PARENT SESSION
Oral Session #44: Global Change and Climate Change.
Presiding: C. Wessman
Tuesday, August 6. 1:00 PM to 4:45 PM. Grand Ballroom West, Radisson.

Assessing northern red oak sensitivity to global warming using tree-ring based growth-climate correlations.

Terrell, Mark^*,1, LeBlanc, David¹, ¹ Indiana University-Purdue University-Indianapolis, Indianapolis, IN

ABSTRACT- Computer simulation models predict significant forest decline, changes in forest composition and structure, and shifts in the spatial distribution of forests in response to global warming. However, these models assume that the influence of climate controls the spatial distribution of forests, predicting maximal growth in the center of a species range and zero growth (100% mortality) at the range limits and beyond. Using standard dendrochronological procedures, seventy-one northern red oak (Quercus rubra) tree-ring chronologies were prepared to test the spatial nature of the tree growth-climate assumptions inherent in these models. Correlational analyses measured the association between annual ring width and 120 climate variables. Tree growth was strongly correlated and spatially consistent with early growing season variables, including 1) actual-to-potential evapotranspiration (AE:PE) ratio, 2) mean temperature, 3) drought, and 4) total precipitation. These growth-climate correlations were used as indicators of tree sensitivity to variations in historical climate. GIS analysis and partial Mantel tests were used to document and measure the association between the spatial variation in tree-sensitivity indices and the spatial patterns along latitudinal and longitudinal climatic gradients, while controlling for spatial autocorrelation among site locations. No significant spatial association was found between the spatial pattern of precipitation and spatial variation of tree sensitivity to precipitation. Spatial variation in tree sensitivity indices for temperature, AE:PE ratio, and drought were only marginally correlated (r = 0.2) with spatial patterns of the corresponding climate gradients. These low r-values strongly suggest that climate does not completely control tree-growth, mortality, and geographic distribution of northern red oak, as assumed by forest simulation models. Therefore, current algorithms used in these models need to be re-calibrated with data from empirical modeling studies to improve their predictive validity, may over-estimate the role of climate in controlling tree growth and mortality, and may exacerbate the negative impacts of warming on North American deciduous forests.

KEY WORDS: Forest, Spatial, Mantel, Climate