Sources of variation in ecophysiological parameters in Douglas-fir and grand fir canopies

Forest process models predict ecosystem responses from climate variables and physiological parameters. The parameters describe key ecosystem attributes, often as lumped averages. However, the sources and magnitude of variation in these physiological parameters are unknown, which complicates sampling if models are to be parameterized with field measurements. We measured several key parameters, which had been identified by sensitivity analyses of three models, in Abies grandis (Dougl.) Lindl. and Pseudotsuga menziesii var. glauca (Beissn.) Franco trees throughout the growing season. Trees were sampled at eight sites across the interior northwest of the USA. At each site, fertilized and control plots were sampled. The design provided statistical replication for the analysis of variance within a site, allowing us to draw inferences about a regional population of stands. Specific leaf area (SLA) varied by canopy position and treatment (P = 0.0003), by date of sampling (P < 0.0001) and by species (P = 0.0188). Mass-based foliar nitrogen concentration (%N) increased during the summer in both species (P = 0.0019), but at a faster rate in P. menziesii var. glauca than in A. grandis. Sun foliage had a higher mean %N (1.00, SE = 0.02%) than shade foliage (0.92 +/- 0.01%). Apparent quantum yield (Phi) varied among treatments, between species and between canopy positions; each of these variables interacted with date of sampling (P = 0.0207, P < 0.0001 and P = 0.0344, respectively). In A. grandis, mean Phi values (+/- SE) were 0.048 +/- 0.006 mol CO2 (mol incident photon)-1 for sun foliage and 0.067 +/- 0.007 mol CO2 (mol incident photon)(-1) for shade foliage. In P. menziesii, the corresponding mean Phi values were 0.032 +/- 0.003 and 0.047 +/- 0.004 mol CO2 (mol incident photon)-1. Parameters SLA, %N and Phi were all influenced by date, fertilizer treatment, species and crown position. We discuss methods of inferring quantum yields from light response curves and their utility for parameterizing process models. Parameter mean values are presented for each site; these tables provide a documented data set for the parameterization of models describing the western interior forests of the USA.