Skip to content

Development of a simplified method for estimating grain size distribution of pyroclastic fall deposits using image analysis

Yasuaki Kaneda 1, Kuniaki Nishiki1, Yoshimi Hiroi1

  • Affiliations: ^ 1^Division of Research for Earthquake and Tsunami, Nuclear Regulation Authority, 1-9-9 Roppongi, Minato-ku, Tokyo 106-8450, Japan. 

  • Presentation type: Poster

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

  • Poster Board Number: 108

  • Programme No: 3.15.13

  • Theme 3 > Session 15

Abstract

The grain size distribution (GSD) of pyroclastic deposits provides insight into the processes of fragmentation, transport, and deposition of explosive eruptions. Sieving and laser diffraction methods are quick and straightforward methods for determining GSDs. Unfortunately, these methods cannot handle the weathered fragile deposits and weakly-welded tuff. Image analysis offers a potential solution to these limitations. However, the technical challenges and accuracy of the image analysis remain unclear, as only a few cases have been studied using this method. To evaluate the applicability of the image analysis to pyroclastic fall deposits, we compared the GSDs obtained by the image analysis with those obtained by the combined method of sieving and laser diffraction applied to the same pyroclastic fall deposit ejected from Tarumae volcano, Japan. Compared to the GSDs obtained using the combined method, those from the image analysis were either similar or shifted toward the coarser side on the histograms. The difference in median grain size between the cases with the image analysis and combined method shows a clear trend: the coarser sample, the larger the discrepancy. To correct these discrepancies, a stereological analysis was performed using "ImageJ" and "CSD corrections" to estimate the 3D GSDs from 2D images. The discrepancies between the estimated 3D GSDs and those obtained by the combined method were successfully reduced. In conclusion, this study not only identified the systematic discrepancies of GSDs between the image analysis method and combined method, but also presented a simple method to successfully correct these discrepancies using existing applications.