The Marine Icing model for the Norwegian Coast Guard (NoCG; MINCOG) formulated in 2016 by Eirik Samuelsen of the Norwegian Meteorology Institute includes more metocean data, plus selected ship and course details based on the accretion rates observed onboard three Norwegian Coast Guard vessels of the class NoCG “Nordkapp” navigating within the Barents Sea during the 1983–1998 timeframe.
Investigations into fatal icings of the three vessels above revealed considerable discrepancies between real life icing rates and hindcasts made using the parameters of the Overland and MINCOG methods, and the conclusion is that—no matter how many parameters they include—empirical models cannot yield predictions valid for any vessel under all circumstances.
They build on a handful of ship types and few observations of spray icing. Global warming is also altering the basic condition compared to when the observations were made. New, expensive, and global measurements on a wide range of ships are required to produce an effective model for any type of vessel and sea weather conditions.
Laboratory experiments
To obviate the lack of experimental data, Sujay Deshpande, then a Ph.D. student, simulated in 2020 the icing of sea spray in a laboratory of the Arctic University of Norway, where seawater with different temperatures from a cooling chamber reproduced the conditions of the sea, and was sprayed onto plates of different materials inside a freezing chamber reproducing the atmosphere. The sticking point in all prediction models is determining spray flux data consistently: Spray data gained on a real ship can be used only for that ship.
Deshpande identified seven critical parameters—air and water temperature, salinity, and wind speed are readily measurable, while spray flux, duration, and frequency are not. He used computational fluid dynamics to simulate the latter spray parameters to feed an artificial intelligence (AI) predictor. But reconstructing the real-life behavior of droplets ranging from 10 µm to millimeters under different sea, gravity, and wind conditions and by diverging temperatures and different liquid water content, remains of paramount importance for formulating prediction models that work.
In the past, mechanical and electronic devices based on some form of physical collection or registration of spray samples from a fraction of the cloud were installed on decks for field work to monitor sea spray live. By all their merits, a quite limited picture of the processes in the freezing cloud resulted and the data sets thus collected fell short in grasping the multifaceted phenomenon, calling for the implementation of a more advanced technology.
A game-changer: Near-range LiDAR
There is a long history of using elastic backscatter LiDARs with expensive high-energy pulsed lasers for atmospheric characterization at great distances. Recent advancements in high pulse repetition frequency lasers and eye-safe photodetectors widen the scope of atmospheric LiDAR applications to include shorter ranges. Further, Eduard Gregorio and colleagues at the University of Leida in Spain demonstrated that a near-range compact system based on light-emitting diode (LED) light sources and suitable avalanche photodetectors enable better observations and more efficient characterizing of pesticide dispersion at affordable costs in comparison to direct field particle sampling.
Based on this idea, Professor Sushmit Dhar at the Arctic University of Norway reasoned that a similar compact system could provide more comprehensive information on how water spray clouds behave and freeze on a ship’s deck and superstructures. With the goal of unprecedented quality of sea spray analysis, Dhar designed the MarSpray LiDAR (MSL) prototype for onboard icing research (see Figs. 3 and 4).