Intellectual merits. The cycling of fluids into and out of subduction zones plays an integral role in the formation of volcanic arcs. The Central American subduction zone is an ideal location for studying the effects of fluid on the arc and underlying mantle wedge and subducted slab due to its varying inputs and outputs. Existing geochemical studies indicate wide variability in degree of slab hydration along this arc, with maximum and minimum slab contributions to magma generation localized beneath Nicaragua and Costa Rica, respectively. Geochemical studies suggest that slab fluid contributions are also present beneath El Salvador and Guatemala, but diminish to the northwest. Geophysical studies are consistent with a high degree of slab serpentinization and fluid release beneath Nicaragua, versus comparatively little beneath Costa, but otherwise the rest of the arc is poorly characterized geophysically - primarily because the slab is not well-resolved beneath El Salvador and Guatemala. The present study will utilize all available seismic data to extend the velocity model developed for Costa Rica and Nicaragua into the northwestern section of the arc, allowing regional analysis of the relation between seismic structure and fluid distribution. Data reported to the International Seismic Centre will be combined with additional data from local seismic networks from Guatemala through Costa Rica, which has never been integrated to study the entire Central American subduction zone. This study will obtain a more extensive seismic image of the slab and mantle wedge within the Central American subduction zone through the use of a double difference velocity tomography and relocation method.
Broader Impacts. The proposed project will primarily support the professional development of a young female scientist (Syracuse), who has not has prior NSF support. Syracuse will participate in the UW-Madison DELTA program for Integrating Research, Teaching, and Learning. DELTA supports current and future faculty in science, technology, engineering and math in their improvement of student learning.
The primary focus of this project was analyzing seismic data for the Central American subduction zone from a variety of international sources (International Seismic Centre, UK; Central American Seismic Center, Costa Rica; the TUCAN array, a previous NSF-funded project) and extending the existing body-wave velocity model of Syracuse et al. (2008) to northwestern the Central America. We used a state-of-the-art travel time tomography algorithm (Lin et al., 2010, Zhang and Thurber, 2003) with the goal of assessing along-arc differences in seismicity and water content in the slab and overlying mantle wedge, which could then be tied to differences in volcanism throughout Central America. Ten years of origin time and P- and S-wave arrival times were compiled for approximately 8000 earthquakes occurring between 2000 and 2009. We were unable to obtain uncertainty estimates for many of these arrival times, and as a result, their quality needed to be carefully addressed before accepting and interpreting the results. The types of uncertainties that affect arrival times, and therefore tomographic inversions, include timing errors (e.g., a seismometer’s clock shifted by as little as a few milliseconds, or a uniform delay in one or more stations’ data due to a delay caused by telemetering data to a central location) and analyst errors (e.g., picking an onset from a noisy signal). In order to identify individual stations with potential widespread timing errors, we tested the contribution of each station to the tomography; if a station was found to have a disproportionate effect on the inverted model, it was considered suspect. Out of approximately 300 stations, three were found to have possible timing errors and were removed from further analysis. While the quality of individual arrival times could not be assessed, a variety of tests were designed to remove the effects of poorly recorded earthquakes. These tests included placing minimum thresholds on the number or distance range of stations recording an earthquake, the number of neighboring earthquakes that can be used to constrain a location, among many others. The velocity model obtained using other tomographic codes were also compared to ensure that the results were not entirely unique to the particular algorithm being used. While we ultimately found that the data were too noisy to be successfully used in an inversion and the inversion results were not reliable enough to be confidently published, this project did provide a variety of benefits to those involved. It provided the PI her first experience leading a grant, and this experience has helped her write other successful proposals that have funded other more successful projects. It has also helped her guide other first-time PIs through grant writing and provide a graduate student with hands-on experience performing travel time tomography, which is a technique central to the field of seismology.
