Processing Settings

If you were ever wondering what the different options in the processing meant (both for Storm and QuickWave), then have a look below. " />'>

Method  This is the method or algorithm used for the directional wave processing.  MLMST is the default. This is the Maximum Likelihood Method with Surface Tracking.  This is a special type of array processing that is well suited for AWAC wave measurements.  

The MLM method is the same but without surface tracking, and is used for AWAC data that does not have AST.  

The PUV method is suited for instruments that collect this special “triple point” type of measurement.  This includes the Aquadopp, Aquadopp Profiler, and the Vector.  

The SUV method is a special method that is a hybrid between the PUV and AST.  It is particularly well suited for AWAC deployments on a subsurface buoy.  The method may be used for bottom mounted deployments even if data was collected without the SUV mode turned on in the AWAC’s deployment configurations.  The method is also better suited (than MLM) for deployments where the waves are exposed to large mean currents.  Mean currents can present a Doppler shift on the wave field and introduce error in the directional and non-directional estimates if not corrected; The SUV method does not require this correction.



Spectrum This choice specifies which measurement will be used to calculate the non-directional or Energy spectrum.  The AWAC has three possible measurements to choose: Pressure, Velocity, and Acoustic Surface Tracking (AST).  When the Optimized option is selected, the software will use the AST as the primary means of calculating the energy spectra. If the software deems that the AST data have more than 10% outliers (bad detects), then it will revert to pressure (depth less than 20 meters) or velocity (depth greater than 20 meters) for the energy spectra.  With the other methods there is no reverting to another spectrum type. If you don't select Optimized, then you can select which single data set you wish to use for the spectral calculation (pressure, velocity, AST).



Direction Wave directions are always reported as direction “from”.  The four check boxes here allow the user to deselect a beam if it experiences complications during the deployment (for example blocked by a structure).  A minimum of 3 beams must be used for directional processing.  Beams 1-3 are velocity measurements and Beam 4 is the AST measurement.



Smoothing  The value chosen here specifies how much averaging is used at each of the discrete frequencies of the spectra.  More specifically, the number here specifies how many FFT bins are used at each frequency.  The greater the number of bins, the more smooth the spectra will appear; conversely the lower the number of bins, then the more the spectra will appear to be “spikey”.  The number of bins specified does not change the total energy, but may change the distribution of energy slightly.



Frequency Range  This sets the frequency range of the non-directional (energy) spectra.  Note that this will also be the limit for the directional spectra.  If the user selects a upper frequency that is out of range, then the software will use the Nyquist limit (sampling frequency/2) as the upper frequency.  Note that the AST samples at twice the rate as the pressure and velocity so this can be twice as high as the directional spectra limit; directional spectra is limited by the sampling frequency of the velocity measurements.  The lower limit is set to 0.02 Hz by default.  The smallest frequency step is 0.005 Hz, but is set to 0.01 as a default.



Offsets  The option to correct for offsets allows the user to adjust for external influences or differences that are discovered during the data analysis.  

The Pressure Offset can be used if the user discovered that the pressure offset was not set prior to deployment or if the pressure readings show a known bias.  

The Compass Offset is most commonly used to correct for the Earth’s magnetic field.  Wave direction should always be reported relative to true north.  

The Mounting Height is the distance the instruments head (the pressure sensor location) is over the sea floor; this is used to correctly calculate the transfer functions used for the pressure and velocity. Transfer functions for the velocity and pressure measurements are used when estimate standard wave parameters




regards,
Torstein

Hi Torstein

I note your comments relating to the merits of the different processing methods and in particular about currents possibly affecting the results. What do you mean by 'large mean currents'? Would a velocity of 0.4m/s be of significance? Which method is better to use for bottom-mounted AWAC-AST data - SUV or MLMST?

Cheers

Jon

Dear Jon

Long question ...

  is the "Doppler shift" equation for surface waves (from Wikipedia) and it describes how the current affects the wave length.    I just wrote a long answer before the internet connection was lost and the posting disappeared (when is ever someone going to write an app. that saves whatever you write on the computer so you don't have to do things over??) but I will give you the short version:

The effect of your 0.4 m/s current depends on the wave length.  For short waves, the effect may change a 1.5 second wave from 5 m to 4 m (assuming the current is in the opposite direction of the wave direction) and this can be quite dramatic in a small boat.  The same current can change a 100m wave to a 95m wave, which is less relative change and certainly less dramatic when being out on the ocean in a boat.

Which is better, MLM or SUV? Wouldn't we all like to know.....   Actually, they both do quite well in most conditions but I think Torstein will agree that SUV may be a little better for long waves and MLM is better for short waves (long/short is here relative to the spacing between the wave velocity cells).

Best regards, Atle Lohrmann

Hello

I,ve been processing lots of AWAC data series, and as a part of th job I have to deliver internal quality parameters for both waves and currents. I´ve been using Amplitude for currents and AST quality for waves. 

However, I realize that this is not the way Storm works when processing wave data. When choosing to use AST or pressure data for calculating the spectra, it looks for the "bad detects" as a quality criteria. I´d really like to know what are this "bad detects", since I've been trying to reproduce this on MATLAB with different methods (count values, standart deviation, outliers, etc), without any success. When I process the data on STORM using only AST, it will insert gaps on this low quality data positions, but it won´t display then on the wave count graph. I´ve also tried to process data on two different STORM versions: 1.05 and 1.07, obtaining different results.

Could you help me with this matters? I´ve been working on data obtained from a equipment that tilted 13º during the deployment and now I have to exclude bad AST wave data, justifying for it.

 

Thanks a lot,

Pedro Ribeiro

Hi Pedro,

 

We do two checks; one based on standard deviation and one acceleration check.

We start by calculating the standard deviation of the time series of the AST. Then we reject the samples that are outside 8 times standard deviation. We repeat this process after removing the out-layers at the same time as we reduce the n-times standard deviation criteria. We repeat until we've reach 4 times standard deviation.

The acceleration check is checking if the change of wave height is physically possible taking gravity into consideration.

Best regards

Jonas Røstad