Recognizing and interpreting the longest wavelength lithospheric magnetic signals obscured by overlap with the core field 2004 Fall AGU: GP31A-0821 Michael Purucker, Raytheon Geodynamics Branch GSFC, Greenbelt, MD USA Kathryn Whaler, School of Geosciences, University of Edinburgh, West Mains Rd, Edinburgh, EH9 3JW, UK References, and suggested readings: GMT, 2004, v..4, P. Wessel and W. Smith. Jackson, A., Accounting for crustal magnetization in models of the core magnetic field, Geophys. J. Int., 103, , Langel, R. and Hinze, W., The magnetic field of the Earth’s lithosphere, Cambridge Univ. Press, Cambridge, 429 pp, 1998 Nataf, H., and Ricard, Y., 3SMAC: an a priori tomographic model of the upper mantle based on geophysical modeling, Phys. Earth Plan. Int., 95, , Maus, S., et al., Earth’s crustal magnetic field determined to spherical harmonic degre 90 from CHAMP satellite measurements, Geophys. J. Int, submitted, Available electronically at Mayhew, M. and Estes, R.,, J. Geomagnetism and Geoelectricity, 35, , 1983 Mayhew, M. and Estes, R., Equivalent source modeling of the core magnetic field using Magsat data, J. Geomagnetism and Geoelectricity, 35, , 1983 Parker, R.L., Shure, L. and Hildebrand, J.A., The application of inverse theory to seamount magnetism, Rev. Geophys., 25, 17-40, Parker, R.L., Geophysical Inverse Theory, 386 pp., Princeton University Press, Princeton, Purucker, M.E.,T.J. Sabaka, and R.A. Langel, Conjugate gradient analysis: a new tool for studying satellite magnetic data sets, Geophys. Res. Lett., 23, , Purucker, M, Maus, S., and Luehr, H., From validation to prediction: Lithospheric field studies from Magsat to Swarm, Earth Planets and Space, available electronically at Reigber, C., Luhr, H., Schwintzer, P., and Wichert, J. (editors), Earth Observation with CHAMP, Springer, Heidelberg, Sabaka, T.J., Olsen, N., and Purucker, M., Extending Comprehensive models of the Earth’s magnetic field with Orsted and CHAMP data, Geophys. J. Int., 159, , Shure, L., Parker, R.L. and Backus G.E., Harmonic splines for geomagnetic modelling, Phys. Earth Planet. Int., 28, , Stauning, P., et al. (editors), Proceedings of the 4th Orsted International Science Team Conference, Danish Met. Inst., Copenhagen, 2003, 300 p. Whaler, K. and Langel, R., Minimal crustal magnetization from satellite data, Phys. Earth Planet. Int., 98, , Abstract: We recognize and characterize two distinctive patterns evident in new maps of the lithospheric magnetic field from the CHAMP satellite, and new minimum amplitude magnetization models that we deduce. The boundaries of these patterns define long-wavelength features in the lithospheric field not previously recognized because they were obscured by overlap with the core field. These boundaries correspond to known crustal thickness variations. The major exceptions, the Sahara and most of South America south of the Equator, are regions where direct estimates of crustal thickness and heat flow are sparse. Satellite-based magnetic field maps (MF-3) In order to minimize time-variable fields associated with the interaction of the solar and terrestrial dynamos, we usually utilize spherical harmonic models built from data gathered during magnetically quiet times, rather than the field data directly. Both CM4 (Sabaka et al., 2004) and MF-3 (Maus et al., 2004) are models of this type. We prefer to use MF-3 for the purpose of this exercise because it goes to higher spherical harmonic degree (90 vs 65). MF-3 is a lithospheric field model only, and extends from degree 16 to 90. The CHAMP magnetic field satellite input to MF-3 has had removed an internal field model to degree 15, an external field model of degree 2, and the predicted signatures from eight main ocean tidal components. Additional external fields are subsequently removed in a track-by-track scheme. Because of its design philosophy, MF-3 can be considered a minimum estimate of the lithospheric magnetic field, one in which there will be some suppression of along- track magnetic fields, which are N-S in equatorial and mid-latitudes. Regularization has been applied to degrees higher than 60 to extract clusters of spherical harmonic coefficients that are well- resolved by the data. The highest noise levels remain in and around the auroral zones, and we will defer characterization of the fields in those areas because of the very band-limited nature of the lithospheric signal in those areas. Introduction Features of the lithospheric magnetic field with wavelengths in excess of 3000 km (spherical harmonic degree 13) are completely obscured by overlap with the core field. Between 2600 and 3000 km both core and lithospheric signatures are present, hindering efforts at separation. Previous efforts (see for example Mayhew and Estes, 1983) at separation of the two fields have failed, and there is strong reason to believe that the two fields are not separable unless the core field is shut off, or changed signficantly. However, new higher resolution models of the crustal field are becoming available (Maus et al., 2005). In order