Observation and analysis of spatio-temporal characteristics of surface ozone and carbon monoxide at multiple sites in the Kathmandu Valley, Nepal

Abstract. Residents of the Kathmandu Valley experience severe particulate and gaseous air pollution throughout most of the year, even during much of the rainy season. The knowledge base for understanding the air pollution in the Kathmandu Valley was previously very limited but is improving rapidly due to several field measurement studies conducted in the last few years. Thus far, most analyses of observations in the Kathmandu Valley have been limited to short periods of time at single locations. This study extends the past studies by examining the spatial and temporal characteristics of two important gaseous air pollutants (CO and O 3 ) based on simultaneous observations over a longer period at five locations within the valley and on its rim, including a supersite (at Bode in the valley center, 1345 m above sea level) and four satellite sites: Paknajol (1380 m a.s.l.) in the Kathmandu city center; Bhimdhunga (1522 m a.s.l.), a mountain pass on the valley's western rim; Nagarkot (1901 m a.s.l.), another mountain pass on the eastern rim; and Naikhandi (1233 m a.s.l.), near the valley's only river outlet. CO and O 3 mixing ratios were monitored from January to July 2013, along with other gases and aerosol particles by instruments deployed at the Bode supersite during the international air pollution measurement campaign SusKat-ABC (Sustainable Atmosphere for the Kathmandu Valley – endorsed by the Atmospheric Brown Clouds program of UNEP). The monitoring of O 3 at Bode, Paknajol and Nagarkot as well as the CO monitoring at Bode were extended until March 2014 to investigate their variability over a complete annual cycle. Higher CO mixing ratios were found at Bode than at the outskirt sites (Bhimdhunga, Naikhandi and Nagarkot), and all sites except Nagarkot showed distinct diurnal cycles of CO mixing ratio, with morning peaks and daytime lows. Seasonally, CO was higher during premonsoon (March–May) season and winter (December–February) season than during monsoon season (June–September) and postmonsoon (October–November) season. This is primarily due to the emissions from brick industries, which are only operational during this period (January–April), as well as increased domestic heating during winter, and regional forest fires and agro-residue burning during the premonsoon season. It was lower during the monsoon due to rainfall, which reduces open burning activities within the valley and in the surrounding regions and thus reduces sources of CO. The meteorology of the valley also played a key role in determining the CO mixing ratios. The wind is calm and easterly in the shallow mixing layer, with a mixing layer height (MLH) of about 250 m, during the night and early morning. The MLH slowly increases after sunrise and decreases in the afternoon. As a result, the westerly wind becomes active and reduces the mixing ratio during the daytime. Furthermore, there was evidence of an increase in the O 3 mixing ratios in the Kathmandu Valley as a result of emissions in the Indo-Gangetic Plain (IGP) region, particularly from biomass burning including agro-residue burning. A top-down estimate of the CO emission flux was made by using the CO mixing ratio and mixing layer height measured at Bode. The estimated annual CO flux at Bode was 4.9 µ g m −2 s −1 , which is 2–14 times higher than that in widely used emission inventory databases (EDGAR HTAP, REAS and INTEX-B). This difference in CO flux between Bode and other emission databases likely arises from large uncertainties in both the top-down and bottom-up approaches to estimating the emission flux. The O 3 mixing ratio was found to be highest during the premonsoon season at all sites, while the timing of the seasonal minimum varied across the sites. The daily maximum 8 h average O 3 exceeded the WHO recommended guideline of 50 ppb on more days at the hilltop station of Nagarkot (159 out of 357 days) than at the urban valley bottom sites of Paknajol (132 out of 354 days) and Bode (102 out of 353 days), presumably due to the influence of free-tropospheric air at the high-altitude site (as also indicated by Putero et al., 2015, for the Paknajol site in the Kathmandu Valley) as well as to titration of O 3 by fresh NO x emissions near the urban sites. More than 78 % of the exceedance days were during the premonsoon period at all sites. The high O 3 mixing ratio observed during the premonsoon period  is of a concern for human health and ecosystems, including agroecosystems in the Kathmandu Valley and surrounding regions.