Attempted Logical Approach
This is why Mcrit should change with change in altitude->
At FL200 ,Assume Aerofoil 'A' 's airflow reaches M1.0 on the upper surface at forward speed of M 0.85, This defines Mcrit of A as Mcrit=M0.85, which means the upper surface of the wing has accelerated the air by M0.15.
At FL400, the air is thinner, ie, less dense, ie, is easier to accelerate (F=mXa, where m has dropped due to drop in density). Because the air is easier to accelerate, it would accelerate over the upper surface of the wing quicker, which means the upper surface can accelerate the airflow by, say, M0.16, effectively dropping Mcrit to M0.84 (For arguments sake)
Conclusion-Mcrit should ideally drop, but it probably drops insignificantly, so it is probably calculated at the service/absolute ceiling and is fed to us pilots as one number easier to remember.....
My language should have made it clear that it is a theory backed by nothing but previously studied material.