Salvage logging after windthrow alters microsite diversity, abundance and environment, but not vegetation Downloaded from https://academic.oup.com/forestry/article/81/3/361/655733 by guest on 30 September 2021

CHRIS J. PETERSON 1 * and ANDREA D. LEACH 2

1 Department of Plant Biology, University of Georgia, 2502 Plant Science Building, Athens, GA 30602, USA 2 Department of Cellular and Molecular Biology, University of Georgia, Athens, GA 30602, USA * Corresponding author. [email protected]

Summary An increasing number of researchers propose that disturbance effects in forests are mediated through ‘ legacies’ , which are organisms or organically derived structures that persist after a disturbance. Much controversy currently surrounds the potential impact of post-disturbance salvage logging, in part, because of the potential for such actions to alter post-disturbance legacies and thus forest regeneration. Microsites (e.g. downed tree crowns, boles, stumps, treefall pits and mounds) created by natural disturbances are a subset of the broader concept of legacies, but the effect of windthrow + salvage logging on microsites and their environment and vegetation has not been previously examined. In a wind-damaged forest in western Tennessee, USA, we documented microsite diversity and abundance, environmental conditions, and initial vegetation regeneration in salvaged and unsalvaged areas. We found that salvaged areas had signifi cantly greater variety of microsites, altered microsite abundance, higher soil temperature and greater canopy openness relative to unsalvaged areas. However, 2 years after the storm, herbaceous cover and species richness and tree seedling density and species richness did not differ between salvaged and unsalvaged areas. Soil moisture also was unaffected by salvaging. In contrast, environmental conditions and vegetation characteristics differed signifi cantly among microsite types, with treefall mounds being warmer and drier than other microsites. This intermediate- severity wind disturbance, followed by moderate intensity of salvaging, created microsites that differed in environment and vegetation, and although the salvaging altered microsite diversity, abundances and conditions, the initial vegetation did not show detrimental effects of the salvage operations. We suggest that primary determinants of the consequences of salvaging after natural disturbance are the severity of the natural disturbance, and intensity of salvage operations. Detrimental effects of salvaging may accrue only if some combined severity threshold is exceeded.

Introduction 2006; Newton et al. , 2006; Baird, 2006) have brought the discussion of post-disturbance sal- Recent natural disturbances (e.g. extremely large vage logging to a heated intensity (Floyd, 2006). wildfi res in numerous locations worldwide) and Currently, most forest areas subject to large natu- controversial research fi ndings (Donato et al. , ral disturbances (e.g. fi re or wind) subsequently

