Yesterday I re-ran Craig's SRCL to DARM transfer function and his analysis code which calculates the circulating arm power from the RIN to DARM transfer function. This measurement was done ~10 hours into the lock.

Results: 

Arm QPD      Gain [DARM m/RIN]    Arm Power Inferred [kW]
------------------------------------------------------------------
XA           1.55e-6              183
XB           1.58e-6              188
YA           1.65e-6              195
YB           1.72e-6              204

X arm power = 185 kW
Y arm power = 200 kW

Average arm power = 192.4 +/- 7.9 kW

The fit of arm power RIN to DARM is shown in the attached figure. If we remove the two lowest frequency points we will probably improve the uncertainty on the arm power.