Local topography controls temperature regimes in subarctic fell area, northern Finland

Climate is the most important factor influencing the relationships between earth surface processes, soil, vegetation and hydrological properties at broad geographical scales in subarctic areas. The basic idea is that climate, as a source of energy and moisture, acts as the primary control for the geographical systems. However, the primary controls over the climatic effects change with scale. At the local scale, the landscape pattern is controlled by microclimate and ground conditions, especially soil moisture availability.

In this study a topography-based generalized linear model (GLM) of the spatial distribution of local temperature is presented in relation to a number of topographical variables using GIS tools. The study region of ca. 4000 km2 is located in the zone of discontinuous permafrost in the northern part of Finnish Lapland (69°30'N, 26°E). Elevations in the area range from 60 m to 640 m a.s.l. The specific aims of this study are to determine: 1) the correlation between single topographical variables and local temperature, 2) the extent to which the local temperature regime can be modelled solely on the basis of topographical variables; and 3) the relative importance of different topographical variables in multivariate models in influencing the distribution of local temperature. The measurements were based on 36 temperature loggers.

In the study area, the mean temperature varied from –1.0 to 0.4ºC and the minimum temperatures varied from –22.7 to –38.0ºC from 15th August 2001 to 14th August 2002. The lowest temperature of -38 ºC was measured in the bottom of the Utsjoki river valley at the elevation of 78 m above sea level, whereas the highest minimum temperature of –22.65ºC was measured at the elevation of 426 m at the top of the fell summit. Separate testing of the relationship between the minimum temperature and all explanatory variables based on GLMs showed that only two variables of the four explanatory variables were significantly (p<0.05) related to the minimum temperature, namely elevation and wetness index. The multivariate GLM model indicated that the minimum temperature increased with elevation and decreased with wetness index. The GLM models explained 86% of the variation in the minimum temperature. It should be noted that the elevation variable explained 76% of the variation alone. Finally, as a tentatively exercise we analysed the distribution of main land cover types in relation to modelled minimum temperatures of the winter 2001-2002 in the studied fell area. The minimum temperatures between the land cover types varied considerably: the minimum temperature of the alpine vegetation zone was only ca. -26°C, whereas the temperatures of the river valleys and low-lying peat plateaus were often more than ten degrees colder.

Air temperature is a crucial environmental driver for frost processes although usually only indirect information, such as a model of solar radiation or elevation, is available for spatial analyses. In the study area the activity of frost processes is concentrated on the lower altitudes. The negative association of frost activity and elevation can be explained by climatological factors. During the winters, the valleys and lower altitudes are colder due to temperature inversion and cold air drainage. Moreover, the drier soil conditions at higher altitudes are unsuitable for frost activity.

In conclusion, simple topographical parameters produced from digital elevation model can explain almost 90% of the variation in yearly minimum temperatures in subarctic areas. We conclude that temperature information derived from spatial modelling has great potential in local scale climatological studies, particularly in remote areas with scattered meteorological observation network. The application of a topoclimatic modelling approach as used here provides an efficient method to extrapolate climatological information to unsurveyed areas.