Nitrogen cycling following mountain pine beetle disturbance in lodgepole pine forests of Greater Yellowstone

Widespread bark beetle outbreaks are currently affecting multiple conifer forest types throughout western North America, yet many ecosystem-level consequences of this disturbance are poorly understood. We quantified the effect of mountain pine beetle (Dendroctonus ponderosae) outbreak on nitrogen (N) cycling through litter, soil, and vegetation in lodgepole pine (Pinus contorta var. latifolia) forests of the Greater Yellowstone Ecosystem (WY, USA) across a 0-30 year chronosequence of time-since-beetle disturbance. Recent (1-4 years) bark beetle disturbance increased total litter depth and N concentration in needle litter relative to undisturbed stands, and soils in recently disturbed stands were cooler with greater rates of net N mineralization and nitrification than undisturbed sites. Thirty years after beetle outbreak, needle litter N concentration remained elevated; however total litter N concentration, total litter mass, and soil N pools and fluxes were not different from undisturbed stands. Canopy N pool size declined 58% in recent outbreaks, and remained 48% lower than undisturbed in 30-year old outbreaks. Foliar N concentrations in unattacked lodgepole pine trees and an understory sedge were positively correlated with net N mineralization in soils across the chronosequence. Bark beetle disturbance altered N cycling through the litter, soil, and vegetation of lodgepole pine forests, but changes in soil N cycling were less severe than those observed following stand replacing fire. Several lines of evidence suggest the potential for N leaching is low following bark beetle disturbance in lodgepole pine.     

Risk rating for mountain pine beetle infestation of lodgepole pine forests over large areas with ordinal regression modelling

The mountain pine beetle Dendroctonus ponderosae Hopkins is endemic to lodgepole pine, Pinus contorta var. latifolia Engelmann, forests in western Canada. However, the current beetle epidemic in this area highlights the challenges faced by forest managers tasked with prioritizing stands for mitigation activities such as salvage harvesting and direct control methods. In western Canada, the operational risk rating system for mountain pine beetle is based on biological knowledge gained from a rich legacy of stand-scale field studies. Due to the large spatial (millions of hectares affected) and temporal (over 10 years) extents of the current epidemic, new research into large-area mountain pine beetle processes has revealed further insights into the landscape-scale characteristics of beetle infested forests. In this paper, we evaluated the potential for this new knowledge to augment an established system for rating the short-term risk of tree mortality in a stand due to mountain pine beetle. New variables explored for utility in risk rating include direct shortwave radiation, site index, diameter at breast height, the temporal trends in local beetle populations, Biogeoclimatic Ecosystem Classification and beetle–host interaction variables. Proportional odds ordinal regression was used to develop a model for the Vanderhoof Forest District in west-central British Columbia. Prediction on independent data was assessed with the area under the receiver operator curve (AUC), indicating good discriminatory power (AUC = 0.84) for predicting levels of mountain pine beetle-caused pine mortality.     

Susceptibility of ponderosa pine, Pinus ponderosa (Dougl. ex Laws.), to mountain pine beetle, Dendroctonus ponderosae Hopkins,attack in uneven-aged stands in the Black Hills of South Dakota and Wyoming USA

Mountain pine beetle, Dendroctonus ponderosae Hopkins can cause extensive tree mortality in ponderosa pine, Pinus ponderosa Dougl. ex Laws., forests in the Black Hills of South Dakota and Wyoming. Most studies that have examined stand susceptibility to mountain pine beetle have been conducted in even-aged stands. Land managers increasingly practice uneven-aged management. We established 84 clusters of four plots, one where bark beetle-caused mortality was present and three uninfested plots. For all plot trees we recorded species, tree diameter, and crown position and for ponderosa pine whether they were killed or infested by mountain pine beetle. Elevation, slope, and aspect were also recorded. We used classification trees to model the likelihood of bark beetle attack based on plot and site variables. The probability of individual tree attack within the infested plots was estimated using logistic regression. Basal area of ponderosa pine in trees ≥25.4 cm in diameter at breast height (dbh) and ponderosa pine stand density index were correlated with mountain pine beetle attack. Regression trees and linear regression indicated that the amount of observed tree mortality was associated with initial ponderosa pine basal area and ponderosa pine stand density index. Infested stands had higher total and ponderosa pine basal area, total and ponderosa pine stand density index, and ponderosa pine basal area in trees ≥25.4 cm dbh. The probability of individual tree attack within infested plots was positively correlated with tree diameter with ponderosa pine stand density index modifying the relationship. A tree of a given size was more likely to be attacked in a denser stand. We conclude that stands with higher ponderosa pine basal area in trees >25.4 cm and ponderosa pine stand density index are correlated with an increased likelihood of mountain pine beetle bark beetle attack. Information form this study will help forest managers in the identification of uneven-aged stands with a higher likelihood of bark beetle attack and expected levels of tree mortality.     

