Modeling the health and productivity of Oak Savannas in central USA

Oak species have a long history of domination in eastern North America and their present distribution in various regions exceeds that recorded in the original forests at the time of European settlement. The increase in oak during the late 18th and 19th centuries can be attributed to historical changes in disturbance regimes in the eastern biome. The expansion in oak distribution has occurred on xeric or nutrient-poor sites, which indicates the stress tolerance capabilities of many oak species. The aim of this research is to assess the health and productivity of the fragmented oak savannas that span from Texas north to the Canada border using statewide GAP, climate and MODIS data and the 3PGS (Physiological Principles Predicting Growth using satellite data) ecosystem process model. 3-PGS is a simple big-leaf productivity model that sets upper limits on monthly gross primary productivity (GPP) by determining the amount of photosynthetically active radiation absorbed (APAR) by vegetation and the photosynthetic capacity (LUE) of the canopy. The utilized portion of APAR is calculated by reducing total PAR by an amount determined by the most constraining of a series of environmental modifiers that affect gas exchange through stomata. These include: (a) high daytime atmospheric VPD; (b) soil water availability; and (c) the frequency of sub-freezing temperatures (<-2 °C). Climate data including day length, precipitation, average temperature and vapor pressure deficit are obtained from the Daymet daily gridded surface data from Oak Ridge National Laboratory (ORNL). The fraction of APAR is obtained from the MODIS terra/aqua combined product. Monthly surfaces were derived for the study period 2003-2008 spanning available MODIS and Climate datasets. Soil properties for the entire U.S.A., derived from luster analysis of STATSGO soil parameters, terrain and climate observations were obtained from ORNL. Oak savannas within this region are identified using the statewide Gap Analysis Program data provided by USGS. Several landsat scenes selected across the study region provided an estimate of the seasonal variation in fPAR and distinction between the tree and understory grass signals.