Files
Abstract
Light interception is a key ecological indicator that is strongly related to forest productivity. Its application has increased over the last few decades with the advent of process based models. Evaluating light interception requires an assessment of the amount of leaves present in a given forest plantation. This assessment is done with indirect methods such as spectral indices derived from satellite data, that empirically relate different satellite bands with ground based leaf area measurements. These indices are not exempt from measurement error, due to the many factors affecting light reflection, refraction, and scattering in its path between the leaves and the satellite sensor. As a result, the assumption of error free predictors is violated, resulting in a biased estimator for leaf area. This is also known as an error-in-variable problem. To bridge this gap, a modeling framework is presented that account for stochastic deviations in the independent predictors, resulting in an improved model when compared to other published research.
Leaf area, as an indicator of site productivity is a function of climate and soil factors. Many of the relations between these environmental predictors are described through complex relations inside process based models, with assumptions like carbon partitioning fractions, instantaneous carbon allocation and fixed respiration costs. These models trade simplicity for comprehensiveness, finding low acceptance from practitioners interested in operational applications. To bridge this gap, a semi-empirical model was developed to allow the description of foliage display in loblolly pine, as a function of environmental variables. The model was able to predict leaf area display, including a parameter responsible for foliage carrying capacity. The model was parameterized using plot level observations from two research studies across 24 locations across the southeast US.
Finally, a mechanistic model is presented that utilizes the foliage display model and relates it with a dominant height equation for loblolly pine. The model was fit as a system of simultaneous differential equations using local and global parameters. The model was able to correctly describe dominant height over age, bridging the gap between an empirical growth and yield approach and an important ecological indicator, estimated with an unbiased equation from satellite time series.
Leaf area, as an indicator of site productivity is a function of climate and soil factors. Many of the relations between these environmental predictors are described through complex relations inside process based models, with assumptions like carbon partitioning fractions, instantaneous carbon allocation and fixed respiration costs. These models trade simplicity for comprehensiveness, finding low acceptance from practitioners interested in operational applications. To bridge this gap, a semi-empirical model was developed to allow the description of foliage display in loblolly pine, as a function of environmental variables. The model was able to predict leaf area display, including a parameter responsible for foliage carrying capacity. The model was parameterized using plot level observations from two research studies across 24 locations across the southeast US.
Finally, a mechanistic model is presented that utilizes the foliage display model and relates it with a dominant height equation for loblolly pine. The model was fit as a system of simultaneous differential equations using local and global parameters. The model was able to correctly describe dominant height over age, bridging the gap between an empirical growth and yield approach and an important ecological indicator, estimated with an unbiased equation from satellite time series.