Negative apparent magnetic susceptibility values
Hi,
I have been trying to derive near surface apparent magnetic susceptibility from aeromagnetic survey data. In order to do this I have been applying high pass butterworth filters (experimenting with different wavelengths) to try and isolate the residual field and then performing an apparent susceptibility calculation (SUSC) on it.
My problem is that I am getting negative apparent susceptibility values as a result. Does anyone have experience with why this is the result?
I have been trying to derive near surface apparent magnetic susceptibility from aeromagnetic survey data. In order to do this I have been applying high pass butterworth filters (experimenting with different wavelengths) to try and isolate the residual field and then performing an apparent susceptibility calculation (SUSC) on it.
My problem is that I am getting negative apparent susceptibility values as a result. Does anyone have experience with why this is the result?
Tagged:
0
Comments

I have never tried this but my guess is that your HP filtered data has negatives in it (or at least negatives relative to the average value). A quick look at the SUSC filter suggests that high pass filtering your data before is not a good idea. The filter does a downward continuation that will roughen the gridded data. It also says that there should be no remanance which I read to mean no negatives after RTP. Which points back to my initial guess. Not sure how to fix it but I hope this gives you a few ideas.
Sean1 
@WilliamMcNeice have you considered using VOXI to model the magnetic susceptibility?
Even when doing using VOXI for the inversion, you'll have to consider what @SeanWalker mentioned... Are there negative RMI values? The inversion will want to deal with these by using negative susceptibilities  but these are negative relative to the background you removed to create the residual.
They may also be negative due to noninduced magnetization (i.e. remenance, shape anisotropy, etc.). You'll have to consider what you know about the geology of the area to provide further insight.
You can learn more about getting started with VOXI here: https://my.geosoft.com/elearning/path/459Customer Success Manager  Geophysical Modelling0 
As the other comments suggest, the negative susceptibility could represent negative RMI values. However in my experience, the SUSC filter produces susceptibility distributions with negative values if the source thickness is not infinite (that is one of the assumption in the method).
The concept of a "magnetic annihilator" (see Blakely's book or https://academic.oup.com/gji/article/155/2/509/598173) means that most susceptibility distributions are nonunique. This means you can add certain susceptibility distributions (annihilators) to your model without changing the magnetic response. So if your susceptibility model is a single layer of constant thickness, you can simply add a constant to your susceptibility model to force all the values to be positive without changing the magnetic field response. Remember that an infinite, constant thickness slab of constant susceptibility produces no magnetic response (i.e. the trivial annihilator).Senior Scientist  Potential Fields3 
I have the same issue, I have aeromagnetic survey and would like to calculate the App. Mag. Susc at a certain depth (e.g. 500 m) to be used as lateral distribution Susc. grid in GMSys modeling. The App.Mag.Susc.500.grd has negative values as WilliamMcNeice mentioned.
One more question, I think the magmap produce App. Mag. Susc. in SI unit. In GMSys, should I convert it into MicroCGS unit?
One more question, If I would like to estimate the Mag. Susc. at 500 m depth from the ground, the flight height is 300 m, I guess I have to insert 800 m, am I right?0 
If you are running the filter in MagMap and then importing the results into GMSYS 3D, you should convert SI to microcgs. If you are using the GMSYS 3D inversion, it takes care of that for you.
In MagMap, it is the distance below the survey level. In the GMSYS 3D Lateral Susceptibility inversion you specify 1) the name of the layer to be inverted, 2) the elevation of the “nominal top of surface” being inverted, & 3) “nominal mean of surface” which is the top of the layer. These last 2 entries essentially act as a filter. If you put the top of the layer too close to the survey elevation, the invesion can “blow up”. If that happens, move those down. Note that the Apparent Susceptibility filter is used in the GMSYS 3D inversion in the “sharpening iterations”.Senior Scientist  Potential Fields0
This discussion has been closed.