Missing methane emissions from urban sewer networks.

Methane emissions from sewer networks are an important source of anthropogenic greenhouse gases (GHGs) but are not currently reflected in the national GHG inventory. We found significant CH4 emissions of approximately 573 [395-831] CH4 t y-1 from sewer networks in the old residential and commercial areas of Seoul (Gwanak district) using an electric vehicle-based atmospheric GHG monitoring platform. The majority of ethane-to-methane ratios (<0.005) from the observations further suggest that distinctive CH4 emissions from sewer networks are likely related to microbial activity rather than to simple natural gas leakage. Because over 90% of the sewer network in Seoul is a gravity drain type of combined sewer network, where both wastewater and stormwater flow through the same pipes, resulting in the generation of methane emissions from the microbial activity and the manholes and rain gutters, which are directly connected to the combined sewer networks are major sources of atmospheric methane emissions. This study suggests that appropriate treatment of sewer networks can mitigate missing methane emissions in cities that were not originally included in GHG inventory of South Korea.

Machine learning based estimation of urban on-road CO2 concentration in Seoul.

The urban on-road CO2 emissions will continue to increase, it is therefore essential to manage urban on-road CO2 concentrations for effective urban CO2 mitigation. However, limited observations of on-road CO2 concentrations prevents a full understanding of its variation. Therefore, in this study, a machine learning-based model that predicts on-road CO2 concentration (CO2traffic) was developed for Seoul, South Korea. This model predicts hourly CO2traffic with high precision (R2 = 0.8 and RMSE = 22.9 ppm) by utilizing CO2 observations, traffic volume, traffic speed, and wind speed as the main factors. High spatiotemporal inhomogeneity of hourly CO2traffic over Seoul, with 14.3 ppm by time-of-day and 345.1 ppm by road, was apparent in the CO2traffic data predicted by the model. The large spatiotemporal variability of CO2traffic was related to different road types (major arterial roads, minor arterial roads, and urban highways) and land-use types (residential, commercial, bare ground, and urban vegetation). The cause of the increase in CO2traffic differed by road type, and the diurnal variation of CO2traffic differed according to land-use type. Our results demonstrate that high spatiotemporal on-road CO2 monitoring is required to manage urban on-road CO2 concentrations with high variability. In addition, this study demonstrated that a model using machine learning techniques can be an alternative for monitoring CO2 concentrations on all roads without conducting observations. Applying the machine learning techniques developed in this study to cities around the world with limited observation infrastructure will enable effective urban on-road CO2 emissions management.