The present paper describes a model for frost assessment in mountainous are
as in relation to forest management. Data were collected at 38 locations wi
thin a 625 km(2) region, which is characterised by a diverse topography and
vegetation cover. Air temperature measurements were performed during the p
eak of the growing season. (July-August 1996) in the southern Swedish mount
ains at elevations between 500 and 1200m a.s.l. The variation in nighttime
minimum temperatures was analysed in relation to the prevailing weather and
terrain type of the measuring sites.
From the analysis of the temperature and weather conditions, i.e., wind and
radiation, it was concluded that more than 90% of the frost situations, oc
curring during the study period, were of the radiation type. It was further
concluded that the variation among the studied stations was closely relate
d to the terrain type during these situations. Frost occurred most frequent
ly in narrow valleys, then in concave and flat locations. Elevated and conv
ex areas were found to have very few situations of radiation frost.
Local terrain information was used together with a calculated frost index f
or assessment of the spatial variation in frost risk. Furthermore, a grid n
et was applied to the study area and the pixels were given a terrain form o
f the type convex, slope, flat, wide concave or narrow concave according to
the dominating terrain curvature. For each pixel a frost index value was e
stimated from the recorded temperatures at the field stations. A cluster an
alysis was used to group the terrain types according to the index, whereby
six obvious clusters were obtained each with clearly differentiated frost i
ntensity. The analysis showed that this kind of treatment is a suitable met
hod for assessing the spatial variation in frost risk. (C) 2000 Elsevier Sc
ience B,V. All rights reserved.