Satellite-based evidence highlights a considerable increase of urban tree cooling benefits from 2000 to 2015

Tree planting is a prevalent strategy to mitigate urban heat. Tree cooling efficiency (TCE), defined as the temperature reduction for a 1% tree cover increase, plays an important role in urban climate as it regulates the capacity of trees to alter the surface energy and water budget. However, the spatial variation and more importantly, temporal heterogeneity of TCE in global cities are not fully explored. Here, we used Landsat- based tree cover and land surface temperature (LST) to compare TCEs at a reference air temperature and tree cover level across 806 global cities and to explore their potential drivers with a boosted regression tree (BRT) machine learning model. From the results, we found that TCE is spatially regulated by not only leaf area index (LAI) but climate variables and anthropogenic factors especially city albedo, without a specific variable dominating the others. However, such spatial difference is attenuated by the decrease of TCE with tree cover, most pronounced in midlatitude cities. During the period 2000- 2015, more than 90% of analyzed cities showed an increasing trend in TCE, which is likely explained by a combined result of the increase in LAI, intensified solar radiation due to decreased aerosol content, increase in urban vapor pressure deficit (VPD) and decrease of city albedo. Concurrently, significant urban afforestation occurred across many cities showing a global city-scale mean tree cover increase of 5.3 +/- 3.8% from 2000 to 2015. Over the growing season, such increases combined with an increasing TCE were estimated to on average yield a midday surface cooling of 1.5 +/- 1.3 degrees C in tree-covered urban areas. These results are offering new insights into the use of urban afforestation as an adaptation to global warming and urban planners may leverage them to provide more cooling benefits if trees are primarily planted for this purpose.

Automated summary

Tree planting is an effective strategy for mitigating urban heat. This study examines the spatial variation and temporal heterogeneity of tree cooling efficiency (TCE) in global cities, and identifies the potential drivers of TCE using remote sensing data and a machine learning model. The results show that TCE is regulated by factors such as leaf area index, climate variables, and anthropogenic factors, and there is a decreasing trend of TCE with tree cover.