Monitoring Inflation and Emplacement During the 2014-2015 Kilauea Lava Flow With an Unmanned Aerial Vehicle

Abstract:

<span" roman"="Roman"" new="New">Unmanned aerial vehicles (UAVs) provide a powerful new tool for collecting high resolution on-demand spatial data over volcanic eruptions and other active geomorphic processes. These data can be used to improve hazard forecasts and emergency response efforts, and also allow users to economically and safely observe and quantify lava flow inflation and emplacement on spatially and temporally useful scales. We used a small fixed-wing UAV with a modified point-and-shoot camera to repeatedly map the active front of the 2014-2015 Kīlauea lava flow over a one-month period in late 2014, at times with a two-hour repeat interval. An additional subsequent flight was added in July, 2015. We used the imagery from these flights to generate a time-series of 5-cm resolution RGB and near-infrared orthoimagery mosaics and associated digital surface models using structure from motion. Survey-grade positional control was provided by ground control points with differential GPS. Two topographic transects were repeatedly surveyed across the flow surface, contemporaneously with UAV flights, to independently confirm topographic changes observed in the UAV-derived surface models. Vertical errors were generally 10 cm. Inside our 50 hectare study site, the flow advanced at a rate of 0.47 hectares/day during the first three weeks of observations before abruptly stalling out <200 m from Pahoa Village road. Over 150,000 m³of lava were added to the study site during our period of observations, with maximum vertical inflation >4 m. New outbreak areas, both on the existing flow surface and along the flow margins, were readily mapped across the study area. We detected sinuous growing inflation ridges within the flow surface that correlated with subsequent outbreaks of new lava, suggesting that repeat UAV flights can provide a means of better predicting pahoehoe lava flow behavior over flat or uneven topography. Our results show that UAVs can generate accurate and digital surface models quickly and inexpensively over rapidly changing active pahoehoe lava flows.