Avian Home >> Research and Monitoring >> Publications

USE OF A LANDBIRD MONITORING DATABASE TO EXPLORE EFFECTS OF PARTIAL-CUT TIMBER HARVESTING

Jock S. Young and Richard L. Hutto

This paper has been published and should be cited as:
Young, J. S., and Hutto, R. L. 2002. Use of a landbird monitoring database to explore effects of partial-cut timber harvesting . Forest Science 48(2): 373-378

DRAFT EXCERPTS:

Long-term monitoring projects provide a wealth of data that can be used for many purposes in the short term, before the analysis of long-term trends is possible.When vegetation data are collected at each point, correlative relationships can be studied for any habitat variable measured.Although such data sets are not the products of studies designed to test the effects of specific management practices, they include a much greater number of sites than are typically included in before-after/treatment-control studies, so they have the potential to provide meaningful information about management effects.For example, bird and vegetation data collected from about 6,000 long-term monitoring points established in association with the U. S. Forest Service (USFS) Northern Region Landbird Monitoring Program (NRLMP) have already provided the information needed to establish bird-habitat relationships among natural and human-modified cover types (Hutto and Young 1999).

In the NRLMP, long-term monitoring is conducted across the region on a biennial basis, while in the intervening years we conduct more focused, single-year projects that address current information needs, especially as they relate to known management issues and potential population declines that may one day reveal themselves through the long-term monitoring efforts.One such study in 1997 examined the effects of partial-cut timber harvesting on bird abundance in mid-elevation, conifer forests on 3 National Forests in northwestern Montana and northern Idaho (Hutto and Young, unpublished data).In this paper we examine the same question using data from the long-term monitoring database to see whether opportunistic use of a long-term database might expose the same general patterns that appeared from more carefully designed studies of management treatment effects.

METHODS

Vegetation treatment categories.For the comparison in this paper, we selected points that were surrounded by unlogged “Mature” or early post-harvest “Partial Cut” cover types for at least a 100-m radius (no edge nearby) and were within mixed-conifer, Douglas-fir, or grand fir (Abies grandis) dominated stands.We used only those points from west of the Continental Divide.We included all types of partial-cut harvesting (e.g., thinning, shelterwood, seed tree), as long as the stands still included at least 1% cover of mature conifer trees.Some transects were not run every year, and many points did not meet the above criteria in all years that they were visited because observers in different years classified the same points differently with respect to cover type, canopy cover, or presence of edge.Bird survey methods.The bird counts followed recommendations discussed by Ralph et al. (1995) and methods described by Hutto et al. (1986).A 10-min point count was conducted at each of the 10 sampling points along a transect.Points were visited once each breeding season between mid-May and mid-July.All birds seen or heard within the count period were recorded.Field observers generally began counts about 15 min after sunrise (after the pre-dawn chorus), and completed counts between 0630 and 1100 hrs (MST).Counts were not conducted on days with continual rain or high winds.

Analysis methods. We used logistic regression (Hosmer and Lemeshow 1989) to test for differences between the proportion of points with detections in uncut and partial-cut stands for all bird species detected on at least 60 points.We then used logistic regression to test the effect of canopy cover to determine if that variable may be a more useful predictor of bird occurrence than the categorical treatment variable. If there were no residual treatment effects after canopy cover was present in the model, this would suggest that the effect of logging was due entirely to the reduction in canopy cover. For these analyses we used a pooled dataset of the 3 years from 1994 to 1996. Repeat visits to individual points were treated as independent counts. Although not strictly correct for confirming significance, this method gave an indication of the relative effects of treatment and canopy cover.

RESULTS

Both treatment categories included a highly diverse mix of forest structures and species compositions. There was a large overlap in percent canopy cover between the treatments, with an overall mean canopy cover for 1994-1996 of 34.3% (SD = 19.0) for the unlogged sites and 21.1% (SD = 15.7) for the partial-cut sites (Figure 1). Tall shrub cover also varied tremendously, with 12.9% (SD = 14.6) for the unlogged sites and 14.7% (SD = 15.8) for the partial-cut sites.

Figure 1. Distribution of estimated canopy cover within each treatment type.

