AGE-DEPENDENT TREE-RING GROWTH-RESPONSES OF SUB-ARCTIC WHITE SPRUCE TO CLIMATE

We test one of the fundamental assumptions of most dendroclimatological research, that the radial growth response of trees to climate does not vary with age once the biological growth trend has been removed. Picea glauca (Moench) Voss from three sites in the western Northwest Territories, Canada, are disaggregated into age-classes, and their response to climate examined through response function and linear regression analyses. These data are then used in multiple regression analyses to estimate June-July temperatures at Norman Wells, N.W.T., from 1909 to 1989 using both age-dependent and standard (age-independent) models. The response function and regression analyses suggest that the response of Picea glauca radial growth to climate differs between trees greater than 200 years old and less than 200 years old. These results suggest that the assumption of an age-invariate climate-growth function is therefore invalid at these sites. These apparent age-dependent responses are site specific and may reflect physiological changes related to aging. One possible causal factor of age dependence is that the trees are becoming increasingly moisture stressed with age owing to a reduction in the efficiency of water and nutrient translocation mechanisms. The estimation of June-July temperatures based on an age-dependent model produced improved calibration and verification statistics as compared with a reconstruction based on standard dendroclimatic modelling. If present, age-dependent climate-growth relationships may result in less accurate reconstructions of past climate, particularly during the early portions. However, age-dependent responses could also be used to increase the number of tree-ring based predictor variables for dendroclimatic reconstructions.