USING SEM AND FRACTAL ANALYSIS TO FINGERPRINT VOLCANIC ASH DEPOSITS

The shapes of volcanic ash particles reflect the eruption that produced them, and analysis of ash morphology can be useful for discriminating between different volcanic deposits, understanding processes associated with different styles of eruption, and assessing volcanic hazards. However, quantitative analysis of volcanic particle shapes has proven to be a difficult task, with no unifying methodology due to their complexity and variability. This study uses scanning electron microscopy (SEM), fractal analysis, and multivariate statistics to produce a unique quantitative fingerprint for a given volcanic ash deposit. Focusing on basaltic volcanism, our samples represent lava-fountain activity (Kilauea Iki, Hawaii, 1959), phreatomagmatic activity (Kilauea Keanakakoi, Hawaii, 1790), and explosive activity driven by magmatic volatiles (Kilauea Kulanaokuaiki, Hawaii, A.D. 400-1000, and Masaya San Judas Fm., Nicaragua, >2 ka). For each sample, SEM images were captured from five sieve fractions (>32µ, >63µ, >125µ, >250µ, and >500µ), and then converted to binary by thresholding. Fractal dimension of each particle’s 2-D outline was calculated with Fovea Pro software (Minkowski method based on Euclidian distance map or EDM). Fractal data was summarized using principal components analysis, allowing depiction of multiple particles as a single point (component score) on a binary plot. For a given sample, component scores for the different size fractions plot in unique positions. When connected with a line, the component scores form a fingerprint-like pattern that represents this volcanic sample and might be characteristic of this style of volcanic activity. One of our goals is to determine if these patterns are useful for identifying specific volcanic deposits. In addition, analysis of different morphological features at different scales, suggested by these patterns, may yield new insights into magma fragmentation processes during explosive basaltic eruptions.