Spring frost and growing season length co- control the cold range limits of broad- leaved trees

AimThe aim of this study was to test, based on biological theory, which facet of temperature is most closely associated with the elevational and latitudinal low-temperature limits of seven European broad-leaved tree species. We compared three temperature-related potential constraints across three study regions: (1) absolute minimum temperature within 100years; (2) lowest temperatures during the period of bud-break; and (3) length and temperature of the growing season. LocationWestern and Eastern Swiss Alps (1165-2160m a.s.l.) and southern Sweden (57 degrees N-59 degrees N). MethodsIn situ temperature was recorded at the high-elevation and high-latitude limits of seven broad-leaved tree species and correlated with temperatures at the nearest weather stations, in order to reconstruct the temperature regime for the past 50years. By applying generalized extreme value distribution theory, we estimated the lowest temperatures recurring during the life span of a tree. ResultsAt their high-elevation limits, five out of the seven tree species experienced winter minimum temperatures considerably warmer than their known maximum freezing resistance in winter. For the bud-break period, potentially damaging temperatures occurred at both the elevational and the latitudinal limits and for all four species for which phenological data were available. Three out of five species for which a latitudinal replicate was available showed a similar length of growing season at their respective elevational and latitudinal limits. The mean temperature during the growing season was always warmer at a species' latitudinal limit than at its elevational limit, and hence this variable does not bear general explanatory power for the range limit. Main conclusionsLow-temperature extremes during bud-break are the most likely candidates for controlling the elevational and latitudinal limits of broad-leaved tree species. The absolute minimum temperature in winter and the mean temperature during the growing season are unlikely to constrain the cold limits of these species. Thus, the results call for the use of temperature data (extremes) during key stages of spring phenology when attempting to explain the low-temperature range limits and to predict the potential range shifts of deciduous tree species.