How would you predict the effect of a slight decrease in the temperature gradient at the surface of a spherical heat generator, on the temperature distribution in the core?

Would you do it like it's done in appendix four of www.nujournal.net/core.pdf or in a different way and why?