Global climate change has precipitated recurrent extreme events, presenting formidable challenges to society and the environment (Tripathy et al., 2023; Cheng et al., 2023; Peng et al., 2024). These challenges encompass threats to human health, degradation of ecosystems, and heightened energy demands (Deroubaix et al., 2021; Kotcher et al., 2021; Outhwaite et al., 2022). Temperature, being the preeminent parameter within meteorological variables, serves as an instrument for monitoring climate fluctuations and is imperative for the formulation of policies and the implementation of appropriate response measures (Yang et al., 2020; Yin et al., 2023; Peng et al., 2020a). Nevertheless, the genuine awareness of human and their surroundings, denoted as human comfort, assumes greater significance in a comprehensive evaluation of the influence of environmental conditions (Gobo et al., 2022). Human-perceived cold or thermal stress is intricate, intimately linked to various meteorological variables. For instance, wind speed can either amplify or mitigate perceived body temperature, humidity can modulate the efficiency of evaporation, and solar radiation can elevate perceived temperature when exposed to sunlight (K. Zhang et al., 2023; Fahad et al., 2021). Consequently, while a solitary meteorological variable, namely temperature, remains crucial, an index that amalgamates multiple meteorological variables is better poised to mirror the authentic human perception of the ambient environment.
Till now, several indices pertaining to human comfort have been widely adopted, encompassing the heat index, wet-bulb temperature, and humidity index (Vargas Zeppetello et al., 2022; Freychet et al., 2020). The Universal Thermal Climate Index (UTCI), a novel index of human comfort, excels in portraying human responses to thermal and cold stress more accurately (Bröde et al., 2012). The UTCI hinges on the concept of equivalent temperature, defined as the temperature within a standardized reference environment, furnishes a more comprehensive and precise portrayal of human perceptions under diverse meteorological circumstances (Bröde et al., 2012). By integrating a gamut of meteorological variables, including temperature, humidity, wind speed, and solar radiation, the UTCI aptly characterizes comfort levels across varying environments (Park et al., 2014). As an advanced biometeorological index, the UTCI has objectivity in assessing the impact of the atmospheric milieu on the human organism (Zare et al., 2018). Currently, the UTCI is extensively employed in studies concerning short-term repercussions of atmospheric conditions on humans and urban bioclimatology and evaluations of the urban heat island effect (Hwang et al., 2022; Kyaw et al., 2023; S. Zhang et al., 2023). Consequently, the UTCI, with its incorporation of multiple meteorological variables and hallmark objectivity and comprehensiveness, can characterize the thermal and cold stresses experienced by humans well.
Several datasets encompassing human comfort indices have been produced for global or localized domains (H. Zhang et al., 2023; Dong et al., 2022). However, the quantity and the quality of UTCI datasets are insufficient, which hinders in-depth research and the application of the UTCI. The existing UTCI datasets predominantly exhibit low spatial resolutions, such as the ERA5-HEAT with a spatial granularity of 0.25° (encompassing the globe) and the HiTiSEA with a spatial granularity of 0.1° (encompassing East and South Asia) (Di Napoli et al., 2021; Yan et al., 2021). These prevailing UTCI datasets are often inadequate for urban and landscape scale investigations, given that these studies necessitate data of higher spatial resolution to accurately capture intra-urban meteorological variations and human perceptions (Peng et al., 2021; Yang et al., 2021; Cao et al., 2022). Therefore, the development of a UTCI dataset that is globally accessible, has a long time series, and has a high spatial resolution is imperative. This initiative will address the existing void in UTCI data availability and enhance the precision and practicability of the UTCI for urban and landscape-scale investigations.
To facilitate the widespread future applications of UTCI data, we have produced GloUTCI-M, a monthly UTCI dataset characterized by global coverage, a long time series, and high spatial resolution. This work involves establishing a systematic process for generating and describing the UTCI dataset and relying on machine learning models that incorporate multiple covariates as well as exploratory data analysis. Several key contents include (1) examining the relationship between the UTCI and various covariates, utilizing multiple machine learning models; (2) employing the optimal machine learning model to produce a monthly high-spatial-resolution UTCI dataset that spans the entire globe, known as GloUTCI-M; (3) analyzing the global spatiotemporal characteristics and pattern evolution of the UTCI based on GloUTCI-M; and (4) comparing GloUTCI-M with existing UTCI datasets.
