Abundance of Western Red-backed Salamanders (Plethodon vehiculum) in the Washington Coast Range after headwater stream-buffer manipulation.
Key words: alternative forestry, amphibians, Douglas-fir forests, forest buffer, Pacific Northwest, riparian buffer, streamside habitat
Plethodon or woodland salamanders have been advocated for monitoring biodiversity and ecosystem integrity following forest disturbance (Welsh and Droege 2001; Welsh and Hodgson 2013). Salamander responses to small stream buffers that retain trees in the Pacific Northwest, however, show varied results. Western Red-backed Salamanders (Plethodon vehiculum) in small stream buffers were found to be sensitive to adjacent clear-cuts <5 y old in the Oregon Coast Range (Vesely and McComb 2002); but in a study of experimentally thinned riparian buffers in Washington, Hawkes and Gregory (2012) found increased abundance of P. vehiculum within 2-y post-harvest, and again, a decade later. In western Oregon, thinned buffers of varying widths (but [greater than or equal to] 6 m) by headwater streams showed no evidence of treatment effects on P. vehiculum (Olson and others 2014).
We assessed the initial response of bank-dwelling P. vehiculum to alternative forest buffers along headwater streams in western Washington. Our objective was to determine initial treatment effects of buffer manipulation on P. vehiculum abundance and on down-wood as salamander habitat. Our research goal was to increase understanding of the effectiveness of alternative buffering strategies in the conservation and protection of headwater streams during timber harvest.
We sampled 6 study sites located across the Willapa and Black Hills, Washington, that ranged from 60 to 400 m in elevation (Fig. 1, Table 1). At all sites, upland forests were naturally-generated 2nd-growth, consisting of Douglas-fir (Pseudotsuga menziesii), Western Hemlock (Tsuga heterophylla), and Western Redcedar (Thuja plicata); tree ages ranged between 61 and 108 y. We sampled 23 streams with catchments ranging in size between 1.9 and 8.1 ha ([bar.[chi]] = 4.7 m, [s.sub.[bar.[chi]]] = 0.4); length of stream reaches ranged between 111 and 480 m ([bar.[chi]] = 252 m, [s.sub.[bar.[chi]] = 19); and channel gradient ranged between 11 and 46% ([bar.[chi]] = 29%, [s.sub.[bar.[chi]]] = 2) (Bisson and others 2013). Stream bank-full widths were generally <3 m and flow was mostly perennial, but 1 stream had a well-defined seasonally-dry bed.
We examined 3 riparian buffer designs that were incorporated into the timber harvest prescriptions (Fig. 2, Table 1): (1) continuous strip buffers (n = 7 streams), initially varying in widths between about 12 and 26 m on each side of the stream ([bar.[chi]] = 21 m, [s.sub.[bar.[chi]]] = 2); (2) discontinuous patch buffers, which were <0.6-ha leave tree clumps (patch leave portion) within a clear-cut landscape (patch cut portion) (n = 3 streams), configured to protect headwalls, stream sources, and unstable soils, with widths or diameters similar to strip buffers; and (3) nobuffer clear-cuts (n = 7 streams), that had a few scattered or clumped residual trees (hereafter, we use the term "buffers" to refer to the logged streams collectively, including the strip, patch, and no-buffer clear-cut treatments). Streams in uncut forest served as controls (n = 6 streams) (Table 1).
The selection of streams to include in the study and the assignment of harvest treatments were largely guided by forest management operational concerns. They were not random, but were semi-systematic. All control streams were located either adjacent to or in close proximity to the associated treated streams. Site conditions, logistics, and logging operations at individual streams resulted in a wide mix of dimensions and design structure in the strip and patch buffers. Treatments were initiated in September 2003 and ended in early 2005. The study spanned 4 y (2003 to 2006) and comprised: (1) a pretreatment year (2003); (2) the treatment year (harvest, 2004); (3) post1 (1 y after harvest, 2005); and (4) post2 (2 y after harvest, 2006).
In the Pacific Northwest, pit trapping seems to be a more efficient approach to capturing surface active amphibians than by ground searches or cover boards (Bury and Corn 1988; McDade and Maguire 2005). We trapped P. vehiculum in number 10 single-deep cans buried to rim level. We added nearby litter and moss to "season" traps. We used inner tapered plastic sleeves to prevent salamanders from escaping, and small sticks to allow small mammals to exit traps (Perkins and Hunter 2002). We also used the sticks to prop open square pressboard can lids for deflecting rain and debris. Openings were 40 to 50 mm above the rims and simulated crevice refugia that might entice salamanders to the traps.
