As a preliminary feasibility study of a proposed 3D tomographic inversion of new onshore-offshore seismic data, we have carried out a forward modeling exercise of traveltimes using realistic 3D velocity models for Kilauea Volcano. Approximately 2000 km of airgun shots set off during a 1998 marine seismic survey on board the R/V Maurice Ewing over the southern and eastern offshore region of the island of Hawaii were recorded by the permanent array of on-land seismic stations maintained by Hawaii Volcano Observatory - US Geological Survey. This data set provides the unique opportunity to study regions of Kilauea Volcano that have been poorly imaged by previous passive tomographic inversions.

Several 3D velocity models were created to portray the major subsurface features of Kilauea Volcano. In this study, a 1D double-gradient velocity-depth model proposed by Klein (1981) serves as the reference model. Even this simplified model yields good correspondence between predicted and recorded traveltimes for actual shot-receiver geometries. Velocity perturbations based on recent passive seismic tomographic studies of Kilauea's summit and rift zones (e.g., Okubo et al., 1997; Haslinger et al., 2001) were then introduced into the reference model in key regions within the Kilauea subsurface. For example, high Vp regions were constructed beneath the active rift zones and summit with velocities of up to 7.6 km/sec and 8.0 km/sec respectively, representing intrusive dikes and cumulate deposits, and a low Vp zone with a maximum of only 5.4 km/sec was placed in the coastal and offshore region, denoting either deep seated slump faults or clastic deposits generated by shoreline crossing lava flows. Differences in traveltimes between the reference and alternate models were calculated using actual shot-receiver geometries for several seismic stations, sampling key regions of Kilauea and the island of Hawaii. As expected, coastal stations were unaffected by the addition of high Vp zones in the cores of Kilauea and Mauna Loa Volcanoes, while summit and inland stations showed traveltime drops up to 300 ms for nearshore shots. The opposite occurred with the addition of the coastal and offshore low Vp zone, which resulted in significant traveltime increases up to 1200 ms for nearshore shots and up to 500ms at greater shot-station offsets. The observed traveltime differences for the various shot profile-station combinations reflect unique raypath trajectories through the volcanic edifice and underlying ocean crust and mantle. Calculated traveltime perturbations from this study indicate that a full 3D tomographic analysis of the complete onshore-offshore data set will be able to resolve details of many key geologic structures of the dynamic Kilauea volcano.

(submitted to Fall AGU Meeting in San Francisco, 2001)