Contribution of environmental factors to temperature distribution at different resolution levels on the forefield of the Loven Glaciers, Svalbard
ABSTRACT The climate and its components (temperature and precipitation) are organised according to different spatial scales that are structured hierarchically. The aim of this paper is to explore the dependence between temperature and deterministic factors at different scales on a 10 km 2 study area...
| Published in: | Polar Record |
|---|---|
| Main Authors: | , |
| Format: | Article in Journal/Newspaper |
| Language: | English |
| Published: |
Cambridge University Press (CUP)
2007
|
| Subjects: | |
| Online Access: | http://dx.doi.org/10.1017/s003224740700678x https://www.cambridge.org/core/services/aop-cambridge-core/content/view/S003224740700678X |
| Summary: | ABSTRACT The climate and its components (temperature and precipitation) are organised according to different spatial scales that are structured hierarchically. The aim of this paper is to explore the dependence between temperature and deterministic factors at different scales on a 10 km 2 study area on the northwestern coast of Svalbard. A GIS was developed which contained three sources of information: temperature, remotely sensed imagery and digital elevation models (DEM), and derived raster data layers. The first layer, temperatures, was acquired at regularly observed temporal intervals from 53 stations. The second layer comprised remotely sensed images (aerial photography and SPOT imagery) and DEM data at 2 m and 20 m resolution, respectively. From these, a windowing procedure was applied to derive several spatial subsets of different spatial resolutions (6, 14, 30, 60, 140, and 300 m). The third layer comprised slope, aspect, and a theoretical solar radiation value derived from the DEM, and a vegetation index derived from the remotely sensed imagery. Linear regressions were then systematically conducted on the datasets, with temperature as the dependent variable, and each of the other data layers as the independent variables. By using graphical analysis, we link the correlation coefficients obtained for each factor, from the smallest spatial resolution (6 m) to the largest resolution (300 m). The results indicated that each explanatory variable and scale brings a specific contribution to changes in temperature. For example, the effect of elevation remains constant for all spatial resolutions, reflecting a quasi ‘non-scalar’ pattern of this variable. For other variables however, the effect of spatial scale can have a strong effect. In the case of solar radiation, a maximum of explanation was obtained for spatial resolutions of 14 m and 60 m; for vegetation index the optimum contribution was related to the 300 m resolution. Thus, different environment characteristics may have significant effects on changes in ... |
|---|