That’s right. And it can be seen that in figure 11 the FWM power for 0 dBm input power is near -100 dBm. So I am guessing (and it’s just a guess) that spectrum in figure 9 has been shown with system settings similar to one of the cases in figure 11, and not with the system settings whose results are there in table 2. Only this way makes sense to me. If the authors want to take 0 dBm as -100 dBm, then why not using the -100 dBm itself?