The concept of subduction seismogenic zone, where megathrust seisms occur in a layer between an upper and a lower limit is well documented in the 1990/early 2000 literature :
Some example :
Pacheco, J. F., Sykes, L. R., & Scholz, C. H. (1993), Nature of seismic coupling along simple plate boundaries of the subduction type. Journal of Geophysical Research: Solid Earth, 98(B8), 14133–14159. http://doi.org/10.1029/93JB00349
Hyndman, R. D., Yamano, M. and Oleskevich, D. A. (1997), The seismogenic zone of subduction thrust faults. Island Arc, 6: 244–260. doi:10.1111/j.1440-1738.1997.tb00175.x
But is this concept still accurate and used in the scientific community now ? i.e. after the megaseisms of Indonesia in 2004 and Tohoku in 2011, where huge areas break up to the surface and generate big megatsunamis. More exactly, is the notion of upper limit still making sense and, if yes, what are the currently-supposed parameters which control it ?