Clarification on heights Assertion in get_annual_mb and Monthly MB Handling

[Ruitangtang]

Hi David @drounce , I had a question about the assertion check: np.testing.assert_allclose(heights, fl.surface_h)(

PyGEM/pygem/massbalance.py

Line 202 in 3590de9

np.testing.assert_allclose(heights, fl.surface_h)

) .

I’m saving monthly climatic mass balance (MB) along a flowline using OGGM’s flowline model, but got troubles when saving the monthly climatic mb along the flowline.

My questions are :

1. Why is this check necessary? Does it ensure consistency between heights and the flowline’s surface elevation (fl.surface_h)?
2. Could this cause issues if monthly MB is computed using a previous height (e.g., height = surface_h_previous[fl_id]) instead of the current fl.surface_h?
3. To compute monthly MB, I’m passing surface_h_previous[fl_id] as the height and adjusting the year (self.yr - 1/12): val = self.get_mb(surface_h_previous[fl_id], self.yr - 1/12, fl_id=fl_id) , Should fls (or fl) also be passed as an input to get_mb for consistency? Or is the current approach sufficient?

Appreciate any insights into the design rationale or suggestions for handling monthly MB correctly. Thanks for your time!

[drounce]

That line is from a long time ago. Yes, I believe it's to ensure that you've passed the correct heights associated with the flowlines.

Yes, it likely would cause an issue if you were computing it with a previous height. I'm not sure why you wouldn't want to use the present height though? Although I see you are doing that in Question 3. Why? and I'm not sure what self.get_mb is - do you mean self.get_annual_mb? Or are you creating a new function?

[drounce]

Yes, so you're running into the issue of the frontal ablation scheme living in OGGM. If you're trying to update the surface heights monthly, then you need to update the flowlines monthly as well. Currently, everything is run annually. Do that make sense?

[Ruitangtang]

Yes, I run the model monthly and update/save the flowline data monthly. The issue arises specifically when saving the climatic mass balance (climatic_mb) to the flowline.

Assertion Check: Would it be safe to simply comment out the assertion (or is there a better workaround)?

Thresholds & Errors: Could the random errors be related to internal thresholds for updating ice thickness (thickness) or surface height (surface_h)? For example, does OGGM/PyGEM enforce a minimum change threshold before applying updates?

[drounce]

If it's just saving the climatic mass balance, then wouldn't that be related to modifying the output files? I don't see why it wouldn't pass the assert statement if you're properly updating your flowlines each month.

Unfortunately, the code is not set up yet to run glacier geometry updates monthly and I'm heading into the field tomorrow, so I don't have bandwidth to troubleshoot this at the moment.

No, there's no minimum change threshold used to update the geometry.

[Ruitangtang]

You're right! However, the flowline itself doesn't have the climatic_mb_myr attribute, which is why we need to save it in a separate DataFrame. That’s also the reason get_annual_mb is called with both the previous height and the flowline. I'm going to try saving the previous flowline — in theory, that should do the trick! Will update you then!

Thanks a lot.
Have a safe and nice fieldwork.

[drounce]

Sounds great. Let me know how it goes.