© Institute of Chartered Foresters, 2008. All rights reserved. Forestry, Vol. 81, No. 3, 2008. doi:10.1093/forestry/cpn007 For Permissions, please email: [email protected] Advance Access publication date 19 March 2008 362 FORESTRY have salvage harvesting activities that aim to re- reduced in salvaged areas. To our knowledge, coup fi nancial losses from the natural disturbance Elliott et al. (2002) and Rumbaitis Del Rio (2006) ( Lindenmayer et al. , 2004). These salvage logging represent the only studies available on the effect operations are widespread and common, but we of windthrow + salvaging on regeneration. Con- know very little about how such management ac- sequently, there is yet exceedingly little informa- tions infl uence forest regeneration, biodiversity tion about forest response to salvaging after wind and risk of later disturbance. disturbance, particularly in eastern North Amer- One of the potential infl uences of salvage log- ica. Therefore, we report here on the microsite ging on forest recovery and diversity might occur abundances and very early regeneration follow- Downloaded from https://academic.oup.com/forestry/article/81/3/361/655733 by guest on 30 September 2021 via alterations in the variety and abundance of ing moderate windthrow in hardwood forests of microsites (Beatty, 1984; Peterson et al. , 1990; central Tennessee, USA. Following Lindenmayer Kuuluvainen and Juntunen, 1998), which are et al. (2004), we hypothesized that salvaging one component of the broader concept of biotic might destroy treefall mounds and remove tree legacies (Franklin et al. , 2000; Lindenmayer, trunks and crowns in addition to exposing bare 2006). Lindenmayer et al. (2004, see also Foster soil in skid trails and depositing pruned branches and Orwig, 2006 ; Lindenmayer and Noss, 2006) in slash piles. Therefore we expected that sal- suggested that one way disturbances maintain vaging reduces microsite diversity and alters ecosystem diversity and function is through the microsite abundance (hypothesis 1). We further variety of legacies resulting from disturbances, expected that since salvaging removes downed and that salvaging reduces legacy diversity. It is stems and crowns potentially capable of sprout- indeed well known that differing microsites in ing and surviving, it would increase canopy open- wind-disturbed forests have distinct environ- ness at ground level (hypothesis 2). Finally, we ments (Beatty, 1984; Beatty and Sholes, 1988; expected that microsites with exposed mineral Carlton and Bazzaz, 1998a), and differentially soil (pits, mounds, bare soil patches) and more infl uence tree seedling germination, growth and open conditions (salvaged areas) would contain vulnerability to herbivores (Long et al. , 1998; a higher proportion of shade-intolerant, early- Peterson and Pickett, 1995, 2000; Krueger and successional species (hypothesis 3) and higher Peterson, 2006 ). These effects often result in species richness (hypothesis 4). vegetation variation among microsites, often with greater abundance of pioneer species in the more severely disrupted microsites (e.g. Peter- Materials and methods son et al. , 1990; Peterson and Campbell, 1993), Study site although such among-microsite variation is not always found (Webb and Scanga, 2001 ). Thus, This research was conducted at Natchez Trace decreases in the variety or abundance of micro- State Park and Forest (NTSF), an 18 245-ha sites might decrease forest regeneration poten- natural area located in west-central Tennessee tial or species diversity in the regenerating area. (35º N 88º W). The area lies within the East Gulf However, we know of no studies of microsite Coastal Plain section of the Coastal Plain phys- abundances, composition or diversity after natu- iographic province (Braun, 1950). Braun (1950) ral disturbance and salvaging. classifi ed the vegetation of western Tennessee as The existing information about consequences belonging to the Mississippi Embayment section of post-fi re or post-windthrow salvaging in North of the Western Mesophytic Forest Region. The America are mostly restricted to the north-west topography consists of gently rolling uplands (e.g. Donato et al. , 2006) or boreal sites in Canada separated by broad fl oodplains; elevation aver- ( Greene et al. , 2006; Macdonald, 2007). Donato ages 165 m above sea level, ranging from 137 to et al. (2006) found reduced tree seedling density 182 m ( Smalley, 1991). Soils are derived from the and diversity in salvaged areas of the 2002 Biscuit McNairy sands geologic formation and consist of Fire in Oregon, and Greene et al. (2006) report well-drained fi ne sandy loams and silt loams with that although asexual reproduction by Populus occasional fragipans and a discontinuous loess was unaffected by salvaging in central Quebec, cap ( Kupfer and Franklin, 2000 ). The climate is the regeneration from seed by conifers was much humid continental, with short, mild, wet winters SALVAGING AFTER WIND DISTURBANCE 363 and long, hot, dry summers and an average grow- hectares), perfect randomization and replication ing season of 202 days. Mean annual precipita- would be diffi cult to achieve. tion is 1240 mm, mostly rain falling in the late To quantify relative abundances of microsite winter and early spring (Flowers et al. , 1960). types, we sampled along four parallel 30-m lines Much of the park was settled and cleared for 10 m apart in each large plot. At 0.5-m inter- agriculture in the 1800s and early 1900s, but the vals along these lines, we dropped a pin and re- land was abandoned in 1935 and was deeded to corded microsite information, yielding a total of the state of Tennessee in 1955 (Smalley, 1991). 244 points per plot (244 points × 32 plots; n = The second-growth forests consist of Quercus 7,808 sampling points). We classifi ed points into Downloaded from https://academic.oup.com/forestry/article/81/3/361/655733 by guest on 30 September 2021 species (oaks), Carya species (hickories) and other 13 microsite categories on the basis of the sub- mixed mesophytic species, including Pinus taeda strate conditions and the nature of overlying ma- L. (loblolly pine), in the uplands. The understory terials; this classifi cation is the basis for Figure consists primarily of Cornus fl orida L. (fl owering 2 , but does not utilize vegetation in the microsite dogwood), Nyssa sylvatica Marsh. (blackgum) defi nition (see ‘ focal microsites’ below). Substrate and Sassafras albidum (Nutt.) Nees. (sassafras). was recorded as organic or mineral, and as intact On 5 May 1999, a downburst (straight-line or disrupted. If disrupted, the disruption was re- thunderstorm) damaged ~ 3000 ha at NTSF (Roy corded as natural (treefall pits or mounds, other Ward, State Forest Supervisor, personal com- natural disruption) or anthropogenic (skid trails, munication). As recorded in nearby Lexington, other anthropogenic). Any material overlaying Tennessee, USA, the thunderstorm lasted over the substrate was recorded as leaves, needles, fi ne 1.5 h with sustained wind speeds over 90 km woody debris (includes the distal portion of tree h − 1 and produced wind gusts exceeding 145 km crowns and fi ne debris piles in salvaged areas), h − 1 (NOAA, 1999). Forest damage was patchy, coarse woody debris (includes tree trunks), moss and basal area losses ranged from 26.6 per cent or nothing (bare). We also recorded the identity to 49.5 per cent in our study area. All of the of any vascular plants touching the pin below ‘ downed’ trees were harvested with the exception 1.5 m (ferns, sedges, grasses, to genus for shrubs of two areas (6 ha each, ~ 2.5 km apart) that were and to species for trees). From this inventory, set-aside for research purposes and left unsal- we calculated the total absolute area covered by vaged. Because two pairs of salvaged/unsalvaged each of the following microsite types: intact bare locations were available at NTSF, we sampled at mineral soil, intact bare organic soil, treefall root both locations and refer to them as Site A and pits, treefall mounds, other naturally disturbed Site B. Salvage operations removed 105 stems (73 soil (e.g. eroded stream bank), intact leaf-covered stems ha− 1 ) and 10.88 m2 (7.55 m 2 ha − 1 ) of basal soil, intact needle-covered soil, fi ne woody debris, area in the areas that we sampled. coarse woody debris, stumps, moss, skid trails and other anthropogenically disturbed soil. We investigated the vegetation and environ- Sampling design mental conditions in detail within the fi ve most During June– August 2001, sampling plots were common wind-related microsite types (focal mi- established within salvaged (S) and unsalvaged crosites): treefall pits, treefall mounds, downed (U) portions of wind-damaged forest at Sites A tree crowns, patches of bare soil (intact soil, and B (‘ stands’ refers to the four site × treatment sparse litter, no vegetation) and Vitis rotundifo- combinations, i.e. unsalvaged portion of Site lia patches. Note that in contrast to the general A = ‘ UA’ , etc.). Eight large (30 m × 30 m) plots microsite abundance inventory above, here we were used to quantify forest damage in each of include a vegetation-defi ned microsite: Vitis ro- the four stands. In the strictest sense, plots may tundifolia was the most locally abundant ground- not qualify as replicates because they are nested layer vegetation and therefore used as a vegetated within stands, and salvage treatments were as- microsite category. Our aim was to use micro- signed haphazardly rather than randomly to site types similar to those used in other studies stands. However, we perform our statistical anal- ( Nakashizuka, 1989; Carlton and Bazzaz, 1998b ; yses below as if plots were true replicates; at the Kuuluvainen and Juntunen, 1998). Because most scale of this study (e.g. stands cover many tens of crowns were removed from salvaged areas, we 364 FORESTRY substituted slash piles for tree crowns in salvaged 1999). Photographs were taken at 1.25 m above areas when necessary since they were the closest the center of each quadrat in late June 2002 using analog to downed tree crowns. a Nikon 885 Coolpix digital camera fi tted with To examine groundlayer (<2 m tall) vegetation an FC-E8 fi sheye converter lens (Nikon Corpora- and environmental characteristics of microsites, tion, Tokyo, Japan). We used a gimbal mounted we established 0.5-m2 circular quadrats in the on a monopod to level the camera to a zero zenith fi ve microsite types within each large plot. Quad- angle and oriented it to true north prior to taking rats were placed in up to three replicates of each photos. All photographs were taken in the early of the fi ve microsite types in a stratifi ed random morning and early evening hours to minimize Downloaded from https://academic.oup.com/forestry/article/81/3/361/655733 by guest on 30 September 2021 manner, although three replicates were not always the solar glare that distorts the estimates of true available for all fi ve microsite types in each plot. canopy cover. Per cent canopy openness was de- The total number of microsite quadrats was 369 termined from the digital images using Gap Light (see Table 1 , for microsite quadrat frequencies). Analyzer software ( Frazer et al. , 1999), which Within each quadrat, we recorded the species divides pixels into sky and non-sky classes and identity and measured the height of all woody calculates canopy openness based on per cent sky stems up to 2 m tall, and visually estimated per pixels relative to non-sky (canopy) pixels ( Frazer cent cover of herbs and vines. Per cent cover was et al. , 1999). noted for all plant material overhanging the quad- We measured soil temperature and soil mois- rat, whether or not the stem was rooted in the ture simultaneously on two occasions in June plot (thus total cover could exceed 100 per cent). 2002. Within each quadrat, we measured tem- When species identities were unknown, voucher perature of the upper 15 cm of soil at three points specimens were collected for identifi cation. Spe- using Reotemp stem thermometers that were al- cies nomenclature follows the USDA Plants list lowed to equilibrate for 45 s. Per cent soil mois- ( USDA and NRCS, 2002). ture (volumetric water content) of the top 12 We characterized canopy openness using hemi- cm of soil was determined using a Hydrosense spherical photography (Robison and McCarthy, time-domain refl ectometer (Decagon Devices,

