On chemical reaction models in the mesosphere and thermosphere in FWHIST

[koichi omi]

I ran two simulations, one with and one without aircraft emission data in the input data, to investigate the effect of aircraft emission on the atmosphere.
When we compared the results of these simulations, they seemed to be physically wrong. The details of the simulations are as follows.

CASE1 (simulation with aircraft emission):
compset: FWHIST
grid: f09_f09_mg17
simulation period : 3 years
input data file: default

CASE2 (simulation without aircraft ):
compset: FWHIST
grid: f09_f09_mg17
simulation period : 3 years
input data file: following data files were removed from input data.

emissions-cmip6_bc_a4_aircraft_vertical_1750-2015_0.9x1.25_c20170608.nc​

emissions-cmip6_NO2_aircraft_vertical_1750-2015_0.9x1.25_c20170608.nc​

emissions-cmip6_num_bc_a4_aircraft_vertical_1750-2015_0.9x1.25_c20170608.nc​

emissions-cmip6_SO2_aircraft_vertical_1750-2015_0.9x1.25_c20170608.nc​

The average NOx concentrations of the last year of the three-year simulation results were compared.
Look at the attached figure. This figure shows the NOx concentration change (<annual longitude average of CASE1> ー <annual longitude average of CASE2>)
From this figure, we can find NOx concentration in the mesosphere and thermosphere was affected by aircraft emission.
Since the maximum altitude for aircraft exhaust is about 14 km, I think it is not physically correct. In addition, this result is not consistent with the trend of results from other references. Is this the correct behavior?

 

[marsh]

Hello Koichi Omi,

Sorry for the delay. It is not likely the changes are due to changes in aircraft emissions but instead are from dynamical variability. As you probably know, FWHIST WACCM is a free running climate model and not a CTM, so there will be considerable year to year variations in the winds and constituent transport. You can test this yourself by looking at differences between each year in each case. To test if differences are due to forcing differences rather than interannual variability, you can do a Student's t-test. You will likely need to run more than 3 years to determine the standard deviation of case.

If you are only interested in stratospheric/tropospheric chemical differences you may want to consider using the specified dynamics (SD) compsets.

best regards,

Dan

 

[koichi omi]

Thank you so much for your kind advice!
I tried FCSD instead.
However, another problem occurred when running the simulation.
I'll post another thread about the problem in infrastructure.