Modelling minimum air temperature in partially and clear felled forests

Effects of tall shelterwood and soil preparation on minimum air temperature over undulating terrain were modelled for clear and calm nights. A stepwise linear regression model describing the spatial variability of minimum air temperature at a height of 25 cm above grass-covered clear-cut areas accounted for 87% of the measured temperature variability. The explanatory variables parameterise drainage, pending and stagnation of cold air and were retrieved from a digital elevation model in a GIS. There was no influence from the azimuth orientation of slopes on the minimum air temperature. A model for both shelterwood of three stem densities and clear-cut explained 89% of the spatial variation in minimum air temperature where variables related to sky view, topography and presence of scarified ground were used. Topographic influences increased in parallel with increasing sky view. Simple topagraphy related measures within shelterwood were successfully used in the model where detailed topographic mapping was impeded by forest stands. For flat terrain, a maximum shelterwood effect of +4.6 degrees C for a density of 275 stems/ha was estimated for nights with moderately strong radiative cooling. An effect of +1.6 degrees C was estimated for scarification of clear-cut ground. Data sets collected separately for validation for each model compared well with estimated values. (C) 1998 Elsevier Science B.V. All rights reserved.