Updating to NumPy > 2.0

[pat-schmitt]

In this PR, I address test failures and warnings related to NumPy > 2.0.

Two tests were failing due to a silent conversion from float to int during the fixed geometry spinup. Here's a small example illustrating the issue:

test_array = np.array([1, 2, 3])
test_nr = np.float64(1.2)
test_array[0] = test_nr
print(test_array)

will print array([1, 2, 3]), whereas

test_array = np.array([1., 2., 3.])
test_nr = np.float64(1.2)
test_array[0] = test_nr
print(test_array)

will print array([1.2, 2. , 3. ]).

This behaviour is tricky to catch and highlights the importance of being mindful of the dtype when working with NumPy > 2.0.

Most of the warnings I fixed are related to xarray (likely to be resolved in future updates to xarray), and there are still some warnings from pandas that need to be addressed within pandas itself.

• Tests added/passed
• Fully documented
• Entry in whats-new.rst

[pep8speaks]

Hello @pat-schmitt! Thanks for updating this PR. We checked the lines you've touched for PEP 8 issues, and found:

There are currently no PEP 8 issues detected in this Pull Request. Cheers! 🍻

Comment last updated at 2024-09-10 18:13:37 UTC

[fmaussion]

Thanks!