We established 18 trapping stations per side of each stream (n = 36 per stream) for each treatment type and control. Arrays were centered at about midpoint of the entire stream reach between the uppermost point and the downstream confluence. The configuration covered about 85 m of a study reach (Fig. 2). We used single-deep cans with escape sticks because a nearby experiment showed that captures of salamanders were similar between single- and double-deep cans, and with or without sticks (Aubry and Stringer 2000). On each stream side we placed stations about 5 m apart, and alternated their positions between the stream edge and about 2 m away from the bank in a zigzag pattern (Fig. 2).
After logging, trap arrays that were in the discontinuous patch buffers could be in the leave tree portion or in the clear-cut portion of the treatment, or partially in both portions of the patch buffer treatments each year (Fig. 2). In strip buffer treatments, all trap arrays were inside the buffers. Across streams in treatments, logging slash and wind-throw buried some trapping stations or slightly altered portions of individual stream courses. For these, we relocated stations close to their original positions. We checked traps 8 times/stream/year for a 15- or 16-d period. Each check was conducted at 2-d intervals; however, in some years the interval between the 7th and final check at some streams was 1 d due to the logistics of returning salamanders to their capture locations during the final check. In total, we made 687 stream visits with 2-d trap-check intervals, and 17 stream visits with a 1-d interval. We removed all caught amphibians to a refrigerator and returned them back to original capture sites unmarked on the last day of trapping each year. Because of access restrictions, 2 streams partially sampled in pre-treatment (1 strip and 1 nobuffer clear-cut), and 2 streams not trapped in the harvest year (2 no-buffer clear-cuts) were not included in the analysis for those years. All trapping was conducted during the same periods of time each year, ranging between 29 September and 17 October.
To assess habitat effects that may influence P. vehiculum abundances, we used pre- and post-treatment measurements of tree canopy cover (%) and down-wood log density (n/ha) (hereafter, down-wood). Wind-thrown trees with green foliage were not tallied as down-wood. These data were collected by random sampling within treatments both inside and outside of trapping arrays, and not directly at established trapping stations (Fig. 2). Canopy point samples and down-wood sampling were conducted in each of two 10- X 10-m plots (= 1 habitat sampling array) on each stream side; downwood within each array was based upon 20 m of transect; and riparian management zone buffers were excluded (see Fig. 2). We measured down-wood using line-intercept sampling. The diameter of each piece of wood at the point of intersection was [greater than or equal to] 10 cm, and the length of each piece was at least 0.3 m (Waddell 2002). We recorded wood species and uniquely marked all dead trees [greater than or equal to] 10 cm diameter at breast height. We measured canopy cover using hemispherical photographs from 4 photo points (Fig. 2) that were 1.2 m above ground with a fisheye converter on a digital camera with a 20% mask of the sky (Frazer and others 1999).
We assessed weather conditions for each trapping session for each year. We used average mean daily temperature (average of high and low temperatures over a 24-h period) and 24-h rainfall from data recorded at 2 nearby weather stations, Huckleberry Ridge and the Olympia Airport (Fig. 1).
We made between-year treatment comparisons using paired sample f-tests for P. vehiculum abundance and habitat variables, and assumption-free Wilcoxon signed-rank related-sample tests for both weather variables. For within-year, between treatment comparisons, we used independent sample f-tests for P. vehiculum abundance and habitat variables, and Wilcoxon rank-sum tests for weather variables. Paired-sample f-tests focus on the differences between compared variables, whereas independent sample f-tests focus on the amounts. We computed the relative effect size r-statistic for all habitat comparisons, and for statistically significant outcomes of P. vehiculum abundance using the benchmarks of Cohen: 0.1 (small effect, explains 1% of the total variance); 0.3 (medium effect, 9% of variance); and 0.5 (large effect, 25% of variance) (Nakagawa and Cuthill 2007; Field 2013). We used [alpha] = 0.1 for statistical significance because of our small sample size and high variability in the data. In order to compare salamander relative abundances across unequal numbers of buffer treatments, we took the total number of salamander captures at each stream as an average within each treatment. These were salamander captures/stream. To statistically analyze these data, we first normalized individual stream capture data before analysis by taking the [log.sub.e](y + 1), where y = total capture count at each stream (McCune and Grace 2002), and then conducted statistical tests on these transformed data. Canopy cover (%) was square root-transformed. We assessed the variability of the relative abundance of salamanders (transformed) with coefficients of variation ([CV.sub.x100]). We used IBM SPSS software 21 (2012 version) for data processing and analysis.
