ICMNS 2016: Comparative petrographic analysis of volcanic rocks in 5 different areas using R

​[ABS-366] Comparative petrographic analysis of volcanic rocks in 5 different areas using R

Jejen Ramdani (a), Dasapta Erwin Irawan (b), Achmad Darul (a)

(a) Institut Teknologi dan Sains Bandung; (b) Institut Teknologi Bandung

Abstract

Petrographic is a method to determine detailed description of the rock, related to composition and classification based on minerals composition in the volcanic rock. Five locations have been selected based on public domain datasets from Center for Geological Survey of Indonesia: Mt. Lamongan, Banyuresmi – Bogor area, Mt. Kromong , Jember area, and Mt. Wangi Purworejo.

We took notes of the mineral composition of the 75 samples in the dataset and analyze it using basic and multivariate statistics — principal component analysis (PCA), cluster analysis (CA), and multiple regressions (MR). The size of the dataframe was 75 x 20. R package were used to implement an open science work flow. Final code and datasets will be made available online.

Our preliminary results show the separate clustering of Tertiary and Quarternary volcanic samples suggesting significant differences between the two types of rock. Our trials using PCA, CA, and MR produce a consistent results. We identify a seemingly “pure” volcanic rock domain in Mt. Lamongan, Mt. Kromong, and Mt. Wangi, while the Banyuresmi and Jember areas show a slightly different cluster that due to sedimentary components influence, as shown by the occurrence of organic matters. As our analysis is still developing, currently we don’t find any major drawbacks in applying this method to classify rock samples based on their petrographic properties.

Keywords: Petrographic, statistics, volcanics

Topic: Earth and Space Sciences