Application of Low-cost Sensor Network in Extreme Environment: A Case Study of Ozone Pollution in Tibetan Plateau

Low-cost sensor networks (LCSNs) have attracted widespread attention as valuable monitoring tools for environment science and air quality management, revolutionizing traditional air monitoring systems. However, previous studies concentrated on the urban pollution hotspots monitoring, leaving a gap in understanding the performance and application of LCSN for long-term regional monitoring in extreme environments.

This study presents a novel and challenging application of LCSN to investigate the causes and sources of ozone pollution in the unique and remote Tibetan Plateau, characterized by rapid temperature and humidity fluctuations and significant diurnal variations. Since 2023, a real-time and high-density LCSN with 40 sites has been gradually established in Tibet, covering 6 cities including Nagri and Lhasa. The sensors are designed to mitigate the influence of temperature and humidity from both hardware and algorithm adjustments. Calibration and side-by-side tests against reference-grade instruments are conducted to characterize the sensitivity and baseline of the sensors based on the variation in pollutant concentrations in the real-world atmosphere. The spatial-temporal characteristics, production and transport patterns of ozone were preliminarily interpreted in combination with meteorological data and the HYSPLIT model.

The results show that the optimized and calibrated low-cost sensors were consistent with standard equipment in long-term monitoring, indicating the accuracy and stability of sensors in the extreme plateau environment. LCSN can provide reliable real-time data with high temporal and spatial resolution to support the exploration of regional ozone and its precursor production and transport patterns. Tibet ozone pollution is dominated by regional contributions with favorable weather conditions, while locally generated pollution has little impact. Different cities and regions show variability in ozone pollution pattern. For instance, in Nagri, located near the western border, approximately 70% of ozone pollution is attributed to regional transport, which is affected by the cross-border transport accompanied by westerly winds; while Lhasa, with a larger population and more transportation and industrial activities, has an increased proportion of local contributions influenced by precursors. These findings reveal the robustness and applicability of LCSN in extreme environments, showcasing its potential to provide actionable insights into dynamic air pollution. This study highlights the opportunities and challenges of utilizing LCSN for air quality monitoring in remote regions and extreme environments, providing scalable solutions for global air quality management and sustainable development strategies.