Abstract: Urban land cover composition and structure directly affect land surface temperature (LST) and urban heat island intensity, which has become a main focus of urban planning and urban ecology. Taking Beijing, Tianjin, and Shijiazhuang as study areas, urban land cover and LST were extracted and retrieved from Landsat-8 OLI/TIRS remote sensing images, respectively. Various geospatial approaches, including urban-rural gradient analysis, multi-scale analysis and landscape pattern index, were used to analyze the impacts of urban land cover characteristics on LST. The results showed that the area of impervious surface area (ISA) in Beijing, Tianjin, and Shijiazhuang was 1103.86, 610.33, and 326.75 km², respectively. LST of bare soil was the highest, followed by ISA and green vegetation (GV), and that of water body was the lowest. The ISA ratio (positively) and GV ratio (negatively) significantly correlated with LST along the urbanrural gradient. LST would decrease by 0.4-0.6 ℃ with an 10% increase of GV. With the increase of spatial analysis scale, the correlations between the proportion of urban ISA and GV and LST first rised and then stabilized. The patch size, shape and aggregation of urban ISA and GV all had significant effects on LST, with the aggregation having the strongest effect. Thus, urban GV plays an important role in urban heat mitigation. We suggest that reasonable layout and ratio of ISA and GV in urban planning should be implemented to improve urban ecosystem services and environmental quality of human settlement.

Key words: fertilization pattern, greenhouse vegetable, soil enzyme activities.