of site index and stand biomass were greater than the preharvest ranges.
|
|
Prior site index or biomass had the greatest effect on current site index or biomass followed by bedding and several inherent site and soil factors. Bedding enhances microsite drainage and restores altered soil physical properties (Schultz and Wilhite, 1974; Terry and Hughes, 1975; Gent et al., 1983; McKee et al., 1985; Pritchett and Fisher, 1987; Morris and Lowery, 1988; Tippett, 1992; Aust et al., 1995), which enhances survival and growth of young loblolly pine (Baker and Langdon, 1990; Schultz, 1997). The effects of harvesting were relatively small with distance to landing playing a relatively minor role.
The models predict that sites with higher initial site index are more likely to lose rank relative to other plots. This result is contrary to reports by authors who argue that higher quality sites are in fact more resilient to disturbance (Burger and Scott, 2001). However, that generalization usually has to do with fertility rather than physical properties. On these sites soil-site quality, as reflected by site index, is primarily a function of drainage rather than fertility. Intuitively, some changes in soil physical properties, such as increased bulk density caused by compaction, have a maximum limit. Other soil properties may have thresholds, where tree growth becomes restricted once a parameter drops below a certain limit. Therefore, there may be a limit to how bad poor sites may become due to conventional logging practices. Sites that have high physical quality have more to lose as a result of trafficking (Scheerer 1994, Aust et al., 1995).
Another reason the models might predict that sites with higher initial site index are more likely to lose rank relative to other plots is because the chance that the rank of sites with high initial site indexes or biomass will remain the same or decrease is greater than the chance that rank would increase. However, an analysis of variance of the mean RCSI, RCTB, and RCSB values for the four quartiles of initial site quality indicated there were no significant differences in the magnitude of the observed changes. Additionally we performed regressions that included only data points from the second and third quartiles so that the rank for a single observation had a chance to both increase as well as decrease. Although the model coefficients changed slightly from the original models, the same factors were selected. Therefore, we conclude that the relationship between rank change and prior rank is due to site quality, silvicultural, and harvesting effects. It stands to reason that high-quality sites, which derive their quality from good physical properties, would be most at risk since low-quality sites essentially have little to lose. High soil physical quality is a product of pedogenic processes that require many years of natural development (Stone, 1975). Good structure, porosity, hydraulic conductivity, and ideal texture are not properties that would necessarily be corrected by a bedding plow that only affects a limited portion of the soil profile.
A further question would be whether bedding differentially affects sites with high initial site quality compared to sites of low initial site quality. We tested this by performing regressions using a subset of data restricted to the bedded sites and from the second and third quartiles. Prior site index was still significant in the RCSI model and still had a negative coefficient, as was prior biomass in the RCTB and RCSB models. Therefore, bedding on higher quality sites either does not enhance soil quality or does not fully remediate sites with high initial site quality. Higher quality sites, where drainage may already be adequate, probably have less to gain by bedding than lower quality sites, where drainage may be poor and soil properties may be limiting. Again, we feel it is unlikely that a bedding plow is capable of restoring properties that are the result of pedogenic processes; therefore, bedding has little effect on high quality sites in terms of physical soil quality.
Stand density is an important determinant of biomass production and like site index can be another reflection of site quality (Zedaker et al., 1987). Stand density was only included in the RCSB model but has an overwhelming influence on the response compared to other site factors. Survival was excellent on all sites; however, there were 3 yr of drought up to age 5 (Eisenbies et al., 2004) which may have improved survival on what are typically wet sites and affected which model components were significant.
Relative elevations varied by 2 m within our blocks, and drainage class is very influential on the productivity of wet pine flats (Shoulders, 1976; Fisher and Garbett, 1980; Haywood et al., 1990; Hauser et al., 1993). Although relative elevation was calculated for several neighborhood sizes (4–40 ha), the most significant was the 30-ha neighborhood. The flow accumulation regressor, although not included in the final regression models, also indicated that 30 ha was an important size for expressing landscape-level drainage. The connotation of this size appears to relate to stream and drainage density, while not being so large as to only mimic regular elevation. Average water tables rose 14 to 21 cm within the first year after harvesting, although they returned close to preharvest conditions within 2 yr (Xu et al., 2002). Saturated hydraulic conductivities decreased by as much as 72%, and total porosities by 8% (Xu et al., 2000).
