NOAA/WDS Paleoclimatology - Fennoscandia 1250 Year Height Increment Summer Temperature Reconstruction

STUDY NOTES: 1250 year reconstruction of June - August temperature from height increment of Scots pine at the northern timberline in Fennoscandia. Annual height increments of 60 living Scots pine (Pinus sylvestris L.) trees from the northern timberline were measured. Data collection, preparation and measurement were carried out according to the guidelines of Aalto and Jalkanen (1998). The Laanila site is located at 68°28' - 68°31'N, 27°16' - 27°24'E, 220-310 m a.s.l., North Finland, about 80 km south of the pine tree line. The Sodankylä meteorological station is located c. 140 km from the sampling site and provides a climate record from 1908 to present. Height-increment chronology was built from samples from 152 trees. Indexing by 33% splines. A 1-year lag is built into the data table, because current height growth responds strongly to temperature of the previous summer. Mean June - August temperature at Sodankylä meteorological station was used in simple linear regression models. Final model: y = 9,4615 + 3,1389 x Calibration and verification statistics: Calibration period: 1908 – 2007 (100 years) Early period: 1908 – 1957 (50 years) Late period: 1958 – 2007 (50 years) RE = Reduction of error CE = Coefficient of efficiency Early-calibration Late-calibration Whole period (late-verification) (early-verification) R2 0.28 0.35 0.31 RE 0.33 0.219 CE 0.32 0.217
ABSTRACT SUPPLIED BY ORIGINATOR: New data have allowed us to extend a previous height-increment chronology of Scots pine (Pinus sylvestris L.) at the northern Fennoscandian timberline 817 years backwards in time, from 1561 to 745. Our final transfer model accounts for 31% of the dependent instrumental (mean June-August) temperature variance between 1908 and 2007. According to the 1263 yr long summer temperature proxy, the most severe summers were experienced in 1601, 1790 and 782. Correspondingly, the summers of 1689, 885 and 1123 were the most favourable for growth. Two drastic shifts in temperature variability were also found. The twentieth century experienced a multidecadal change as the cold 1905-1914 period was immediately followed by a warm period from 1915 to 1944. An even more prominent shift occurred in the Middle Ages, as the most severe cold spell during 1135-1164 was preceded by the warmest period only a decade earlier, during 1115-1124. The Fourier spectrum of the reconstruction shows significant concentrations of variance around 33.3, 23.3 and 11 years, and between 2.6 and 3.0 years. The wavelet spectrum was able to date several centres of fluctuating periodicities between 745 and 2007. Furthermore, daily temperature records allowed us to define the major growth forcing climatic factor in more detail than in previous response analyses. The mean temperature during a 53 day season from 14 June to 6 August produced the strongest positive growth response (r2 = 0.36).