Table 1 : MANOVA and univariate ANOVAs for treatment and microsite effects on environmental variables

Source of variation Num df Den df Pillai’s traceF P

MANOVA Treatment 3 357 0.368 69.41 <0.0001 Microsite 12 1077 0.384 13.19 <0.0001 Treatment × microsite 12 1077 0.054 1.65 0.0729 ANOVA Type III SS Temperature Treatment 1 0.012 84.40 <0.0001 Microsite 4 0.020 35.77 <0.0001 Treatment × microsite 4 0.001 1.61 0.1705 Error 359 0.049 Moisture Treatment 1 0.000 0.00 0.9880 Microsite 4 2.178 17.55 <0.0001 Treatment × microsite 4 0.048 0.39 0.817 Error 359 11.133 Per cent openness Treatment 1 6.441 158.61 <0.0001 Microsite 4 0.296 1.82 0.1243 Treatment × microsite 4 0.606 3.73 0.006 Error 359 14.579 df = Degrees of freedom; Num = numerator; Den = denominator. SALVAGING AFTER WIND DISTURBANCE 365

Inc., Pullman, Washington, USA) at three points rial MANOVA. Environmental data were log- within each microsite quadrat. All measurements transformed to meet assumptions of normality were taken within 2 h of solar noon. Because soil and homogeneity of variances. Signifi cant main temperature and moisture can vary dramatically effects or interactions from MANOVA were ex- temporally and spatially, readings were averaged amined more closely with univariate two-factor within and between sampling periods. ANOVAs.

Vegetation We conducted separate two-way (fi ve microsite types × two treatments) ANO- Downloaded from https://academic.oup.com/forestry/article/81/3/361/655733 by guest on 30 September 2021 Statistical analysis VAs on herbaceous cover and species richness, Microsite diversity and abundance Because most and woody seedling density and richness. Data plots were represented by 244 pin-drop points were log10 (x + 1) or square root transformed classifi ed into the 13 microsite categories, we fi rst to meet normality and equal variance assump- calculated richness (number of microsite catego- tions. Signifi cant ANOVAs were followed by ries) and diversity (Shannon’s H’ ) of microsites Tukey’s studentized range test (HSD) multiple per plot. These values were then compared among comparison procedure to determine which site/treatment combinations to test for a treatment means differed signifi cantly from one another. (salvaging) effect, using site (A or B) as a block- Correlations, MANOVA and ANOVAs were ing variable and n = 8 plots as replicates. Because performed using SAS statistical software (SAS natural windthrow levels were not equal across all Institute, 1995). plots, and severity of natural disturbance is likely We conducted G -tests of independence to to infl uence microsite characteristics, we used the test for differences among treatments in relative severity of natural disturbance (proportion of pre- abundances of microsite types. We also assigned disturbance trees fallen) as a covariate. all tree species to one of three shade-tolerance We examined microsite ‘ composition’ of plots classes (tolerant, intermediate or intolerant) based in two ways: in an exploratory step, we per- on information in Burns and Honkala (1990) and formed detrended correspondence analysis ordi- Brown and Kirkman (1990) and conducted G - nation (McCune and Mefford, 1999) to reveal tests of independence to determine whether suc- similarity and differences among plots of various cessional status (based on proportions of tolerant, treatments in terms of their microsites. To test for intermediate or intolerant) of the seedling layer differences in relative abundances, we performed or abundance of the fi ve most common seedling G -tests of homogeneity among treatments, using species differed between salvaged and unsalvaged data from all plots pooled within a treatment. areas or among the fi ve microsite types (Sokal and Rohlf, 1995 ). Species without published tolerance Abiotic conditions To determine whether en- values were excluded from analysis, but because vironmental conditions (canopy openness, soil these were rare, their exclusion is unlikely to have temperature and soil moisture) varied predictably infl uenced conclusions. among microsite types and/or treatments, we used We used canonical correspondence analysis multivariate analysis of variance (MANOVA). (CCA) to examine similarities in species compo- This was done because when dependent vari- sition among quadrats and correlations between ables are correlated, it is more appropriate to use compositional variation and resource levels MANOVA than multiple univariate ANOVAs; ( Palmer, 1993). Herbs and woody seedlings were MANOVA controls for Type I error (which is analyzed separately because we used different infl ated when conducting multiple univariate measures of abundance (percent cover for her- ANOVAs) and adjusts for correlations between baceous vegetation and height for woody seed- dependent variables. Pillai’s trace test statistic is lings). Quadrats not containing any herbaceous extremely robust and its use is suggested when de- or woody seedling species or lacking complete signs are unbalanced, sample sizes are small and environmental information were removed from covariances are non-homogeneous. We analyzed analysis. We included 344 of the 369 quadrats in light, temperature and moisture simultaneously the herbaceous analysis and 208 in the seedling using a 2 (treatment) × 5 (microsite type) facto- analysis. Environmental variables used in both 366 FORESTRY analyses include soil temperature, soil moisture and per cent canopy openness; all environmen- tal variables were log-transformed to reduce the infl uence of outliers ( Palmer, 1993 ). We used PC-Ord software (McCune and Mefford, 1999) to conduct the CCAs, and all environmental vari- ables (soil temperature, soil moisture and per cent canopy openness) were log-transformed to reduce the infl uence of outliers. In the follow- Downloaded from https://academic.oup.com/forestry/article/81/3/361/655733 by guest on 30 September 2021 ing, all reports of signifi cant tests are signifi cant at P < 0.05.

