Estimation of potential forest productivity across the Oregon transect using satellite data and monthly weather records

Detailed physiological and micrometeorological studies have provided new insights that greatly simplify the prediction of gross photosynthesis (P_G) and the fraction of production that goes into above-ground net primary production (NPPA). These simplifications have been incorporated into a process-based forest growth model called 3-PGS (Physiological Principles Predicting Growth with Satellite Data). Running the model requires only monthly weather data, an estimate of soil texture and rooting depth, quantum efficiency (α), and a satellitederived Normalized Difference Vegetation Index (NDVI) correlated with the fraction of visible light intercepted by foliage. The model was originally tested in Australia where seasonal variation in NDVI is extreme. In Oregon, NDVI varies much less seasonally and fully stocked coniferous stands maintain nearly constant canopy greenness throughout the year. We compared 3-PGS estimates of P_G and NPPA across a steep environmental gradient in western Oregon where groundbased measurements at six sites were available from previous studies. We first tested the simplification in data acquisition of assigning the same quantum efficiency (α=0.04 mol C/MJ APAR) and available soil water storage capacity (=226 mm) to all sites. With these two variables fixed, the linear relation between predicted and measured P_G was y=1.45 x +2.4 with an r²=0.85. When values of  were adjusted to match seasonal measurements of predawn water potentials more closely, and the quantum efficiency was increased to 0.05 mol C/MJ absorbed photosynthetically active radiation (APAR) on the most productive site, predicted and observed values of P_G and NPP_A were in near 1:1 agreement with r²=0.92. Because maximum greenness (NDVI) reflects the seasonal availability of water, limits on soil water storage capacity can be inferred from calculated water balances derived following the onset of summer drought. The simplifications embedded in the 3-PGS model, along with the need to acquire only one midsummer estimate of maximum greenness, make the approach well suited for assessing the productive capacity of forest lands throughout the Pacific Northwest, USA.