Mollisols (Santee soil series) lost productivity relative to the three other soil types found on these study sites, which we believe is related to drainage. However, Mollisols are a relatively unique feature of these sites and represent approximately 13% of the study area. The Santee soil series (fine, mixed, active, thermic Typic Argiaquolls) is generally found in broad depressional areas and differs from the other three soil series represented by this study mainly by its mollic epipedon and slightly coarser textures. The pedogenesis of their formation in wet pine flats is due to poor drainage leading to the retention of carbon (Stuck, 1982).
Distance to landing of a harvested site was the most significant disturbance factor at the polypedon scale. The fact that distance to landing was selected rather than other indicators of rutting and organic matter disturbance implies harvest deck location, harvest system pattern, and arrangement of traffic are as important as the presence or absence of physical disturbance. The number of passes, disturbance depth, rut orientations (Greene and Stuart, 1985, Burger et al., 1989; Carter et al., 1989), and harvest residues (Hall, 1999) influence the cumulative effect of traffic near the landing. Furthermore, disturbance on the less trafficked areas may have already recovered during the first 5 yr to the point that growth of young pines was not affected, as has been shown on similar sites (Scheerer, 1994, Lacey and Ryan, 2000).
Distance to landing had a relatively high P value in our regression (P = 0.045), which led us to consider how localized the effect was to the landing. When we model with the observations within 40 m of the landing removed, distance to landing becomes nonsignificant (P = 0.765). Within 40 m of the landing soils receive repeated trafficking in random directions, perhaps altering surface drainage so there is no clear path away from those locations. Maintaining surface drainage is a central issue for these sites, as subsurface drainage can be very slow and most water leaves sites similar to these primarily via evapotranspiration during the growing season (Konyha et al., 1988, Aust et al., 1993).
Considerations Using Change in Rank to Evaluate Long-term Productivity
Change in rank is a new approach that enhances our ability to evaluate the effects of intensive forestry practices on productivity. Some of its advantages have already been noted, but some further discussion of its use is still warranted. The key advantage of using rank as a diagnostic is that it provides a means to evaluate changes in productivity between rotations rather than relying on current production to make this assessment. When comparing the productivity of successive rotations, change in rank reduces the influence of confounding biotic and abiotic factors discussed earlier (e.g., genetics, silvicultural technology, and climate). In addition, basing the assessment only on current production in this study might lead to the false conclusion that there is some benefit associated with the combination of wet harvesting and bedding. A second advantage is that large-magnitude changes that might occur at the tails of normal distributions have the same change in rank as small-magnitude changes near the median. As a result, the influence of outliers from the untransformed dataset is suppressed using a rank distribution.
Although the change in rank method has certain advantages for evaluating treatment effects on changes in productivity, there are some limitations. First and foremost, a relative ranking is used in lieu of a more intuitive and absolute measure of production (e.g., site index, tree biomass, stand biomass), which are subject to the confounding factors that limit their interpretation. Second, statements about potential decreases in long-term productivity such as those attributed to second and third rotation decline (Powers et al., 1990; Kimmins, 1996; Worrell and Hampson, 1997) are not possible using rank as used here because such declines represent ubiquitous effects. This method can only provide a partial, albeit valuable, solution to the problem of evaluating changes in productivity caused by forest management: specifically, treatments can be compared within the context of intensive forestry, but the effect of intensive forestry itself cannot be evaluated. Finally, change in rank is probably most useful when put in the context of more intuitive and absolute values on which they are based (height, site index, biomass, leaf area, etc.). While rank could remain significant far into the rotation, the practical differences in actual site index or biomass could become less important. A final limitation to this approach is the need for a large number of subsamples within each treatment area. As fewer subsamples are used, rank changes become increasingly discrete and would no longer adhere to normality assumptions.
Certain considerations should be made before using this approach. Since the results are not in concrete biological terms, and rely on relative performance of the treatments, care should be taken not to introduce confounding factors that prevent meaningful interpretation. (1) The bounds of the neighborhoods are critical. Each neighborhood must contain all treatments, and sites should be similar enough that a separation in rank is a possibility. (2) Site history should also be similar. Neighborhoods should include plots that are the same age and have the same prior treatments. (3) It is very important to include a reference treatment that is similar to the preharvest treatments, or one that can be assumed to not alter site productivity, to provide a meaningful contrast. In the case of this study, the DB sites were the reference; although the DF sites could also have been used for the purpose as they probably better reflect the "natural" condition. (4) A limitation of the regression approach is that results apply only to the area studied, and only to the particular soil and topographic conditions sampled within the study area (Carmean, 1975). It would be possible for the landowner to use resource maps to evaluate risk, but the extent of the valid inference space is not known. (5) The inclusion of too many treatments relative to the number of subsamples may reduce the resolution with which treatment differences can be detected at the operational scale. (6) If a new species or genotype is used in the following rotation, each should probably be suited for the same site conditions. (7) Finally, it is still necessary to wait a full rotation before definitive statements are made about negative changes in soil-site quality associated with specific treatments because the sites may recover at different rates.
