Volumetric characteristics of lava flows from interferometric radar and multispectral data: the 1995 Fernandina and 1998 Cerro Azul eruptions in the western Galápagos.

Rowland, Scott K. and Harris, Andrew L. and Wooster, Martin J. and Amelung, Falk and Garbeil, Harold and Wilson, Lionel and Mouginis-Mark, Peter J. (2003) Volumetric characteristics of lava flows from interferometric radar and multispectral data: the 1995 Fernandina and 1998 Cerro Azul eruptions in the western Galápagos. Bulletin of Volcanology, 65 (5). pp. 311-330. ISSN 1432-0819

Abstract

We have used a suite of remotely sensed data, numerical lava flow modeling, and field observations to determine quantitative characteristics of the 1995 Fernandina and 1998 Cerro Azul eruptions in the western Galápagos Islands. Flank lava flow areas, volumes, instantaneous effusion rates, and average effusion rates were all determined for these two eruptions, for which only limited syn-eruptive field observations are available. Using data from SPOT, TOPSAR, ERS-1, and ERS-2, we determined that the 1995 Fernandina flow covers a subaerial area of 6.5×10 6 m 2 and has a subaerial dense rock equivalent (DRE) volume of 42×10 6 m 3. Field observations, ATSR satellite data, and the FLOWGO numerical model allow us to determine that the effusion rate declined exponentially from a high of ~60–200 m 3 s -1 during the first few hours to <5 m 3 s -1 prior to ceasing after 73 days, with a mean effusion rate of 4–16 m 3 s -1. Integrating the ATSR-derived, exponentially declining effusion rate over the eruption duration produces a total (subaerial + submarine) DRE volume of between 27 and 100×10 6 m 3, the range in values being due to differing assumptions about heat loss characteristics; only values in the higher part of this range are consistent with the independently derived subaerial volume. Using SPOT, TOPSAR, ERS-1, and ERS-2 data, we determine that the 1998 Cerro Azul flow is 16 km long, covers 16 km 2, and has a DRE volume of 54×10 6 m 3. FLOWGO produces at-vent velocity and effusion rate values of 11 m s -1 and ~600 m 3 s -1, respectively. The velocity value agrees well with the 12 m s -1 estimated in the field. The mean effusion rate (total DRE volume/duration) was 7–47 m 3 s -1. Dike dimensions, fissure lengths, and pressure gradients along the conduit based on magma chamber depth estimates of 3–5 km produce mean effusion rates for the two eruptions that range over nearly four orders of magnitude, the range being due to uncertainty in the magma viscosity, dike dimensions, and pressure gradient between magma chamber and vent. Although somewhat consistent with mean effusion rates from other techniques, their wide range makes them less useful. The exponentially declining effusion rates during both eruptions are consistent with release of elastic strain being the driving mechanism of the eruptions. Our results provide independent input parameters for previously published theoretical relationships between magma chamber pressurization and eruption rates that constrain chamber volumes and increases in volume prior to eruption, as well as time constants of exponential decay during the eruption. The results and theoretical relationships combine to indicate that at both volcanoes probably 25–30% of the volumetric increase in the magma chamber erupted as lava onto the surface. In both eruptions the lava flow volumes are less than 1% of the magma chamber volume.