09:30 〜 09:45
[SCG48-03] 海底音響反射強度データと海底画像の統合的解析に基づく南鳥島周辺マンガンノジュールの分布様態の特徴
キーワード：レアメタル、マンガンノジュール、後方散乱強度調査、ピークフィット解析、南鳥島、海底鉱物資源
Securing the stable supply of critical metals such as Co and Ni is essential to realize a sustainable society. Ferromanganese nodules on deep sea floor concentrate these metals and are attracting worldwide attention as a promising source of the strategic elements. In 2016, a dense field of ferromanganese nodules were discovered to be widely distributed in the Japanese Exclusive Economic Zone (EEZ) around Minamitorishima Island, and various research activities are underway.
In our previous study [1], a new method was established to estimate the distribution of ferromanganese nodules very efficiently using the acoustic backscatter intensity of the seafloor. In the study, acoustic backscatter intensity data obtained by a multi-narrow beam echo sounder during seven cruises were processed and integrated into a unified data set. Analysis of the histograms of the integrated data yielded several threshold values corresponding to the change of distributional pattern of ferromanganese nodules in the study area. Sasaki et al. (2023 JpGU) [2] further discussed how the acoustic data reflects the characteristics of the seafloor surface by separating multiple peaks that appeared in the histogram of the combined data.
To consider the correspondence between the backscatter intensities and seafloor features more in detail, we performed a peak-fitting analysis using the Expectation-Maximization (EM) algorithm [3] on histograms of acoustic backscatter intensities of each cruise before coupling, as well as the previously examined coupled intensity dataset, around Minamitorishima Island. The backscatter intensity histograms were decomposed into multiple peaks, suggesting detection of different seafloor features in each dataset. Moreover, to clarify how these peaks correspond to the actual seafloor facies, the seafloor images observed by the SHINKAI 6500 was analyzed using a machine-learning-based object detection model. In the presentation, we will discuss the relationship between each peak and the distribution patterns of ferromanganese nodules on the seafloor around Minamitorishima Island.
[1] Machida et al. Mar. Georesour. & Geotec. 39(3), 267-279 (2021).
[2] Sasaki et al. JpGU2023, SCG52-P18 (2023).
[3] Matsumura et al. Sci. Technol. Adv. Mat. 20, 733–745 (2019).
In our previous study [1], a new method was established to estimate the distribution of ferromanganese nodules very efficiently using the acoustic backscatter intensity of the seafloor. In the study, acoustic backscatter intensity data obtained by a multi-narrow beam echo sounder during seven cruises were processed and integrated into a unified data set. Analysis of the histograms of the integrated data yielded several threshold values corresponding to the change of distributional pattern of ferromanganese nodules in the study area. Sasaki et al. (2023 JpGU) [2] further discussed how the acoustic data reflects the characteristics of the seafloor surface by separating multiple peaks that appeared in the histogram of the combined data.
To consider the correspondence between the backscatter intensities and seafloor features more in detail, we performed a peak-fitting analysis using the Expectation-Maximization (EM) algorithm [3] on histograms of acoustic backscatter intensities of each cruise before coupling, as well as the previously examined coupled intensity dataset, around Minamitorishima Island. The backscatter intensity histograms were decomposed into multiple peaks, suggesting detection of different seafloor features in each dataset. Moreover, to clarify how these peaks correspond to the actual seafloor facies, the seafloor images observed by the SHINKAI 6500 was analyzed using a machine-learning-based object detection model. In the presentation, we will discuss the relationship between each peak and the distribution patterns of ferromanganese nodules on the seafloor around Minamitorishima Island.
[1] Machida et al. Mar. Georesour. & Geotec. 39(3), 267-279 (2021).
[2] Sasaki et al. JpGU2023, SCG52-P18 (2023).
[3] Matsumura et al. Sci. Technol. Adv. Mat. 20, 733–745 (2019).