WINDOW SIZE IN EULER DECONVOLUTION

QuentinMarcANABAFOTZ

hello
please I wish to know how to find/calculate the right window size to perform Euler deconvolution

GaudPouliquen

Hi @QUENTINMARCANABAFOTZ
The rule of thumb is that the Euler window needs to be large enough for the wavelength of the anomalies that you’re trying to resolve.
The window size is set as number of cells and not ground units. The maximum size is 20 cells, and this means that depending on the wavelength you’re interested in, you might need to re-grid prior to applying Euler.
Let’s say that the wavelength of the anomalies you’re interested in is ~60 km and that your grid cell size is 1 km. If you set your window size to 20 cells, the window size is 20 km, so shorter than the anomalies wavelength and could yield erroneous depth estimates. So in this case you would need first to regrid your magnetic with a 4 km grid cell size, and use 20 for the window size.
I hope this helps, you might also want to look at this article about setting Euler parameters for deep sources.

Alternatively, you can also try to use the located Euler and Located An-Eu which will automatically determine the window size.

Or use the Easy_tilt script, which relies on the tilt derivative to derive depth solution for vertical contacts. You can find the script file in this forum post

PankajKumarDas

excellent

QuentinMarcANABAFOTZ

awesome
thank you so much

QuentinMarcANABAFOTZ

@GaudPouliquen how to know the wavelength of the anomaly of interest

GaudPouliquen

Hi @QUENTINANABAFOTZE , you can have a simple empirical approach and measure the wavelength directly on your grid (use the measure distance tool). Or interactively extract a profile from your grid across your anomaly of interest (Grid and Image > Utilities > Grid Profile), and measure the wavelength in the profile panel of your database. You can also calculate the radially averaged power spectrum of your anomaly to identify dominant wavelengths in your grid: https://my.geosoft.com/elearning/lessons/#/reading/893.

QuentinMarcANABAFOTZ

Thank you very much @GaudPouliquen