Limitations in the accuracy of satellite-derived Park Cooling Distances revealed by high-resolution airborne thermal data: A case study from Soest, Germany.
Abstract
Urban parks are widely regarded as important mitigation elements for reducing urban heat stress, and their cooling influence is commonly quantified using Park Cooling Distance (PCD) derived from satellite-based land surface temperature (LST) data. However, the physical interpretation and robustness of these metrics remain uncertain because moderate-resolution satellite imagery may smooth fine-scale thermal heterogeneity and confound park-induced cooling with surrounding urban structure. This study critically evaluates the robustness and physical interpretation of satellite derived Park Cooling Distance (PCD) using very-high-resolution airborne thermal imagery in Soest, Germany. A multi-scale dataset consisting of 11 Landsat 8/9 summer scenes (30 m resolution) and airborne thermal imagery (20 cm resolution) from comparable meteorological conditions was analyzed. PCD and Park Cooling Intensity (PCI) were calculated using polynomial decay models for undifferentiated land surface temperature (LST) and also for ground-impervious surface temperatures which are more relevant to pedestrian heat exposure. Results show that isotropic PCD derived from Landsat and airborne data are similar (∼310 m). But directional PCD analysis showed distinct anisotropies with PCDs ranging from 165 to 401 m for both platforms. The analyses of ground-impervious surfaces revealed substantially shorter cooling distances (<200 m) and pronounced fine-scale thermal variability, largely caused by instantaneous shading. Multiple regression analyses of directional PCD segments demonstrated that vegetation cover, building density, building height, and shadow fractions explain most spatial temperature gradients, whereas distance from the park contributed only marginally. The results suggest that satellite-derived PCD may be strongly influenced by urban structure gradients surrounding parks. Consequently, part of the cooling distance identified in coarse-resolution thermal imagery may reflect underlying urban spatial heterogeneity in addition to physical park cooling processes. Similarly, satellite-derived PCD metrics may substantially overestimate effective street level cooling benefits from shading which are most relevant for human thermal comfort.