9:00 AM - 10:30 AM
[MGI29-P05] Statistical characteristics of geochemical data and propriate lithological classification of Japan arc
Keywords:geochemical database, Seamless geological map, lithological division, logarithm distribution, geochemical characteristics of Japan arc
Since the completion of geological maps of the entire Japan arc, the distribution of "what kind of geology" and "in what proportions" in the Japan arc has been studied in various ways according to trends in geological research in each era. The geological and geochemical processes of the Japan arc have also been studied, reflecting trends in chemical analysis and the construction and revision of growth models in each era. A large amount of geological and geochemical data has been accumulated that are the basis of these studies. And technology for using these data has also progressed. Analysis of the geochemical development process of the Japan arc through these studies is still an ongoing research topic.
Haraguchi et al. (Geological Society of Japan Autumn Meeting 2019, JpGU-AGU Joint Meeting 2020) focused on the "East-West difference" in geochemical characteristics of Japan arc, and considered origin of this difference in relation to the geological development, used the seamless geological map issued by the Geological Survey of Japan and the database “DODAI” mainly containing geochemical data published in Japan. Furthermore, Haraguchi et al. (JpGU, Geological Society of Japan Autumn Meeting 2022) focused on the "area ratio and data distribution density" of geological components, and discussed the differences in the characteristics of data distribution and trend of each chemical researches. Through these reports, we introduced the current situation and problems in considering the geochemical characteristics of the Japan arc based on the classification of the lithological division and geochemical characteristics of the Japan arc. In this report, we focus on the "statistical concept" of the "chemical data distribution" for each lithofacies based on former reports.
Geochemical data are divided into the major classifications of “volcanic rocks”, “plutonic rocks”, “sedimentary rocks” and “metamorphic rocks”, and subdivided into mafic, intermediate, felsic, etc. based on petrological descriptions. The chemical data of these lithological groups shows a distribution in which many elements start at 0 wt% and ppm, have a distribution peak on the low-concentration side, and the number of data decreases monotonously toward the high-concentration side. When considering data distribution using statistical values, normal distribution is the standard for statistical values such as skewness and kurtosis. In such a monotonically decreasing distribution, the numerical values often deviate greatly from the standard range. On the other hand, when looking at the data distribution with "logarithm", many elements with monotonically decreasing distribution show "normal distribution". In addition, it is possible to collectively treat extremely high concentration values due to "concentration of elements by differentiation" by taking logarithms. Based on these characteristics and problems of logarithmic distribution, when we look at the classified data, we find that many lithology and elements show normal distributions, and it is interpreted that they also have the meaning of chemical "groups". Among them, some lithologies and elements show bimodal characteristics. Especially in volcanic rocks, the grouping of high-content of incompatible elements such as alkali elements, which reflects the "alkaline series", is remarkable. We consider statistical positioning of such "elements that cause distortion of data distribution", "validity of lithofacies classification" and "positioning of lithofacies in the tectonic development of the Japan arc".
Acknowledgments
We would like to thank Mr. Yoshiharu Nishioka (AIST) for his cooperation in calculating the area ratio for each lithology based on the Seamless Geological Map.
Haraguchi et al. (Geological Society of Japan Autumn Meeting 2019, JpGU-AGU Joint Meeting 2020) focused on the "East-West difference" in geochemical characteristics of Japan arc, and considered origin of this difference in relation to the geological development, used the seamless geological map issued by the Geological Survey of Japan and the database “DODAI” mainly containing geochemical data published in Japan. Furthermore, Haraguchi et al. (JpGU, Geological Society of Japan Autumn Meeting 2022) focused on the "area ratio and data distribution density" of geological components, and discussed the differences in the characteristics of data distribution and trend of each chemical researches. Through these reports, we introduced the current situation and problems in considering the geochemical characteristics of the Japan arc based on the classification of the lithological division and geochemical characteristics of the Japan arc. In this report, we focus on the "statistical concept" of the "chemical data distribution" for each lithofacies based on former reports.
Geochemical data are divided into the major classifications of “volcanic rocks”, “plutonic rocks”, “sedimentary rocks” and “metamorphic rocks”, and subdivided into mafic, intermediate, felsic, etc. based on petrological descriptions. The chemical data of these lithological groups shows a distribution in which many elements start at 0 wt% and ppm, have a distribution peak on the low-concentration side, and the number of data decreases monotonously toward the high-concentration side. When considering data distribution using statistical values, normal distribution is the standard for statistical values such as skewness and kurtosis. In such a monotonically decreasing distribution, the numerical values often deviate greatly from the standard range. On the other hand, when looking at the data distribution with "logarithm", many elements with monotonically decreasing distribution show "normal distribution". In addition, it is possible to collectively treat extremely high concentration values due to "concentration of elements by differentiation" by taking logarithms. Based on these characteristics and problems of logarithmic distribution, when we look at the classified data, we find that many lithology and elements show normal distributions, and it is interpreted that they also have the meaning of chemical "groups". Among them, some lithologies and elements show bimodal characteristics. Especially in volcanic rocks, the grouping of high-content of incompatible elements such as alkali elements, which reflects the "alkaline series", is remarkable. We consider statistical positioning of such "elements that cause distortion of data distribution", "validity of lithofacies classification" and "positioning of lithofacies in the tectonic development of the Japan arc".
Acknowledgments
We would like to thank Mr. Yoshiharu Nishioka (AIST) for his cooperation in calculating the area ratio for each lithology based on the Seamless Geological Map.