When I understood you correctly, you maybe meant that if the lognormal uncertainty of an exchange is selected for the first time, the pre-filled value for gmean is the amount of the exchange itself.
When you give your exchanges an uncertainty value and there was no uncertainty entry before, the software will put inside the gmean field for log-normal distributions the amount of the exchange as a default. And you are right, that usually the gmean value is equal to the exchange amount.
But if these two values would be fixed and coupled, there would be less flexibility for users to set exchange amounts and mean values independently, since this condition is not true for all uncertainty distributions.
When you set-up a process and you change the amounts, it does not take much time to also change the gmean, no? Or are you looking for a parameterization of the uncertainty values?
Best wishes,
Conrad