Peak power for nuclear explosions

[valhallast]

I was reading this article http://news.yahoo.com/ap-exclusive-graph-suggests-iran-working-bomb-161109665.html on Iran's nuclear programme and at first thought the numbers on the left (relating to kilotons of energy per second) were off, but then I realized that it relates to peak power output at that specific period of time. How would one calculate this normally? I know all the power equations relating to energy/time, but peak power output other than for peak current and peak voltage in alternating current systems.

 

[mfb]

The energy graph looks like the result of a simulation, and power is simply its derivative.

 

[Astronuc]

The power pulse looks like a idealized (Gaussian) pulse, similar to what one uses for a reactivity insertion transient in a conventional reactor. There the pulse for a reactor would be on the order of several milliseconds. If the plot has microseconds, this would be an indication of a nuclear weapon.

The energy is just the pulse amplitude integrated over time. The ordinate could be written in W, in which case the energy would be written in J. One only needs to convert J to kT_eq.