Stand characteristics and downed woody debris accumulations associated with a mountain pine beetle (Dendroctonus ponderosae Hopkins) outbreak in Colorado

Lodgepole pine (Pinus contorta Dougl. ex Loud.)-dominated ecosystems in north-central Colorado are undergoing rapid and drastic changes associated with overstory tree mortality from a current mountain pine beetle (Dendroctonus ponderosae Hopkins) outbreak. To characterize stand characteristics and downed woody debris loads during the first 7 years of the outbreak, 221 plots (0.02 ha) were randomly established in infested and uninfested stands distributed across the Arapaho National Forest, Colorado. Mountain pine beetle initially attacked stands with higher lodgepole pine basal area, and lower density and basal area of Engelmann spruce (Picea engelmannii [Parry]), and subalpine fir (Abies lasiocarpa (Hook.) Nutt. var. lasiocarpa) compared to uninfested plots. Mountain pine beetle-affected stands had reduced total and lodgepole pine stocking and quadratic mean diameter. The density and basal area of live overstory lodgepole declined by 62% and 71% in infested plots, respectively. The mean diameter of live lodgepole pine was 53% lower than pre-outbreak in infested plots. Downed woody debris loads did not differ between uninfested plots and plots currently infested at the time of sampling to 3 or 4–7 years after initial infestation, but the projected downed coarse wood accumulations when 80% of the mountain pine beetle-killed trees fall indicated a fourfold increase. Depth of the litter layer and maximum height of grass and herbaceous vegetation were greater 4–7 years after initial infestation compared to uninfested plots, though understory plant percent cover was not different. Seedling and sapling density of all species combined was higher in uninfested plots but there was no difference between infested and uninfested plots for lodgepole pine alone. For trees ≥2.5 cm in diameter at breast height, the density of live lodgepole pine trees in mountain pine beetle-affected stands was higher than Engelmann spruce, subalpine fir, and aspen, (Populus tremuloides Michx.), in diameter classes comprised of trees from 2.5 cm to 30 cm in diameter, suggesting that lodgepole pine will remain as a dominant overstory tree after the bark beetle outbreak.     

Southern pine beetle regional outbreaks modeled on landscape, climate and infestation history

The southern pine beetle (Dendroctonus frontalis, SPB) is the major insect pest of pine species in the southeastern United States. It attains outbreak population levels sufficient to mass attack host pines across the landscape at scales ranging from a single forest stand to interstate epidemics. This county level analysis selected and examined the best climatic and landscape variables for predicting infestations at regional scales. The analysis showed that, for a given county, the most important factor in predicting outbreaks was that the county was classified as in outbreak status in the previous year. Other important factors included minimum winter temperature and the greatest difference between the average of daily minimums and a subsequent low temperature point, precipitation history either seasonally in the previous year or difference from average over the previous 2 years, the synchronizing effect of seasonal temperatures on beetle populations and the relative percentage of total forest area composed of host species. The statistical models showed that climatic variables are stronger indicators of outbreak likelihood than landscape structure and cover variables. Average climatic conditions were more likely to lead to outbreaks than extreme conditions, supporting the notion of coupling between a native insect and its native host. Still, some extreme events (i.e., periods of very low temperature or very high precipitation) did precede beetle infestation. This analysis suggested that there are predisposing and inciting factors at the large scale but the driving factors leading to individual infestations operate at smaller scales.     

