CR 13:125-139 (1999)  -  doi:10.3354/cr013125

ABSTRACT: This study deals with the magnitude of influence of various geographical factors and meteorological variables on the air temperature difference between a park and a built-up area (DeltaT_u-p). The statistical analysis focuses on the time of nocturnal maximum DeltaT_u-p in summer in Göteborg, Sweden, during May to October 1995. The geographical factors include sky view factor, height above sea level and distance from the park border. The meteorological variables considered include wind, cloud cover and type, global radiation, air temperature, subsurface temperature and humidity. Principal component analysis is used to identify temporal and spatial patterns of the temperature anomaly along a transect across the built-up area and green area. The first principal loading is found to explain 71% of the total variance. It represents a kind of mean pattern and depicts the dominant distribution of the anomaly along the transect, which is interpreted as being created by the differences in surface characteristics. Distance from the park border accounts for 86% of the spatial variation in air temperature. The magnitude of this pattern is mainly modulated by the average wind speed and the average cloud index, i.e. cloud cover and type, from sunset until the time of interest, i.e. 27 and 13% respectively of the variation in DeltaT_u-p can be explained by these variables. Subsurface temperature and urban-park vapour pressure difference can explain minor parts of the variation in DeltaT_u-p. The wind direction affected both the spatial pattern and the magnitude of DeltaT_u-p. The influence on magnitude under different wind directions was probably caused by the difference in relative sizes between the warm built-up area and the colder park and suburban area.

KEY WORDS: Air temperature · Wind · Cloud · Global radiation · Air humidity · Sky view factor · Principal component analysis · Summer