Models of low temperature and high irradiance and their application to explaining the risk of seedling mortality

Low temperature during the growing season is known to be a leading cause of stress and damage to tree seedlings and interactive effects with high irradiance have been recognised as enhancing the damage. Spatial variation in low temperature and irradiance for mountainous terrain was, therefore, hypothesised to give rise to spatially variable potential for regeneration of tree seedlings and expansion of forest. Combined exposure to low temperature and high irradiance appears most likely on clear mornings following frosty nights. Geographic information system-based models of the spatial variation in minimum air temperature over clear-felled terrain and potentially intercepted direct radiation energy were constructed and were used together with measured values of minimum air temperature to explain observed spatial variation in mortality rates for planted tree seedlings in central Sweden. A local-scale multiple regression model explained 78% of the measured spatial variability in minimum air temperature over 625 km(2) and a corresponding, but univariate, microscale model explained 54% of the measured spatial variation in minimum air temperature over 4 ha. Multiple logistic regression models for the mortality rates of seedlings showed there was no significant effect of potentially intercepted amounts of direct radiation energy during the morning for either scale. There was, however, a significant effect (p<0.0001) of measured minimum air temperature on seedling mortality rates for the microscale study, but no corresponding significant effect for the local-scale study. This confirms that low temperature is an important determinant for seedling mortality. It is likely that other factors than low temperature may be important. However, our results indicate that effects of local-scale variability in minimum air temperature, represented by the minimum air temperature at standard height above the ground (1.8 m), may be over-shadowed by minimum air temperature variability at seedling height in complex terrain. This has implications for the use of standard meteorological data for studies into the responses of plants to abiotic factors. (C) 2000 Elsevier Science B.V. All rights reserved.