I am helping to develop the coupling of oggm with an energy balance/snow model, and this has been implemented by creating a new class that inherits from the MassBalanceModel class.

The snow model has memory, such that in consecutive years the snow model and ice temperature state is preserved. An issue arises when get_specific_mb is meant to be called multiple times with a specific instance. As far as i can tell, this is resulting in apparent_mb_from_any_mb throwing an error in _check_terminus_mass_flux, where the exact residual has been found.. but then is no longer the correct residual since a subsequent call at this line. As I have not explored enough functionality, Im not sure but similar issues may arise elsewhere.

My question is -- what would be the best way to handle this. I could envision:

1. adding a kwarg, e.g. reset_state=false to massbalancemodel.get_specific_mb (and get_annual_mb) to "reset" the state, and using the kwarg where appropriate (there will likely be multiple locations). It would do nothing for all exisiting MB classes.
2. re-instantiating the mass balance model where a state reset is needed.

(1) seems better, though will look a little odd until (hopefully) our energy mb model is merged. if you are OK with this, i will create a branch and add the kwarg in all the places i think are appropriate, though this will likely not be exhaustive without further testing.