When international researchers began formally conducting deep seismic meetings in 1984, they were asking ``What are the origins of deep reflections? Can we observe consistent features around the world?.'' The answer to the last question was generally ``Yes, however these features often serve to confuse us.'' Twelve years later, the issue of crustal reflectivity remains a complex one. Illustrative of some advances, intra-crustal and sub-crustal water is no longer believed to be a primary source of major reflections that it once was. Reflections are now regularly attributed to intrusive sills and basalts, perhaps deformed in structural shear zones. Furthermore, global analysis shows reflectivity can be profoundly modified by successive episodes of lithospheric deformation.
In recent times, the application of rock physics principles to deep seismic data has proven to be a useful technique in understanding the petrology and rheology of the crust, however, significant problems remain. An example is the discrepancy that exists in reconciling the different composition of continental and oceanic crust. Whereas the oceanic crust is fundamentally uniform and predictable in composition, the (much older) continental crust is poorly understood. The chemical composition of the continents is largely intermediate, in striking contrast to the mafic oceanic crust and the ultramafic mantle. Results of modeling the crust as a stochastic medium are consistent with the intermediate model (Levander et al., 1994). The problem is whether these compositional differences are primarily due to different mechanisms of crustal extraction from the mantle or to different mechanisms of differentiation and alteration of newly formed continental crust (Rudnick, 1996).
Answers to many other fundamental questions are also not clear. The fate of subducting slabs in the lower lithosphere has never been determined, yet in an ad hoc way the dipping reflections that are infrequently observed in the upper mantle are typically interpreted as representative of such slabs. Many deep reflections are well documented; however, it has never been properly explained why the lower crust is so anomalously reflective and electrically conductive. On this note, the crust/mantle boundary and seismic transition called the ``Moho'' remains poorly understood. Based on both reflection and petrological considerations, questions have arisen regarding whether the Moho really does represent or coincide with the crust-mantle boundary between rocks of crustal origin and tectonized mantle peridotite, whether in subducted/accreted crust multiple Moho horizons exist, and whether the Moho transition incorporates a large-scale heterogeneous upper mantle ``zone''. These uncertainties are compounded by the complexity of the recorded seismic wavelets. For example, the seismic Moho is often resolved at different depths by complementary seismic reflection and refraction studies. Seismic modeling ambiguously predicts that both discrete and multi-layered structures can yield the seismic signatures observed on deep data. Such problems need to be solved. Walter Mooney from the USGS has recently shown that the upper-mantle has a surprisingly large control upon near-surface processes. Therefore, these formerly obscure deep geological features are suddenly more significant than they once were.