The pollen record of a 20th century spruce beetle (Dendroctonus rufipennis) outbreak in a Colorado subalpine forest,USA

The frequency and intensity of ecosystem disturbance, including outbreaks of forest insects and forest fires, is expected to increase in the future as a result of higher temperatures and prolonged drought. While many studies have concentrated on the future climatic impacts on fire, little is known about the impact of future climate on insect infestation. Paleoecological techniques are important in this regard in identifying the potential relationships between climate and insect outbreaks in the past, as a predictive tool for the future. We examine a high-resolution 20th century record of spruce beetle (Dendroctonus rufipennis) infestation from a small, subalpine lake, comparing the paleoecological record to the historical and tree-ring record of the event. An extensive spruce beetle outbreak occurred in northwestern Colorado during the 1940s and 1950s, causing widespread mortality of mature Picea engelmannii. Pollen analysis of this period documents the decline of Picea and its replacement locally by Abies lasiocarpa, paralleling age and composition studies of modern forest stands in the region. This study is a proof of concept that, when applied to longer sedimentary records, could produce a detailed record of infestation for the Late Holocene or older time periods. This information will be useful to forest managers in efforts to plan for the effects of D. rufipennis infestations, and subsequent succession within high elevation conifer forests.     

Effect of management on the spatial spread of mountain pine beetle (Dendroctonus ponderosae) in Banff National Park

To evaluate control measures, the spread of mountain pine beetles, Dendroctonus ponderosae, in management and monitoring regions in Banff National Park was analyzed for years 1997 to 2004. The Park allowed mountain pine beetles to follow their natural course in a monitoring zone (74,041 ha), whereas in a management zone (45,997 ha) an extensive eradication program was established in 2001 which included baiting mountain pine beetles and cutting and burning all colonized trees. We used data collected from an annual aerial survey and the geo-referenced location of trees that were cut and removed to assess if the area colonized and the spatial extent of mountain pine beetles differed between the two zones. After 4 years, the control measures did not significantly affect the area colonized by mountain pine beetles, and in 2004 mountain pine beetles were still expanding in both zones, although at a slow rate (1.23 ha per year). The spatial extent and the rate at which 500 m × 500 m cells (25 ha) were colonized were much reduced in the management zone. Thus, the management program appeared to have reduced the success of long distance movement as measured by colonizing new 25 ha cells, but short distance dispersal remained successful as indicated by the continued increase in area colonized. The effectiveness of control measures was probably limited by the number of colonized trees that were missed and by survival rates determined by other untested factors.     

Woodpeckers increase in abundance but maintain fecundity in response to an outbreak of mountain pine bark beetles

Many temperate woodpecker species are thought to be highly conservative in their fecundity with little response to fluctuations in availability of resources. In a 15-year field study in interior British Columbia, we evaluated responses in abundance and fecundity of six species of resident and migrant woodpeckers (downy woodpecker [Picoides pubescens], hairy woodpecker [Picoides villosus], American three-toed woodpecker [Picoides dorsalis], pileated woodpecker [Drycopus pileatus], northern flicker [Colaptes auratus], and red-naped sapsucker [Sphyrapicus nuchalis]) to a large-scale outbreak of mountain pine bark beetles that resulted in a strong positive pulse in food supply. Population densities of woodpecker species increased during outbreak years. Despite the year-round multi-annual increase in food resources, and in contrast to the strong increases in fecundity shown by nuthatches and chickadees, annual fecundity (as indicated by clutch size and number of nestlings that fledged) did not change for any woodpecker species over the study. Similarly, we found no changes in fecundity in response to selective forest harvesting despite numerical increases for woodpeckers at these sites. Our study confirms that these woodpecker species are conservative in their reproductive investment patterns even during strong multi-annual increases in food. Our findings indicate woodpecker populations are regulated numerically through variable survival and/or greater immigration rates, which can result in higher breeding densities temporarily during resource pulses.     

