Microclimate moderates plant responses to macroclimate warming.

Recent global warming is acting across marine, freshwater, and terrestrial ecosystems to favor species adapted to warmer conditions and/or reduce the abundance of cold-adapted organisms (i.e., "thermophilization" of communities). Lack of community responses to increased temperature, however, has also been reported for several taxa and regions, suggesting that "climatic lags" may be frequent. Here we show that microclimatic effects brought about by forest canopy closure can buffer biotic responses to macroclimate warming, thus explaining an apparent climatic lag. Using data from 1,409 vegetation plots in European and North American temperate forests, each surveyed at least twice over an interval of 12-67 y, we document significant thermophilization of ground-layer plant communities. These changes reflect concurrent declines in species adapted to cooler conditions and increases in species adapted to warmer conditions. However, thermophilization, particularly the increase of warm-adapted species, is attenuated in forests whose canopies have become denser, probably reflecting cooler growing-season ground temperatures via increased shading. As standing stocks of trees have increased in many temperate forests in recent decades, local microclimatic effects may commonly be moderating the impacts of macroclimate warming on forest understories. Conversely, increases in harvesting woody biomass--e.g., for bioenergy--may open forest canopies and accelerate thermophilization of temperate forest biodiversity.

Quantifying canopy complexity and effects on productivity and resilience in late-successional hemlock-hardwood forests.

The regrowing forests of eastern North America have been an important global C sink over the past 100+ years, but many are now transitioning into late succession. The consequences of this transition are unclear due to uncertainty around the C dynamics of old- growth forests. Canopy structural complexity (CSC) has been shown to be an important source of variability in C dynamics in younger forests (e.g., in productivity and resilience to disturbance), but its role in late-successional forests has not been widely addressed. We investigated patterns of CSC in two old-growth forest landscapes in the Upper Peninsula of Michigan, USA, to assess factors associated with CSC and its influence on productivity and disturbance resilience (to moderate-severity windstorm). CSC was quantified using a portable below-canopy LiDAR (PCL) system in 65 plots that also had long-term (50-70+ years). inventory data, which were used to quantify aboveground net primary productivity (ANPP), disturbance history, and stand characteristics. We found high and variable CSC relative to younger forests across a suite of PCL-derived metrics. Variation in CSC was driven by species composition and size structure, rather than disturbance history or site characteristics. Recent moderate severity wind disturbance decreased plot-scale CSC, but increased stand-scale variation in CSC. The strong positive correlation between CSC and productivity illustrated in younger forests was not present in undisturbed portions of these late-successional ecosystems. Moderate severity disturbance appeared to reestablish the positive link between CSC and productivity, but this relationship was scale and severity dependent. A positive CSC-productivity relationship was evident at the plot scale with low-severity, dispersed disturbance, but only at a patch scale in more severely disturbed areas. CSC does not appear to strongly correlate With variation in productivity in undisturbed old-growth forests, but may play a very important (and scale/severity-dependent) role in their response to disturbance. Understanding potential, drivers and consequences of CSC in late-successional forests will inform management focused on promoting complexity and old-growth conditions, and illustrate potential inipacts of such treatments on regional C dynamics.

Multi-decade biomass dynamics in an old-growth hemlock-northern hardwood forest, Michigan, USA.

Trends in living aboveground biomass and inputs to the pool of coarse woody debris (CWD) in an undisturbed, old-growth hemlock-northern hardwood forest in northern MI were estimated from multi-decade observations of permanent plots. Growth and demographic data from seven plot censuses over 47 years (1962-2009), combined with one-time measurement of CWD pools, help assess biomass/carbon status of this landscape. Are trends consistent with traditional notions of late-successional forests as equilibrial ecosystems? Specifically, do biomass pools and CWD inputs show consistent long-term trends and relationships, and can living and dead biomass pools and trends be related to forest composition and history? Aboveground living biomass densities, estimated using standard allometric relationships, range from 360-450 Mg/ha among sampled stands and types; these values are among the highest recorded for northeastern North American forests. Biomass densities showed significant decade-scale variation, but no consistent trends over the full study period (one stand, originating following an 1830 fire, showed an aggrading trend during the first 25 years of the study). Even though total above-ground biomass pools are neither increasing nor decreasing, they have been increasingly dominated, over the full study period, by very large (>70 cm dbh) stems and by the most shade-tolerant species (Acer saccharum and Tsuga canadensis). CWD pools measured in 2007 averaged 151 m(3)/ha, with highest values in Acer-dominated stands. Snag densities averaged 27/ha, but varied nearly ten-fold with canopy composition (highest in Tsuga-dominated stands, lowest in Acer-dominated); snags constituted 10-50% of CWD biomass. Annualized CWD inputs from tree mortality over the full study period averaged 1.9-3.2 Mg/ha/yr, depending on stand and species composition. CWD input rates tended to increase over the course of the study. Input rates may be expected to increase over longer-term observations because, (a) living biomass is increasingly dominated by very large trees whose dead trunks have longer residence time in the CWD pool, and (b) infrequent major disturbances, thought to be important in the dynamics of these forests, have not occurred during the study period but would be expected to produce major, episodic pulses in CWD input. Few fragments of old-growth cool-temperate forests remain, but such forests can constitute a very large carbon pool on a per-area basis. The carbon sink/source status of these forests remains unclear. While aboveground living biomass at this study site shows no strong aggrading or declining trend over the last half-century, this remains a modest span in the innate time-scale of late-successional forest. The effects of rare disturbances, long-term shifts in composition and size structure, and changes in soil carbon and CWD pools may all influence long-term carbon status.