Spectral peak of a 2D wave spectrum

How is the spectral peak of a 2D wave spectrum defined when calculating e.g. the peak period or the mean direction at the spectral peak?

Do you a) pick the point in the two dimensional frequency-direction space with the most energy, or do you b) integrate over all direction (thus turning the whole thing to a 1D-spectrum, and then pick the frequency with the most energy from there?

For example in the spectrum with mock values:

0 deg 90 deg 180 deg 270 deg

0.1 Hz 0 0 0 0

0.5 Hz 0.1 0.3 0.5 0.2

1.0 Hz 0 0 1 0

method a) whould result in a peak period of 1s (1 Hz), and method b) in a peak period of 2s (0.5 Hz). The mean direction at the spectral peak would also differ, because it would be calculated over different frequancy bands.

Edited December 1, 201312 yr by tmpst