Eighty-five bird species were detected within 100 m of an observer using region-wide monitoring points that fell within either uncut or partial-cut forest stands, with 33 species detected on at least 60 points over 3 years (Table 1). The abundances of 20 of these 33 species were significantly different between uncut and partial-cut forest stands (Figure 2), with 5 species (brown creeper, winter wren, golden-crowned kinglet, varied thrush, and Townsend’s warbler) being more abundant on points in uncut stands, and 15 species more abundant on points in cut stands.

The Figure below shows selection indices for the 20 bird species with significant differences between treatments (the black bars; a selection index of 1.0 would indicate the species was only in uncut stands, and in index of -1.0 that it was only in cut stands). The grey bars show the selection index for the same species in the 1997 study, which was based on 35 uncut stands and 37 partial-cut stands.

When the effect of canopy cover was included in the regression models along with the treatment effect (Table 1), it was apparent that the treatment effect could be explained almost entirely by canopy cover changes for many species (e.g., mountain chickadee, brown creeper, golden‑crowned kinglet, ruby‑crowned kinglet, varied thrush, and Townsend’s warbler). Furthermore, some species showed strong correlations with canopy cover but almost no treatment effect (e.g., chestnut-backed, hermit thrush, black-headed grosbeak, and pine siskin), suggesting that canopy cover may be the more useful variable. On the other hand, many species still had strong treatment effects after canopy cover was accounted for (e.g., Cassin's vireo, winter wren, yellow‑rumped warbler, western tanager, chipping sparrow, dark-eyed junco, and brown-headed cowbird), suggesting that the treatment effect was due to some other variable associated with the logging treatment rather than the decrease in canopy cover itself.
______________________________________________________________________________

Table 1. Results of logistic regression analyses (chi-square statistic for likelihood-ratio test and significance level) of the effects on bird occurrences of partial cutting (treatment effect), percent canopy cover, and each of these variables after controlling for the other (data pooled over 3 years, 1994-1996)Results are shown for all species with at least 60 detections over 3 years.

Table 1, cont.

^a   n = number of points with detections out of 2,243 point counts, 1994-1996

^b “effect” includes the direction of each effect (treatment or canopy) and the significance level;   

* = P < 0.05;  **   = P < 0.01;  *** = P < 0.00                       

DISCUSSION

Of the 5 species that were more abundant in uncut stands, all but the varied thrush also had significant differences in a separate control-treatment study (Hutto and Young, unpublished data). These 5 species were also among the most consistent species showing such relationships in numerous studies reviewed by Hejl et al. (1995), and are species of potential concern because mature forests are more likely to be converted to partial cuts in the future, rather than the reverse. Many of the bird species that were more abundant in the partial-cut stands, such as the hairy woodpecker, mountain chickadee, yellow‑rumped warbler, and western tanager, are

open-forest species that might be expected to be more common in thinned conifer forests than anywhere else, and may be of little concern, especially because many of these species are common in all types of harvested forest (Hutto and Young 1999). However, we still do not know if these species are doing as well in these newly created habitats as their relative abundance suggests, or if cutting practices might be creating unusual forest structures that might, in turn, act as “ecological traps” by providing the birds with cues for settling without providing habitat that is otherwise suitable. One concern is that the brown-headed cowbird is much more likely to occur in partially cut than in uncut forests (Figure 2; Young and Hutto 1999) and the presence of this nest parasite may contribute to the creation of unsuitable but otherwise attractive habitat.

The results from this study indicate significant changes in the bird community as a result of partial-cut timber harvesting. Many of these changes can be ascribed to the decrease in canopy cover (Table 1). Because of the almost infinite variability in timber harvest prescriptions currently conducted on the landscape, continuous variables such as canopy cover may actually be more useful tools than categorical treatment variables for discerning management effects. However, there were several species that showed treatment effects without being greatly influenced by canopy cover, so treatment per se may still be an important variable. These species appear to be affected by something associated with the treatment process itself, whether resulting from differences in understory or tree architecture, or any number of other possible explanations. Interestingly, all but one of these species (the winter wren) showed positive, rather than negative, effects of treatment.

