Fix CHELSA-W5E5 precipitation scaling

[juseg]

I overlooked the different precipitation units in CHELSA-W5E5 data when implementing #86. This issue was brought to my attention by the following line in hyoga.open.atmosphere(), which currently fails not because of different units, but because the 'time' dimension is called 'month' in the new source.

hyoga/hyoga/open/reprojected.py

Lines 202 to 203 in 48391fc

	# convert precipitation to kg m-2 day-1
	ds['precipitation'] /= xr.DataArray(months, coords={'time': ds.time})