Multi-scale nest-site selection by black-backed woodpeckers in outbreaks of mountain pine beetles

Areas of mountain pine beetle (Dendroctonus ponderosae Hopkins) outbreaks in the Black Hills can provide habitat for black-backed woodpeckers (Picoides arcticus), a U.S. Forest Service, Region 2 Sensitive Species. These outbreaks are managed through removal of trees infested with mountain pine beetles to control mountain pine beetle populations and salvage timber resources. To minimize impacts to black-backed woodpeckers while meeting management objectives, there is a need to identify characteristics of these areas that support black-backed woodpeckers. We examined the habitat associations of this species nesting in areas of beetle outbreaks in the Black Hills, South Dakota in 2004 and 2005. We used an information theoretic approach and discrete choice models to evaluate nest-site selection of 42 woodpecker nests at 3 spatial scales—territory, nest area, and nest tree. At the territory scale (250 m around nest), availability and distribution of food best explained black-backed woodpecker selection of beetle outbreaks versus the surrounding forest. Selection at the territory scale was positively associated with densities of trees currently infested by mountain pine beetles and indices of wood borer (Cerambycidae and Buprestidae) abundance, and was greatest at distances of 50–100 m from the nearest patch of infestation. At the nest-area scale (12.5 m radius around the nest), densities of snags positively influenced nest-area selection. Finally, at the nest-tree scale, aspen (Populus tremuloides) and 3–5-year-old ponderosa pine (Pinus ponderosa) snags were important resources. The association between abundant wood-boring insects and black-backed woodpeckers creates a difficult challenge for forest managers. In the absence of fire, areas of beetle outbreak might serve as the only substantial source of habitat in the Black Hills. Regulating insect populations via salvage logging will reduce key food resources to black-backed woodpeckers during nesting. Therefore, given the relatively infrequent occurrence of large-scale fire in the Black Hills, management should recognize the importance of beetle-killed forests to the long-term viability of the black-backed woodpecker population in the Black Hills.     

Prescribed fire effects on bark beetle activity and tree mortality in southwestern ponderosa pine forests

Prescribed fire is an important tool in the management of ponderosa pine (Pinus ponderosa Dougl. ex Laws.) forests, yet effects on bark beetle (Coleoptera: Curculionidae, Scolytinae) activity and tree mortality are poorly understood in the southwestern U.S. We compared bark beetle attacks and tree mortality between paired prescribed-burned and unburned stands at each of four sites in Arizona and New Mexico for three growing seasons after burning (2004–2006). Prescribed burns increased bark beetle attacks on ponderosa pine over the first three post-fire years from 1.5 to 13% of all trees, increased successful, lethal attacks on ponderosa pine from 0.4 to 7.6%, increased mortality of ponderosa pine from all causes from 0.6 to 8.4%, and increased mortality of all tree species with diameter at breast height >13 cm from 0.6 to 9.6%. On a per year basis, prescribed burns increased ponderosa pine mortality from 0.2% per year in unburned stands to 2.8% per year in burned stands. Mortality of ponderosa pine 3 years after burning was best described by a logistic regression model with total crown damage (crown scorch + crown consumption) and bark beetle attack rating (no, partial, or mass attack by bark beetles) as independent variables. Attacks by Dendroctonus spp. did not differ significantly over bole heights, whereas attacks by Ips spp. were greater on the upper bole compared with the lower bole. Three previously published logistic regression models of tree mortality, developed from fires in 1995–1996 in northern Arizona, were moderately successful in predicting broad patterns of tree mortality in our data. The influence of bark beetle attack rating on tree mortality was stronger for our data than for data from the 1995–1996 fires. Our results highlight canopy damage from fire as a strong and consistent predictor of post-fire mortality of ponderosa pine, and bark beetle attacks and bole char rating as less consistent predictors because of temporal variability in their relationship to mortality. The small increase in tree mortality and bark beetle attacks caused by prescribed burning should be acceptable to many forest managers and the public given the resulting reduction in surface fuel and risk of severe wildfire.     

