Well, my first impulse is to actually hang the two lines and measure
the resultant impedance at the source end...
I am assuming the lines are spread apart so that they are not closely
coupled...
One approach to get you in the ballpark would be to model the 30
footer alone to get a source impedance, then model the 20 footer
likewise alone, then model the resultant of those two complex
impedance in parallel...
An interesting question...

denny / k8do