Problem of Mass Conservation
§Imperfect conservation of mass (spectral dynamics)
§Empirical correction
qSum up CO2 forcing from the whole surface
üA=·(CO2 forcing) for every timestep (40 min)
qAccumulate total mass of CO2 in the simulation: B
üB= · (CO2 concentration) over all grids for every timestep
qMultiplying the ratio of A/B to every grid
üA>B (loss of mass in SPEEDY)
ü 
q
During this work, I found that SPEEDY does not conserve the mass perfectly, which results from the numerical problem of spectral dynamics.  So, I just corrected this problem empirically.  Since I know total amount of source from the surface, I multiplied the ratio of total mass that the model should have to the total mass that the model has into every gridpoint.