A high-resolution monitoring approach of urban CO2 fluxes. Part 2-surface flux optimisation using eddy covariance observations

Achieving climate neutrality by 2050 requires ground-breaking technological and methodological advancements in climate change mitigation planning and actions from local to regional scales. Monitoring the cities' CO2 emissions with sufficient detail and accuracy is crucial for guiding sustainable urban transformation. Current methodologies for CO2 emission inventories rely on bottom-up (BU) approaches which do not usually offer information on the spatial or temporal variability of the emissions and present substantial uncertainties. This study develops a novel approach which assimilates direct CO2 flux observations from urban eddy covariance (EC) towers with very high spatiotemporal resolution information from an advanced urban BU surface flux model (Part 1 of this study, Stagakis et al., 2023) within a Bayesian inversion framework. The methodology is applied to the city centre of Basel, Switzerland (3 x 3 km domain), taking advantage of two long-term urban EC sites located 1.6 km apart. The data assimilation provides optimised gridded CO2 flux information individually for each urban surface flux component (i.e. building heating emissions, commercial/industrial emissions, traffic emissions, human respiration emissions, biogenic net exchange) at 20 m resolution and weekly time-step. The results demonstrate that urban EC observations can be consistently used to improve high-resolution BU surface CO2 flux model estimations, providing realistic seasonal variabilities of each flux component. Traffic emissions are determined with the greatest confidence among the five flux components during the inversions. The optimised annual anthropogenic emissions are 14.7 % lower than the prior estimate, the human respiration emissions have decreased by 12.1 %, while the biogenic components transformed from a weak sink to a weak source. The root-mean-square errors (RMSEs) of the weekly comparisons between EC observations and model outputs are consistently reduced. However, a slight underestimation of the total flux, especially in locations with complex CO2 source/sink mixture, is still evident in the optimised fluxes.

Automated summary

This study develops a novel approach that combines urban eddy covariance (EC) tower observations of CO2 flux with an advanced urban bottom-up (BU) surface flux model within a Bayesian inversion framework. The methodology provides optimized gridded CO2 flux information for each urban surface flux component at high spatial and temporal resolution. The results show improved estimations of flux variabilities and reductions in anthropogenic emissions, with traffic emissions having the greatest confidence among the flux components.