See references in Salem et al. (2014, Tectonophysics). See Bansal et al. (2011? Geophysics) on how to use fractal model for computations. They use small windows with the previously specified windows. Also, if you use centroid method, you cannot determine fractal parameter of the “field", you have to assume one. The higher the fractal parameter, the lower the obtained depth.
See Bouligand et al. (2009, JGR) to see how one may construct a fractal model to test whether the method is working. Bouligand et al. also have another parameter search method for fractal modeling, but once again the fractal parameter of “magnetization” (in this case), must be assumed. The minimum RMS between observed and modeled spectra depends generally underestimates the depth. Carol Finn of USGS is trying to get the Geosoft’s Curie depth/magnetic GX make available publicly (Claire Bouligand and I gave our previous programs for this). She said you need to ask Dima at Geosoft and she will request the permission for making the module available on a case by case basis. (Both Dima and Carol are going to hate me for writing this in a reply on the listserver since this could mean added work for it - so if anyone doesn’t anticipate using this in the next year or so might want to wait. But I would presume the rest of you will love me for it. :-) These methods cannot determine fractal parameter. The only one that does is the Salem et al (2014, Tectonophysics). My model studies show that windows need to be about 500 km unless the magnetic bottom is on the order of 10 km (in which case it could be lower, like 300 km). There is a lot of spectral averaging involved in getting these estimates and one needs to “interpret” the estimates further in complex geological situations - many factors come into play). It takes me almost an hour (sometimes more) to be happy with a single estimate and one also needs the knowledge of regional geology and geophysics to make sense of it (or “interpret” the result). Long wavelength anomaly fidelity is also very important in all this, which means if one is processing data pre-2005 IGRF, you need to strip off IGRF and use the Comprehensive Model of the magnetic field (CM3, and CM4, Sabaka et al., 2002, 2004, GJI). Removing degree 15 main field in the CM is very useful. If one is not willing to do all this (including model studies to understand the limitations), one should avoid this problem. Cheers, tiku On May 19, 2015, at 5:25 AM, solomonjanjaro <[email protected]> wrote: > Dear all. > Can anyone help me with the procedure and steps on how to use Fractal method > to calculate curie surface using aeromagnetic data. > Thank you. > Solomon Nehemiah Yusuf > PhD student (Mautech) > > On 19 May 2015 07:21, "solomonjanjaro <[email protected]>" > <[email protected]> wrote: > Thank you for your contributions and also sharing from your wealth of > experience. I think I will quickly revisit my window size and also follow > all these suggestions on this forum to achieve the desired result. > Grateful > Solomon Nehemiah Yusuf > PhD student > Mautech, Nigeria > > On 19 May 2015 03:55, "Dhananjay Ravat <[email protected]>" > <[email protected]> wrote: > Thank you, Steve, for the paper recommendation :-) > > Here are a couple of really important things, I think, to keep in mind. > (There are a lot more, but these are really really important, keeping the > original question in mind.) > > One needs to use large windows (quite a bit larger than what people have been > using - otherwise one ends up looking at “some” layer in the upper crust when > the true Curie depth might be deeper. My experience with a lot of fractal > and other model studies is that window size needs to at least 10 times "the > expected Curie depth-to-observation elevation distance" (I have applications > where data are at 20-30 km elevations). Upward continuing is not a good idea > as a result with spectral methods I think - if anything one is making > response of the magnetic bottom even longer wavelength and smaller in > amplitude. The smaller the window size the larger the variance of the > solutions; so with small windows (less than 200-300 km) one could sometimes > get a close to correct result for deep magnetic bottoms, but the variance of > the solutions is really large to be usable - but one doesn’t know that if one > is using the centroid method. Despite large windows, one needs to use some > tapering or other spectral estimation strategies in order reduce the variance > of the low-wavenumber spectral estimates. > > Taking into account the fractal parameter of the field is very important (if > the underlying magnetization is fractal - true in most cases). Salem et al. > (2014, Tectonophysics, The Defractal Method) show how to go about estimating > fractal parameter by modeling the