In the present study, the iceberg drafts and iceberg-seabed interaction process were simulated using the random forest regression (RFR) algorithm for the first time. Initially, utilizing the parameters governing the iceberg drafts and the iceberg-seabed interaction process in the sandy seabed, a set of RFR models were developed. To train and test the RFR models, a comprehensive dataset was subsequently constructed by using the field and experimental values reported in the published literature. By performing a sensitivity analysis, the premium RFR model and the most significant input parameters were introduced.
Approximately 22% of the Earth's undiscovered hydrocarbons are stored in the Arctic region. Therefore, the Arctic area is one of the best resources to grow oil and gas loading equipment. Moreover, the Arctic offshore regions with rich wind culture have a high potential for the development of offshore wind farms. This combined with climate change and global warming means an increased number of ice management operations to protect the subsea pipelines, power cables, and offshore and subsea structures. The schematic layout of the iceberg free-floating and iceberg scouring in cold waters is shown in Figure 1.
As seen, the iceberg is in a free-floating situation if the ocean depth is greater than the iceberg draft; otherwise, the seafloor is scoured, and the seabed soil shear resistance causes the soil displacement to extend deeper than the iceberg tip threatening the buried subsea assets.
The efficient iceberg management designs and the guaranteed operational integrity of the sea bottom-funded infrastructure against the berg attacks in the ice-prone areas demand the appropriate iceberg draft appraisal, which may lead to a potential decrease in operating expenses and downtime. Earlier investigations have tended to focus on modeling the iceberg draft by using the iceberg length or iceberg mass (Robe and Farmer 1976; Hotzel and Miller 1983; Barker et al. 2004). King et al. (2016) performed a field investigation to calculate the iceberg rolling rate. The iceberg drafts were estimated utilizing a calving analysis, with a calculated standard deviation of draft variations from 19% to 34%. The iceberg drafts corresponded with the mass of the icebergs. In another investigation, Turnbull et al. (2018) proposed a model for the drift estimation of moving icebergs on the Grand Banks of Newfoundland. This model approximated the draft of icebergs roughly 1.3 times more than the real values.
