Based On GEOS-Chem Model Simulation Of Atmospheric Carbon Dioxide Cycle

One effective method in studying carbon cycle is multi-scale atmospheric transport model.It play an important role in understanding nautural ecosystems and predicting the change of future climate. We simulate the atmospheric carbon dioxide cycle taking advantage of advanced atmospheric transport model (GEOS-Chem(v8-02-01)) using top-down approch. The following is the main content.In this paper, the forward model is driven by GEOS-5 using ORCHIDEE model to establish a basic function. The global carbon dioxide fluxes in 22 regions are calculated on the basis of observational atmospheric carbon dioxide concentration in 109 stations in GLOBALVIEW-CO2 (2009) adopting two kinds of method (Bayes theorem and Ensemble Kalman Filter)during the period of 2003 to 2006 and then analysis their changes in various regions. The result show that: atmosphteric carbon dioxide fluxes in most land areas are increasing year by year and their rising rates are different, the Eurasian temperate zone was of the largest increasing rate, followed by the tropical Asia and Europe zones, North American Temperate zone declined slightly in 2006, the North American boreal zone, South Africa, and marine areas were little changed significantly. However, atmospheric carbon dioxide fluxes in 22 regions from the model are positive or zero, that means they are carbon sources, or they are of their own balancing, which indicates that atmospheric carbon dioxide fluxes from the model simulation are higher than their real values.In this paper, the prior fluxes of atmospheric carbon dioxide are basically consistent with the posterior fluxes calculated by two ways which are Bayes Theorem and the Ensemble Kalman Filter and their covariances are quite small. There is a correlation analysis between the model data and corrreponding observational data in 109 stations during the period of February to December in 2003 to 2006. When the confidence level is 99%, 71.56% of the observation stations are correlation signifcantly; when the confidence level is 95%, it is 77.98, which can validate the reliability of the model.The sensitivity of local qualitative analysis method is applied in this paper to study the respond of the GEOS-Chem model results to observational error, model error, prior covariance and prior fluxes in land areas. The results indicate GEOS-Chem model respond to the observatinal error and model error insensitive,to prior land fluxes sensitive. With the prior covariance increasingh, the sensitivity of gthe model is stronging. When the prior covariance less than 5GtC/y, the respond of model to it insensitive, when the prior coavariance greater than 10GtC/y, the respond of model to it sensitive apparently.In addition, there is a correlation analysis of atmospheric carbon dioxide concentration to atmospheric temperature and the total of energy consumption. Studies have shown that atmospheric carbon dioxide concentration is apparently relevant to temperature in daily, monthly and quarterly scale and the total of energy consumption , but not to annual temperature. The next step will study the relationship between the atmsopheric carbon dioxide concentration and temperatur in a smaller time scale.