The impact of temporal variability in prior emissions on the optimization of urban anthropogenic emissions of CO₂, CH₄ and CO using in-situ observations

Highlights

• Specific temporal profiles capture more of the intrinsic variability in emissions.

• Temporal profiles have strong impact on monthly top-down emission estimates.

• Temporal profiles affect the source-receptor relationship.

• Accurate temporal profiles are essential for extrapolation of monthly emissions.

• Observational sampling strategy affects monthly top-down emission estimates.

Abstract

Constraining urban emissions is gaining more attention because of the important role of cities in reaching national climate mitigation targets. Urban inverse modelling studies could constrain emissions of large hotspots, but still face many challenges. It has been argued that more detailed information is needed on both atmospheric transport and prior emissions when moving to a higher spatial and temporal resolution. In this work we focus on the description of temporal variability in the prior emissions and examine how it impacts the optimization of urban emissions of CO₂, CH₄ and CO on a monthly time scale representative for a measurement campaign. Currently, temporal profiles based on long-term average activity data are often applied. However, these average temporal profiles are unable to capture a realistic variability in the emissions, such as those imposed by environmental conditions. Therefore, we created a set of location- and time-specific temporal profiles and compared the optimized emissions using these average and specific temporal profiles. We find that using the specific temporal profiles increases the optimized CO₂ emissions with 19%, even though the prior monthly emissions are the same. This suggests a change in the source-receptor relationship that affects comparison of the observed and simulated mixing ratios, leading to a different emission estimate. The impact is also large (~40%) for CH₄, but this is mainly due to the increase in prior emissions caused by redistributing agricultural emissions over all months of the year. Moreover, we show that extrapolating monthly emission estimates to annual estimates, required for reporting, using the various sets of temporal profiles can result in differences of max. 26% for CO₂, 101% for CH₄, and 13% for CO. Therefore, we conclude that an accurate representation of the temporal variability is essential for urban inverse modelling studies.