Search

1. Boreal forest soils contain large stocks of soil carbon (C) that may be sensitive to changes in climate and disturbance. Destablization of boreal forest soil C through changes in C inputs, belowground C pools and/or wildfire could feedback to accelerate rising atmospheric CO2 concentration. Additionally, increasing frequency of severe fires may be changing the dominant forest types and reshaping aboveground C stocks. 2. Although controls on ecosystem C pools have received considerable attention, many studies have been limited to locations near the road system, leading to uncertainty in current and future C stocks across boreal Alaska. Here, we leveraged 545 randomly selected and spatially balanced Forest Inventory and Analysis (FIA) plots across ~13.5 million hectares in interior Alaska to examine the factors governing soil and live tree C pools. 3. Forest type mediated the effects of mean summer air temperature and the probability of near-surface permafrost on soil C. Meanwhile, forest type, stand age and aspect were the primary drivers of live tree C. Overall, plots with a known history of wildfire during the past 70 years did not have significantly different soil C stocks than plots without a known history of fire, likely due to the historical predominance of low severity fires. 4. Where wildfire likely initiated a transition to deciduous trees (19% of plots), live tree and soil C pools were reduced by 16% and 20%, respectively. Ecosystem C likely recovered over time, as maturing deciduous stands rapidly gained C in live trees. Deciduous stands without a known fire had comparatively very large live tree C stocks, suggesting a significant change in the distribution of ecosystem C following severe fire. 5. Synthesis. Our results highlight the nuanced interactions among wildfire, landscape position and forest type that will play important roles in shaping future boreal ecosystem C stocks. [ABSTRACT FROM AUTHOR]

Long-term experiments provide a way to test presumed causes of successional or environmentally driven vegetation changes. Early-successional nitrogen (N)-fixing plants are widely thought to facilitate productivity and vegetation development on N-poor sites, thus accounting for observed vegetation patterns later in succession. We tested this facilitative impact on vegetation development in a 23-yr field experiment on an Interior Alaska ( USA) floodplain. On three replicate early-successional silt bars, we planted late-successional white spruce ( Picea glauca) seedlings in the presence and absence of planted seedlings of an early-successional N-fixing shrub, thinleaf alder ( Alnus incana). Alder initially facilitated survivorship and growth of white spruce. Within six years, however, after canopy closure, alder negatively affected spruce survivorship and growth. Our three replicate sites followed different successional trajectories. One site was eliminated by erosion and supported no vegetation development during our study. The other two sites, which differed in site moisture, diverged in vegetation composition. Structural equation modeling ( SEM) suggested that, in the drier of these two sites, alder inhibited spruce growth directly (presumably by competition) and indirectly through effects mediated by competition with other woody species. However, at the wetter site, alder had both positive and negative effects on spruce growth, with negative effects predominating. Snowshoe hares ( Lepus americanus) in alder thickets further reduced height growth of spruce in the wetter site. We conclude that net effects of alder on white spruce, the late-successional dominant, were primarily inhibitory and indirect, with the mechanisms depending on initial site moisture. Our results highlight the importance of long-term research showing that small differences among initial replicate sites can cause divergence in successional trajectories, consistent with individualistic distributions of species and communities along environmental gradients. This divergence was detectable only decades later. [ABSTRACT FROM AUTHOR]

Increasing temperatures have resulted in reduced growth and increased tree mortality across large areas of western North American forests. We use tree-ring isotope chronologies (δ13C and δ18O) from live and dead trees from four locations in south-central Alaska, USA, to test whether white spruce trees killed by recent spruce beetle ( Dendroctonus rufipennis Kirby) outbreaks showed evidence of drought stress prior to death. Trees that were killed were more sensitive to spring/summer temperature and/or precipitation than trees that survived. At two of our sites, we found greater correlations between the δ13C and δ18O chronologies and spring/summer temperatures in dead trees than in live trees, suggesting that trees that are more sensitive to temperature-induced drought stress are more likely to be killed. At one site, the difference between δ13C in live and dead trees was related to winter/spring precipitation, with dead trees showing stronger correlations between δ13C and precipitation, again suggesting increased water stress in dead trees. At all sites where δ18O was measured, δ18O chronologies showed the greatest difference in climate response between live and dead groups, with δ18O in live trees correlating more strongly with late winter precipitation than dead trees. Our results indicate that sites where trees are already sensitive to warm or dry early growing-season conditions experienced the most beetle-kill, which has important implications for forecasting future mortality events in Alaska. [ABSTRACT FROM AUTHOR]