Applicability of methods for detecting volcanic ash distribution based on the relationship between volcanic ash depth and coherence loss using multi-temporal SAR imageries
Masaki MIZUNO, Hironori TSUTSUMI, Satoshi OKAZAKI, Takashi SHIBAYAMA,
Ikushi HIRATA, Takeshi HONDA, Jun SUGIMOTO and Shin-ichiro HAYASHI
If volcanic ash from explosive eruptions accumulates on a slope, debris and mud flows may occur owing to subsequent light rainfall before the accumulation. Mitigating damage from such debris and mud flows for residents and properties in downstream areas necessitates the immediate identification of affected areas and thickness distribution of volcanic ash to enable emergency structural countermeasures and evacuation of residents. Coherence analysis using satellite SAR (synthetic aperture radar) imageries is a useful method to detect areas of volcanic ash accumulation, because it can safely observe wide areas without the influence of volcanic smoke and even during bad weather and at night. Previous studies of coherence analysis have estimated the area of ash fall, but they have not quantitatively estimated the thickness of the fallen ash. In this study, we analyze the relationship between estimated volcanic ash fall areas obtained by coherence analysis using two or three ALOS-2 SAR imageries before and after the eruptions and the thickness of fallen ash measured through field surveys after Mt. Ontake (2014) and Mt. Aso (2016) eruptions. The results of this study were as follows. 1) The deeper depth of accumulated volcanic ash measured through field surveys corresponded to lower coherence value. 2) The coherence value for coherence analysis using two imageries from 0.7 to 0.6 was suitable for the detection of areas where volcanic ash accumulation exceeded 1 cm. 3) Approximately 0.2 of the coherence difference value for coherence analysis using three imageries was suitable to detect areas where volcanic ash accumulation exceeded 2 cm. 4) The coherence analyses using two and three imageries with short imagery capture intervals were found to be better than these with longer imagery capture intervals to estimate volcanic ash fall areas.
volcanic ash, synthetic aperture radar, coherence, ALOS-2