Investigation of two pack ice besetting events on the Umiak I and development of a probabilistic prediction model

A probabilistic model for fracture events of Petermann ice islands under the influence of atmospheric and oceanic conditions

“…To date, no previous research has adopted a Bayesian approach to predict the probability of ice island fracture events under the influence of the metocean conditions that control these events, likely due to the lack of reliable data. However, several studies have adopted a probabilistic approach using a Bayesian belief network and hydrometeorological variables for navigational risk assessment of ships (Zhang et al, 2013) or to estimate the conditional probability of ship besetting in sea-ice-covered waters (Turnbull et al, 2019;Fu et al, 2016;Montewka et al, 2015Montewka et al, , 2013. This study uses the CI2D3 database and adopts a similar methodology to that used in these besetting studies to present a probabilistic fracture model for ice islands as a function of the metocean conditions.…”

“…While a Bayesian network has never been employed for forecasting ice island fracture events, the probabilistic model presented in this paper was developed based on the methodology used in Turnbull et al (2019) and Fu et al (2016), where the Bayesian approach was used to predict vessel besetting events in pack ice. The study by Fu et al (2016) used a Bayesian network to investigate the inter-relationship between nine variables (i.e., ship speed, engine power, wind speed, air temperature, low visibility, sea temperature, ice concentration, ice thickness, and wave height), as well as their influence on the probability of a ship getting stuck in ice while navigating through the Northern Sea Route.…”

“…The study by Fu et al (2016) used a Bayesian network to investigate the inter-relationship between nine variables (i.e., ship speed, engine power, wind speed, air temperature, low visibility, sea temperature, ice concentration, ice thickness, and wave height), as well as their influence on the probability of a ship getting stuck in ice while navigating through the Northern Sea Route. Using a similar Bayesian approach, Turnbull et al (2019) studied two pack ice besetting events of the Umiak I and developed a probabilistic forecast model for future besetting events experienced by Umiak I under the influence of nine ice and metocean variables (i.e., ice concentration, ice thickness, floe size, minimum coast distance, wind-coast direction, wind speed, current-coast direction, current speed, and wind divergence). While this approach has not been used in a past iceberg fracture model, there are some deterioration models such as the one by Kubat et al (2007) that account for the influence of metocean variables (e.g., wind speed, current speed, water temperature, wave height, and wave period) on the calving events of the overhanging slabs resulting from the repeated action of waves.…”

Toward a method for downscaling sea ice pressure for navigation purposes

“…The McGill model has, over the years, been extensively tested (e.g. Lemieux et al, 2014;Bouchat and Tremblay, 2017;Williams and Tremblay, 2018). A few simple experiments were conducted in order to validate the implementation of the new stress boundary conditions.…”

“…Probabilistic models for ship performance in sea ice and likelihood of besetting events have also been developed (e.g. Montewka et al, 2015;Turnbull et al, 2019). Turnbull et al (2019) argue that the primary cause of the besetting events they studied were the relatively large ice floes encountered by the vessel.…”