Hi Elizabeth I tried these steps and also tried some other constraints. Although I did not set Reweighting constraint to IRI=2 . What exactly does that constraint do? The thing I find strange is the inversion always gives very low contrast (background values) near the surface, so I get a low density that fans out rather than a basin shape that dips towards the centre.
I think a big problem is getting an accurate residual from this data. You would suggest removing a linear background? I tried the feature in Voxi that removes linear background and I tried polynomials with the database. A lot of the anomaly is only half defined by parts of the survey making it hard to remove background. Is there a way to constrain the inversion to only have 2 possible values? 2.0 g/cm3 (or -0.45) in valley and 2.45 (or 0) for bedrock. I see you can use upper and lower limit voxels but you always end up with an even distribution of densities within those limits. Thanks, Doug On Fri, Apr 12, 2013 at 4:03 AM, Elizabeth Baranyi < [email protected]> wrote: > Hello Doug**** > > ** ** > > Indeed GM-SYS would have been the approach to take with your model, the > alternative approach with VOXI would be:**** > > **· **Remove a linear background from your gravity data. > Consequently you will be modelling the gravity contrast as opposed to > absolute densities.**** > > **· **First set the Reweighting constraint to IRI=2 and run the > inversion. If you do not find that the solution represents your geological > knowledge then try the approach below.**** > > **· **Set the upper bound constraint using the basin depth **** > > **o **Create a grid that covers your area of interest of a constant > value equivalent to the 650 m depth on your Z axis. **** > > **o **Generate an upper bound constraint using this grid and set the > upper limit above this grid (basin fill) to the upper density contrast > limit of the basin fill**** > > **o **Set the upper limit of the volume below the 650 m (bedrock) to a > much larger value. **** > > **o **Set the Reweighting constraint to IRI=2**** > > **· **Run the inversion**** > > Let me know how the above works, and if would you like further > recommendations.**** > > Regards**** > > *Elizabeth Baranyi > *Technical Product Manager, Geophysical > > *GEOSOFT INC. > **freedom to explore > *T +1 416.369.0111 #356 > C +1 647.504.6006 > F +1 416.369.9599 > > Visit our site at www.geosoft.com**** > > ** ** > > *From:* [email protected] [mailto:[email protected]] *On > Behalf Of *Doug Perkin > *Sent:* Thursday, April 11, 2013 4:12 PM > *To:* Geosoft Oasis montaj and Target > *Subject:* [geonet] Gravity Modelling**** > > ** ** > > Hi**** > > Does anyone have experience using Voxi for 3D modeling of basins using > gravity? I'm interested in getting a rough model with basin depth and a > rough idea of its general geometry. I want to create a simple model with > two density values: basin fill and bedrock. I also already know that the > basin is at least 650m deep. I have tried using some of the constraints in > Voxi, but have not gotten a reasonable model. **** > > I guess the hardest thing is just separating the response due to the basin > from the rest of the Bouguer anomaly. I have tried polynomials, second > order looks best. But just wondering if anyone else here has experience > attempting this? I see GM-Sys is probably a better approach but don’t have > a license for that yet.**** > > ** ** > > Thanks,**** > > ---**** > > Forum archives can be accessed here: > http://lyris.geosoft.com/read/?fourm=geonet**** > > You are currently subscribed to geonet as: [email protected].* > *** > > To subscribe or unsubscribe from any of our forums, select the User Forum > selection on the Geosoft Community page: > http://www.geosoft.com/support/community**** > > --- > > Forum archives can be accessed here: > http://lyris.geosoft.com/read/?fourm=geonet > > You are currently subscribed to geonet as: [email protected]. > > To subscribe or unsubscribe from any of our forums, select the User Forum > selection on the Geosoft Community page: > http://www.geosoft.com/support/community > --- Forum archives can be accessed here: http://lyris.geosoft.com/read/?fourm=geonet You are currently subscribed to geonet as: [email protected]. To subscribe or unsubscribe from any of our forums, select the User Forum selection on the Geosoft Community page: http://www.geosoft.com/support/community
