If something is 10-20 years in the future in these types of models, this has:
a) Already happened unless b) there are major changes / interactions that scientific models have not had factored in while c) Anthropic actions cannot change this, but it can change x+n where X is original time and n is the 'down ramp' from your curve.
For instance: if your Co2 is 400ppm now, then the effects have already happened 20-50 years into the future. It can be made much much worse through events before you hit that time (let us say either Yellowstone or the entire of India / China buying a car per family) but your actions in the interim are merely altering the effects after time X.
In the case of water we can say: model presents X+n time > if action: such as massive investment in desalination and/or new osmosis materials (positive to time change) minus climate impacts we've not noticed yet (negatives to time change) where n is less than continued effect without any other imputs.
This is a rather loopy way to say:
People think of these types of announcements as future based predictions: they're not, they're present events if (and only if) your models don't change.
[Note: this isn't to say they're scientifically incorrect - but this inability to understand time in these types of reports fuels a lot of ignorance from both "sides" of Climate / Ecological debates]