Results Microsite diversity Figure 1 . Richness of microsites across four stand types. Mean ± standard deviation of number of mi- The richness of microsite types per plot ranged crosites encountered in n = 8 plots per stand type. from 3 to 11 and was signifi cantly greater in sal- Sal = salvaged; Uns = unsalvaged. vaged than in unsalvaged areas ( Figure 1 ) when controlling for blowdown severity (ANCOVA, site was not signifi cant). Evenness of microsites age, there was more coverage by Vitis microsites was signifi cantly greater in salvaged than unsal- in unsalvaged areas than in salvaged areas (data vaged areas (ANCOVA, block was again not not shown). In salvaged areas, rank order of signifi cant), and microsite diversity measured as microsite abundance was Vitis , bare, crown, H ′ was marginally greater in salvaged areas than mound and pit microsite types (in descending in unsalvaged. Considering all 32 plots, simple order), whereas in unsalvaged areas, it was Vitis , linear regression of severity of wind damage vs crowns, mounds and pits, and bare soil micro- richness of microsites produced a signifi cant sites. Salvaged areas contained a much higher positive relationship (R 2 = 0.260, n = 32). percentage of bare soil microsites (24.8 per cent vs 0.7 per cent) and mound microsites (12.2 per cent vs 2.7 per cent) than unsalvaged areas. Microsite abundance However, unsalvaged areas contained more crown coverage (27 per cent vs 14.7 per cent) Intact soil with leaf litter cover was the most than salvaged areas. abundant of the microsites, followed by fi ne woody debris/tree crown microsites (Figure 2, abundances based on soil and debris microsites, Microsite environmental conditions and thus do not consider vegetation microsites). Among the fi ve focal microsites that were stud- Soil temperature and soil moisture were nega- ied in more detail (and which include vegetation- tively correlated (r = − 0.324, n = 369) while tem- defi ned microsites), patches of Vitis and downed perature and canopy openness were signifi cantly tree crown microsites were the most abundant, positively correlated (r = 0.397). Although mois- accounting for 58.1 per cent and 22.8 per cent ture and canopy openness were not correlated of 2450 sampling points falling within one of ( r = 0.032), a MANOVA was appropriate be- the fi ve focal microsites. Bare soil patches (9 cause both were correlated with temperature. per cent), mounds (6 per cent) and pits (4.1 per Though the sampling design included two sites cent) combined accounted for 19.1 per cent of (statistical blocks), conducting separate MANO- total coverage. The relative abundances of the VAs for each site gave similar results. Therefore, we fi ve focal microsite types differed signifi cantly pooled data from the two sites to increase statisti- between treatments (G -test). Although Vitis mi- cal power (n = 369 quadrats). The two-treatment crosites accounted for the greatest overall cover- (salvaged vs unsalvaged) × fi ve microsite type SALVAGING AFTER WIND DISTURBANCE 367 Downloaded from https://academic.oup.com/forestry/article/81/3/361/655733 by guest on 30 September 2021

Figure 2 . Relative abundance of soil + debris — defi ned microsites across four stand types. Abundance based on 244 pin drops in each of eight 30 m × 30 m plots per stand type (1952 points); a few points were inacces- sible and therefore no data were recorded. Sal = salvaged; Uns = unsalvaged; min = mineral; CWD = coarse woody debris; FWD = fi ne woody debris.

MANOVA revealed a signifi cant multivariate Microsite vegetation main effect for both treatment and microsite Groundlayer vegetation Total amount of her- type on environmental variables, but no effect baceous cover differed between microsite types for treatment × microsite interaction (results for but did not differ between treatments (Figure 5); MANOVA and ANOVAs are summarized in Vitis microsites had the highest cover, and crown, Table 1). Thus, environmental differences among open and pit microsites had the lowest. However, microsites are consistent between the salvaged because ANOVA revealed a signifi cant treatment and unsalvaged treatments. × microsite type interaction, differences in herba- Because the MANOVA revealed signifi cant ceous cover among microsite types were inspected multivariate main effects, we examined the re- within each treatment separately. In the salvaged sponses of the three environmental variables sep- areas, Vitis (as expected) had the highest per cent arately in univariate ANOVAs. Soil temperature cover followed by the other microsite types, which varied signifi cantly among both treatment and did not differ in amount of cover. However, in un- microsite type (Table 1). Soil temperature was salvaged areas, herbaceous cover was signifi cantly higher in salvaged areas than in unsalvaged areas higher in mounds and Vitis patches, compared (Figure 3). Mounds were warmer than all other with crowns, pits and bare microsites, which did microsite types, bare soil patches were interme- not differ (Figure 5). Herbaceous species richness diate in temperature and pits, crowns, and Vitis differed between microsite types but did not dif- patches were coolest (Figure 4 ). Soil moisture did fer between treatments ( Figure 6 ), nor was there not differ between treatments (Figure 3) but did a signifi cant treatment × microsite interaction. signifi cantly differ among microsite types, with Species richness of herbaceous vegetation was pits having the wettest soils, mounds having the highest in bare soil microsites and lowest in driest soils and bare, Vitis , and crown microsites crown microsites (Figure 6 ). having intermediate soil moisture levels (Figure Seventy-fi ve groundlayer species were present 4). While canopy openness was higher in salvaged in the microsite quadrats. Species with the highest areas than in unsalvaged areas (Figure 3 ), it did overall cover included V. rotundifolia (41.3 per not differ between microsite types (Figure 4). cent), Parthenocissus quinquefolia (10.8 per cent), 368 FORESTRY Downloaded from https://academic.oup.com/forestry/article/81/3/361/655733 by guest on 30 September 2021

Figure 4 . Resource levels by focal microsite. (A) Soil Figure 3 . Resource levels by treatment. (A) Soil temperature. (B) Soil moisture. (C) Per cent canopy temperature. (B) Soil moisture. (C) Per cent canopy openness. Mean ± standard deviation. Different let- openness. Mean ± standard deviation. Different let- ters denote signifi cant differences at level α = 0.05. ters denote signifi cant differences at level α = 0.05.

Smilax spp. (5.3 per cent; includes Smilax glauca ration of salvaged and unsalvaged quadrats along and Smilax rotundifolia ) and Lonicera japonica Axis 1 (data not shown), but the eigenvalue for (4.1 per cent). Disturbance specialists or pioneer the fi rst axis was only 0.257, accounting for a species— Rubus spp. (includes Rubus argutus and meager 1.1 per cent of the variance in species Rubus fl agellaris ), L. japonica , S. glauca , Toxico- composition. dendron radicans , and Pueraria montana — were more abundant in salvaged areas while P. quin- Tree seedlings Total tree seedling density (all quefolia and V. rotundifolia were more abundant species pooled) was highest in bare soil and pit in unsalvaged areas. microsites and lowest on mounds ( Figure 7 ) The CCA of environmental conditions and while crown and Vitis microsites exhibited inter- groundlayer vegetation revealed substantial sepa- mediate density levels; these differences among SALVAGING AFTER WIND DISTURBANCE 369 Downloaded from https://academic.oup.com/forestry/article/81/3/361/655733 by guest on 30 September 2021

Figure 6 . Groundlayer species richness by (A) treat- ment and (B) focal microsite. Least square mean ± standard error. Different letters denote signifi cant differences at level α = 0.05.

