Chi Squared One-Sigma Limit

natski

Consider you have a very complex model with many input parameters. The best fit model gives you a chi squared value = x. Now imagine you wish to find the one-sigma error in several of the input parameters. By changing just a isngle input parameter, chi squared will obviously increase.

My question is, is there some % by which chi squared can be increased to which signifies the input parameter being changed has reached its one sigma limit?

Thanks
Natski