PNNL

Abstract

Vegetation growth models have been coupled with data from remotely sensed imagery and surface meteorological networks to monitor terrestrial production and ecosystem-atmosphere carbon exchange across a wide range of spatial and temporal scales (e.g., MODIS, CASA, GLO-PEM). Many of these diagnostic models are based on a light-use efficiency equation and two-component model of whole-plant growth and maintenance respiration, which have been parameterized for functionally distinct vegetation types and biomes. This study was designed to assess the robustness of these parameters for predicting interannual plant growth and carbon exchange, and more specifically, to address inconsistencies that may arise during forest disturbances and loss of canopy foliage. A model based on the MODIS MOD17 algorithm was parameterized for a mature upland hardwood forest by inverting CO2 flux tower observations during years when the canopy was not disturbed, and used to make predictions during a year when the canopy was 37% defoliated by forest tent caterpillars. To accurately capture interannual variability during all years, algorithms needed to be modified to scale for the effects of diffuse radiation and loss of leaf area. Photosynthesis and respiration model parameters were found to be robust at daily and annual time scales, and differences in net ecosystem production in the presence and absence of large numbers of defoliating insects was approximately 2 g C m-2 d-1 and