Plethodon vehiculum (n = 104) comprised >60% of all salamanders (n = 164) and amphibians (total n = 172) captured. We also captured the Rough-skinned Newt (Taricha granulosa), Ensatina (Ensatina eschscholtzii), Northwestern Salamander (Ambystoma gracile), Coastal Tailed Frog (Ascaphus truei), and, in the Willapa Hills only, the Columbia Torrent (Rhyacotriton kezeri), Van Dyke's (Plethodon vandykei), Dunn's (P. dunni) and Cope's Giant (Dicamptodon copei) salamanders. The species composition seemed representative for the area.
We observed the following significant differences in P. vehiculum relative abundances in the treatment buffers. (1) No-buffer dear-cuts, between years: (a) the mean difference in P. vehiculum relative abundance was significantly lower in the harvest year (P = 0.072), and in the post2 year (P = 0.051), compared to the pretreatment year (Table 2); (b) the mean difference in abundance was significantly lower in the post2 year, compared to the post1 year (P = 0.047); and, (c) the mean differences in P. vehiculum abundance in the harvest year was lower than in the post1 year (no statistic) (Table 2, but see Fig. 3). (2) No-buffer clear-cuts, within year and between treatments: P. vehiculum mean relative abundance in the post2 year was significantly smaller compared to the patch buffer treatments (P = 0.020), and compared to control streams (P = 0.003) (Table 2, Fig. 3). (3) Strip buffer treatments, between years: the mean difference in P. vehiculum relative abundance was significantly lower in both the pre-treatment year (P = 0.086) and the harvest year (P = 0.060) than in the post1 year (Table 2). All effect sizes were large (r > 0.5) (Table 2).
In the pre-treatment year, P. vehiculum abundances varied 20.0% across all 4 treatment groupings. Across the 3 post-harvest years, we observed the lowest variation in relative abundance in controls ([CV.sub.x100] = 28.9%), followed by strip buffers ([CV.sub.x100] = 33.8%) and patch buffers ([CV.sub.x100] = 34.5%), with more year-to-year fluctuation in the no-buffer dear-cuts ([CV.sub.x100] = 88.7%) (Fig. 3).
We found canopy closure (%) significantly reduced in harvested treatments; whereas, down-wood increased in all categories compared to the pre-treatment year and all effect sizes were large (r > 0.5) (Table 2). Besides logging slash accumulations inherent in harvest operations, tree blow-down contributed to increased loadings of down-wood and untallied live down-wood. Average amounts of down-wood in the uncut areas of strip buffers, patch buffers, and controls were 1.5 to 1.9 times greater after logging compared to the pretreatment year. The averaged amounts of down-wood in cut areas along streams (patch-cut portion and no-buffer clear-cut treatments) were 2.6 to 5.7 times greater than in the pretreatment year (Table 3).
Ambient temperatures in the harvest year ([bar.x] = 12.1[degrees]C, [s.sub.[bar.x]] = 0.4) were warmer than in the post1 ([bar.x] = 10.5[degrees]C, [s.sub.[bar.x]] = 0.5) ([T.sub.16] = 14, P = 0.016) and post2 years ([bar.x] = 10.9[degrees]C, [s.sub.[bar.x]] = 0.4) ([T.sub.16] = 26, P = 0.030). Pre-treatment year temperatures were similar to other years ([bar.x] = 11.9[degrees]C, [s.sub.[bar.x]] = 0.8). Rainfall in the pre-treatment year ([bar.x] = 5.3 mm, [s.sub.[bar.x]] = 1.9) was greater than in the post2 year ([bar.x] = 2.1 mm, [s.sub.[bar.x]] = 1.3) ([T.sub.17] = 23, P = 0.064). Rainfall in the post1 year ([bar.x] = 5.5 mm, [s.sub.[bar.x]] = 1.4) was also greater than in the post2 year ([T.sub.17] = 103.5, P = 0.066). Harvest year rainfall was similar to other years ([bar.x] = 4.2 mm, [s.sub.[bar.x]] = 2.4). Rainfall of [greater than or equal to] 0.01 mm occurred on 43% of the days in the harvest year, and 41% of the days in the post2 year. In the pre-treatment and post1 years, rain occurred on 72 and 84% of the days, respectively.
