Atkinson tables not implemented correctly beyond rMax

[pmpowers-usgs]

For example, using AB06', ground motions beyond 500km should use the rMax value derived from the appropriate ground motion table adjusted by a scale factor that scales as 1/R. However, when outputting mean ground motion values in the amean11 function, the log-mean ground motion (un-adjusted) starts to climb instead of staying constant with increasing distance.