ness did not differ between treatments (Figure 8), nor was there a signifi cant treatment × microsite interaction. Figure 5 . Per cent cover of groundlayer vegetation The microsite quadrats contained seedlings of by (A) treatment, (B) focal microsite in salvaged 28 tree species. There were higher percentages of areas and (C) focal microsite in unsalvaged areas. shade-tolerant seedlings (29.6 per cent vs 23.8 Least square mean ± standard error. Different letters per cent) and intolerant seedlings (60.9 per cent denote signifi cant differences at level α = 0.05. vs 53.9 per cent) in salvaged areas (G -test; Table 2) than in unsalvaged areas, whereas there were more intermediates in unsalvaged areas (22.6 per microsites were highly signifi cant. Seedling den- cent vs 9.4 per cent). Several shade-intolerant sity did not differ between treatments, and there species such as P. taeda , Quercus stellata , Betula was no treatment × microsite interaction. Tree nigra and Ulmus alata were entirely absent from seedling species richness was highest in bare soil microsites in unsalvaged areas. Oaks, particu- and pit microsites and lowest on mounds (Figure larly Quercus alba and Quercus velutina , ac- 8 ), and crown and Vitis having intermediate spe- counted for most mid-tolerant species (Table 2), cies richness; these among-microsite differences but since oaks accounted for a higher propor- were again highly signifi cant. Tree seedling rich- tion of pre-disturbance canopy in the unsalvaged 370 FORESTRY Downloaded from https://academic.oup.com/forestry/article/81/3/361/655733 by guest on 30 September 2021

Figure 7 . Tree seedling density by (A) treatment and Figure 8 . Tree seedling species richness by (A) treat- (B) focal microsite. Least square mean ± standard ment and (B) focal microsite. Least square mean ± error. Different letters denote signifi cant differences standard error. Different letters denote signifi cant at level α = 0.05. differences at level α = 0.05. areas, we cannot attribute this fi nding solely to common species, Q. alba and S. albidum were salvaging. most abundant in unsalvaged areas, whereas G -tests of shade-tolerance categories revealed A. rubrum and L. tulipifera were most abundant that proportions of seedlings in these categories in salvaged areas (signifi cant G -test; Table 3). differed signifi cantly among microsites. Mounds Pinus taeda, Q. stellata, L. tulipifera and Liquid- had the highest percentage of shade-intolerant ambar styracifl ua were often found in skid trails seedlings (65.5 per cent) followed by bare (63 per (A. D Leach, personal observation). cent) and pit (53.5 per cent) microsites (Table 2). In pit microsites, L. tulipifera (51.7 per cent of Crowns had the highest percentage of seedlings all seedlings) and A. rubrum (42.5 per cent) seed- intermediate in shade tolerance (29.2 per cent) lings were the most abundant while Q. alba (42.5 followed by bare soil microsites (21.2 per cent). per cent) was the most abundant within crown Pits (44.6 per cent) and mounds (29.1 per cent) microsites, S. albidum (41.9 per cent) within Vitis had the highest percentage cover by shade-tolerant microsites and P. serotina (27.6 per cent) within seedlings. bare soil microsites. The abundances of these fi ve Liriodendron tulipifera accounted for 16.1 individual species differed signifi cantly among per cent of total seedling density (Table 3) fol- microsites (G -test, Table 3 ). lowed by Q. alba (13.4 per cent), S. albidum The CCA of seedling and environmental data (12.2 per cent), Acer rubrum (12.0 per cent) (data not shown) revealed strong separation of and Prunus serotina (11.2 per cent). Of these quadrats by treatment, but the total variance SALVAGING AFTER WIND DISTURBANCE 371

Table 2 : Frequencies of tree seedlings in different shade-tolerance categories, within treatments and microsite types

Tolerant Intermediate Intolerant TotalG -tests

(A) Treatment Unsalvaged 60 57 135 252 Salvaged 69 22 142 233 Total 129 79 277 485G 2 = 16.1, df = 2, P = 0.0003 Downloaded from https://academic.oup.com/forestry/article/81/3/361/655733 by guest on 30 September 2021 (B) Microsite Bare 29 39 116 184 Crown 18 21 33 72 Mound 16 3 36 55 Pit 45 2 54 101 Vitis 21 14 38 73 Total 129 79 277 485G 2 = 58.62, df = 8, P = < 0.0001 df = Degrees of freedom.

Table 3 : Frequencies of the fi ve most common tree seedling species within microsite quadrats

