The objective of this study is to map the pyroclastic deposits resulting from the Semeru eruption in 2021 and to analyse the process of pyroclastic deposit formation, including the tephra and lithification phases, using multi-sensor data. The distribution of pyroclastic deposits was mapped using machine learning algorithms, Sentinel 2 and Sentinel 1 imagery on Google Earth Engine. Data acquisition in the Semeru volcano area was conducted both prior to and following the 2021 eruption. Additionally, field investigations were conducted to analyse the distribution of pyroclastic deposits, including the characteristics of individual layers, the lithification process of pyroclastic layers, and their relationship with the characteristics of the eruption. The results demonstrate that pyroclastic deposits in the proximal facies are composed of lithic, lapilli tuff, lava and tuff. The mapping results also indicate the presence of volcanic ash deposits in the central facies or summit area of Semeru, exhibiting distinctive characteristics associated with volcanic ash. The 2021 eruption has also resulted in rapid changes to river morphometry, as evidenced by notable shifts in a significant decrease in pixel values for NDVI (−0.0781-0.0625), a significant increase in pixel frequency of EVI (1800–6000), and MNDWI (1700–3000), respectively. The change of pyroclastic fall into consolidated tephra occurred from December 2021 to May 2023, where in May 2023 volcanic ash had changed into tuff and lapilli had changed into lapilli tuff. Furthermore, data from Sentinel-1 indicated that pyroclastic flows could be identified using both vertical and horizontal polarization. It can therefore be concluded that the 2021 eruption of Semeru exhibited distinctive characteristics of pyroclastic deposits, where some of the characteristics of the pyroclastic distribution can be monitored rapidly with Google Earth Engine. Therefore, the use of Google Earth Engine with integration of radar and optical data can be recommended to mapping pyroclastic deposits rapidly.
Read full abstract