1.1 Background
Urban areas are very significant sources of atmospheric pollutants and greenhouse gases (GHGs), including carbon dioxide (CO2). In 2020, it was estimated that urban areas were responsible for approximately 70 % of global CO2 emissions (Lwasa et al., 2022). Increased population densities and intensive energy consumption can result in CO2 urban domes, where CO2 is enhanced by a few parts per million (ppm) to tens of ppm in and around the urban extent (Xueref-Remy et al., 2023). Reducing CO2 emissions and the decoupling of carbon emissions from economic growth are priorities for most national and subnational governments in order to avoid some of the worst negative consequences of anthropogenic climate change (IPCC, 2023). The importance of CO2 emissions from urban areas has driven top-down analysis methods, where observations of CO2 are combined with atmospheric inversion modelling systems to validate bottom-up emission inventory-based estimates. These two approaches (top-down and bottom-up) are complementary, and their reconciliation is expected to yield the most reliable emission estimates to allow for potential management.
The monitoring of CO2 in urban areas has been a lesser priority when compared to the monitoring of traditional air pollutants because of the lack of legal standards for CO2. Therefore, high-quality CO2 time series are generally confined to isolated or remote locations where immediate emission sources are absent. These sites are suitable for capturing long-term and large-scale processes, but they are unable to resolve the dynamics of CO2 sources and sinks within urban areas (Hernández-Paniagua et al., 2015). In addition, the technology used for high-accuracy monitoring of CO2 measurement remains expensive (Mao et al., 2012; Martin et al., 2017), and, therefore, the deployment of several CO2 analysers in a city is usually considered cost-prohibitive. An alternative approach is to deploy lower-cost CO2 sensors, and several research groups have deployed monitoring networks in this context (Maag et al., 2018). Although such sensors have lower measurement performance, their poorer accuracy can be offset by being deployed in larger numbers, and thus, they offer the possibility of resolving spatial and temporal patterns at a smaller scale (Peltier et al., 2021). Therefore, the utility of lower-cost sensors can still be high (Bart et al., 2014; Casey and Hannigan, 2018).
Prominent urban CO2 monitoring networks include the Berkeley Environmental Air-quality and CO2 Network (BEACO2N) located across the San Francisco Bay area (Shusterman et al., 2016; Turner et al., 2016; Kim et al., 2018; Delaria et al., 2021), the Indianapolis (INFLUX) Urban Test Bed (Turnbull et al., 2015; Davis et al., 2017), the Los Angeles Megacity Carbon Project (Verhulst et al., 2017), the Northeast Corridor tower network (Karion et al., 2020), networks in Paris (Arzoumanian et al., 2019; Lian et al., 2024), and the Carbosense network across Switzerland (Müller et al., 2020). The nomenclature regarding the cost points for these networks is inconsistent because the definition of what a lower-cost sensor is varies among operators. Here, we discuss a CO2 sensor network that has been defined as “mid-cost” and is in a price range that is comparable to the BEACO2N and Paris networks. The Carbosense network, in contrast, used sensors at a significantly lower price point and, therefore, would be defined as a low-cost CO2 sensor network.