Assessing forest productivity is important for developing effective management regimes and predicting future growth. Despite some important limitations, the most common means for quantifying forest stand-level potential productivity is site index (SI). Another measure of productivity is gross primary production (GPP). In this paper, SI is compared with GPP estimates obtained from 3-PG and NASA’s MODIS satellite. Models were constructed that predict SI and both measures of GPP from climate variables. Results indicated that a nonparametric model with two climate-related predictor variables explained over 68% and 76% of the variation in SI and GPP, respectively. The relationship between GPP and SI was limited (R² of 36%–56%), while the relationship between GPP and climate (R² of 76%–91%) was stronger than the one between SI and climate (R² of 68%–78%). The developed SI model was used to predict SI under varying expected climate change scenarios. The predominant trend was an increase of 0–5 m in SI, with some sites experiencing reductions of up to 10 m. The developed model can predict SI across a broad geographic scale and into the future, which statistical growth models can use to represent the expected effects of climate change more effectively.

To the best of our knowledge, one or more authors of this paper were federal employees when contributing to this work. This is the publisher’s final pdf. The published article is copyrighted by Society of American Foresters and can be found at: http://www.safnet.org/publications/jof/index.cfm/Tree biomass is typically estimated using statistical models. This review highlights five limitations of most tree biomass models, which include the following: (1) biomass data are costly to collect and alternative sampling methods are used; (2) belowground data and models are generally lacking; (3) models are often developed from small and geographically limited data sets; (4) simplistic model forms and predictor variables are used; and (5) variation is commonly averaged or grouped rather than accounted for. The consequences of these limitations are highlighted and discussed. Several recommendations for future efforts are presented including the following: (1) collection of field measurements of tree biomass using consistent protocols; (2) compilation of existing data; (3) continued evaluation and improvement of existing models; (4) exploration of new models; and (5) adoption of state-of-the-art analytical and statistical techniques. Given the increasing importance of accurately estimating forest biomass, there is a critical need to understand, evaluate, and improve current tree biomass prediction methods