Forest thinning and subsequent bark beetle-caused mortality in Northeastern California

The Warner Mountains of northeastern California on the Modoc National Forest experienced a high incidence of tree mortality (2001–2007) that was associated with drought and bark beetle (Coleoptera: Curculionidae, Scolytinae) attack. Various silvicultural thinning treatments were implemented prior to this period of tree mortality to reduce stand density and increase residual tree growth and vigor. Our study: (1) compared bark beetle-caused conifer mortality in forested areas thinned from 1985 to 1998 to similar, non-thinned areas and (2) identified site, stand and individual tree characteristics associated with conifer mortality. We sampled ponderosa pine (Pinus ponderosa var ponderosa Dougl. ex Laws.) and Jeffrey pine (Pinus jeffreyi Grev. and Balf.) trees in pre-commercially thinned and non-thinned plantations and ponderosa pine and white fir (Abies concolor var lowiana Gordon) in mixed conifer forests that were commercially thinned, salvage-thinned, and non-thinned. Clusters of five plots (1/50th ha) and four transects (20.1 × 100.6 m) were sampled to estimate stand, site and tree mortality characteristics. A total of 20 pre-commercially thinned and 13 non-thinned plantation plot clusters as well as 20 commercially thinned, 20 salvage-thinned and 20 non-thinned mixed conifer plot clusters were established. Plantation and mixed conifer data were analyzed separately. In ponderosa pine plantations, mountain pine beetle (Dendroctonus ponderosae Hopkins) (MPB) caused greater density of mortality (trees ha⁻¹ killed) in non-thinned (median 16.1 trees ha⁻¹) compared to the pre-commercially thinned (1.2 trees ha⁻¹) stands. Percent mortality (trees ha⁻¹ killed/trees ha⁻¹ host available) was less in the pre-commercially thinned (median 0.5%) compared to the non-thinned (5.0%) plantation stands. In mixed conifer areas, fir engraver beetles (Scolytus ventralis LeConte) (FEN) caused greater density of white fir mortality in non-thinned (least square mean 44.5 trees ha⁻¹) compared to the commercially thinned (23.8 trees ha⁻¹) and salvage-thinned stands (16.4 trees ha⁻¹). Percent mortality did not differ between commercially thinned (least square mean 12.6%), salvage-thinned (11.0%), and non-thinned (13.1%) mixed conifer stands. Thus, FEN-caused mortality occurred in direct proportion to the density of available white fir. In plantations, density of MPB-caused mortality was associated with treatment and tree density of all species. In mixed conifer areas, density of FEN-caused mortality had a positive association with white fir density and a curvilinear association with elevation.     

The role of germination microsite in the establishment of sugar pine and Jeffrey pine seedlings

Mature shrubs can provide microhabitats that are beneficial to tree seedling growth and development. Sugar pine trees (Pinus lambertiana) grow in a narrow zone on the eastern slope of the Carson Range in extreme western Nevada, whereas Jeffrey pine (Pinus jeffreyi) is the dominant tree species in the region, an area extensively disturbed by wild fire. This study compares seedling establishment of sugar pine and Jeffrey pine relative to mature shrubs. In the fall of 2002 (cohort 1) and 2003 (cohort 2), 13,600 seeds of both species were planted in wire mesh enclosures, at three sites, under a variety of microhabitat treatments: under shade and in the open, under two species of shrub cover, and with and without plant litter. Seedlings were monitored for survival through two growing seasons. Even though more sugar pine seedlings emerged, more Jeffrey pine seedlings survived, and Jeffrey pine was the more drought tolerant species, better suited for the xeric climate found in the Carson Range. Litter slightly hindered seedling emergence but had no effect on survival and there was no significant species × litter interaction. Supplemental water facilitated survival in all treatments with highest survival in shade treatments. Sugar pine seedlings showed a significant increase in survival over Jeffrey pine seedlings with the addition of water, particularly in open treatments and more of both species survived under manzanita shrubs with water. The highest seedling mortality occurred when shrub canopy was removed, and seedlings experienced the effect of full sun and competition for soil water. For either species, microhabitat is a significant factor in determining success or failure in rehabilitation efforts after disturbance.     

