Applying random forest to oil and gas exploration in Central Sumatra basin Indonesia based on surface and subsurface data

• Published in: Remote sensing applications 2023-11, Vol.32, p.101039, Article 101039
• Main Authors: Susantoro, Tri Muji, Wikantika, Ketut, Suliantara, Suliantara, Setiawan, Herru Lastiadi, Harto, Agung Budi, Sakti, Anjar Dimara
• Format: Article
• Language: English
• Subjects: basins; drainage; environment; geophysics; gravity; Indonesia; Landsat; oils
• ISSN: 2352-9385; 2352-9385
• DOI: 10.1016/j.rsase.2023.101039

Oil and gas exploration in Indonesia currently requires novel methods during the initial screening of exploration areas that are effective, inexpensive, and utilize open data. This is expected to increase exploration interest, which has continued to decrease, and many oil and gas exploration areas will be returned to the government (up to 100 working areas). The study aimed to examine the utilization of random forest classification for oil and gas exploration in the Central Sumatra Basin based on subsurface and surface data. Subsurface data included gravity data, basement structure maps, seismic interpretation maps (6 variables), and surface data, including Landsat 8 OLI data, SRTM DEM, surface geological maps, drainage pattern maps, and topographical maps (32 variables). The method used in this research is a random forest with sample training in oil- and gas-proven and potential areas (6 classes) and non-oil- and gas-potential areas (5 classes). Based on the Ntree 600 parameter, Mtry k, and node size 2, unexplored oil and gas potential areas were identified, consisting of (1) a very high potential area in the south of the Bengkalis graben; (2) three locations of high potential areas located on the banks of the Kiri trough and close to the Aman and Balam grabens; (3) two locations of medium potential category II located east and northwest of the Bengkalis graben; and (4) a medium potential area category I located east of the Bengkalis graben.