Abstract
Accurate measurements of volcanic ash morphology are critical to improving both our understanding of fragmentation processes and our ability to predict particle behaviour. In this study, we present new ways to choose and apply shape parameters relevant to volcanic ash characterisation. First, we compare shape measurements from different imaging techniques, including cross-sectional (2-D) and projected area images, and discuss their respective applications. We then focus on specific information that can be obtained from shape analysis of 2-D images. Using cluster analysis as an unbiased method to identify key controls on particle morphology, we find that four shape parameters – solidity, convexity, axial ratio, and form factor – can effectively account for the morphological variance within most ash samples. Importantly, these parameters are scaled to values between 0 and 1, and therefore contribute evenly to discrimination diagrams. In particular, co-variation in convexity and solidity can be used to distinguish different juvenile ash components based on characteristic bubble properties. By reducing observations of natural samples to simplified ash geometries, we quantify morphological changes associated with variations in the relative size and shape of bubbles and particles. Using this relationship, we assess the potential application of size-dependent shape analysis for inferring the underlying bubble size distribution, and thus the pre-fragmentation conditions. Finally, we show that particle shape analysis that includes the full range of available grain sizes can contribute not only measurements of particle size and shape, but also information on size-dependent densities.
Original language | English |
---|---|
Pages (from-to) | 14–30 |
Number of pages | 17 |
Journal | GeoResJ |
Volume | 8 |
Early online date | 6 Oct 2015 |
DOIs | |
Publication status | Published - Dec 2015 |
Keywords
- Shape parameters
- Image analysis
- Volcanic ash
- Fragmentation
Fingerprint
Dive into the research topics of 'Optimising shape analysis to quantify volcanic ash morphology'. Together they form a unique fingerprint.Datasets
-
Liu et al 2015 ash shape data
Liu, E. (Creator), Cashman, K. (Creator) & Rust, A. C. (Creator), University of Bristol, 21 Oct 2015
DOI: 10.5523/bris.wgxxhae8ptdc1he0g2cqq0q1i, http://data.bris.ac.uk/data/dataset/wgxxhae8ptdc1he0g2cqq0q1i
Dataset