LITU QUAL SAAL ACRU PRSE TotalG -tests

(A) Treatment Unsalvaged 36 49 46 14 45 190 Salvaged 43 17 14 45 10 129 Total 79 66 60 59 55 319G 2 = 64.3, df = 4, P < 0.0001 (B) Microsite Bare 32 33 23 17 40 145 Crown 3 17 10 3 7 40 Mound 12 3 8 6 2 31 Pit 31 1 1 24 3 60 Vitis 112189 343 Total 79 66 60 59 55 319G 2 = 132.16, df = 16, P = < 0.0001 Abbreviations are fi rst two letters of genus and species name (Liriodendron tulipifera , Quercus alba , Sassafras albidum , Acer rubrum , and Prunus serotina ); df = degrees of freedom. explained was again very small. Axis 1 was heav- Discussion and conclusions ily infl uenced by canopy openness ( r = 0.926). Quadrats in salvaged areas containing light- In this study, we documented greater microsite rich- demanding species, such as Quercus coccinea , ness in salvaged areas compared with unsalvaged Q. stellata and P. taeda , were separated from un- windthrow areas, along with signifi cant changes in salvaged quadrats with lower canopy openness microsite relative abundances and microsite envi- and dominance by more shade-tolerant species, ronments among treatments. However, vegetation such as C. fl orida , Fagus grandifolia and N. syl- did not signifi cantly respond to salvage activities vatica . Nevertheless, the eigenvalue for the fi rst in this study 2 years after the disturbance. Indeed, axis was again small at only 0.295, accounting for we found greater variation among microsites than 1.8 per cent of the variance in the species data, so between treatments for both environmental con- we make no further interpretation of this result. ditions and regeneration vegetation. We are not 372 FORESTRY aware of any previous studies that have reported primary factor affecting differential establish- (unconfounded) effects of salvage logging in wind- ment was higher moisture levels in pits, and this damaged forests, although a number of previous may also be true for pits at NTSF. studies document microsite effects (Nakashizuka, In contrast to our fi ndings, Nakashizuka (1989) 1989; Peterson et al. , 1990; Clinton and Baker, found that seedling richness and diversity were 2000; Harrington and Bluhm, 2001). In a related highest on mounds in old-growth temperate for- paper (Peterson and Leach, 2008), we report what ests in Japan because mounds provided competi- we believe to be the fi rst attempt to quantitatively tion-free growing space for tree species with small, relate combined (natural blowdown + anthropo- wind-dispersed seeds. The raised, mineral seedbed Downloaded from https://academic.oup.com/forestry/article/81/3/361/655733 by guest on 30 September 2021 genic salvaging) disturbance severity to vegetation condition of mounds is suitable for germination response at the plot level. of these light-seeded species (Denny and Good- It is not surprising that environmental condi- lett, 1956 ; Lyford and MacLean, 1966 ; Peterson tions within individual microsites differ, due to and Campbell, 1993 ; Oliver and Larson, 1996) their unique modes of formation and topography. and may provide protection from mammalian her- Peterson et al. (1990) found that soil moisture bivory ( Long et al. , 1998; Krueger and Peterson and light levels differed between intact soil and 2006). Additionally, pits may be subject to litter pit microsites. We found signifi cant microsite ef- and water accumulation as well as soil deposition fects for soil temperature and moisture; treefall from mound erosion ( Putz, 1983 ). In cooler and/or mounds were signifi cantly warmer and drier than wetter climate zones, harsh environmental condi- pits, which were cooler and moister ( Figure 4). tions on mounds may be ameliorated. However, In contrast to the present study, though, Peterson the harsh environmental conditions coupled with and Pickett (1995) found that light and tempera- high herbaceous cover on mounds may have led ture were greatest in open (=bare: intact soil, no to lowered seedling density and species richness at litter or vegetation) and mound microsites but that NTSF. moisture did not differ. We suspect that the lesser Although none of the vegetation characteristics wind damage severity and therefore lesser canopy differed signifi cantly between treatments, herba- openness in the present study, compared with ceous cover and richness and tree seedling den- Peterson and Pickett (1995) , may have limited the sity and richness differed between microsite types among-microsite differences that have been docu- ( Figures 5– 8 ). Bare soil and pit microsites exhib- mented in more severely wind-damaged sites. ited high herbaceous richness and tree seedling Most studies of wind disturbance that con- density and richness but low herbaceous cover. sider microsites have focussed on differential Therefore, bare soil and pit microsites may be seedling establishment and survival between pits instrumental to maintaining species diversity of and mounds. The pits and mounds created by both the seedling and herbaceous layers. uprooted trees often increase or maintain tree Vitis microsites supported high levels of herba- species diversity by supporting species that are ceous cover but low tree seedling density, perhaps rare in undamaged forest ( Peterson et al. , 1990) because V. rotundifolia shaded or outcompeted and may be important in determining future spe- tree seedlings. Like Vitis patches, mounds in cies composition. We found that pits had greater unsalvaged areas also had high herb cover but tree seedling density (Figure 7) and species rich- low seedling density. In parallel to our fi ndings, ness (Figure 8) than mounds, but in unsalvaged Nakashizuka (1989) found that undisturbed soil areas, mounds had higher herbaceous cover (Fig- patches beneath gaps (equivalent to ‘ bare’ in this ure 5). Herbaceous species richness did not dif- study) did not contribute to maintenance of pio- fer between pits and mounds (Figure 6 ). Peterson neer species, possibly due to inhibition by dwarf and Pickett (1990) found that pits had greater bamboo. Similarly, the fern understory in New species richness and diversity, total biomass and England forests has been shown to alter biotic and stem abundance per square metre than mounds abiotic conditions, thereby decreasing tree seed- because mounds are more susceptible to ero- ling emergence and/or survival (Horsley, 1993; sion and are subject to more extreme tempera- George and Bazzaz, 1999). Similar mechanisms tures (Beatty, 1984 ; Carlton and Bazzaz, 1998a). may be operating within Vitis and mound micro- Peterson and Pickett (1990) proposed that the sites at NTSF. SALVAGING AFTER WIND DISTURBANCE 373

