I've tried searching for this, but it doesn't seem to be documented anywhere except for a paragraph on how it works in VB6:
The amount of albedo change is based on the extent of the ice sheet (F9 again) plus a random factor.
Is this still the case in C#? I've constructed a spreadsheet to calculate optimal terraforming, but as of now I need to recalculate when ice sheets melts. This would be unnecessary if I knew beforehand how the albedo would change :-)
I like you already.
I have the same spreadsheet. In my experience, the albedo doesn't change by much.
So, I don't worry about it when I'm planning.
When the sheet melts, I update the albedo value and make sure that my end-state temperature hasn't gone outside my hab range. Only because I'm paranoid. It's never even come close.
In cases where I'm adding a lot of GH to reach hab range, the albedo increase will allow me to get to the hab range sooner, but I've never seen it reduce the needed terraforming by much.
It does mean that, in cases where your final ATM will be greater than 1.0 (and therefore your final Oxygen will be more than the minimum 0.1), you want to make the sheet melt before you add oxygen, so that you don't add more oxygen than you actually needed for the final amount of GH.