Importance of Scale
From the operational scale we can conclude that as long as bedding is utilized, there should be no negative effects of wet-weather harvesting compared to our operational norm. However, the polypedon-scale RCSI model indicates that proximity to the landing is also an issue, but this effect is not unique to either the dry- or wet-weather harvested sites. At the pedon-scale (0.001 ha), Kelting (1999) showed that bedding did not completely remediate compression and rutting after 2 yr; however, these disturbances only represent about one third of the total area. Obviously these important stand and disturbance scale effects are not manifested at the operational scale.
|
CONCLUSIONS |
|---|
|
TOPNOTESABSTRACTReview of Harvesting Disturbance...Evaluating Changes in Forest...Evaluating Changes in Soil-Site...MATERIALS AND METHODSRESULTS AND DISCUSSIONCONCLUSIONSREFERENCES |
|---|
Past recommendations suggest that logging during wet seasons should be limited to better-drained sites. In the case of more fertile wet pine flats such as these, it appears that in some cases wet-weather logging could be conducted on these sites as long as bedding will be applied as a site preparation. Due to the substand (0.008 ha) variability, it may be very difficult to target specific operational sites that may be at risk. Ultimately, the ability to preserve or restore drainage on these sites should be an important factor in site selection. In addition, site preparation should be a consideration during BMP evaluations that include the effect of rutting and soil disturbance on long-term productivity.
Change in soil-site productivity on fertile, wet pine flats is a function of silvicultural treatments, inherent site factors, and harvesting disturbance after 5 yr. Regression analysis indicates that sites in low relative elevations and residing within about 40 m of the harvest landing are most at risk. Sites with higher initial site quality may either be more at risk or may receive less benefit from bedding because bedding primarily affects surface drainage, and soil physical properties (porosity, water retention) are less likely to be affected by bedding. All components in the model indicated that site drainage is a principal factor influencing the change in site productivity, although more site-specific factors will have to be examined in future studies. Proximity to the harvest landing was associated with decreased soil-site productivity. It may be necessary to take extra measures to restore productivity in these areas. Improving bed quality, surface drainage, and subsurface drainage by double bedding or mole-plowing may be an option for areas close to logging decks. However, mole-plowing did not significantly increase tree performance at the operational scale.
Using change in rank as a diagnostic to evaluate site productivity is a new approach that has some limitations but also offers definite advantages over other approaches. While the method may not be able to establish widespread productivity change, it does reflect how the sites have changed from the prior rotation. It is therefore able to determine which treatments are relatively benign compared to an operational norm. This determination can be made independently of the confounding factors (genetics, climate, technology) normally associated with comparing productivity between rotations. Change in rank is probably most valuable when used in conjunction with other methods. While benign treatments can be identified, it will still take a full rotation to know the true long-term effects of potentially negative treatments and disturbance on carrying capacity. Given time, both flat-planted treatments may recover as well. Rank will also be an effective means to compare productivity change at the end of the rotation as well; however, large rank differences in the future might be small from the standpoint of absolute amounts of wood volume, biomass, or height growth.
|
ACKNOWLEDGMENTS |
|---|
|
NOTES |
|---|
|
TOPNOTESABSTRACTReview of Harvesting Disturbance...Evaluating Changes in Forest...Evaluating Changes in Soil-Site...MATERIALS AND METHODSRESULTS AND DISCUSSIONCONCLUSIONSREFERENCES |
|---|
Received for publication February 18, 2004.
|
REFERENCES |
|---|
|
TOPNOTESABSTRACTReview of Harvesting Disturbance...Evaluating Changes in Forest...Evaluating Changes in Soil-Site...MATERIALS AND METHODSRESULTS AND DISCUSSIONCONCLUSIONSREFERENCES |
|---|
