Comment on "deep resistivity cross section of the intraplate Atlas Mountains (NW Africa): New evidence of anomalous mantle and related Quaternary volcanism"

[1] Tectonic interpretation of images and models of the subsurface derived from geophysical data must be based on a firm foundation of the following four precepts; (1) the data must be shown to be internally consistent, (2) the dimensionality assumed for modeling must be shown to be valid, and if 2D then the adopted strike direction must be shown to be consistent with the data, (3) the final model must fit the data to within their statistical limits, and (4) resolution of interpreted features must be demonstrated. Particularly in a highly nonlinear problem, such as electromagnetic geophysics, without examining model resolution it is impossible to know the veracity of the interpretation. [2] In their paper, modeling magnetotelluric (MT) data across the Atlas Mountains, Anahnah et al. [2011] undertake a tectonic interpretation of a derived electrical resistivity model focusing particularly on a region of purported anomalously low resistivity in the uppermost mantle. However, the data and resulting model on which their interpretation is based are highly suspect, particularly at the longer periods pertaining to lithospheric mantle depths, meaning that the corresponding interpretation has not been proven by the extant data and cannot be trusted. We have little issue with the crustal part of Anahnah et al.'s [2011] model; we contend that the mantle part of their model is highly suspect and that the evidence for thinned lithosphere is simply nonexistent in their data, and indeed is refuted by their data, but appears to be contrived to fit preconceived notions. Our views are entirely based and stated on objective criteria, and suggestions otherwise are refuted and are attempts to detract from our criticisms. [3] At issue is that the data used to construct the two-dimensional (2D) model presented byAnahnah et al. [2011, Figure 4] are questionable, the adopted strike direction is likely invalid, the model does not meet the dimensionality of the data, the model does not fit the data adequately, and the model was not subsequently appraised sufficiently for resolution. 2D modeling is appropriate at those periods sampling the crust, and is also reasonably appropriate for those periods sampling the mantle, but as the crust and mantle have different geoelectric strike directions, 2D modeling of the whole bandwidth of crustal and mantle periods is not appropriate. [4] In their reply, Anahnah et al. [2012], who introduce new material not in the original, including a new model, and that took 6 months to compose, attempt to address some of the points raised in this comment but fail to address the primary one, that is, that the model misfit is completely unacceptable, mainly due to the model responses not fitting the data at all at long periods. Hence their model, and consequent interpretation, of the lithospheric mantle is highly suspect at best and likely erroneous. Therefore, its tectonic interpretation cannot be believed and should not be accepted.