Wood specific gravity (SG) is correlated with many physical and mechanical properties and hence is a widely-used indicator of wood quality. Since dry wood contains around 49% carbon, the role of wood SG has emerged in assessment of carbon sequestration potential. Within tree stems, SG shows radial and vertical patterns of variability that are related with tree age, climate and various site conditions. Our goal was to develop a model that estimates ring SG of loblolly pine (Pinus taeda) as a function of cambial age and environment variables as independent variables. Pith-to-bark SG density profiles were obtained on 12-mm increment cores collected from 269 trees from 13 stands across the Southeastern U.S. using X-ray densitometry. We used the spatial coordinates of the stands to obtain interpolated climate variables for each year corresponding to the year of each growth ring. A four-parameter logistic base model was used to predict SG with cambial age as the explanatory variable. The base model was updated and modified using environmental covariates including annual indices for temperature and water deficit, and soil characteristics including permeability and depth to water table. The inclusion of environmental covariates improved the prediction by reducing root mean square error by 3.5 percent and decrease in AIC by 2% compared with the base model. Temperature positively affected both the upper asymptote and the point of inflection in the SG model, and water deficit was negatively related to SG. The model developed can be used to predict SG under contrasting conditions of climate and soil.