spectral peak. But one needs to do model > studies, and more model studies with 2D and 3D fractal models, to get > experience on how to assess fits of fractal models - before applying the > method (this is especially true because the RMS minimum of the fit is not the > correct solution most times - I think this happens because it is difficult to > pick perfect range of wavenumbers over the fit needs to be performed for > every spectra and also because of the noise in low-wavenumber spectral > estimates. > > Finally, for certain geologic situations there are multiple layers: e.g., > magnetic, non-magnetic/less-magnetic and once again deeper magnetic layers, > in these situations the answers are some sort of weighted average of depth > solutions, or dipping magnetic layers, and therefore the solution may not > correspond to a real interface. > > Despite all this, I am still using the methods, more and more with the > defractal method with large windows because of the ability to estimate the > fractal parameter by evaluating results with a range of fractal parameters, > and selecting manually minimum/maximum solution range, and the best solution. > Then cross-checking with heat flow or whatever other geophysical models > available in the area. All this is fairly time-consuming…. > > I think I have long surpassed the length for an effective e-mail > communication. > > I hope this helps. > > -tiku > > > > On May 18, 2015, at 3:43 PM, Steve Sheriff <[email protected]> > <[email protected]> wrote: > >> Take a look at Ravat et al., 2007, Geophys J. Int'l v169, p 421-434 "A study >> of spectral methods of estimating the depth to the bottom of magnetic >> sources from near-surface magnetic anomaly data" >> >> On 5/18/2015 10:10 AM, solomonjanjaro <[email protected]> wrote: >>> ED CUNION, >>> In a situation where I have only aeromagnetic data to use.....which other >>> methods do I use? Please suggest. >>> Solomon >>> >>> On 18 May 2015 17:05, "edcunion <[email protected]>" >>> <[email protected]> wrote: >>> Beside longer wavelength magnetic data other available datasets that might >>> be tried to crosscheck the mag result depths are regional gravity, >>> satellite gravity datasets are freely available on the web with Grace being >>> just one example, seismic surface wave shear velocity models (coarse in >>> some areas) may also help, some of these data too are obtainable on the >>> web, and regional MT studies, Alan Jones and his teams have published >>> several papers with images from different continents, there may be some >>> regional depth inferences in these MT papers. >>> >>> Ed Cunion >>> Red Rocks Geophysical Consulting >>> 13224 W Utah Circle >>> Lakewood, CO 80228 USA >>> email: [email protected] >>> >>> >>> >>> On Monday, May 18, 2015 6:49 AM, "solomonjanjaro >>> <[email protected]>" <[email protected]> wrote: >>> >>> >>> Dear all. >>> Please I need help on different methods of determining crustal thickness >>> using aeromagnetic data. >>> 1. I have upward continued filter on my residual data and window a block of >>> 66x66 km and I have 16 blocks of this size over my study area and I have >>> also used this formula Zb=2Zo - Zt >>> Where Zb = curie depth >>> Zo = depth to centroid >>> Zt= depth to top >>> But want other methods to confirm result. >>> Thank you >>> Solomon Nehemiah Yusuf >>> PhD student( Mautech Nigeria) >>> --- >>> Forum archives can be accessed here: >>> http://lyris.geosoft.com/read/?forum=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/forums/register/ >>> >>> >>> --- >>> >>> Forum archives can be accessed here: >>> http://lyris.geosoft.com/read/?forum=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/forums/register/ >>> >>> --- >>> >>> Forum archives can be accessed here: >>> http://lyris.geosoft.com/read/?forum=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/forums/register/ >>> >> >> -- >> >> Steven Sheriff >> Professor Emeritus of Geophysics >> University of Montana >> Missoula, MT USA 59812 >> www.umt.edu/geosciences >> --- >> >> Forum archives can be accessed here: >> http://lyris.geosoft.com/read/?forum=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/forums/register/ >> > > --- > > Forum archives can be accessed here: > http://lyris.geosoft.com/read/?forum=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/forums/register/ > > --- > > Forum archives can be accessed here: > http://lyris.geosoft.com/read/?forum=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/forums/register/ > > --- > > Forum archives can be accessed here: > http://lyris.geosoft.com/read/?forum=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/forums/register/ > --- Forum archives can be accessed here: http://lyris.geosoft.com/read/?forum=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/forums/register/
