We are trying to use a simple extension of the temperature rise method used by the IPCs but the calculated values are not accurate to the reality values.
There is no simple equation for this problem. Interpreting IPC as "Institute for Printed Circuits" they should have rules-of-thumb - being on the "safe side".
Once working on a related issue I found a DIN (German Industry Standard) and there may also exist an ISO about current carrying capabilites of PCB tracks. They listed 'considerations', but no values.
Why so? Because the "ambient" (read like "rest of the board") greatly influences each and every track - due to heat dissipation via solid (but not homogeneous) matter.
The "standard approach" at this point is "try and measure". (That is, if you have no prior experience.)
Thanks Dreher, the standard approach was my initial thought and we are close to complete the study for different ambient temperatures (from 25C to 120C).
I don't know, but the equation I developed for on line over load time limit computation and control, for power transformers, on basis of measured hot spot temperature, could give you the direction....