1. Our aim was to quantify the effects of a warmer winter climate on the egg mortality (and, consequently, the risk of an outbreak) of Epirrita autumnata, a serious defoliator of mountain birch.
2. The egg survival of E. autumnata was studied in northernmost Finland from 1993 to 1996, when minimum temperatures were also recorded. Utilizing a digital elevation map with a GIS (Geographical Information System), models were developed for egg survival and minimum temperature, taking into account altitude, local topography and temperatures at a meteorological station in the study area.
3. Areas of high egg mortality of E. autumnata can be modelled in the GIS with reasonable accuracy by the model describing areal distribution of minimum winter temperatures, taking account of the fact that the critical limit for egg survival of E. autumnata in midwinter is about −35·5°C. The approach can be applied in short-term forecasting of outbreak areas when previous year population densities and minimum winter temperatures at meteorological stations are known.
4. In our study area, from 1961 to 1991 more than one-third of the years were egg-killing winters in the mixed birch-pine forests zone, and 15% of years in the birch forest zone. In order for an outbreak to take place, at least three consecutive seasons with a high population growth rate are needed; in the mixed birch-pine forests the possibilities of an E. autumnata outbreak during the period 1961–91 were very rare, while in the birch forest zone outbreak possibilities were also noticeably restricted by winter minimum temperatures.
5. If winter warming takes place according to the scenarios presented, by the middle of the next century the birch-growing areas in our study area protected by low temperatures from outbreaks of E. autumnata will be only one-third as great as those of today, and by the end of the next century only one-tenth of those of today.
6. The models and methods described in this paper can be applied to other ecological processes. The modelling strategies and techniques described here serve as intermediate tools for linking large scale climate models (general circulation models, GCMs) and ecological processes happening at a smaller scale.