North American Magnetic Bottom/Curie Depth estimates and their significance for lithospheric temperatures and magnetization

Spectral magnetic methods have been used to derive the regional layered magnetic structure of the conterminous U.S. and parts of Canada from the North American full spectrum magnetic anomaly dataset. Large window sizes of 500 km were used to have adequate depth sensitivity and the depths were evaluated at 250 km spacing. The primary method used for the determination was the centroid slope method of Bhattarcharyya and Leu (1975). The spectral peak method (Ross et al., 2006; Ravat et al., 2007) and the one layer fractal method (Maus et al., 1997; Bouligand et al., 2009) were used to corroborate the results according to the nature of the spectra in appropriate locations. The depth resolution is poor where magnetic bottoms are deep because spectral slopes are steep, and consequently the depth errors are likely greater than 10 km. The derived magnetic bottom variation is related to a combination of factors including geology, tectonic province, mantle heat flow, and the crustal thickness. Only in a few cases was the derived magnetic bottom significantly deeper than the Moho. In most regions, including the cold Archean and Proterozoic provinces, magnetization appears to primarily lie in the crust and mantle may be non-magnetic. The large window size and spacing necessary for the methods of this study are not conducive to detection of serpentinized mantle in active subduction zones (e.g., Blakely et al., 2005). The deepest crustal magnetic layers must have strong magnetization (2-5 A/m), aided by viscous magnetization acquired since the last geomagnetic reversal and by the Hopkinson effect in order to reconcile rock-magnetic properties and satellite derived magnetization estimates.