Thickets of Prunus spp. and Rubus spp. are factors should be interpreted cautiously. Other often common following wind disturbance in factors that were not measured, such as seed eastern North America, due to buried pools of availability, nutrient availability or soil organic viable seeds, and can act to accelerate succes- matter, may be more important. sion if shade-intolerant species are unable to es- A number of studies (reviewed in McIver and tablish ( Peterson and Carson, 1996 ). However, Starr, 2001) led us to expect higher soil tempera- high cover by herbaceous plants did not lead to tures, lower soil moisture and higher light levels higher proportions of shade-tolerant species in in the more open, salvaged stands since selective Vitis patches and mounds. Instead, while shade- logging and the use of heavy machinery removes Downloaded from https://academic.oup.com/forestry/article/81/3/361/655733 by guest on 30 September 2021 intolerant seedlings (Table 2) were most abun- woody biomass from disturbed areas. Indeed, we dant within all microsite types, pits had the most found that the temperature and canopy openness shade-tolerant seedlings (44.6 per cent). Because were signifi cantly higher in salvaged areas but pits and mounds contain high proportions of both treatment did not affect soil moisture. Most of shade-tolerant and shade-intolerant species, they the studies reviewed in McIver and Starr (2001) may be important for maintaining or increasing appear to have had more severe disturbance from species diversity. fi re than in our study, in addition to more thor- Treefall pits and mounds contained very few ough salvaging. It is likely that such differences intermediate-tolerance seedlings (2 per cent and in combined disturbance severity explain some 5.5 per cent, respectively) because few oaks were of the discrepancies between the reviewed studies present (Tables 2 and 3 ). Mounds may have and our fi ndings. low retention of heavy and/or round seeds (i.e. Overall, microsite type was far more important acorns), perhaps because they roll off the elevated than treatment in infl uencing vegetation parame- microsites (Peterson et al. , 1990; Carlton and ters. It is likely that the wind disturbance at NTSF Bazzaz, 1998a ). Because crown and bare micro- (with or without salvaging) increased diversity of sites contained the highest proportion of seed- the seedling layer by allowing germination and lings intermediate in shade tolerance (primarily survival of shade-intolerant species within dis- Q. alba, Q. rubra and Q. velutina ), crowns and turbance-created microsites. The importance of bare soil patches may be important for maintain- post-windthrow salvaging derives from potential ing oak populations. However, whether this is alteration of environmental conditions (particu- due to decreased mammalian herbivory, moder- larly soil temperature and canopy openness) and ate environmental conditions or increased seed microsite diversity, abundance and composition. input from downed and dying oak crowns is un- Contrary to our expectations, pits and mounds certain. Oaks are not represented in the seedling were not destroyed by salvaging operations and layer in proportion to canopy abundance (data more existed in salvaged areas due to higher lev- not shown) and removing crowns by salvaging els of uprooting in the natural windthrow (A.D. may reduce already low oak regeneration. Leach and C.J. Peterson, unpublished data). Differences in resource levels among microsites However, salvaging operations were of very low may be important if they directly affect vegeta- intensity at NTSF and this generalization may not tion characteristics. That the fi rst three axes of hold following higher severity disturbance and/ the CCAs accounted for only a miniscule por- or higher intensity logging operations. This leads tion of the variation in species composition of the us to suggest that the impacts of salvage logging both herbaceous (Axis 1 = 1.1 per cent, Axis 2 = on forest regeneration potential and diversity are 0.5 per cent, Axis 3 = 0.2 per cent) and seedling likely to vary along a combined severity gradi- (Axis 1 = 1.8 per cent, Axis 2 = 1.3 per cent, Axis ent— one that sums the combined severity of the 3 = 0.4 per cent) layers may suggest that soil tem- natural and the anthropogenic disturbances. This perature, soil moisture and canopy openness did is consistent with some recent conceptual treat- not strongly infl uence species composition. How- ments that predict forest response to combined ever, our characterization of soil temperature and natural and anthropogenic impacts by summing moisture was coarse and may have missed ex- the effects if the impacts occur in quick succes- tremes that could infl uence vegetation; thus, the sion (Frelich and Reich 1999; Roberts, 2004). In apparent limited infl uence of these environmental this perspective, the combined disturbances are 374 FORESTRY unlikely to cause drastic shifts in composition Burns , R.M. and Honkala , B.H. 1990 Silvics of North and diversity if summed severity is low, but the America: 1. Conifers; 2 . Hardwoods Agriculture risk of such deleterious changes will increase as Handbook 654 . U.S. Department of Agriculture, severity increases (Roberts, 2004), perhaps with a Forest Service , Washington, D.C. threshold (e.g. Frelich and Reich, 1999). In some Carlton , G.C. and Bazzaz , F.A. 1998a Regeneration cases, unexpected and undesirable consequences of three sympatric birch species on experimental may result ( Paine et al. , 1998). hurricane blowdown microsites . Ecol. Monogr. 68 , Research documenting microsite and vegeta- 99 – 120 . tion characteristics after wind disturbance and Carlton , G.C. and Bazzaz , F.A. 1998b Resource Downloaded from https://academic.oup.com/forestry/article/81/3/361/655733 by guest on 30 September 2021 both before and after salvage logging, across congruence and forest regeneration following an a range of combined disturbance severities, experimental hurricane blowdown . Ecology . 79 , would allow determination of whether and how 1305 – 1319 . salvaging truly affects microsite abundance, Clinton , B.D. and Baker , C.R. 2000 Catastrophic environmental conditions and vegetation char- windthrow in the southern Appalachians: character- acteristics. istics of pits and mounds and initial vegetation re- sponses . For. Ecol. Manage. 126 , 51 – 60 . Denny , C.S. and Goodlett , J.C. 1956 Microrelief result- Funding ing from fallen trees . United States Geological Sur- vey Prof Paper . 288 , 59 – 66 . USDA Forest Service Southern Research Station and by Donato , D.C. , Fontaine , J.B. , Campbell , J.L. , Robinson , grants from the University of Georgia Department of Plant Biology. W.D. , Kaufman , J.B. and Law , B.E. 2006 Post- wildfi re logging hinders regeneration and increases fi re risk . Science . 311 , 352 . Acknowledgements Elliott , K.J. , Hitchcock , S.L. and Krueger , L. 2002 Veg- We thank Chris Goetz and Roy Ward of Natchez Trace etation response to large scale disturbance in a south- State Park and Forest for allowing use of the fi eld sites. ern Appalachian forest: Hurricane Opal and salvage Paul Stufel, Lisa Krueger and Charles Cowden helped logging . J. Torrey Bot. Soc. 129 , 48 – 59 . with fi eldwork. Lisa Kruse helped with plant identifi ca- Flowers , R.L. , Wiliams , L.D. , Walker , D.D. and Phillips , tion. Rebecca Sharitz, Bruce Haines and Eleanor Par- J.A. 1960 Soil Survey of Henderson County, Tennes- dini provided useful comments on the manuscript. see . Series 1954. No. 9. U.S. Department of Agricul- ture, Soil Conservation Service , Washington, D.C. Floyd , D. 2006 Science, censure, and the reignited de- Confl ict of Interest Statement bate over the 2002 Biscuit fi re . J. For. 104 , 57 . None declared. Foster , D.R. and Orwig , D.A. 2006 Preemptive and salvage harvesting of New England forests: when doing nothing is a viable alternative . Conserv. Biol. 20 , 959 – 970 . References Franklin , J.F. , Lindenmayer , D.B. , Macmahon , J.A. , Baird , B.N. 2006 Comment on “ Post-wildfi re logging Mckee , A. , Magnusson , J. and Perry , D.A. 2000 hinders regeneration and increases fi re risk” . Science . Threads of continuity: ecosystem disturbances, bio- 313 , 615b . logical legacies and ecosystem recovery . Conserv. Beatty , S.W. 1984 Infl uence of microtopography and Biol. Pract. 1 , 8 – 16 . canopy species on spatial patterns of forest under- Frazer , G.W. , Canham , C.D. and Lertzman , K.P. 1999 story plants . Ecology . 65 , 1406 – 1419 . Gap Light Analyzer (GLA): Imaging Software to Beatty , S.W. and Sholes , O.D.V. 1988 Leaf litter ef- Extract Canopy Structure and Gap Light Transmis- fect on plant species composition of deciduous forest sion Indices from True-Colour Fisheye Photographs, treefall pits . Can. J. For. Res. 18 , 553 – 559 . Users Manual and Program Documentation . Simon Braun , E.L. 1950 Deciduous Forests of Eastern North Fraser University , Burnaby, B.C. America . The Blakiston Company , Philadelphia, PA . Frelich , L.E. and Reich , P.B. 1999 Neighborhood ef- Brown , C.L. and Kirkman , L.K. 1990 Trees of Georgia fects, disturbance severity, and community stability and Adjacent States . Timber Press , Portland, OR . in forests . Ecosystems . 2 , 151 – 166 . SALVAGING AFTER WIND DISTURBANCE 375

