CR 10:143-153 (1998)  -  doi:10.3354/cr010143

ABSTRACT: Observations from the Norwegian Arctic show positive trends in annual mean temperatures from 1912 to the 1930s and from the 1960s to 1996. Between these periods there was a negative trend, and there is no statistically significant trend in the record as a whole. The present temperature is approximately the same as in the 1920s, and lower than during the 1930s and 1950s. Spring is the only season which shows a statistically significant warming from 1912 to 1996. Annual precipitation, on the other hand, has increased in the Norwegian Arctic. At Spitsbergen the measurements show a statistically significant increase in annual and in spring, summer and autumn precipitation. Monthly values of mean sea level pressure of 4 grid points were used to develop models for monthly mean temperature and monthly precipitation at Spitsbergen. During the period 1912 to 1993 the temperature model accounts for 30 to 45% of the variance in the seasonal mean temperatures. The correlation between observed and modelled values is at a minimum in the summer and at a maximum in the autumn. The precipitation model accounts for 15 to 35% of the variance in seasonal precipitation sums. The correlation between observed and modelled values is lowest in winter, when the problems with drifting and blowing snow are greatest. Even though the observed and modelled seasonal values in most cases are better correlated for temperature than for precipitation, the precipitation model accounts for more of the decadal scale variability and long-term trends. The precipitation model reproduces the observed positive precipitation trends on both a seasonal and annual basis. Concerning decadal scale variability, most of the main observed features are also modelled satisfactorily. It is concluded that the major observed features concerning decadal scale variability and trends in precipitation at Spitsbergen are connected to variability in the atmospheric circulation pattern. The temperature model reproduces reasonably well the observed positive trend during the last 3 decades of the series. The very low temperature before 1920 and the high values in the 1930s and the 1950s, on the other hand, are not modelled satisfactorily. Thus, while the temperature increase of the later decades may mainly be explained as a result of changes in advection, the temperature increases in the Norwegian Arctic from the beginning of the measurements to the 1930s cannot be explained in this way.

KEY WORDS: Climate variation · Arctic · Atmospheric circulation · Temperature · Precipitation