Remote sensing of net primary production in boreal forest stands

A series of LANDSAT greenness vegetation index (Gi) images, calibrated to a common reference image, were used to model the foliar phenological dynamics of 30 quaking aspen (Populus tremuloides) and 29 black spruce (Picea mariana) stands in the boreal forest of Northeast Minnesota. The fitted phenology models were used in a simplified canopy radiative transfer model to estimate the amount of incident photosynthetically active radiation (PAR) intercepted annually. The relationship between annual estimated PAR interception (IPAR) and measured annual above-ground net primary production (AANPP), that is, the dry matter yield of Sigma IPAR (epsilon(i)) was examined. Annual PAR interception among aspen stands was relatively high (600-1100 MJ m(-2) yr(-1)), with just three stands intercepting less than 600 MJ m(-2) yr(-1). Stand age affected the relationship between Sigma IPAR and AANPP in aspen, with Sigma IPAR in young stands being strongly related to AANPP (r(2) = 0.77, p < 0.002) yet weakly related in mature stands (r(2) = 0.18, p < 0.05). This observation is consistent with increased maintenance respiration demands as the ratio of total to foliar biomass increases with stand age. epsilon(i) for above-ground production in aspen stands varied from 0.44-1.29 gMJ(-1), averaging 0.92 (+/- 0.22) gMJ(-1). In spruce, Sigma IPAR varied widely among stands (100-1100 MJ m(-2)), regardless of age, and explained 76% of the variance in AANPP (p < 0.0001). epsilon(i) for above-ground production in spruce varied from 0.17-0.89 g MJ(-1), averaging 0.49 (+/- 0.17) gMJ(-1). These values are within the range reported for other forest ecosystems. Estimates of AANPP from Sigma IPAR, assuming median values of epsilon(i), were used to estimate a carbon drawdown of 256 Mg C km(-2) for the above-ground component of general ecosystem cover-types over the 2280 km(2) study area. Variability in Ei among forest stands, both within and between species, is discussed in the context of evolutionary optimization of resource use efficiency. Results suggest that factors in addition to light interception are required to estimate stand productivity in boreal forest ecosystems, particularly in stands that are subject to periodic moisture stress and increased respiration demands with age.