Leave this open. An important next step will be to populate masked regions with something that does not impact the final power spectrum. The key test is to overplot the input power spectrum (e.g. -2.9) onto the plot to see if you can actually recover the input. It might be that it is trivial, but it might not be. See code here for example: https://healpy.readthedocs.io/en/latest/tutorial.html . Anyways, we can address this when you get back.
Leave this open. An important next step will be to populate masked regions with something that does not impact the final power spectrum. The key test is to overplot the input power spectrum (e.g. -2.9) onto the plot to see if you can actually recover the input. It might be that it is trivial, but it might not be. See code here for example: https://healpy.readthedocs.io/en/latest/tutorial.html . Anyways, we can address this when you get back.