RECENT VOLCANO-TECTONIC EVOLUTION OF THE WESTERN GALÁPAGOS

Abstract

Calderas are the most active and dangerous volcanic systems on Earth. All active calderas experience periods of unrest characterized by changes in the baseline monitoring parameters, such as ground deformation, degassing and seismicity, which can last from hours to years. While only some unrest episodes culminate in eruption, most eruptions are preceded by unrest. Therefore, understanding the nature of unrest becomes fundamental to assess volcanic hazard and represents one of the main problems in volcanology. Active mafic calderas undergo repeated unrest episodes, often characterized by multiple and regular cycles of uplift, with pre-eruptive seismicity followed by eruption and co-eruptive deflation. Conversely, felsic calderas are restless for decades or show only isolated and short unrest episodes and erupt infrequently. Mafic calderas, therefore, provide simpler conditions to study unrest. In this group are the six active calderas of the western Galápagos (Ecuador), which share similar characteristics and experienced repeated unrest episodes in the last decades. This Ph.D. Thesis aims at widening the knowledge on the unrest episodes at the western Galápagos calderas and at extending such results to other mafic calderas with similar characteristics. To pursue this aim, I first analysed eruptive and non-eruptive unrest episodes at two of the less studied calderas of the western Galápagos: Alcedo (article published in Journal of Geophysical Research-Solid Earth) and Cerro Azul (article submitted to Journal of Geophysical Research-Solid Earth). In these studies, I investigated the non-eruptive unrest episodes occurring from 2007 to 2011 at Alcedo and the 2008 eruptive unrest at Cerro Azul. As ground deformation is a constant feature in all the unrest episodes at western Galápagos calderas, I studied surface displacements as measured by Synthetic Aperture Radar Interferometry (InSAR). Then, by modelling the geodetic data, I placed constraints on the sources responsible for these episodes of unrest, as has been done for most of the previous studies of unrest episodes at other western Galápagos calderas. The new results highlight how Alcedo experienced two episodes of uplift due to new magma injections in its shallow magmatic system, separated by an episode with a limited lateral propagation of magma, probably interrupted by the lack of new magma supply to the magmatic system. The same results also hint at a possible relationship between these short-term unrest episodes and the longer-term process of caldera resurgence at Alcedo. The 2008 eruptive unrest of Cerro Azul started with eight months of pre-eruptive uplift of the caldera floor, followed, from May to June 2008, by two eruptive phases related to the lateral propagation of a radial dike. To interpret InSAR phase in terms of a geophysical model it is necessary to estimate the integer ambiguities in the phase through the phase unwrapping process. This is generally done prior to geophysical inversion, but any phase-unwrapping errors will bias the resulting geophysical parameters. To overcome the limiting presence of unwrapping errors affecting some of the InSAR data of Cerro Azul we have developed a new method to directly invert the wrapped interferometric phase by estimating the integer ambiguities simultaneously with the geophysical parameters. Finally, I reviewed all the geodetically monitored unrest episodes at the western Galápagos calderas, mainly considering the estimated injected volumes of magma and the related intrusive rates (results included in a not-yet-submitted manuscript). These data were compared to those from other mafic calderas sharing similar characteristics, to find a consistent behaviour. Results highlight a relationship between the rates at which magma is injected inside the shallow magmatic system and the possibility to trigger a magmatic intrusion. In particular, unrest episodes with injection rates > 5 x10-2 km3/year seem to promote the propagation of a dike (eruptive or not) in < 1 year, while rates <9 x10-3 km3/year do not trigger any dike propagation in <1 year and, even over longer time spans, the formation of a dike seems uncommon. Therefore, these data suggest that the rates at which magma is supplied in the shallow system may provide an important informative parameter for the forecast of how the unrest will evolve in a mafic caldera with a shallow magmatic system.