Within the first 3 y of harvest we observed lower abundances of P. vehiculum only in the clear-cut no-buffers. It thus appears that some level of protection was provided by strip and patch buffers. Dupuis and others (1995) and Dupuis and Bunnell (1999) found few P. vehiculum in clear-cuts, and Patrick and others (2006) cite several studies with low amphibian abundances in clear-cuts. Clear-cutting and associated reductions of down-wood are believed to be part of the reason for local reductions or extirpations of 7 salamander species in the Pacific Northwest (Bunnell and Houde 2010). Salamander species responses may vary with site conditions retained after harvest, such as canopy cover and down-wood that help maintain important microhabitats and microclimates (see Bunnell and Houde 2010). Landform (for example see Wilkins and Peterson 2000), possibly associated with stream order, may also influence abundances. Plethodon vehiculum abundances in no-buffer clear-cut treatments were lower than their abundances in patch treatments in the harvest year and in the post2 year, and were also lower than in the controls in the final year, but no differences in abundances occurred in the post1 year (Fig. 3). Weather conditions, particularly during the transitional period to the rainy season, were likely an important influence on salamander surface activity. When harvests reduce canopy cover, rainfall becomes critical to mediate ambient moisture loss. The shift from no effect in post1 year no-buffer clear-cut treatments to lower abundances in the harvest year and post2 year likely reflected the influence of "more suitable" weather conditions in the post1 year trapping session. Higher relative moisture and relatively cooler conditions may have lessened the potential detrimental treatment effects in the clear-cut treatments during that period. Lower P. vehiculum abundances in the no-buffer clear-cut treatments in the harvest year, particularly, and in the post2 year occurred during drier than average conditions and less frequent rain days. In clear-cuts, rainfall may affect salamander activity more than in mature forests because the infusion of ambient moisture facilitates salamander movement and foraging (Fraser 1976; Johnston and Frid 2002; Perkins and Hunter 2006).
In the Pacific Northwest, it has been suggested that the maintenance of microclimate is key to buffer effectiveness for P. vehiculum (Kluber and others 2009; Hawkes and Gregory 2012; Olson and others 2014). Tree canopy cover helps to maintain forest floor moisture and microclimates in riparian forests (Brosofske and others 1997; Anderson and others 2007; Rykken and others 2007). Microclimates in clear-cuts are warmer, drier, and more variable than in mature unmanaged forests due to the unimpeded penetration of solar radiation and wind to the forest floor (Chen and others 1999; Rykken and others 2007). In the absence of frequent rainfall, these conditions result in a drying of upper soil and litter layers forcing Plethodon salamanders to retreat to moist refugia to avoid dehydration (Jaeger 1978, 1980; Feder 1983). As a result of harvest, the period of suitable ambient conditions for surface activity is shortened and the capacity of the forest to support salamanders is decreased (Welsh and Droege 2001). In our study, the low surface abundance of salamanders in the no-buffer clear-cut treatment may reflect the displacement of individuals by harvest effects that restrict them to cooler and moister microhabitats and time periods (see Peterman and others 2011).
Studies from the Pacific Northwest demonstrate the importance of canopy closure and decayed logs to salamander persistence (Dupuis and others 1995; Grialou and others 2000; Vesely and McComb 2002). Recruitment of down-wood with partial canopy retention may also provide viable habitat for plethodontid salamanders (Kluber and others 2009). Thus, the anticipated negative effects of logging on P. vehiculum may be lessened by retaining pre-harvest down-wood, with new down-wood, as a result of harvest operations and wind-throw. Down-wood helps ameliorate the effects of forest thinning along headwater streams because it helps retain cool, moist conditions, which provide refugia to terrestrial amphibians (Rundio and Olson 2007). Logs of wide size range, as well as soils also protect plethodontid salamanders against thermal extremes reached by air temperatures in thinned stands with limited overstory (Kluber and others 2008, 2009). In our post-treatment patch buffers there were large reductions of canopy closure in patch-cuts with concurrent large additions of down-wood logs (Table 3). Together, with lesser canopy reduction and increases in down-wood in the uncut patches, there was no overall effect on P. vehiculum abundances in patch buffers, which seems consistent with the above-cited studies. In no-buffer clear-cuts, there was no consistent similarly-moderating influence (no effect) on abundances from retention of large amounts of down-wood except when mediated by higher rainfall and cooler conditions. Therefore, it appears that maintaining amounts of dead down-wood in no-buffer clear-cut treatments and in the cut areas of patch buffers that averaged 3 to 6 times greater than occurred before buffer creation along these headwater streams (Table 3) may help lessen initial treatment effects on these woodland salamanders.
