We welcome the opportunity to discuss the geodynamics of the Rif-Betic region and will address the points in the order raised in the comment.
1. There is no rigid Alboran plate. Instead, a portion of the upper plate (the Gibraltar block) is pulled westward by hydrostatic suction, induced by the rollback of the steeply subducting slab (see analog models by Shemenda, 1994). This motion is observed at several Global Positioning System (GPS) stations, with vectors of 5–10 mm/yr in a west-southwest direction (with respect to Eurasia), independent of the 4 mm/yr northwest motion of African plate stations in northwest Morocco (Reilinger et al., 2001). This leads to active opening of the western Alboran Sea, consistent with recent east-west extensional focal mechanisms reported here (Stich et al., 2003), and with the strong crustal thinning (<10 km) and very thick Neogene sediments (>10 km) observed here.
2. The >600-km-long slab is not necessarily Miocene in age. It formed partly by north-northwest–directed subduction of Tethyan lithosphere (since 35 Ma), followed by rapid rollback to the west, since the Miocene. Slab length is not completely expressed as surface deformation. Crustal shortening (as recorded by sedimentary and basement deformation) in the Andes ranges from 100 to 400 km (Kley and Monaldi, 1998), yet thousands of kilometers of the Farallon-Nazca slab have subducted there.
3. Lithospheric delamination and deblobbing models were based largely on north-south tomographic cross sections showing a high P-wave velocity body (cold, apparently detached lithosphere) beneath a low P-wave velocity layer (warm asthenosphere) (see Calvert et al., 2000, their Fig. 12C; Gutscher et al., 2002, their Fig. 4D). Eastwest–oriented cross sections (Calvert et al., 2000, their Fig. 12B; Gutscher et al., 2002, their Fig. 4, A and B) demonstrate westward continuity with Atlantic oceanic lithosphere, effectively ruling out the deblobbing model. The steep east-dipping slab has a curved, horseshoe shape in three dimensions, similar to the Calabria slab (see the 200 km horizontal cross section of Wortel and Spakman, 2000, their Fig. 2A). Our southern cross section (Gutscher et al., 2002, Fig. 4B) shows a gracefully flexed slab, with a thickness of ~100 km (narrowest here due to its three-dimensional geometry), that in no way resembles a “blob.”
4. Nowhere else in the world does intermediate depth and deep focus seismicity occur in the absence of a subducting slab of oceanic lithosphere. Both types of seismicity occur beneath the Gibraltar arc and southern Spain (Buforn et al., 1991). Numerous subduction zones show a gap in seismicity between ~200 km and 600 km depth (e.g., Colombia, Peru).
5. Active deformation in the accretionary wedge cannot be gravity driven, since the basement and décollement are shown by wide angle seismic data to be eastward dipping (Gutscher et al., 2002, their Fig. 3A). Indeed, Flinch interprets the external Betic-Rif units to be “an accretionary complex . . . (with) some similarity to the Aleutian accretionary wedge” (Flinch, 1993, p. 247). Coeval normal faulting can occur arcward of subduction related thrusts.
6. New geochemical data indicate subduction-related magmatism in the Gibraltar arc until 4.8 Ma, with a shift in geochemistry from calc-alkaline to potassic before extinction, a trend also observed in the Aeolian-Calabrian magmatic arc (Duggen et al., 2003).
7. Syncollisional exhumation (Chemenda et al., 1995) can produce pressure-temperature-time paths similar to those observed in the peridotites, a process not considered earlier. This mechanism is consistent with recent U-Pb dating of zircons, indicating Miocene subduction prior to exhumation (Sanchez-Rodriguez and Gebauer, 2000).
Additionally, active mud volcanoes have now been mapped in all surveyed areas of the accretionary wedge (about two-thirds), but none have been mapped outside (Pinheiro et al., 2002).
In conclusion, subduction and rollback of a narrow east-dipping slab (in an overall north-south convergent system) is an attractive unifying model, able to explain a variety of phenomena: surface deformation, kinematics, magmatism, exhumation, backarc basin formation, deep mantle processes, and the horseshoe shape of the Betic-Rif orogen. More importantly, it makes testable predictions. We invite Platt and Houseman (and others) to propose tests to discriminate between the delamination model and the subduction model. We plan bathymetric surveys in the Gulf of Cadiz to examine the morphology and continuity of active thrust faults in the accretionary wedge and to complete the inventory of mud volcanoes. We eagerly await acquisition of additional GPS data. We suggest, let's collect the necessary evidence and then put the case to rest.