The influence of environmental conditions on tree-ring series of Norway spruce for different canopy and vitality classes

Abstract

A dendroecological study of Norway spruce trees was undertaken by Eckstein et al. (1989) as part of an interdisciplinary project investigating the influence of air pollution on tree growth in northern Germany. They analyzed individual trees stratified according to three different canopy classes and three different vitality classes, using response function analysis (Fritts 1976). Eckstein et al. conclude that there is a significant positive correlation between precipitation during the growing season and the tree-ring series, and that the observed forest decline is probably caused by a severe unknown stress factor (but clearly not a climate factor) during the mid-1950s. The use of response functions has the limitation of time invariance and the separated adjustment of trend and climate variables for the tree-ring series. In this article a model is applied for simultaneous adjustment of trend and explanatory variables for tree-ring chronologies. Time-dependent or stochastic response functions are obtained by writing these models in the so-called state-space form and analyzing them with the discrete Kalman filter. Consequences for chronology building are discussed. After analyzing the tree-ring chronologies of the Norway spruce by these more sophisticated methods, it is found that: 1. the stochastic response functions for precipitation during the growing season are principally constant over time and positive, 2. the stochastic response functions for the temperature in February and March are not constant and become significant after the mid- 1950s for trees from the classes 'not dominant-vital,' 'not dominant- moderately vital,' 'dominant-not vital,' 'moderately dominant-not vital,' and 'not dominant-not vital,' 3. the influence of climate factors on tree-ring development is higher if the vitality and dominance of the trees is lower, 4. addition of an SO 2 emission curve as explanatory variable, simultaneously with trend and climate variables for tree-ring chronologies, did not reveal a significant air pollution effect on tree-ring development.