Continuous measurements of atmospheric CO2 over the continents are potentially powerful tools for understanding regional carbon budgets, but our limited understanding of the processes driving the high-frequency variability in these measurements makes interpretation difficult. In this paper we examine the synoptic variability (∼days) of surface CO2 concentrations in four continental records from Europe and North America. Three source functions corresponding to the ocean, land biosphere and anthropogenic sources and sinks for CO2 have been implemented in a regional atmospheric transport model. In previous carbon studies, monthly biospheric fluxes have typically been used, but here high spatiotemporal (daily, 1°× 1°) resolution biospheric fluxes are obtained from the NCAR Land Surface Model (lsm). A high-pass filter is used to remove atmospheric variability on time scales longer than 2 months, and the resulting simulated concentration fields replicates reasonably well the magnitude and seasonality of the synoptic variability across the four observation sites. The phasing of many of the individual events are also captured, indicating that the physical and biogeochemical dynamics driving the model variability likely resemble those in nature.

The observations and model results show pronounced summer maxima in the synoptic CO2 concentration variability at the two stations located in North America, while a slightly different seasonality with high variability throughout fall and winter is observed at the European sites. The mechanisms driving these patterns are studied and discussed based on correlations between the concentration anomalies and the driving atmospheric physical variables and surfaces fluxes in the simulations. During the summer, the synoptic variability over the continents has a significant contribution from variations in regional net primary production, which in turn is modulated by regional, synoptic temperature variability. In winter the synoptic variability is partitioned about equally between biospheric and anthropogenic CO2 and is mainly driven by local vertical mixing and synoptic variations in atmospheric circulation working on the large-scale atmospheric gradient. This study highlights the importance for future modeling work of improved high temporal resolution (at least daily) surface biosphere, oceanic and anthropogenic flux estimates as well as high vertical and horizontal spatiotemporal resolution of the driving meteorology.