Monitoring, forecasting collapse events, and mapping pyroclastic deposits at Sinabung volcano with satellite imagery

Abstract

During the ongoing (2013–present) eruption of Sinabung volcano, north Sumatra, we have routinely used a variety of satellite remote sensing data to observe and forecast lava dome and lava flow collapse events, to map the resulting pyroclastic deposits, and to estimate effusion rates. In this paper, we focus on the first two years of the current eruption (September 2013–December 2015), and we summarize major events in 2016. We divide the eruption into 5 major phases: 1) phreatomagmatic (July 2013–18 December 2013), 2) first dome growth and collapse (18 December 2013–10 January 2014), 3) lava-flow (10 January 2014–mid-September 2014), 4) second lava dome and collapse (mid-September 2014–July 2015), 5) lava dome collapse and ash explosion phase (August 2015–present). Throughout the eruption, remotely sensed information has been instrumental in assessing the stability of the lava dome and flow and to forecast collapse events that produce pyroclastic density currents (PDCs: block-and-ash flows, co-ignimbrite surges, and blasts). Forecasts based on remote sensing data in combination with seismic, geodetic and gas-monitoring data have also helped inform decisions related to alert levels and evacuations. Relatively unusual aspects of the Sinabung eruption include the transition from dome to flow morphology (phase 2 to phase 3 transition) and the frequent occurrence during phase 3 of collapses from the lava flow-front and flow-margins—collapses that produced extensive pyroclastic density currents. By analogy to the well-known “Merapi type” collapses and pyroclastic deposits, we propose that lava flow-front and flow-margin collapses with associated PDCs be known as “Sinabung type.” Although detailed study of deposits has not been possible due to continuing hazards, our observations suggest that the transition from lava dome to lava flow and the occurrence of flow-front and flow-margin collapses reflect a particular combination of lava viscosity and steepness of slope. Our observations also show clear evidence of at least one slope-parallel high-velocity and dilute PDC (a “blast”) that emanated from a lava-margin collapse site 500 m downslope from the vent. This 1 February 2014 blast downed and singed a forest out to at least 3.9 km from the collapse site and killed 16 people. We also use a combination of field and remotely sensed data to map the distribution of Sinabung deposits. We estimate eruptive volumes and extrusion rates by combining sequential measurements of lava surface and pyroclastic flow areas with thickness estimates derived from simple geometric assumptions, oblique photographs and Digital Elevation Models (DEMs) derived from remotely sensed data. Our estimates of short-term effusion rates vary widely on a daily to weekly basis, from <1 to >20 m 3 s −1 . In a few cases, periods of increased extrusion precede lava flow-front collapses by a few days to a week, suggesting delays in transmittance of effusion pulses as lava moves from vent to flow front. We find that, as of 1 January 2016, the total area of deposits is 10 7 m 2 , and their approximate deposit volume is about 0.3 km 3 , equivalent to 0.2 km 3 Dense Rock Equivalent (DRE). We anticipate that our deposit maps will be valuable in the future as a framework for the study of the magmatic and textural evolution of eruptive products through time.