The CVs of P. vehiculum abundance may reflect a continuity of abundance by buffer type, even though wide variation occurred in some individual treatments. The low variation in post-harvest strip buffers may reflect the positive effectiveness of this treatment across a wide range of buffer width dimensions. Initial post-treatment conditions in strip buffers varied from relatively low shading in narrower widths to relatively dense shading in the widest buffers. Canopy cover in strip buffers after treatment ranged between 68 and 100% of pretreatment cover ([bar.x] = 85%, [s.sub.[bar.x]] = 4). Tree wind-throw caused by windstorms also altered strip and patch buffers causing dynamic inter-year structural changes. These changes likely contributed to the high within-year variation in P. vehiculum abundances across individual strip buffers (error bars, Fig. 3). In a long term study of amphibians in Oregon, Olson and others (2014) surmised that mixed widths of thinned headwater stream buffers may be beneficial from a management perspective to hedge the uncertainties of ecological risks to riparian vertebrates. Our data tacitly support this idea because P. vehiculum abundances in the strip and patch buffers were similar across years and similar to controls, except for marginal differences in strip buffers between the pre-treatment and the post 1 years (Table 2). The discontinuous patch buffers may be even more dimensionally and structurally diverse than the strip buffers, manifesting a wide range of site conditions with high tree-fall. The tree-clumped patches in the clear-cut landscape were maintained to protect fragile sites, often having seeps. Wet soils create diverse microhabitats and are more preferred by some salamanders than adjacent streamside habitats (Grover and Wilbur 2002), and inherently attract a diverse ground-dwelling fauna, while being excluded from harvest. The patch buffers, therefore, may also provide a hedge to ecological uncertainties and risks to riparian vertebrates as surmised for the strip buffers.
Although we observed differential effects of alternative buffer designs on abundances of P. vehiculum, the consistent implementation of our experimental approach among sites was challenging (see Bisson and others 2013). This may limit the scope of inference of our findings because: (1) treatments were not assigned randomly and the design was imbalanced owing to operational constraints; (2) the implementation of buffers among years introduced potential time effects among sites (such an effect may also confound our single pre-treatment year of data collection); and (3) large variation in the ultimate dimensional design of strip and patch buffers may have obscured some differences among treatments (Wilk and others 2010; Raphael and Wilk 2013). Log cover along small streams following harvest is influenced by amount and pattern of tree-fall (Grizzel and Wolff 1998). In our experience, there was also a strong element of observer bias in the down-wood data in sampling plots due to the presence of live wind-thrown trees (which were not tallied as down-wood) and logging-induced debris; each added to the variability in down-wood measurements. Overall, forest-floor structure was therefore underestimated for actual down-wood (dead and alive) due to unrecorded live logs. Despite these challenges, our results seem to support the results of other studies and provide a basis for comparisons that seem biologically meaningful (see Johnson 1999), but may not necessarily provide evidence for extrapolating conclusions to areas beyond the scope of this study. Our findings do, however, supplement and highlight the need for additional data to help inform and support headwater stream riparian habitat conservation planning in the Washington Coast Range.
R Bigley coordinated the DNR aspects of the study and partially funded a 2002 pilot study. AS Anderson, G Becker, B Belanger, S Crow, MH Dennis, MC Doyle, BA Evans, K Gridley, MN Halcott, J Heimburg, S Hull, A Hyduke, BT Kavanagh, C Kellum, M Klotz, SM Leach, JK Lockwood, M Moore, A Pauley-Cawley, KR Randall, JA Reed, NC Rice, W Scarlet, R Shetler, H Shelan-Cole, and KI Wilk (WI Wolf) contributed to field data collection. P Cunningham provided statistical guidance. Comments from DH Olson and 3 reviewers markedly improved the paper.
ANDERSON PD, LARSON DJ, CHAN SS. 2007. Riparian buffer and density management influences on microclimate of young headwater forests of western Oregon. Forest Science 53:254-269.
AUBRY KB, STRINGER AB. 2000. Field test of the SMED, a small mammal escape device for pitfall trapping amphibians. Northwestern Naturalist 81:69.