to make some progress on qualitatively understanding the longest wavelengths, we borrow an old idea from the exploration geophysics community, and visually characterize the field, and magnetization solutions deduced from that field. Because of the wider spectral content of the new solutions, we hope that larger patterns will become apparent, patterns that were not obvious when we were examining very band-limited solutions. By way of analogy, we hope to be able to differentiate ‘the forests from the fields’ by characterizing features at smaller spatial scales (like the ‘trees and grasses’). This analogy implies that our ‘imaging’ technique can’t see the ‘forests and fields’, just the ‘trees and grasses’, and that there are features at small scales that give us clues into what is happening at the largest scales. A 3-component magnetization model from MF-3 In order to further characterize the magnetic field, we derive and show a three-component magnetization model from MF-3. Using all three components, we model magnetization as a linear combination of the Green's functions relating magnetization at any point in a 40 km thick magnetized crust to a satellite measurement of the magnetic field. This avoids subjective choices on the arrangement of equivalent source dipoles (Purucker et al., 2004), and produces a spatially continuous magnetization model. Details of the technique are presented below. The field predicted from the damped inversion is shown above, immediately to the right of the MF-3 model. To the right of that can be seen the three component magnetization solution, the calculated scalar magnetization, and the declination and inclination of the magnetization, plotted where those angles are well-determined. All are plotted at the Earth’s surface. Note that we are NOT assuming that the magnetization is in the direction of the core field. Conclusions We recognize and characterize two distinctive patterns evident in new maps of the lithospheric field deduced from CHAMP. The boundaries of these patterns define long- wavelength features in the lithospheric field not previously recognized because they were obscured by overlap with the core field. These boundaries correspond in a general way to known magnetic crustal thickness variations. The major exceptions, the Sahara and most of South America south of the Equator, are regions where crustal thickness and heat flow are poorly known. MF-3 and its recovery MF-3 model evaluated over North and Central America (Maus et al., 2004) Characterizing the MF-3 model Over the North American region, there are two patterns apparent in the vertical component map predicted at 300 km altitude at the left. The first pattern, which we will refer to as ‘C’, encompasses the North American land mass, the Caribbean and Gulf of Mexico, and northernmost South America. The peak-to-trough magnitude of anomalies in ‘C’ typically exceeds 50 nT, and the anomalies are either equidimensional or oriented in a direction subparallel to the nearest coastline or tectonic element. The second pattern, which we will refer to as ‘O’, encompasses the Eastern Pacific, the Cocos plate, and the western Atlantic away from continental North America. The peak-to-trough magnitude of anomalies in ‘O’ is typically less than 30 nT, and the anomalies are commonly narrow and elongate in the direction of the nearest spreading or subduction zone. The ‘C’ pattern can be discerned on the global maps above, when account is taken of the higher altitude. The ‘C’ pattern is characteristic of much of the Asian landmass, a region centered on but more extensive than Australia, and two broad regions within the African landmass. The ‘O’ pattern is seen in the eastern Pacific, the North and South Atlantic, and the Indian oceans. Characterizing the MF-3 based magnetization model The ‘C’ and ‘O’ patterns are evident in the MF-3 based magnetization model. The magnitude of M shows these patterns unambiguously. Regions with magnetizations greater than 0.1 A/m (red regions on above map) correspond to the ‘C’ pattern, and regions with magnetization less than 0.06 A/m (grey regions) correspond to the ‘O’ pattern. Intermediate values of magnetization (between 0.06 and 0.1 A/m, pink on above map) generally envelop regions displaying the ‘C’ pattern. In a general way, the ‘C’ and ‘A’ patterns correspond to regions of thick and thin magnetic crustal thickness, as defined by temperature and seismology in the 3SMAC model (Nataf and Ricard, 1996) and shown immediately to the right of the scalar magnetization map. There are conspicuous exceptions to this generalization. Most of the South American landmass south of the Equator is characterized by the ‘O’ pattern, yet crustal thicknesses are typical of continental crust. The other major exception is the Sahara desert, again characterized by the ‘O’ pattern but with typical continental crustal thicknesses. Acknowledgments: We thank Gauthier Hulot for providing some clarity to our early work on this subject, and Stefan Maus and the CHAMP team at GFZ for MF-3 Details of technique