Effects of rising temperatures on forest composition and biogeochemical cycles.

At Harvard Forest, I am investigating whether juvenile trees exhibit species-specific responses to warming (5C), and what the potential long-term implications are for forest composition and ecosystem functioning.  Thus far, similar to my work on CO2 impacts at Duke Forest, I am finding that the tree species benefiting most from warming are the least-productive taxa.  This work is described in a Mohan, Melillo, et al. manuscript, in review with Nature.  In addition, recruitment of red oak – a major dominant tree in the eastern and central U.S. – is negatively affected by increased soil temperatures, suggesting both a decline in future abundance and a northward range shift for this species relative to maple competitors.
In collaboration with Jerry Melillo I am currently examining the impacts of a 5C increase in temperature on carbon and nitrogen dynamics of temperate forest soils and phenological, growth and survivorship responses of canopy trees.  One of the most important findings from the original soil warming experiment at Harvard Forest, initiated in 1991, was that warming caused a large increase in the flux of CO2 from soils to the atmosphere. But after 10 years, respiration rates in heated plots were equivalent to those of un-heated control plots, indicating enhanced soil respiration was a transient response.  However, that the rate is still holding at control levels suggests that although an initial, highly-labile pool of carbon was respired with warming, microbes in the warmed plots are now able to decompose more recalcitrant carbon at the same rate that more labile carbon is utilized under control conditions.  Further, due to an unexpected lightening-strike in 2005, I have been able to determine that heated plots respire 34% less carbon when they are maintained at the same temperature as control plots.  This suggests that a direct temperature effect, such as increased rates of biogeochemical cycling and/or acclimation by soil microbes, is at least partially responsible for the equivalent rates of respiration in warmed plots.  I am working with Mark Bradford (Univ. Georgia), Kathleen Treseder (Univ. California, Irvine), Matthew Wallenstein (Univ. California, Santa Barbara), and Serita Frey (Univ. Hew Hampshire) investigating the possibility that shifts in microbial communities are also responsible for altered carbon metabolism in the heated soils.
Another response to warming we are observing in both the original and new soil warming experiments is increased net nitrogen mineralization rates, representing nitrogen available to plants.  I am examining the growth and survivorship of canopy trees in the large-scale warming experiment, initiated in 2003, and calculating net carbon accumulation rates in woody biomass, to determine if increased growth of trees will offset the increased carbon flux from the heated soils.  I initiated phenological studies in 2004, and am finding that, contrary to my prediction, canopy trees in the warmed plot are leafing-out about seven days earlier in the spring relative to those in the control plot.  Beginning in 2004 I began documenting increased canopy tree growth, but I am also observing, in accordance with ecological theory, increased mortality rates of canopy trees due to self-thinning in the heated, “fertilized” plot.  The result thus far is that live canopy trees are growing fast enough with warming to accumulate about 50% of the “extra” carbon currently respired from the warmed soils.  The future, however, is less certain.  The additional soil nitrogen availability, which is sustained over time, may result in future forests growing faster and becoming greater carbon sinks (increased soil respiration is a transient response, diminishing after about 10 years).  However, increased tree mortality under warmed conditions suggests the possibility that over time, decomposition of dead woody debris will increase the amount, and perhaps extend the duration, of carbon respired by warmed forest soils, resulting in temperate forests being net sources of atmospheric CO2.
The large-scale warming experiment at Harvard Forest provides a unique opportunity to assess how warming, and accompanying impacts on soil nitrogen and growing season duration, may affect reproductive efforts in forest trees and herbs.  In 2004 I initiated a long-term study of reproductive responses of two common forest herbaceous species – Maianthemum canadensis (Canadian mayflower) and Trientalis borealis (starflower).  My students and I are finding consistent results in 2004-2006; namely, that soil warming and associated increases in soil nitrogen are diminishing overall fruit production and reproductive probabilities for a given plant size in both species.  These results mirror what I am observing for the dominant Quercus rubra (red oak) and Acer rubrum (red maple) tree species.  Both trees produce fewer and smaller seeds under warmed conditions, and red maple seeds from heated-plot trees exhibit reduced germination probabilities even after seed size is taken into account.  These new data are some of the first from an intact forest experiment, and I look forward to pursuing similar reproductive work in other ecosystem-scale global change experiments.
The Canadian Global Change Model predicts mean temperature increases in the southeastern U.S. of up to 5.5C by year 2100, and predicted increases in maximum summer temperatures are the highest in the nation . Yet it is unclear how these changes will impact plant community composition and ecosystem functioning.  I propose to initiate the first ecosystem-scale warming experiment in the southeastern U.S., and in particular to do this in a setting allowing cross terrestrial-riparian-aquatic investigations.  Although the northeast now has three soil warming experiments in forests, globally most warming experiments are located in short-statured plant communities such as alpine meadows, bogs, and tundra.  The large-scale nature of such an experiment (i.e., the Harvard Forest experiment is 1800 m2) would allow an assessment of basic ecological responses ranging from tree growth and survivorship, reproduction and recruitment, phenology, tissue chemistry and decomposition, to shifts in community structure that are simply not possible to investigate in smaller-scale studies.  The large sample sizes of plants and soils in this project would allow hierarchical Bayesian analyses to account for inherent variability between individuals and microenvironments.  This research would permit comparisons with northeastern U.S. and Scandinavian experiments that include differences in climate and soils, but also allows examinations of the effects of species identities and species diversity on overall ecosystem response.  The aquatic component of the experiment invites novel climate change research opportunities.
Until now, vegetative work at most warming experiments, including those at Harvard Forest, focused primarily on aboveground vegetation processes.  I recently initiated an experiment quantifying fine root dynamics in the new, large-scale warming experiment at Harvard Forest, and am currently placing root in-growth cores in the field to assess relative root production and turnover rates in 2006-2007.  During the 2006 summer, I found that fine root biomass in the warmed plot was about 40% less than in the control plot, suggesting increased soil nitrogen availability is causing a shift away from belowground production.  Root in-growth cores will also be placed at the original warming experiment begun in 1991 to compare root dynamics in forests at different stages of warming-induced soil organic matter depletion.  In 2007, I will be collaborating with Lindsey Rustad (Univ. Maine) initiating a minirhizotron study at both sites to further assess dynamics of root growth and turnover.  Additionally, I will be working with Mark Bradford (Univ. Georgia), Kathleen Treseder (Univ. California, Irvine), and Matthew Wallenstein (Univ. California, Santa Barbara) performing ∂13C pulse-chase studies at both the new and the original warming experiments.  In these investigations 13C-labelled compounds will be used to access abundances of different microbial indicators (Treseder, Wallenstein), and rates and amounts of 13C incorporation into root, foliage, and woody biomass (Mohan, Melillo) and soil organic matter (Bradford) pools.

Harvard Forest LTER Website