BISSON PA, CLAESON SM, WONDZELL SM, FOSTER ad, STEEL A. 2013. Evaluating headwater stream buffers: Lessons learned from watershed-scale experiments in southwest Washington. In: Anderson PD, Ronnenberg KL, editors. Density management in the 21st Century west side story. Portland, OR: USDA Forest Service. General Technical Report PNW-GTR-880. p 169-188.
BROSOFSKE KD, CHEN J, NAIMAN RJ, FRANKLIN JF. 1997. Harvesting effects on microclimatic gradients from small streams to uplands in western Washington. Ecological Applications 7:1188-1200.
BUNNELL FL, HOUDE I. 2010. Down-wood and biodiversity--implications to forest practices. Environmental Review 18:397-421.
BURY RB, CORN PS. 1988. Douglas-fir forests in the Oregon and Washington Cascades: Relation of the herpetofauna to stand age and moisture. In: Management of amphibians, reptiles and small mammals in North America. Proceedings of a symposium 17 to 21 July 1988; Flagstaff, AZ. p 11-22.
CHEN J, SAUNDERS SC, CROW TR, NAIMAN RJ, BROSOFSKE KD, MROZ GD, BROOKSHIRE BL, FRANKLIN JF. 1999. Microclimate in forest ecosystem and landscape ecology. BioScience 49:288-297.
DUPUIS LA, BUNNELL F. 1999. Effects of stand age, size and juxtaposition on abundance of Western
Redback Salamanders (Plethodon vehiculum) in Coastal British Columbia. Northwest Science 73:27-33.
DUPUIS LA, SMITH NM, BUNNELL F. 1995. Relation of terrestrial breeding amphibian abundance to tree-stand age. Conservation Biology 9:645-653.
FEDER ME. 1983. Integrating the ecology and physiology of plethodontid salamanders. Herpetologica 39:291-310.
FIELD A. 2013. Discovering statistics using IBM SPSS Statistics. London, UK: Sage Publications Limited. 915 p.
FRASER DF. 1976. Empirical evaluation of the hypothesis of food competition in salamanders of the genus Plethodon. Ecology 57:459-471.
FRAZER GW, CANHAM CD, LERTZMAN KP. 1999. Gap Light Analyzer (GLA): Imaging software to extract canopy structure and gap light transmission indices from true-colour fisheye photographs. Users manual and program documentation, Version 2.0. Burnaby, BC: Simon Fraser University, and Millbrook, NY: Institute of Ecosystem Studies.
GRIALOU JL, WEST SD, WILKINS RN. 2000. The effects of forest clearcut harvesting and thinning on terrestrial salamanders. Journal of Wildlife Management 64:105-113.
GRIZZEL JD, WOLFF N. 1998. Occurrence of wind-throw in forest buffer strips and its effect on small streams in northwest Washington. Northwest Science 72:214-223.
GROVER MC, WILBUR HM. 2002. Ecology of ecotones: Interactions between salamanders on a complex environmental gradient. Ecology 83:2112-2123.
HAWKES VC, GREGORY PT. 2012. Temporal changes in the relative abundance of amphibians to riparian buffer width in western Washington, USA. Forest Ecology and Management 274:67-80.
JAEGER RG. 1978. Plant climbing by salamanders: Periodic availability of plant-dwelling prey. Copeia 1978:686-691.
JAEGER RG. 1980. Microhabitats of a terrestrial forest salamander. Copeia 1980:265-268.
JOHNSON DH. 1999. The insignificance of statistical significance testing. Journal of Wildlife Management 63:763-772.
JOHNSTON B, Frid L. 2002. Clearcut logging restricts the movements of terrestrial Pacific Giant Salamanders (Dicamptodon tenebrosus Good). Canadian Journal of Zoology 80:2170-2177.
KLUBER MR, OLSON DH, PUETTMANN KJ. 2008. Amphibian distributions in riparian and upslope areas and their habitat associations on managed forest landscapes in the Oregon Coast Range. Forest Ecology and Management 256:529-535.
KLUBER MR, OLSON DH, PUETTMANN KJ. 2009. Downed wood microclimates and their potential impact on plethodontid salamander habitat in the Oregon Coast Range. Northwest Science 83:25-34.
MCCUNE B, GRACE B. 2002. Analysis of ecological communities. Gleneden Beach, OR: MJM Software Design. 300 p.
MCDADE KA, MAGUIRE CC. 2005. Comparative effectiveness of three techniques for salamander and gastropod land surveys. American Midland Naturalist 153:309-320.
NAKAGAWA S, CUTHILL IC. 2007. Effect size, confidence interval and statistical significance: A practical guide for biologists. Biological Review 82:591-605.