When comparing results of these analyses to an independent control-treatment study (Hutto and Young, unpublished data), twelve species had significant treatment effects in the same direction (8 positive and 4 negative) and none in opposite directions. The pileated woodpecker and the Swainson’s thrush were the only species that showed effects in the 1997 study but none based on the monitoring database. The detection rate of woodpeckers was higher in the 1997 study due to the addition of supplemental data from outside the 10-min count period. Because woodpeckers were not well sampled using standard point-count methods during the passerine breeding season, they may be difficult to study in many monitoring programs without a modified protocol.

The long-term monitoring database had the advantage of greater sample sizes, but points were sampled without a priori control over site conditions. Consequently, the database included a complex mixture of logging treatment types (Figure 1). The monitoring database also resulted from single visits to each point in a given year. Nonetheless, the long-term monitoring database revealed significant differences between treatments for 10 bird species that did not show such differences in the 1997 study, including 2 species that were detected too infrequently in the latter study. Even if the sample sizes for some of these species were too small for firm conclusions, many of the results undoubtedly indicated true treatment effects, especially for species with large differences in abundance such as the brown-headed cowbird and dusky flycatcher.

The failure to detect differences in the abundances of some species in the more focused, management-effect study, relative to the region-wide database, may have been due to their relatively uncommon occurrence in the forest types included in the 1997 study. The long-term monitoring database may have included more points with these species because it included a greater diversity of forest structures. Species such as the warbling vireo, MacGillivray's warbler, and black-headed grosbeak are bird species of shrubby riparian areas and clearcuts, and do not occur as commonly in partial-cut forests (Hutto and Young 1999), especially recent harvest units without a well-developed shrub layer, as was the case in the 1997 control-treatment study. The fact that there were more shrubs on uncut sites in the 1997 study may have been one reason why neither the warbling vireo nor the MacGillivray's warbler were more common in the partial-cut sites.

Provided that sufficiently accurate vegetation data are coupled with bird abundance data, and provided that enough points are visited to compensate for an uncontrolled design, these results suggest that data from long-term monitoring points can be useful not only for information about long-term population trends, but also in the short-term as a source of information about habitat relationships and the effects of land use on bird populations (e.g., Hutto and Young 1999, Young and Hutto in press). The land use effects revealed here also suggest that regional bird populations may be strongly affected when increasing acreages of partial-cut forestry are summed across a new landscape (Thompson et al. 1995).

LITERATURE CITED

Hejl, S.J., R.L. Hutto, C.R. Preston, and D.M. Finch. 1995. Effects of silvicultural treatments in the Rocky Mountains. P. 220-244 in T. Martin, and D.M. Finch, (eds.). Ecology and Management of Neotropical Migratory Birds. Oxford University Press, New York, NY

Hosmer, D.W., Jr., and S. Lemeshow. 1989. Applied logistic regression. John Wiley and Sons, New York, NY. 306 p.

Hutto, R.L., S.M. Pletschet, and P. Hendricks. 1986. A fixed-radius point count method for nonbreeding and breeding season use. Auk 103:593-602.

Hutto, R.L., and J.S. Young. 1999. Habitat relationships of landbirds in the Northern Region, USDA Forest Service. USDA For. Serv. Gen. Tech. Rep. RMRS-GTR-32. 72 p.

Ralph, C.J., S. Droege, and J.R. Sauer. 1995. Managing and monitoring birds using point counts: standards and applications. P. 161-168 in Ralph, C.J., et al. (eds.). Monitoring bird populations by point counts. USDA For. Serv. Gen. Tech. Rep.PSW-GTR-149.

Thompson, F.R. III, J.R. Probst, and M.G. Raphael. 1995. Impacts of silviculture: overview and management recommendations. P. 201-219 in Martin, T., and D.M. Finch, (eds.). Ecology and Management of Neotropical Migratory Birds. Oxford University Press, New York, NY.

Young, J. S., and Hutto, R. L. 1999. Habitat and landscape factors affecting cowbird distribution in the northern Rockies. Stud. Avian Biol. 18:41-51.

Young, J. S., and Hutto, R. L. In press. Use of regional-scale exploratory studies to determine bird-habitat relationships. In Predicting Species Occurrences: Issues of Scale and Accuracy. Scott, J.M. et al. (eds.). Island Press, Washington, DC