Climate signal age effects - Evidence from young and old trees in the Swiss Engadin

A potential limitation of tree-ring based climate reconstructions is related to climate signal age effects (CSAE). CSAE may arise if the climatic response of young tree-rings differs from that of old tree-rings. This could mean that climatic signals become stronger (or weaker) with tree aging, or that the seasonality of signals or the sensitivity to a specific element (e.g., temperature, precipitation) changes over time. Such changes would affect the interpretation of dendroclimatic reconstructions, as the tree-rings included in these records are generally oldest at the end of a record (e.g., 21 st century)-which is the time period generally used for calibration with instrumental data. We here addressed this concern by analyzing young and old Pinus cembra trees from three high elevation sites in the central European Alps. Core and disc samples were collected in pre-defined plots to allow for a representative analysis of tree ages with tree-ring width (TRW) measurement series categorized into age classes (i) >1880, (ii) 1880-1939, and (iii) 1940-2002. Notably we report on the signal of the very young category (iii) not yet described in literature, and thus allow estimation of climate response and signal strength characteristics during the first years of the trees' lifespans. Comparison of age classes (i)-(iii) revealed differences in TRW coherence and size, but little change in climatic signal. CSAE are in the order of the differences recorded among high elevation sites-a conclusion that holds for inter-annual to decadal scale TRW variations at near-treeline Swiss stone pine. Such data are typically included in regional and larger-scale temperature reconstructions; thus, our results add confidence to long-term climate estimates integrating a range of tree-ring age classes. Other findings, such as the reaction wood in juvenile tree-rings, and sensitivity of the climate signal to sample replication, suggest that comparisons of young and old age classes, and separate calibration of these categories against instrumental climate data might further the estimation of long-term uncertainty changes in tree-ring based climate reconstructions. (C) 2008 Elsevier B.V. All rights reserved.