OLSON DH, LIERNESS JB, CUNNINGHAM PG, STEEL EA. 2014. Riparian buffers and forest thinning: Effects on headwater vertebrates 10 years after thinning. Forest Ecology and Management 321:81-93.
PATRICK DA, HUNTER JR ML, CALHOUN AJK. 2006. Effects of experimental forestry treatments on a Maine amphibian community. Forest Ecology and Management 234:323-332.
PERKINS D, HUNTER JR ML. 2002. Effects of placing sticks in pitfall traps on amphibian and small mammal capture rates. Herpetological Review 33: 282-284.
PERKINS D, HUNTER JR ML. 2006. Effects of riparian timber management on amphibians in Maine. Journal of Wildlife Management 70:657-670.
PETERMAN WE, CRAWFORD JA, SEMLITSCH RD. 2011. Effects of even-aged timber harvest on stream salamanders: Support for the evacuation hypothesis. Forest Ecology and Management 263:23442353.
RAPHAEL MG, WILK RJ. 2013. The riparian ecosystem management study: A response of small mammals to streamside buffers in western Washington. In: Anderson PD, Ronnenberg KL, editors. Density management in the 21st Century west side story. Portland, OR: USDA Forest Service. General Technical Report PNW-GTR-880. p 163-168.
RUNDIO DE, OLSON DH. 2007. Influence of headwater site conditions and riparian buffers on terrestrial salamander response to forest thinning. Forest Science 53:320-330.
RYKKEN JJ, CHAN SS, MOLDENKE AR. 2007. Headwater riparian microclimate patterns under alternative forest management treatments. Forest Science 53: 270-280.
VESELY DG, McComb WC. 2002. Salamander abundance and amphibian species richness in riparian buffer strips in the Oregon Coast Range. Forest Science 48:291-297.
WADDELL KL. 2002. Sampling coarse woody debris for multiple attributes in extensive resource inventories. Ecological Indicators 1:139-153.
WELSH JR HH, DROEGE S. 2001. A case for using plethodontid salamanders for monitoring biodiversity and ecosystem integrity of North American forests. Conservation Biology 15:558-569.
WELSH JR HH, HODGSON GR. 2013. Woodland salamanders as metrics of forest ecosystem recovery: A case study from California's redwoods. Ecosphere 4: article 59.
WILK RJ, RAPHAEL MG, NATIONS CS, RICKLEFS JD. 2010. Initial response of small ground-dwelling mammals to alternative buffers along headwater streams in the Washington Coast Range, USA. Forest Ecology and Management 260:1567-1578.
WILKINS RN, PETERSON NP. 2000. Factors related to amphibian occurrence and abundance in headwater streams draining second-growth Douglas-fir forests in southwestern Washington. Forest Ecology and Management 139:79-91.
RANDALL J WILK
USD A Forest Service, Pacific Northwest Research Station, Olympia Forestry Sciences Laboratory, 3625 93rd Ave SW, Olympia, WA 98512 USA; email@example.com
JEFFREY D RICKLEFS
Washington State Department of Natural Resources, 1111 Washington St SE, PO Box 47106, Olympia, WA 98504 USA
MARTIN G RAPHAEL
USDA Forest Service, Pacific Northwest Research Station, Olympia Forestry Sciences Laboratory, 3625 93rd Ave SW, Olympia, WA 98512 USA
