Climatic response of budburst in the mountain birch at two areas in northern Fennoscandia and possible responses to global change


  • P. S. Karlsson,

  • H. Bylund,

  • S. Neuvonen,

  • S. Heino,

  • M. Tjus

P. S. Karlsson ( and M. Tjus, Abisko Scientific Research Station, The Swedish Royal Academy of Sciences, SE-981 07 Abisko, Sweden (present address of P. S. K.: Dept of Plant Ecology, Uppsala Univ., Villavägen 14, SE-752 36 Uppsala, Sweden, (present address of M. T.: Ingareds skolgata 60, SE-441 65 Alingsås, Sweden). – H. Bylund, Dept of Entomology, Swedish Agricultural Univ., P.O. Box 7044, SE-750 07 Uppsala, Sweden. – S. Neuvonen and S. Heino, Kevo Subarctic Research Inst., Univ. of Turku, FIN-20014 Turku, Finland.


The relationship between the climate and budburst of the mountain birch was evaluated for two areas in subarctic (ca 69°N) Fennoscandia; at Abisko, Swedish Lapland, and at Kevo, Finnish Lapland. Thermal time (TT, degree-day sums) to budburst was calculated for experimental conditions in the laboratory and for in situ observations of budburst. Two types of models predicting leaf emergence in situ were used: 1) TT to budburst for different threshold temperatures based on daily mean (TTMean) or daily maximum (TTMax) temperatures and 2) ecophysiological budburst models. The obtained models were used to estimate effects of a changed climate.
Laboratory experiments of TT to budburst indicated no differences in the thermal requirements at the two areas. Temperature requirements of budburst declined successively during the progression of spring, from ca 250 degree-days (>+2°C) in January to ca 100 in May. No significant trend in the date of budburst was found over the last 70 (Abisko) or 20 (Kevo) yr. There were some differences in the type of model that best explained the date of budburst in situ at the two areas. For Kevo the best prediction (minimum root mean square error, RMSE) of budburst was obtained by a simple thermal time model (TTMean>5.5°C) from 1 January (RMSE=2.1). For budburst at Abisko, models based on daily maximum temperature fitted better than those based on daily means. For Abisko, models based on thermal time accumulation only showed systematic errors in the predicted budburst that were correlated with budburst previous year (BBPY). Including this apparent memory effect in the model decreased the error by 2.4 d. The best prediction for Abisko was thus obtained using TTMAX>6.5 (RMSE=3.1) from 1 January.
Using these models to predict the effect of a changed air temperature climate indicate 3–8 d earlier budburst for a one-degree increase in temperature, the effect being smaller for Kevo than for Abisko. For both areas a change in May temperature has a larger effect on the date of budburst than changes in any other month.