Influence of micro-site conditions on tree-ring climate signals and trends in central and northern Sweden

Tree-ring chronologies are important indicators of pre-instrumental, natural climate variability. Some of the longest chronologies are from northern Fennoscandia, where ring width measurement series from living trees are combined with series from sub-fossil trees, preserved in shallow lakes, to form millennial-length records. We here assess the recent ends of such timeseries by comparing climate signals and growth characteristics in central and northern Sweden, of (1) trees growing at lakeshore micro-sites (representing the source of sub-fossil material of supra-long chronologies), with (2) trees collected in dryer micro-sites several meters inland. Calibration trials reveal a predominating June-September temperature signal in N-Sweden and a weaker but significant May-September precipitation signal in C-Sweden. At the micro-site level, the temperature signal in N-Sweden is stronger in the lakeshore trees compared to the inland trees, whereas the precipitation signal in C-Sweden remains unchanged among the lakeshore and inland trees. Tree-rings at cambial ages > 40 years are also substantially wider in the lakeshore micro-site in C-Sweden, and juvenile rings are more variable (and wider) in the dryer micro-site in N-Sweden (compared to the adjacent micro-sites). By combining the data of the various micro-sites with relict samples spanning the past 1,000 years, we demonstrate that growth rate differences at the micro-site scale can affect the low frequency trends of millennial-length chronologies. For the supra-long chronologies from northern Fennoscandia, that are derived from sub-fossil lake material, it is recommended to combine these data with measurement series from only lakeshore trees.