New directional wave conditions for Sydney

hi craig,
yep, all 7 sites will go up, and there'll be several other analysis going online soon too (though this is my go-to when i'm deciding where to surf). there's been a lot of changes to our telemetry and not all sites can be channeled online yet ... soon hopefully!

the plots should be pretty easy to interpret for anyone that regularly follows surf conditions, and there's an explaination on the site, but i'll look forward to your article too. anything that makes the data accessible to the wider community is welcome

now we just need the swell to come to the party!

the_b: That's on the cards too, either a video that rolls through the last day or an animation where you can select the period. That's down the track.

I'd love to Jeff, any excuse to get back to my home town. But as far as I can tell that buoy is non-directional.

Hi Don, maybe you could talk them into a secondment, I wouldn't mind some time in qld too!

Second question first (deap breath... ) The directional spectrum is determined by first taking the 1D spectrum in the same way as a non-directional buoy. This tells us where in the frequency (period) range the energy is. There is always some energy across all frequencies, but theres a peak or two at the period of the swell(s). This isn't normalised (max of 1.0) but scales with wave height squared. Larger waves mean there is more energy in the spectrum.
Then for each frequency (period) band the direction of the energy is determined. There is also a spread associated with the peak so that we may have a tight peak (long crested waves) or a broad peak (as we would see in an onshore breeze with choppy waves). This is normalised for each frequency band as we want to keep the total energy constant.
But it does mean that if we have two swells at the same period from different directions, the energy will smear across the directions rather than give the two peaks we would like (this doesn't seem to happen very often fortunately).

That how the direcional spectra is made. To determine the wave height we integrate (add) all the energy cells in the plot and convert to height. I've spent a lot of time mulling over how we might separate the peaks to determine the heights from each event. Its realatively easy to choose the peaks by eye, but selection/separation of events is rather tricky to automate for the webpage. For now you can get a sense of the relative height by looking at the peaks (colour depth and area covered), and remember that energy is height squared so you need 4x energy to get 2x height.

Sorry for the long reply, does this answer it?
-Ben

Nice demo plot, thanks Craig.

Energy density is (m2/deg/Hz) so you need to know the size of the bins in direction and frequency to be able to perform the calc (1deg, 0.005Hz bins up to 0.1Hz then 0.01Hz). If you do this for one bin you get a tiny value of wave height. We have to add all the values in a region to get a sensible height associated with a swell event.
So its not really possible to visually convert the information on the plots to an individual wave height for different swells. Maybe down the track we can develop something for that. For now it simply gives a intuitive indication of where the swell energy is.

Hopefully with a bit of practice people will be able to see energy from different areas of the plot light up different breaks. I find it interesting following the energy source, eg a low moving across the tasman and seeing the period, spread, height and direction follow the source.

Hi Dave,
Good suggestions, thanks. Our Byron Bay buoy will go up, is that close to you?

The plots were originally developed for scientists/engineers so some convensions are kind of set from there. So where we have period (4,5,7.1,10,20 seconds) would normally be (0.25, 0.20, 0.15, 0.10, 0.05 Hz). I just changed the label.
The choice of long period at the centre seemed odd to me at first too, but I guess you just get used to it - it really makes more sense to have it at the outer since longer periods have more energy and should take up more space on the plot. I'll have a play and see how they look in reverse.

We have a 48hr loop working now, just waiting on approval to go public.

Cheers

Sydney is up now, and so are the 48hr animations.
Enjoy!
http://new.mhl.nsw.gov.au/data/realtime/wave/DirectionalSpectra-sydney

I'm not at the lab this week so I don't know what's been happening (too busy surfing), but I could take a guess.
Two sets of analysis are performed on the raw data - one on the buoy , the other at the receiving station. Both are sent back and stored at mhl. Wave analysis is a black art and small changes to the fft parameters can give (usually small) differences in the results. What you are seeing are the two peak period results. Sometimes if there are two swell trains at similar energies they will pick different peak periods, though the Hs will be the same.
I'm not sure why both are there. We really only want one to save confusion. Probably a change has been made over the break that will be fixed in a couple of weeks.
Cheers, Ben