Diagonal lines on results of the conventional FK

[seguuu91]

The problem that I've encountered is diagonal lines.
The lines of some results(diag-1.png : bottom left ) are negligible but some(diag-2.png) are not.

The attached text file is the parameters of the diag-2.

Thank you!

[admin]

Several remarks after having a look at your .max file.

• the shape of your array is not optimal: it is rather elongated which provides a good resolution power along direction 25 deg. but rather poor in the perpendicular direction. We always try to get similar responses in all directions which is obtained with round shapes (not necessarily circles). To get the theoretical array response of your geometry, you can use warangps (http://www.geopsy.org/wiki/index.php/Warangps). With this tool, you can get kmin and kmax limits on a dispersion plot. You can copy the layer and paste it on your results in gpviewmax or gphistogram.
• if you plot the theoretical array response over your results, you will see that all results are above the resolution limit of the array (see Wathelet et al. 2008 for instance). With Conventional FK, you will not get any valid dispersion above kmin.
• frequency sampling is defined by 300 values. I prefer defining a constant log step (e.g. 1.025) which remains valid for all sites. I change only the min and max frequency which may be different
• you processed using Conventional FK, did you try high resolution? Do you have 3C for these stations? You may also try RTBF (see Wathelet et al. 2018)
• you probably forced GRID_STEP and GRID_SIZE. Default values are deduced from the theoretical array response and should not be touched unless you know what you are doing. GRID_SIZE can increased in some cases to explore a high frequency.
• I tried to select only azimuth along the elongation of your array (from 0 to 50 deg. and from 200 to 250 deg.), by running "gphistrogram -azimuth", filtering and rexporting into a new .max. You have to reprocess keeping the default value for GRID_SIZE and searching for velocities lower than 300 m/s. With the azimuth selection a sligthly denser area is visible that may be linked to a dispersion curve (see attached figure). The limits shown by black lines are kmin and kmin/2 manually adjusted in warangps to fit the capabilities of the array in the direction 25 deg (kmin=0.11).
• Even for the conventional the number of blocks for the cross-spectrum averaging can be larger that 1. It stabilizes the results. Leave the default value of 0 which implies that the number of blocks is calculted from the number of stations mulitplies by the BLOCK_COUNT_FACTOR (0.5 for conventional FK)