Date of Award


Level of Access Assigned by Author

Campus-Only Thesis

Degree Name

Master of Science (MS)


Forest Resources


Shawn Fraver

Second Committee Member

David Hollinger

Third Committee Member

Robert Seymour

Additional Committee Members

Aaron Weiskittel


Forest productivity is known to vary in response to annual fluctuations in climate. By assessing how closely forest productivity tracks climate on an annual basis, we can gauge climate sensitivity for a given location. However, our understanding of forest productivity is limited, in part, by the complex methods needed to measure annual stand-level productivity. The eddy covariance technique, whereby CO2 exchange (flux) is continuously measured at the canopy-atmosphere interface, has clarified our understanding of stand-level carbon dynamics by addressing productivity and its response to climate fluctuations. Biometric approaches, where forest productivity is estimated from the annual growth of individual trees, can offer a more economical alternative to continuous CO2 flux measurements, while providing annual estimates of CO2 assimilation. However, attempts to link these measurements of productivity (i.e., CO2 flux and tree growth) have produced inconsistent results and demonstrated a need for further research.

Using long-term forest inventory data paired with tree-ring data from a mixed-species conifer forest at the Howland Research Forest, central Maine, we developed annual estimates of productivity at species- and stand-levels. We used annual carbon mass increment (derived from tree-ring data and species-specific wood densities) per unit area as a measure of productivity because it allows direct comparisons with CO2 flux data. Our results demonstrated strong correlations between carbon mass increment and annual CO2 flux measurements summarized from previous-year fall to current-year fall, an improvement from calendar year summaries. Further, our results suggest tree growth was lagged one year behind CO2 flux (i.e., assimilated CO2 was not allocated to growth until the following year) for about the first half of the time-series, but later became synchronized with current year assimilation. We suspect the shift to synchrony reflects the onset of above-average spring temperatures, which shifted carbon allocation from storage to current-year wood formation.

We also explored the use of annual carbon mass increment as an alternative to traditional standardized tree-ring chronologies for assessing climate–growth relationships. Our approach allowed us to assess these relationships across three levels of organization: individual tree-, species-, and stand-levels. We believe this approach is preferable in complex mixed-species forests, as it provides insights unavailable from the traditional standardization approach, which is restricted to the species level. Our results demonstrated that stand-level growth was sensitive to a different set of climate variables than those affecting individual species. We documented previous summer precipitation (higher better) and spring temperature (warmer better) as the most influential variables affecting stand-level growth at Howland Forest. We conclude the use of carbon mass increment has a broad range of applications within studies of forest productivity, and we suggest that future studies using tree-ring data consider the potential benefits of converting to carbon-mass increment. The following thesis highlights the benefits of implementing annual carbon mass increments for the purpose of addressing a broad range of research questions.