Remote Sensing of Strombolian Activity: Frequency-Based Detection and Monitoring Techniques

Strombolian activity is one of the most common types of volcanic activity. When this activity occurs at remote volcanoes it often goes undetected and cannot be monitored easily or safely by direct methods. Satellite remote sensing can be useful in the routine monitoring of this activity. Numerous remote volcanoes in the Aleutian Islands and Kamchatka Peninsula exhibit Strombolian explosive activity. Often this type of activity is a precursor to more vigorous activity, including lava flows and ash producing explosions, which can affect communities and transportation. Time-dependent thermal signals were modeled in the lab using honey as a lava analog in scaled explosions. The thermal signals from the spatter fields were recorded with a FLIR thermal camera and were seen to last tens of seconds. Explosions at Stromboli volcano in Italy were recorded using the same thermal camera. These spatter fields, the likely source of thermal signals, were seen to last tens of seconds to minutes. The transient nature of these thermal features, both scaled and actual, indicates that the probability of capturing strombolian style explosions in a satellite overpass is statistically unlikely. However, these thermal signals are seen in imagery and the number of explosions recorded can be used, along with number of passes per day, viewing geometry, and other constraints, to determine a relative frequency of explosive activity. Factors affecting the visibility of thermal anomalies resulting from small explosions include satellite geometry, crater geometry, weather, and time of image capture. Geometric and timing factors eliminate a significant number of satellite passes; only 10-15% of original passes are useable after considering these issues. The remaining passes are analyzed for weather conditions, which in the N. Pacific is a very limiting factor. Weather is determined from cloud cover in a 40x40 km region around the volcano based on a grading system with five categorizations from clear to cloudy. This is used to calculate a probability of a clear view at the volcano and the likelihood of detecting an explosion. This statistic is used to detect changes in relative frequency of explosive activity and possible periods of heightened activity, leading to heightened awareness of possible hazards. Satellite imagery spanning 2005-2010 from Mt. Chuginadak and Shishaldin in Alaska, and Karymsky on the Kamchatkan Peninsula have been analyzed as well as observatory releases and color code assignments to determine base-line activity at those volcanoes.