Modeling the Synoptic Atmospheric Interaction with Anthropogenic and Biospheric CO2 Emissions
D. Chan, M. Ishizawa and D. Worthy
Environment Canada, Toronto, Ontario M3H 5T4, Canada; 416-739-4852, E-mail: douglas.chan@ec.gc.ca
CO2 concentration measurements at continental sites typically show strong synoptic variability. Figure 1 shows the CO2 concentration time series at the various sites in Canada. There are significant differences in the synoptic variability at these sites. In this study we used a dynamic atmospheric model Global Environmental Multiscale (GEM) with variable horizontal resolution (with minimum delta x of ~15 km) to examine the atmospheric synoptic interaction with the regional anthropogenic and biospheric CO2 fluxes. The results (Figure 2) show the atmospheric transport varies at these sites and the interaction of atmospheric transport and surface CO2 emissions (all sources) leads to different synoptic variability at different sites. Furthermore, the atmospheric interaction with anthropogenic emissions is distinct from the interaction with biospheric emissions, leading to different synoptic characteristics in the CO2 concentrations. Thus the synoptic CO2 concentration characteristics at a site contain information on the regional emission. The model and observations can be used to evaluate different emission estimates.
Figure 1. The Observed CO2 concentration time series at the different continental sites on the map, the gray curve is the hourly concentration, and the red curve is the afternoon (3-6 pm local time) concentration. The synoptic variability varies with locations, indicating different atmosphere and source emission interaction.
Figure 2. Time series from GEM model for December 2007 at the Lac Labiche and East Trout Lake sites. Black curve includes all sources bio+ff+ocn (biosphere + fossil fuel + ocean), green curve is for bio only, and red curve is for ff only. Bio CO2 concentration shows more variability than ff concentration. GEM was initialized with 0ppm CO2 concentration at the initial model integration time, November 1, 2007.