Rapid urbanization and climate change are exacerbating the urban heat island effect (UHI), leading to higher temperatures in cities compared to surrounding areas. This UHI effect negatively impacts human health, energy consumption, and air pollution, posing a significant challenge to sustainable urban development. The UTE is a complex system influenced by numerous interacting factors within a socio-ecological system (SES), encompassing both social (e.g., population density, impervious surfaces) and ecological (e.g., green and blue spaces) subsystems. While remote sensing provides valuable data on land surface temperature (LST) and UHI intensity, understanding the long-term evolution of UTE and its relationship with SES dynamics remains limited. Previous studies often use spatial substitution for time, hindering a complete understanding of the dynamic interactions within SESs. This study aims to bridge this gap by analyzing long-term satellite data (2000–2021) across 136 Chinese cities to explore the spatio-temporal evolution of UTE and its connection to SES interactions, ultimately informing the development of holistic heat mitigation strategies.

Existing literature highlights the escalating UHI effect globally, particularly in rapidly urbanizing regions like Asia and Africa. Studies using remote sensing data have characterized spatial and temporal variations in LST and UHI intensity, revealing greater daytime than nighttime intensity and stronger anomalies during summer months. However, interannual variations and the long-term evolutionary patterns of UTE remain less understood. While some studies suggest a potential link between UHI intensity and stages of urban development, potentially following an inverted U-curve (Environmental Kuznets Curve), this requires further validation. Furthermore, most analyses focus on individual social or ecological factors rather than their complex interactions within the SES framework. The lack of time-series data on SES factors further limits a comprehensive understanding of UTE dynamics and the development of integrated mitigation strategies.

This study analyzed satellite data from MODIS-Terra for 136 Chinese urban areas (population > 1 million) from 2000 to 2021. Data included MOD11A2 LST (8-day composite, 1km resolution) and MOD13A2 EVI (16-day composite, 1km resolution). Nighttime lighting data (1km resolution) were integrated from DMSP-OLS and SNPP-VIIRS. Population density data were sourced from WorldPop, and land cover data (30m resolution) from the China Land Cover Dataset (overall accuracy 79.31%). The UTE intensity (ΔT) was calculated as the difference between LST in urban and reference areas (selected based on land cover and EVI, located within 10–30 km buffers). Summer daytime ΔT (June-August) was analyzed for each year. Ordinary least squares linear regression and piecewise linear regression were used to analyze ΔT trends and identify inflection points indicating phase changes in UTE evolution. Geographic detector models were employed to quantify the effects of social (population density, nighttime light adjusted by vegetation – NLVA, impervious area percentage), ecological (ecological land percentage, ΔEVI), and their interactions on ΔT trends. A social-ecological status index (SESI) was developed to reflect SES structure and status, considering resilience (factors negatively correlated with ΔT) and pressure (factors positively correlated with ΔT) indicators. SESI was used to assess the impact of SES dynamics on UTE. Piecewise linear regression was applied to examine the relationship between SESI and ΔT.

The study revealed a nonlinear intensification of UTE across the 136 cities, with an average ΔT of +2.83 °C from 2000–2021. Spatial heterogeneity was evident, with higher ΔT changes in southeastern cities. A significant increasing trend in mean ΔT was observed, yet inflection points indicating shifts in the rate or direction of ΔT change were present in a substantial proportion (30.88%) of cities. A larger number of cities showed a shift from increasing to decreasing ΔT trends in recent decades, indicating increasing mitigation efforts. Three phases in UTE evolution were identified: rapid increase, slow increase, and decrease phases. Social factors (population density, NLVA, impervious area percentage) showed a positive correlation with ΔT, while ecological factors (ecological land percentage, ΔEVI) showed a negative correlation. Geographic detector analysis revealed economic development (NLVA) and vegetation quality (ΔEVI) as the most influential social and ecological factors, respectively. Significant interactions between social and ecological factors further emphasized the need for an integrated approach. The SESI significantly differed across UTE phases, with the decrease phase exhibiting the highest SESI values, indicating that a higher SES status contributes to UTE mitigation. Piecewise linear regression revealed a threshold effect of SESI on ΔT, with greater mitigation effects at higher SESI values above the threshold (0.414). Spatial analysis of recent SESI showed heterogeneity, with higher values in northeastern China and lower values in southwestern China. Cities with decreasing UTE trends had significantly higher SESI values compared to cities with increasing trends, highlighting the role of SES management in UTE mitigation.

The findings challenge the common perception of continuously increasing UTE intensity, demonstrating the potential for successful mitigation through policy and measures such as "eco-city", "low-carbon city", and "sponge city" initiatives focusing on urban greening and nature-based solutions. The observed three-phase UTE evolution pattern aligns with the Environmental Kuznets Curve, illustrating a potential trade-off between socioeconomic development and environmental protection in the initial stages of urbanization, followed by a shift towards synergistic development where both factors contribute to UTE mitigation. The study’s emphasis on the interaction of social and ecological factors challenges previous findings focusing on individual factors and underscores the importance of holistic, integrated strategies for urban heat mitigation. The results indicate that even with continued socioeconomic growth, ecological enhancement is achievable, leading to a virtuous cycle benefiting both society and the environment. Examples like Beijing, exhibiting decreasing UTE despite economic growth, highlight the feasibility of win-win solutions.

This study reveals the nonlinear evolution of UTE and its strong relationship with SES dynamics, confirming the applicability of the Environmental Kuznets Curve. The key finding emphasizes the crucial role of integrated strategies that consider social and ecological interactions for effective UTE mitigation. Future research should focus on developing more comprehensive datasets encompassing a wider range of social and ecological factors to refine the SESI model and develop more targeted and effective heat mitigation strategies.

Data limitations, particularly regarding the availability of large-scale, long-term datasets on social and ecological factors, constrain the comprehensive characterization of SES interactions. The study's selection of five representative factors may not fully capture the complexity of SES dynamics. Inaccuracies in remote sensing datasets can introduce errors in UTE and SESI quantification. Future research should aim to overcome these limitations by developing more extensive and accurate datasets to achieve more robust results.