George , L.O. and Bazzaz , F.A. 1999 The fern under- McIver , J.D. and Starr , L. 2001 A literature review on story as an ecological fi lter: emergence and estab- the environmental effects of postfi re logging . West. J. lishment of canopy-tree seedlings . Ecology . 80 , Appl. For. 16 , 159 – 168 . 833 – 845 . Nakashizuka , T. 1989 Role of uprooting in composi- Greene , D.F. , Gauthier , S. , Noel , J. , Rousseau , M. and tion and dynamics of an old-growth forest in Japan . Bergeron , Y. 2006 A fi eld experiment to determine Ecology . 70 , 1273 – 1278 . the effect of post-fi re salvage on seedbeds and tree National Oceanic and Atmospheric Administration regeneration . Front. Ecol. Environ. 4 , 69 – 74 . (NOAA) 1999 Storm data . 41 ( 5 ), National Climatic Harrington , T.B. and Bluhm , A.A. 2001 Tree regenera- Data Center , Asheville, NC . Downloaded from https://academic.oup.com/forestry/article/81/3/361/655733 by guest on 30 September 2021 tion responses to microsite characteristics following Newton , M. , Fitzgerald , S. , Rose , R.R. , Adams , P.W. , a severe tornado in the Georgia Piedmont, USA . For. Tesch , S.D. and Sessions , J. 2006 Comment on “ Post- Ecol. Manage. 140 , 265 – 275 . wildfi re logging hinders regeneration and increases Horsley , S.B. 1993 Mechanisms of interference between fi re risk ” . Science . 313 , 615a . hay-scented fern and black cherry . Can. J. For. Res. Oliver , C.D. and Larson , B.C. 1996 Forest Stand Dy- 23 , 2059 – 2069 . namics . Wiley , New York, NY . Krueger , L.M. and Peterson , C.J. 2006 Effects of Paine , R.T. , Tegner , M.J. and Johnson , E.A. 1998 white-tailed deer on Tsuga canadensis regeneration: Compounded perturbations yield ecological sur- evidence of microsites as refugia from browsing . Am. prises . Ecosystems . 1 , 535 – 545 . Midl. Nat. 89 , 151 – 159 . Palmer , M.W. 1993 Putting things in even better Kupfer , J.A. and Franklin , S.B. 2000 Evaluation of an order— the advantages of canonical correspondence ecological land type classifi cation system, Natchez analysis . Ecology . 74 , 2215 – 2230 . Trace State Forest, western Tennessee USA . Landsc. Peterson , C.J. and Campbell , J.E. 1993 Microsite dif- Urban Plan. 49 , 179 – 190 . ferences and temporal change in plant communities Kuuluvainen , T. and Juntunen , P. 1998 Seedling estab- of treefall pits and mounds in an old-growth forest . lishment in relation to microhabitat variation in a Bull. Torrey Bot. Club . 120 , 451 – 460 . windthrow gap in a boreal Pinus sylvestris forest . J. Peterson , C.J. and Carson , W.P. 1996 Generalizing for- Veg. Sci. 9 , 551 – 562 . est regeneration models: the dependence of propa- Lindenmayer , D.B. 2006 Salvage harvesting — past les- gule availability on disturbance history and stand sons and future issues . For. Chron. 82 , 48 – 53 . size . Can. J. For. Res. 26 , 45 – 52 . Lindenmayer , D.B. , Foster , D.R. , Franklin , J.F. , Hunter , Peterson , C.J. , Carson , W.P. , Mccarthy , B.C. and M.L. , Noss , R.F. and Schmiegelow , F.A. 2004 Sal- Pickett , S.T.A. 1990 Microsite variation and soil vage harvesting policies after natural disturbance . dynamics within newly created treefall pits and Science . 303 , 1303 . mounds . Oikos . 58 , 39 – 46 . Lindenmayer , D.B. and Noss , R.F. 2006 Salvage log- Peterson , C.J. and Leach , A.D. 2008 Limited salvage ging, ecosystem processes, and biodiversity conserva- logging effects on forest regeneration after moderate- tion . Conserv. Biol. 20 , 949 – 958 . severity windthrow . Ecol. Appl. 18 , . Long , Z.T. , Carson , W.P. and Peterson , C.J. 1998 Can Peterson , C.J. and Pickett , S.T.A. 1990 Microsite disturbance create refugia from herbivores: an exam- and elevational Infl uences on early forest regenera- ple with hemlock regeneration on treefall mounds? . tion after catastrophic windthrow . J. Veg. Sci. 1 , J. Torrey Bot. Soc. 125 , 165 – 168 . 657 – 662 . Lyford , W.H. and MacLean , D.W. 1966 Mound and Peterson , C.J. and Pickett , S.T.A. 1995 Forest reorga- pit microrelief in relation to soil disturbance and tree nization— a case-study in an old-growth forest cata- distribution in New Brunswick, Canada . Harvard strophic blowdown . Ecology . 76 , 763 – 774 . Forest Paper 15 . Peterson , C.J. and Pickett , S.T.A. 2000 Patch type Macdonald , S.E. 2007 Effects of partial post-fi re sal- infl uences on regeneration in a western Pennsyl- vage harvesting on vegetation communities in the vania, USA, catastrophic windthrow . Oikos . 90 , boreal mixedwood forest region of northeastern Al- 489 – 500 . berta, Canada . For. Ecol. Manage. 239 , 21 – 31 . Putz , F.E. 1983 Treefall pits and mounds, buried seeds, McCune , B. and Mefford , M.J. 1999 Multivariate and the importance of soil disturbance to pioneer analysis of ecological data . MjM Software , Gleneden trees on Barro Colorado Island, Panama . Ecology . Beach, OR . 64 , 1069 – 1074 . 376 FORESTRY

Roberts , M.R. 2004 Response of the herbaceous layer Resort Park, and Wildlife Management Area in to natural disturbance in North American forests . West Tennessee . General Technical Report SO-85. Can. J. Bot. 82 , 1273 – 1283 . U.S.D.A. Forest Service , New Orleans, LA . Robison , S.A. and McCarthy , B.C. 1999 Potential Sokal , R.R. and Rohlf , F.J. 1995 Biometry . 3rd edition . factors affecting the estimation of light availability W.H. Freeman and Company , New York, NY . using hemispherical photography in oak forest un- USDA and NRCS 2002 The PLANTS Database, Ver- derstories . J. Torrey Bot. Soc. 126 , 344 – 349 . sion 3.5 [online] . National Plant Data Center , Baton Rumbaitis Del Rio , C.M. 2006 Changes in understory Rouge, LA . http://plants.usda.gov(accessed on 21 composition following catastrophic windthrow and April, 2003) . Downloaded from https://academic.oup.com/forestry/article/81/3/361/655733 by guest on 30 September 2021 salvage logging in a subalpine forest ecosystem . Can. Webb , S.L. and Scanga , S.E. 2001 Windstorm dis- J. For. Res. 36 , 2943 – 2954 . turbance without patch dynamics: twelve years SAS Institute 1995 SAS/STAT User’s Guide . Cary , NC . of change in a Minnesota forest . Ecology . 82 , Release 6.03 Edition . 893 – 898 . Smalley , G.W. 1991 Classifi cation and Evaluation of Forest Sites on the Natchez Trace State Forest, State Received 25 August 2007