The onine diags were hard coded to look for net_moistening_due_to_machine_learning,column_heating_due_to_machine_learning` when adding the column-integrated dQ2 to various diagnostics. This PR updates the derived variables to add these quantities from the reservoir predictor stepper if they are present in the data.
The onine diags were hard coded to look for
net_moistening_due_to_machine_learning,column_heating_due_to_machine_learning` when adding the column-integrated dQ2 to various diagnostics. This PR updates the derived variables to add these quantities from the reservoir predictor stepper if they are present in the data.