Abstract: Objective　Tropospheric nitrogen dioxide (NO2) is a primary atmospheric pollutant primarily generated from fuel combustion in industrial production, transportation, and residential heating. It plays a vital role in atmospheric photochemistry, contributing to the formation of ozone, secondary aerosols, and regional haze, and poses severe threats to human respiratory health and ecological systems. Due to its short atmospheric lifetime, NO2 exhibits strong spatial heterogeneity, making traditional point-based in-situ monitoring insufficient to capture regional distribution characteristics. Two-dimensional multi-axis differential optical absorption spectroscopy (2D MAX-DOAS), as a passive remote sensing technology, enables simultaneous observation at multiple azimuths and elevation angles, and has distinct advantages in obtaining large-scale horizontal and vertical distribution information of trace gases. This study focuses on the urban area of Taiyuan, China. Taiyuan is a typical industrial city with complex emission sources and unfavorable terrain for pollutant dispersion. By using 2D MAX-DOAS technology to investigate the horizontal distribution characteristics of tropospheric NO2 in Taiyuan, this work aims to validate the reliability of the remote sensing results by comparing with national ground monitoring stations, reveal the vertical variation pattern of NO2 concentration with elevation, identify the spatial correspondence between NO2 hotspots and potential emission sources, and analyze the weekly variation characteristics driven by industrial and traffic emissions. The findings are expected to provide technical support and data basis for regional air pollution source apportionment, refined monitoring, and precise control of NO2 pollution in industrial urban areas like Taiyuan. Methods　The field observation was conducted employing a ground-based 2D MAX-DOAS system in the downtown area of Taiyuan (37° 88′N, 112° 56′E) from September 4 to September 30, 2021. The instrument performed 360° full-circle scanning with 72 azimuths at 5° intervals and 8 elevation angles (1°, 2°, 4°, 6°, 8°, 15°, 30°, 90°), completing a full round scan in approximately 3 hours and 40 minutes. The observed spectral data were processed using the QDOAS software in the wavelength range of 338 – 370 nm to retrieve differential slant column densities (DSCDs) of NO2 and O4 simultaneously. Then based on the measured O4 DSCD, ambient barometric pressure, and temperature, the effective optical path length was calculated to convert NO2 DSCD into volume mixing ratio (VMR). Subsequently, the VMR was further converted into NO2 mass concentration (μg/m³) to facilitate direct comparison with ground-based monitoring data. The 2D MAX-DOAS derived NO2 concentrations were validated against two adjacent national urban air quality monitoring stations (Jiancaoping and Taoyuan). Spatial distribution maps of NO2 VMR at 1° elevation were generated to identify pollution hotspots, and the vertical gradient of NO2 was analyzed by comparing concentrations at different elevation angles. Finally, four typical azimuths were selected to explore the weekly variation patterns of NO2 and distinguish the impacts of industrial and traffic emissions. Results and Discussion　The horizontal distribution of NO2 at 1° elevation presented significant spatial differences across the monitoring domain. Relatively high NO2 concentrations were observed in the west, northwest, and southeast directions of the instrument site, with a maximum average concentration of 13.21 μg/m³. These high-pollution areas were spatially consistent with intensive industrial facilities (including machinery manufacturing plants and large-scale iron and steel enterprises) and high-traffic zones such as cross-river bridges, high-speed railway stations, and bus terminals. In contrast, the lowest concentration of NO2 (9.09 μg/m³) appeared in the eastern direction dominated by mountainous terrain with clean background air masses. The validation against ground national stations showed that the temporal variation trends of 2D MAX-DOAS measurements were in good agreement with in-situ data, with determination coefficients (R²) of 0.8123 for Taoyuan station and 0.7077 for Jiancaoping station, which validates the accuracy and reliability of the remote sensing results. NO2 concentrations exhibited a decreasing trend with the increase of elevation angles, indicating that NO2 was mainly concentrated in the near-surface layer in Taiyuan city, and similar horizontal distribution patterns were maintained at different heights. Analysis of weekly variations in four typical azimuths revealed a remarkable "weekend effect" in the two azimuths without industrial emission sources, where the NO2 concentrations on Sundays were 35% and 42% lower than the corresponding weekday averages, reflecting the dominant control of traffic emissions. However, the azimuths affected by continuous industrial operations did not show such a weekend reduction, and one industrial-influenced direction even showed a 19% increase in NO2 concentrations on Sundays due to uninterrupted factory emissions and increased weekend traffic in recreational and scenic areas. Compared with directly using slant column density, the conversion from DSCD to VMR effectively eliminated the interference of optical path differences and significantly improved the accuracy of source identification. Conclusion　This study confirms that 2D MAX-DOAS is an effective and accurate technical method for obtaining the horizontal distribution of tropospheric NO2 and identifying regional pollution sources. In the urban area of Taiyuan, the distribution of NO2 is jointly affected by industrial emissions and vehicle exhaust, with high-concentration zones spatially matching industrial parks and transportation hubs. The research results of vertical gradient distribution indicate that NO2 is predominantly confined in the near-surface atmospheric layer, which is closely related to ground-based emission characteristics. The contrasting weekly variation patterns between industrial and traffic-dominated areas reflect the different emission rhythms driven by human activities. The satisfactory consistency between MAX-DOAS observations and national station data highlights the complementary role of passive remote sensing in regional air quality monitoring. It is believed that the application of 2D MAX-DOAS can provide high-spatial-coverage data support for refined atmospheric environmental management, source localization, and targeted pollution control in industrial cities with complex emission structures.

Key words: horizontal observation, NO2, volume mixing ratio, Taiyuan