Predisposition to bark beetle attack by root herbivores and associated pathogens: Roles in forest decline,gap formation,and persistence of endemic bark beetle populations

Bark beetles are largely known for their ability to undergo intermittent population eruptions that transform entire landscapes and pose significant economic hardships. However, most species do not undergo outbreaks, and eruptive species usually exert only minor disturbances. Understanding the dynamics of tree-killing noneruptive species can provide insights into how beetles persist at low densities, and how some spatiotemporal patterns of host predisposition may more likely favor breaching eruptive thresholds than others. Elucidating mechanisms behind low-density populations is challenging, however, due to the requirement of long-term monitoring and high degrees of spatial and temporal covariance. We censused more than 2700 trees annually over 7 years, and at the end of 17 years, in a mature red pine plantation. Trees were measured for the presence of bark beetles and wood borers that breed within the primary stem, root weevils that breed in root collars, and bark beetles that breed in basal stems. We quantify the sequence of events that drive this decline syndrome, with the primary emergent pattern being an interaction between below- and above-ground herbivores and their fungal symbionts. This interaction results in an expanding forest gap, with subsequent colonization by early-successional vegetation. Spatial position strongly affects the likelihood of tree mortality. A red pine is initially very likely to avoid attack by tree-killing Ips beetles, but attack becomes increasingly likely as the belowground complex spreads to neighboring trees and eventually make trees susceptible. This system is largely internally driven, as there are strong gap edge, but not stand-edge, effects. Additional stressors, such as drought, can provide an intermittent source of susceptible trees to Ips beetles, and elevated temperature slightly accentuates this effect. New gaps can arise from such trees as they subsequently become epicenters for the full complex of organisms associated with this decline, but this is not common. As Ips populations rise, there is some element of positive feedback, in that the proportion of killed trees that were not first colonized by root organisms increases. This positive feedback is very weak, however, and we propose the slope between beetle population density and reliance on host stress as a quantitative distinction along a gradient from noneruptive through eruptive species. Almost all trees colonized by Ips were subsequently colonized by wood borers, likely a source of negative feedback. We discuss implications to our overall understanding of cross-scale interactions, between-guild interactions, forest declines, and eruptive thresholds.     

Climate and historical stand dynamics in the tropical pine forests of northern Thailand

Forest recruitment is the outcome of local- and regional-scale factors such as disturbances and climate. The relative importance of local- and regional-scale factors will determine the spatial scale at which temporal pulses of recruitment occur. In seasonal tropical forests, where the annual dry-season is a critical bottleneck to seedling survival, multi-year periods of relatively cool, wet dry seasons may be required for successful tree recruitment. Consequently, when such conditions are present, region-wide synchronisation of recruitment may occur. To examine the case for regional synchronisation of forest dynamics in the seasonal tropical pine forests of northern Thailand, we investigated forest age structures at three spatial scales: stand, site and region. We compared forest age structures with instrumental climatic records beginning in 1902. We found significant statistical evidence of synchronous recruitment at the stand- and site-scales, but not at the regional-scale. While correlations between recruitment and climate were not statistically significant, recruitment success was often linked to favourable climatic conditions. For example, recruitment at all sites was associated with multi-year periods of cool-wet dry seasons. The lack of significant correlations between recruitment and climate appears to reflect complex interactions among local disturbance history, regional climate variability and pine recruitment.     

Potential effects of climate change on the growth of lodgepole pine across diameter size classes and ecological regions

We examined 65 lodgepole pine (Pinus contorta Dougl. ex Loud. var. latifolia Engelm.) sites in Alberta using a dendrochronological approach in order to examine the relationships between climate and growth of lodgepole pine across elevational ecoregions and diameter size classes. The 4 elevational ecoregions sampled included the Boreal Highlands (BH: 13 sites); the Foothills (FH: 36 sites); a grouping of the montane and subalpine zones of the southern Rocky Mountains (RM: 12 sites); and the montane zone of the Cypress Hills (CH: 4 sites). The first diameter size class was comprised of the three largest (top) diameter trees at each site. The tree list of each plot was ranked, then divided into three groups of equal basal area and the tree at the midpoint of each group (small, medium, and large) was selected for the other three size classes. Annual growth in basal area of lodgepole pine was generally sensitive to heat and moisture stress in late summer of the previous year, the degree of winter harshness, and the timing of the start of the growing season. Growth was inhibited by low temperature in all winter months at the most northern BH sites which had the coldest winters, but this effect was interrupted in some of the midwinter months in the more southerly sites in the RM, and we postulated this was due to the damaging influences of Chinook winds. Interannual growth patterns were strongly correlated between top diameter trees and the other classes, and trees of all diameter classes generally responded to climate in the same way, which indicated that it is sufficient to sample only the largest diameter trees in a stand to provide insight into growth–climate relationships. The forecasted growth estimates indicate that future climate warming will negatively impact the productivity of lodgepole pine in the FH, the heart of lodgepole distribution and productivity in Alberta.     

