RIASSUNTO
ABSTRACT
The application of Bayesian inference in prediction of meteorological and oceanographic parameters to improve the estimation of sea spray icing in the Arctic region is purposed. Reanalysis data from NOrwegian ReAnalysis 10km (NORA10) during 33 years are applied to evaluate the performance of the model. Consequently, using the 32-year data, the parameters are predicted and compared for the last one-year on a daily basis. The predicted data are considered as input for the newly introduced icing model namely Marine-Icing Model for the Norwegian COast Guard (MINCOG) and the results are evaluated and discussed.
INTRODUCTION
Spray icing is considered a major environmental challenge and a critical risk element for offshore activities in the Arctic waters. Icing may impact offshore operations, reduce safety, operational tempo and productivity, cause malfunction of the operational and communication equipment, slippery handrails, ladders or decks, unusable fire and rescue equipment, and the blocking of air vents (Ryerson, 2011; Dehghani-Sanij, Dehghani, Naterer, and Muzychka, 2017). Icing on vessels may also lead to severe accidents and capsizing (Chatterton and Cook, 2008; Heinrich, 1950). Two main sources of icing are sea spray due to collision of ship and waves, as shown in Fig. 1, and atmospheric icing caused by fog, Arctic sea smoke, high-velocity wind, and rain/drizzle or snow, as shown in Fig. 2 (Rashid, Khawaja, and Edvardsen, 2016).
Wide varieties of techniques and technologies are available to enhance icing safety and protection such as chemicals, coatings, heat, and high-velocity fluids, air, water and steam (Ryerson, 2011; Rashid, Khawaja, and Edvardsen, 2016). However, forecasting the amount and frequency of ice formation aids selection of safety-enhancing strategy and ice protection technologies. Forecasting, also, can aid in tactical preparation prior to an icing event (Ryerson, 2011). However, forecasting icing events and rate is a complicated task due to the chaotic nature of icing and its correlation with large number of parameters. Extensive works have been conducted on historical icing data to predict icing events (Samuelsen, Edvardsen and Graversen, 2017). Accordingly, the data have been examined from different perspectives such as the influence of meteorological parameters on icing rate from statistical point of view (Mertins, 1968), and introducing sea spray algorithms based on the collision of ship and waves considering the environmental data as input parameters (Stallabrass, 1980; Samuelsen, Edvardsen and Graversen, 2017). Due to climate change, direct use of old data does not lead to accurate predictions of the future. One approach to deal with less reliable historical data might be ignoring the older values and only using more recent data. Elsner and Bossak (2001) proposed another alternative applying Bayesian approach, in which different qualities were considered for earlier and more recent data. They applied the approach for the prediction of U.S. land-falling hurricanes. Accordingly, a prior distribution was estimated using earlier data over 1851 to 1899. Then, the remaining data from 1900 to 2000 were used to revise the prior distribution. Bayesian inference is considered a strong approach for deal with the uncertainties in meteorological and oceanographic parameters. Wang et al. (2019) applied Bayesian approach in a model to estimate the uncertainties associated with weather and climate projections (e.g., 2-m temperature, surface radiation fluxes, or wind speed).