Content area
Abstract
My research focuses on understanding the tectonic and upper-mantle structure of Costa Rica through the perspective of passive-source seismology to place constraints on the timing and implications of the actively evolving Costa Rican subduction system. Specifically, I aim to place constraints on three primary geologic observations of southern Costa Rica: 1) the occurrence and youthful uplift of the Cordillera de Talamanca, 2) the cessation of southern Costa Rica volcanism at ~6.5 Ma, and 3) a notable lack of seismicity beneath the Pacific side of southern Costa Rica. In the first contribution, I produced multi-taper spectral correlated receiver functions to measure impedance structure variations within the crust and upper mantle of southern Costa Rica. These measurements are compared to well-relocated Wadati-Benioff zone seismicity to generate detailed interpretations of a complex upper-mantle structure beneath Costa Rica, specifically the Cordillera de Talamanca. In the second contribution, I produced receiver functions from the previously understood back-arc, or Caribbean side, of southern Costa Rica. These measurements are compared to historical seismicity along a previously debated plate boundary to form new impedance-structure constraints for the Caribbean side of southern Costa Rica. For the third contribution, I utilize a suite of seismological tools to further characterize the southern Costa Rican subduction system. I perform a synthetic modeling exercise, using previous studies to attempt a layer stripping approach to probe the lower crust composition and tectonic make-up beneath the Cordillera de Talamanca. I also use the H-Kappa stacking algorithm to attempt to place compositional constraints on the southern Costa Rica crust by means of Vp/Vs estimates.
In the first contribution, I find that the well-relocated Wadati-Benioff zone seismicity and the precisely measured impedance boundaries do not easily align. My receiver function impedance measurements produce clear systematic pulses at 55 – 60 km and at 25 – 30 km depth. Clear Wadati-Benioff zone seismicity can be observed to ~60 km beneath the Cordillera de Talamanca; however, it is clearly offset from my measured impedance changes. This led to an interpretation of a fragment of lithospheric origin existing beneath the Cordillera do Talamanca, likely held in place by buoyancy forces at or near the crust-mantle boundary.
In the second contribution, I find a clear change in impedance structure from one side of the Central Costa Rica Deformed Belt (CCRDB) to the other. To the north of the CCRDB I see a single clear impedance pulse at ~1.5 – 2 seconds; to the south of the CCRDB I observe two impedance pulses, one at ~ 2 – 3 seconds and another at ~ 3 - 4 seconds. The CCRDB is previously noted as a plate boundary separating the Caribbean plate and Panama microplate. The second impedance contrast south of the CCRDB and beneath the Panama microplate overlies historical seismicity and active source seismological imaging of this area and suggests that this second pulse is evidence of a currently subducting Caribbean plate beneath the Panama microplate.
For the third contribution, the layer stripping synthetic modeling experiment produced interesting plots but did not have the resolution to adequately answer the question of lower-crustal composition. However, a need for a synthetically modelled wide-spread low-velocity layer produced an interesting constraint which suggests that there is a hydrated mantle on the Pacific side of the Cordillera de Talamanca that may be ultra-mafic serpentinites. Trial-and-error synthetic modeling for stations on the Caribbean side of Costa Rica agrees with the interpretation of Chapter 3 that there is a tectonic-scale impedance contrast within the upper mantle to the south of the CCRDB but not observed to the north. This feature is likely the subduction of the Caribbean plate beneath the Costa Rican landmass. Results of the H-kappa experiment overall failed to adequately place constraints on the Vp/Vs structure as well. The need for multiply reflected phases which were not observed in the data did not allow for a satisfactory convergence on any Vp/Vs value.
