$y$ is a function of $a, b$ and $c.$
I have data for a, b and c over a period of time.
I am plotting a graph of $y$ against time. At the time $= t_{\text{lost}}$, I start to lose all $c$ values. How do I apply the past standard deviation of $c$ to predict the $y$ values beyond $t_{\text{lost}}$?
Based on your description, I have the feeling that you are trying to extrapolate the variable $c$ from a number of samples.
Extrapolation by an interpolation method such as the Lagrange polynomial will miserably fail, and unless you have a reasonable explanatory model for $c$, anything will fail with such a large extrapolation range.
Your only luck is to have a good parametric model for $c$, fit the model on the available data, and pray.