Can your Smartwatch Measure Ambient Air Temperature?

Abstract

Urban overheating, exacerbated by rapid urbanization and global climate change, stands as a critical environmental challenge for our cities. Despite ongoing efforts to collect high-resolution datasets, a gap persists in human-centric monitoring of urban heat, specifically focusing on the immediate surroundings of individuals as they navigate urban environments. This study aims to develop a reliable prediction model for air temperature in dynamic outdoor settings, providing a comprehensive understanding of how individuals are affected by heat at any given moment and location. The wearable monitoring method, as introduced in Nazarian et al. (2021), integrates environmental and physiological responses to offer a more holistic view of the human experience in urban areas. Smartwatches equipped with sensors measure air temperature and relative humidity at the wrist, along with skin temperature and heart rate. These measurements are then compared with data from a highly-accurate mobile weather station—MaRTy cart—which captures six-directional radiation, wind speed, ambient temperature, and relative humidity. A prediction model for ambient air temperature is proposed based on the wearable datasets. Predicting air temperature using wearable devices introduces a new, comprehensive, and personalized approach to measuring urban heat impacts in cities. Results indicate that ambient air temperature can be predicted with 95 percent accuracy using data from wrist-mounted wearables, achieving a target attainment within ±0.45 °C. In the air temperature prediction model, variables such as air temperature at the wrist (Taw), skin temperature (Ts), and relative humidity at the wrist (RHaw) demonstrated the most significant impact, while Heart Rate (HR) and walking speed (S) had the least influence. These findings underscore the reliability and accuracy of the wearable methodology in predicting air temperature.