Management strategies for a large-scale mountain pine beetle outbreak: Modelling impacts on American martens

We assessed the effectiveness of alternative forest management strategies for maintaining American martens (Martes americana) in a sub-boreal landscape subject to an extensive mountain pine beetle (Dendroctonus ponderosae) outbreak, associated salvage logging, and climate change in north-central British Columbia, Canada. We structured the analysis in a Bayesian network (BN) meta-model that incorporated the results of spatially explicit modelling of landscape conditions (natural and logging disturbance, habitat quality, number of potential territories, and connectedness of territories) with analytical population modelling. The BN meta-model was then used to examine habitat and population size responses (adult females only are presented) to management scenarios, in the context of uncertainty of model parameters, management objectives, and climate change. Status quo management is dominantly clear-cutting with 3–20% of each harvest unit retained as mature patches, with reforestation by planting in the remainder. Management options we examined were: (1) the status quo, (2) varying the total annual timber harvest on the landscape (100%, 80% or 50% of current long-term sustained yield estimates), (3) the protection of understory trees during logging, and (4) 30–70% retention of overstory (partial cutting in distinct patches <1 ha in size) in each harvest unit, for 33% or 50% of the annual timber harvest. We found that marten habitat and population size declined substantively with the beetle outbreak and associated salvage cutting. The choice of management strategy then had a long-term effect on the potential for marten recovery after the beetle outbreak. Partial cutting scenarios had the greatest average long-term marten population levels, followed in order by reduced harvest rates, understory protection, and the status quo. Management scenarios with the best chance of meeting conservation goals without over-protecting habitat (and thus unnecessarily constraining timber management) varied with the population objective chosen. The choice of management strategy will depend on the weighting of marten outcomes against the economic desirability of timber harvest strategies, willingness to gamble on climate change, and the time-frame of interest.     

Canopy disturbance and tree recruitment over two centuries in a managed longleaf pine landscape

Disturbance history was reconstructed across an 11300 ha managed longleaf pine (Pinus palustris Mill.) landscape in southwestern Georgia, USA. Our specific objectives were to: (i) determine forest age structure; (ii) reconstruct disturbance history through the relationship between canopy disturbance, tree recruitment and growth; and (iii) explore the relationship between canopy disturbance and climate. Age structure, canopy disturbance events and initial growth patterns at coring height were examined by randomly sampling 1260 trees in 70 1.3 ha plots. Principal component analysis was used to group plots with similar age structures to gain insight into the dynamics between canopy disturbance and recruitment. Disturbance events were detected by large and rapid increases in radial growth. We tested the following hypothesis to investigate whether these growth increases could have been triggered by improved climatic conditions: precipitation and drought are positively correlated to radial growth releases. Only four stands (comprising <6% of the study area) had an even-aged structure. Further, tree recruitment prior to European settlement indicates that longleaf pine naturally recruited into areas 1.3 ha or less, supporting early-20th century observations that the primary longleaf pine forest was uneven-aged. Contrary to our hypothesis, growing season precipitation and drought was significantly and negatively correlated with canopy disturbance (radial growth releases), which indicates that a reconstruction of disturbance history could proceed with some confidence. Most trees sampled were recruited at coring height from 1910 to 1935. Of the 67 canopy disturbances detected from 1910 to 1935, the average growth release ranged from 139 to 277% per half decade suggesting the occurrence of large canopy disturbances. Rapid initial growth patterns of young trees during these years show evidence of reduced overstory competition and support the detected disturbance intensity. Our reconstruction of stand dynamics is markedly similar to independent records of local oral and written history, which gives an additional set of evidence that the disturbance detection methodology used can be useful in open-canopied forests. Stands with multiple cohorts reveal a mix of continuous minor and major canopy disturbances leading to continual tree recruitment, suggesting their applicability as models for long-term forest management. The significant relationship between climate and disturbance in our data suggests that with the expected warming over the next 100 years, climatic impacts on stand dynamics should be incorporated into long-term longleaf pine forest restoration and management.     

