Urban heat islands (UHIs) significantly impact human health, energy consumption, and urban infrastructure. While urban trees are known to mitigate UHIs through shading and transpiration, their effectiveness compared to treeless green spaces and across diverse climates remains insufficiently explored. This research addresses this gap by analyzing high-resolution satellite data on land surface temperatures (LSTs) in 293 European cities. The study aims to quantify the cooling effect of urban trees relative to other land cover types and to assess how this effect varies across different climatic zones. Understanding these variations is crucial for developing effective and climate-adapted urban heat mitigation strategies. Previous studies have explored the impact of vegetation on urban temperatures, but often lacked the spatial resolution and climatic diversity necessary to distinguish between the effects of different vegetation types (e.g., trees vs. grass) and their regional variation. Some studies focused on specific regions, overlooking broader continental patterns. This research overcomes these limitations by using a large-scale, high-resolution dataset to offer a more comprehensive understanding of the role of urban trees in moderating urban heat across diverse European climates.

Existing literature highlights the importance of urban vegetation in mitigating urban heat islands. Studies using surface urban heat island (SUHI) data and energy-balance models have shown varying cooling effects of vegetation across different climates, particularly limited effectiveness in tropical and dry climates. However, distinguishing the effects of different vegetation types (trees vs. treeless green spaces) on temperature using SUHI data, which often compares city and surrounding temperatures, can be challenging. Research focusing on land-use/land-cover (LULC) changes has demonstrated the climatic impacts of different LULC types, but often lacks focus on the urban environment and regional variations in the effectiveness of urban trees and green spaces. Studies explicitly examining forest types in urban contexts often focus on specific regions, neglecting the broader continental-scale patterns. The current study builds upon this existing research by providing a continental-scale analysis with high-resolution data, allowing for a more nuanced understanding of the regional and climatic variations in the cooling effects of different urban vegetation types.

The study utilized a unique high-resolution dataset of 120,285 Landsat scenes and land cover data from 293 European cities. Land surface temperature (LST) data were obtained from Landsat and Aster satellites, with a focus on Landsat due to greater data availability. Data were pre-processed to account for factors such as time of day and cloud cover. Land cover data categorized areas into urban fabric, urban trees, treeless urban green spaces, rural forests, and rural pastures. Generalized Additive Models (GAMs) were calibrated for each city to estimate temperature differences between these land cover types, considering factors such as LULC fraction, elevation, and aspect. The models allowed for comparisons of LST differences at approximately 10:15 a.m. across different LULC types. Analyses were conducted to compare temperature differences during average summer conditions and hot extremes, considering variations across seasons and regions. Additional data on albedo and evapotranspiration (ET) from MODIS were used to explore the biophysical processes underlying the observed temperature patterns. Multiple linear regression models were fitted to estimate the contribution of different LULC types to albedo and ET. The spatial variation of LST differences was analyzed by calculating trends using GAMs and decomposing LST differences based on background temperature from E-OBS data. The study also accounted for potential limitations such as temporal resolution, cloud cover, spatial arrangement of LULC types, and the exclusion of scattered urban trees.

The study revealed significant regional variation in the cooling effect of urban trees. During hot extremes, urban trees showed lower LSTs than urban fabric in most European cities, except for cities in Southern Turkey, the Mediterranean, and the Iberian Peninsula. The largest LST difference between urban trees and urban fabric was observed in Central Europe (-12 to -8 K), while the difference was smaller in Southern Europe (-4 to 0 K). The cooling effect during hot extremes was not always consistent with average summer conditions; in some regions (e.g., Southern Europe), the cooling effect decreased during hot extremes, while in others (e.g., Scandinavia), it remained similar or even increased. Seasonal variations in cooling were also observed, with higher cooling in spring than summer in Southern European cities, but higher cooling in summer than spring in other regions. Treeless urban green spaces were found to be far less effective in reducing LSTs compared to urban trees (approximately 2-4 times lower cooling effect). The temperature differences between urban trees and urban fabric were closely correlated with the differences between rural forests and urban fabric, suggesting similar underlying biophysical processes. The study found a strong correlation between LST differences and evapotranspiration (ET) levels, particularly for forests. Albedo played a smaller role in explaining LST differences, although it might be more significant in dryer areas. Significant temperature differences were observed between tree-covered areas and green spaces, and between rural forests and pastures, highlighting the importance of vegetation type in determining cooling effects.

The findings demonstrate the substantial cooling potential of urban trees in mitigating urban heat, particularly in Central Europe. The regional variation in cooling effects underscores the importance of considering climatic context when designing urban heat mitigation strategies. While urban trees and forests generally provided cooling, treeless green spaces showed limited or even warming effects in some Southern European regions. The observed decrease in the cooling effect of trees during hot extremes in Southern Europe, potentially linked to reduced stomatal conductance due to water stress, highlights the vulnerability of these regions to future drying. The strong correlation between LST differences and ET emphasizes the importance of maintaining sufficient soil moisture to maximize the cooling benefit of trees. The study also acknowledged the limitations of solely focusing on LST, suggesting that future research should investigate the relationship between LST and air temperature to better understand the impact on human thermal comfort and energy consumption. The limitations of the data regarding the representation of scattered trees and the potential influence of urban morphology also need further research.

This study provides valuable insights into the effectiveness of urban trees in reducing LSTs across Europe, demonstrating significant regional and seasonal variations. Urban trees offer substantially higher cooling potential than treeless green spaces, particularly in Central Europe. However, the cooling effect during heat extremes is less pronounced in Southern Europe due to water stress. Future research should focus on integrating additional data (e.g., leaf area index, heat fluxes, air temperature data) to refine the understanding of biophysical processes and their impact on urban microclimates. Further investigation is needed into the effects of scattered trees and urban morphology to improve the accuracy of modeling efforts.

The study's reliance on remote sensing data introduces limitations regarding temporal resolution and the potential impact of cloud cover on LST measurements. The early observation time (around 10:15 a.m.) could influence predictions of LST differences during hot extremes. The analysis might not fully account for the effects of spatial arrangement of LULC types and the exclusion of scattered urban trees limits the generalizability of the findings. Furthermore, the focus on LST may not fully capture the complex effects of urban trees on air temperature and human thermal comfort.