Skip to content

Thermal remote sensing of lava lakes: a physically-based algorithm

Sofie Rolain1,2, Dries Peumans3,4, Benoît Smets1,2

  • Affiliations: 1 Department of Geography, Vrije Universiteit Brussel, Brussels, Belgium; ^2 ^Natural Hazards and Cartography Service, Department of Earth Sciences, Royal Museum for Central Africa, Tervuren, Belgium; 3 ELEC Department, Vrije Universiteit Brussel, Brussels, Belgium; 4 imec, Leuven, Belgium;

  • Presentation type: Poster

  • Presentation time: Friday 16:30 - 18:00, Room Poster Hall

  • Poster Board Number: 227

  • Programme No: 3.17.14

  • Theme 3 > Session 17

Abstract

Satellite-based thermal remote sensing is an effective tool for volcano monitoring, both for early warning and eruption monitoring. It involves detecting thermal anomalies, or hotspots, and calculating the radiative energy emitted by volcanic activity. Various volcanic hotspot detection algorithms exist in the literature. However, each algorithm has its own strengths and weaknesses, which are influenced by the tradeoffs made during its development, the type of sensor used for data acquisition, and the geometry of image acquisition. Depending on the algorithm used, different results are obtained from the same data and, hence, resulting interpretations might differ in terms of, e.g., energy emitted, effusion rates, and eruption duration. In the present work, we aim at creating a new hotspot detection algorithm that is solely based on physical laws, avoiding any empirical assumptions. We apply it to MODIS --and later on VIIRS-- imagery. This new algorithm is applied to help us unravel the dynamics of thermal emissions from persistent lava lakes, i.e., bassins of lava maintained molten through thermal convection and outgassing. In the present work, we show the first results of our new algorithm on lava lake activity at Kīlauea, Hawaii, and compare them to MODVOLC and ground-based thermal camera imagery openly provided by USGS. Our initial results demonstrate a stronger correspondence with thermal camera data compared to MODVOLC. While we generally detect hotspots on the same days as MODVOLC, we also identify days of activity confirmed by the ground-based camera that MODVOLC missed, showing the greater sensitivity of our approach.