Snow damage in lodgepole pine stands brought into thinning and fertilization regimes

Several heavy wet snowfalls occurred during 2007-2009 across a broad-scale thinning and fertilization experiment to bring overstocked juvenile lodgepole pine (Pinus contorta var. latifolia) in the foothills of Alberta, Canada into an intensive management regime. We examined the bending and breakage of trees in relation to thinning and fertilization and used a multimodel information-theoretic approach to model stand and tree level predictors of snow damage. Fertilized stands suffered the greatest amount of snow damage, and this was most noteworthy when stands were also thinned; here 22% (17% broken stems) of trees were damaged compared to 8% (4% broken stems) in the thinned and unfertilized stands. At the stand level, needle weight and crown cover were reliable predictors of snow damage. At the tree level, separate models were developed for each combination of thinning and fertilization. All models used total tree volume; usually the smaller trees in the stands were more susceptible to damage but in the thinned and fertilized stands larger but slender trees with large asymmetrical crowns tended to be damaged. Also, trees with lower total stem volume were more susceptible to damage. Only in the thinned and fertilized stands were variables related to crown shape and asymmetry important predictors of snow damage. We conclude that snow damage is an important agent for self-thinning in unthinned stands and fertilization tends to exacerbate damage because of increase in foliage size. In areas with regular occurrence of heavy snow, we do not recommend fertilization at the same time as thinning, as the larger and more economically important trees in the stand are at risk.     

Root growth and hydraulic conductivity of southern pine seedlings in response to soil temperature and water availability after planting

Comparison of the root system growth and water transport of southern pine species after planting in different root-zone environments is needed to guide decisions regarding when, and what species to plant. Evaluation of how seed source affects root system responses to soil conditions will allow seed sources to be matched to planting conditions. The root growth and hydraulic conductivity of three sources each of shortleaf, loblolly and longleaf pine seedlings were evaluated for 28 days in a seedling growth system that simulated the planting environment. Across species, an increase in root-zone temperature alleviated limitations to root growth caused by water stress. In the coldest temperature, longleaf pine maintained a higher hydraulic conductivity compared to shortleaf and loblolly pine. Without water limitation, the root growth and hydraulic conductivity of shortleaf and loblolly pine were superior to that of longleaf pine, but as water availability decreased, the root growth of longleaf pine surpassed that of loblolly pine. Hydraulic conductivities of the seed sources differed, and differences were attributed to either new root growth, or an increase in the efficiency of the root system to transport water.     

Radiation use efficiency in adjacent hardwood and pine forests in the southern Appalachians

The efficiency with which trees convert photosynthetically active radiation (PAR) to biomass has been shown to be consistent within stands of an individual species, which is useful for estimating biomass production and carbon accumulation. However, radiation use efficiency (?) has rarely been measured in mixed-species forests, and it is unclear how species diversity may affect the consistency of ?, particularly across environmental gradients. We compared aboveground net primary productivity (ANPP), intercepted photosynthetically active solar radiation (IPAR), and radiation use efficiency (? = ANPP/IPAR) between a mixed deciduous forest and a 50-year-old white pine (Pinus strobus L.) plantation in the southern Appalachian Mountains. Average ANPP was similar in the deciduous forest (11.5 Mg ha⁻¹ y⁻¹) and pine plantation (10.2 Mg ha⁻¹ y⁻¹), while ? was significantly greater in the deciduous forest (1.25 g MJ⁻¹) than in the white pine plantation (0.63 g MJ⁻¹). Our results demonstrate that late-secondary hardwood forests can attain similar ANPP as mature P. strobus plantations in the southern Appalachians, despite substantially less annual IPAR and mineral-nitrogen availability, suggesting greater resource-use efficiency and potential for long-term carbon accumulation in biomass. Along a 260 m elevation gradient within each forest there was not significant variation in ?. Radiation use efficiency may be stable for specific forest types across a range of environmental conditions in the southern Appalachian Mountains, and thus useful for generating estimates of ANPP at the scale of individual watersheds.