Statistical model training data for "Continuous Structural Parameterization: A proposed method for representing different model parameterizations within one structure demonstrated for atmospheric convection"

Gzipped CSV files containing convection scheme inputs and outputs used for training.

Column format of each file:

THETA_IN_1,2,3,4,5,6,7,8,9,10,11,12,13,14,15,16,17,18,19,20,21,22,23,24,25,26,27,28,Q_IN_1,2,3,4,5,6,7,8,9,10,11,12,13,14,15,16,17,18,19,20,21,22,23,24,25,26,27,28,DTHETA_1,2,3,4,5,6,7,8,9,10,11,12,13,14,15,16,17,18,19,20,21,22,23,24,25,26,27,28,DQ_1,2,3,4,5,6,7,8,9,10,11,12,13,14,15,16,17,18,19,20,21,22,23,24,25,26,27,28

where THETA_IN are input values of potential temperature [K], Q_IN are input values of specific humidity [kg/kg], DTHETA are changes in potential temperature due to convection [K], DQ are changes in specific humidity due to convection [kg/kg].

Key:

"llcs" are simulations with Lambert-Lewis.

"gr" are simulations with Gregory-Rowntree.

"4xco2" have 4 x pre-industrial atmospheric carbon dioxide concentration. (Others have 1 x pre-industrial atmospheric carbon dioxide concentration.)

"rh0.7" and "rh0.9" have LLCS RHCRIT set to 70% and 90% respectively.

All are 30 day simulations either for January "jan" or July "jul".