TABLE 1. Study locations (see Figure 1 for code names), riparian buffer assignments, and number of individual stream visits (trap checks) for 23 headwater streams. Riparian buffer design (number of streams): Continuous Discontinuous strip buffer patch buffer No buffer Control LOCATION RO 2 0 1 1 TA 1 2 1 1 SS 1 0 1 1 LR 1 1 1 1 EF 1 0 1 1 SR 1 0 2 1 TRAP CHECKS Pre-treatment 48 24 48 48 Harvest year 56 24 40 48 Post 1 year 56 24 56 48 Post 2 year 56 24 56 48 TABLE 2. Results of analyses on captures-stream ([bar.x] [log.sub.e] [y +1]) of Plethodon vehiculum. * = statistically significant; r- value shown only if P [less than or equal to] 0.1; sequence of table values = t-value (df), P-value, r-value. Between treatments: Strip buffer Patch buffer No-buffer CONTROL Pre-treatment 0.72 (10), 0.49 0.78 (7), 0.46 0.28 (10), 0.79 Harvest year 0.21 (11), 0.83 -0.95 (7), 0.38 1.28 (9), 0.23 Postl year -0.92 (11), 0.38 0.17 (7), 0.87 -0.57 (11), 0.58 Post2 year 1.31 (11), 0.22 0.24 (7), 0.82 3.80 (11), 0.003, 0.75 * STRIP BUFFER Pre-treatment -0.10 (7), 0.92 0.52 (8.8) 0.62 Harvest year 1.45 (8), 0.19 -1.35 (10), 0.21 Postl year -0.91 (8), 0.39 0.60 (9.3), 0.56 Post2 year 0.83 (8), 0.43 -1.28 (12), 0.22 PATCH BUFFER Pre-treatment 0.64 (7), 0.54 Harvest year -3.94 (6), 0.008, 0.85 * Postl year 0.84 (8), 0.43 Post2 year -2.90 (8), 0.020, 0.71 * Between years: Harvest year Post1 year Post2 year PRE-TREATMENT Control 1.05 (5), 0.34 0.86 (5), 0.43 0.18 (5), 0.87 Strip buffer 1.09 (5), 0.32 -2.13 (5), 0.15 (5), 0.88 0.086, 0.69 * Patch buffer -0.84 (2), 0.49 1.00 (2), 0.42 -0.70 (2), 0.56 No-buffer 2.72 (3), 0.072, 0.68 (5), 0.53 2.56 (5), 0.051, 0.84 * 0.75 * HARVEST YEAR Control 0.10 (5), 0.92 -1.26 (5), 0.26 Strip buffer -2.32 (6), -0.14 (6), 0.90 0.060, 0.69 * Patch buffer 1.56 (2), 0.26 1.00 (2), 0.42 No-buffer -- (a) 0.00 (4), 1.00 POST1 YEAR Control -1.18 (5), 0.29 Strip buffer 1.66 (6), 0.15 Patch buffer -1.82 (2), 0.21 No-buffer 2.50 (6), 0.047, 0.71 * (a) t-value could not be computed because the [s.sub.[bar.x]] of the difference was 0. TABLE 3. Measurements of tree canopy and dead down-wood logs ([bar.x] [+ or -] [s.sub.[bar.x]] before and after treatment harvests. Statistical results for canopy cover are from analyses on transformed data. Statistically significant if P [less than or equal to] 0.1; r > 0.5 = large effect size. Control Strip buffer Canopy cover (%) Pre-treatment 77 [+ or -] 1 77 [+ or -] 2 Post-treatment 77 [+ or -] 0.3 66 [+ or -] 4 t (df) -0.15 (5) 3.20 (6) P, r 0.89, 0.07 0.03, 0.79 Dead down-wood (n/ha) Pre-treatment 731 [+ or -] 167 560 [+ or -] 206 Post-treatment 1132 [+ or -] 333 1065 [+ or -] 211 t (df) -1.85 (5) -7.60 (6) P, r 0.12, 0.64 <0.001, 0.96 Patch leave Patch cut Canopy cover (%) Pre-treatment 78 [+ or -] 1 74 [+ or -] 0.3 Post-treatment 56 [+ or -] 7 7 [+ or -] 4 t (df) 2.87 (2) 6.31 (2) P, r 0.10, 0.87 0.02, 0.98 Dead down-wood (n/ha) Pre-treatment 824 [+ or -] 119 466 [+ or -] 46 Post-treatment 1435 [+ or -] 241 2614 [+ or -] 515 t (df) -3.87 (2) -4.06 (2) P, r 0.06, 0.94 0.06, 0.80 No-buffer Canopy cover (%) Pre-treatment 78 [+ or -] 1 Post-treatment 9 [+ or -] 2 t (df) 12.81 (6) P, r <0.001, 0.99 Dead down-wood (n/ha) Pre-treatment 708 [+ or -] 222 Post-treatment 1822 [+ or -] 640 t (df) -2.41 (6) P, r 0.05, 0.70
|Printer friendly Cite/link Email Feedback|
|Author:||Wilk, Randall J.; Ricklefs, Jeffrey D.; Raphael, Martin G.|
|Publication:||Northwestern Naturalist: A Journal of Vertebrate Biology|
|Date:||Sep 22, 2014|
|Previous Article:||Age, region, and temporal patterns of trace elements measured in stranded Harbor Seals (Phoca vitulina richardii) from Washington inland waters.|
|Next Article:||Experimental evidence that scare tactics and effigies reduce corvid occurrence.|