Novel Sensor Networks and Methods for Urban Greenhouse Gas Monitoring

As more than 70% of fossil fuel-based carbon dioxide (CO₂) is emitted in urban areas, urban greenhouse gas (GHG) monitoring plays a crucial role in achieving emission reduction goals. Nowadays, most cities rely on downscaled national data to calculate their total emissions, or on bottom-up methods, where emission factors are multiplied with activity data, such as energy consumption, economic activity, and traffic density. However, at the scale of individual cities, errors of 50-100% in fossil fuel CO₂ emission estimates have been reported. Furthermore, in terms of methane (CH₄), urban emissions are suspected to be substantially underestimated by inventory methods.

Measurements of atmospheric GHG concentrations offer opportunities to identify unknown emission sources and to address biases in urban emission inventories. Urban areas however pose significant challenges to measurement-based emissions quantification, due to the heterogeneous geometry of the cities and the complex atmospheric circulation in this environment. Therefore, representative measurements combined with sophisticated atmospheric models are vital to arrive at robust estimates of urban GHG emissions.

In Munich, Germany, we created an integrated measurement and modeling framework to better understand urban GHG emissions. MUCCnet (Munich Urban Carbon Column network) is a permanent urban GHG sensor network, consisting of five automated ground-based remote sensing systems. It is based on the differential column method (DCM), which features high precision and is relatively insensitive to vertical redistribution of tracer mass and surface fluxes upwind of the city, thus providing favorable input for urban flux inversions. MUCCnet serves to validate satellite measurements, to independently monitor local GHG emissions over the long term, and to detect unknown emission sources.

Using the Munich Oktoberfest as an example, large festivals have been identified as a potentially significant source of fossil fuel CH₄, despite likely being poorly represented in CH₄ emission inventories. In a recent measurement campaign in Hamburg, where DCM was deployed, we have found several significant anthropogenic sources, such as refineries and a farm as well as large area sources such as the River Elbe, whose CH₄ emissions are not yet included in the standard inventories or are highly underestimated.

To assess emissions from the measured concentrations, inverse modeling is an essential tool. We developed a novel Bayesian inversion framework to inversely model emissions using column measurements. We further use mobile in-situ measurements, isotopic measurements, and eddy covariance measurements to enhance the prior knowledge of the emission map.

Within the ICOS Cities project (PAUL), we have been improving the GHG emission assessments in Munich by refining the prior emission localization and timing and by adding additional monitoring capacities, including 100 street-level low-cost CO₂ sensors as well as 20 roof-level mid-cost CO₂ sensors based on the NDIR measurement principle. In addition, we are establishing an autonomous NOx, PM, CO and O₃ network in Munich with 50 stand-alone sensor nodes. This network is used to study the spatial distribution of urban air pollutants and to assess co-emitted species of CO₂ emitters.