AN ARTIFICIAL INTELLIGENCE APPROACH TO PREDICTION OF EXTREME EVENTS: THE CASE OF STORMS IN WESTERN FRANCE

Storms represent an increased source of risk that affects human life, property, and the environment. Prediction of these events, however, is challenging due to their low frequency of occurrence. This paper proposed an artificial intelligence approach to address this challenge and predict storm characteristics and occurrence using a gated recurrent unit (GRU) neural network and a support vector machine (SVM). Historical weather and marine measurements collected from buoy data, as well as a database of storms containing all the extreme events that occurred in Brittany and Pays de la Loire regions, Western France, since 1996, were used. Firstly, GRU was used to predict the characteristics of storms (wind speed, pressure, humidity, temperature, and wave height). Then, SVM was introduced to identify storm-specific patterns and predict storm occurrence. The approach adopted leads to the prediction of storms and their characteristics, which could be used widely to reduce the awful consequences of these natural disasters by taking preventive measures.

Long-gap Filling Method for the Coastal Monitoring Data

Technique for the long-gap filling that occur frequently in ocean monitoring data is developed. The method estimates the unknown values of the long-gap by the summation of the estimated trend and selected residual components of the given missing intervals. The method was used to impute the data of the long-term missing interval of about 1 month, such as temperature and water temperature of the Ulleungdo ocean buoy data. The imputed data showed differences depending on the monitoring parameters, but it was found that the variation pattern was appropriately reproduced. Although this method causes bias and variance errors due to trend and residual components estimation, it was found that the bias error of statistical measure estimation due to long-term missing is greatly reduced. The mean, and the 90% confidence intervals of the gap-filling model’s RMS errors are 0.93 and 0.35~1.95, respectively.

Wind waves in the North Atlantic from ship navigational radar: SeaVision development and its validation with Spotter wave buoy and WaveWatch III

The global coverage of the observational network of the wind waves is still characterized by the significant gaps in in situ observations. At the same time wind waves play an important role into the Earth’ climate system specifically in the air-sea interaction processes and energy exchange between the ocean and the atmosphere. In this paper we present the SeaVision system for measuring wind waves’ parameters in the open ocean with navigational marine X-band radar and prime data collection from the three research cruises in the North Atlantic (2020 and 2021) and Arctic (2021). Simultaneously with SeaVision observations of the wind waves we were collecting data in the same locations and time with Spotter wave buoy and running WaveWatch III model over our domains. Measurements with SeaVision were quality controlled and validated by comparison with Spotter buoy data and WaveWatch III experiments. Observations of the wind waves with navigational Xband radar are in agreement among these three sources of data, with the best agreement for wave propagation directions. The dataset that supports this paper consists of significant wave height, wave period and wave energy frequency spectrum from both SeaVision and Spotter buoy. Currently the dataset is available through the temporary link (https://sail.ocean.ru/tilinina2021/) while supporting dataset (Tilinina et al., 2021) is in technical processing at PANGAEA repository. The dataset can be used for validation of satellite missions as well as model outputs. One of the major highlights in this study is potential of all ships navigating into the open ocean and equipped with X-band marine radar to participate into the development of another observational network for the wind waves in the open ocean once cheap and independently operating version of the SeaVision (or any other system) is available.

SET-UP DUE TO RANDOM WAVES: INFLUENCE OF THE DIRECTIONAL SPECTRUM

The correct estimation of set-up is very important to evaluate coastal hazard and to design coastal structures. In this paper, we derived a mathematical expression for wave set-up in the context of random waves. The solution to this expression assumes straight, parallel depth contours and constant average flow parameters in the longshore direction. We then investigated the effect of different types of sea state taking account of different frequency spectrum and spreading function assumed in the expression on estimates of wave set-up. We found the set-up was highly influenced by the frequency spectrum used. Finally, we applied this expression to estimate set-up values at locations in Italy and in the United States using buoy data provided by ISPRA (Istituto Superiore per la Protezione e la Ricerca Ambientale) and NDBC (National Data Buoy Centre).

SHIP OPERABILITY PREDICTED FROM LONG TERM DIRECTIONAL WAVE RECORDS

Ship operability assessments have traditionally been made using wind and wave data derived from wave atlases, however there are several drawbacks, including the fact that they are usually based on observation rather than measurement, and that spreading or directional effects are lost – such as the separation of sea and swell directions. An alternative approach is demonstrated here, instead of the data summarised in the wave atlas scatter diagram, long term hourly historical wave buoy data may be used. Detailed data sets, including directional wave spectra, are available for a number of specific locations. Direct use of many years’ hourly wave data involves significant computational effort, but results may be achieved within a reasonable time. The technique is demonstrated with the examples of four naval ships and two sites. Analysis considered two main themes, the differences in the ship performance calculated when (a) using wave buoy data rather than wave atlas data for the same sea area and (b) using the most complex available model of the ocean waves compared with the simplified wave descriptions in common use. For (a) the wave buoy data both looked rather different than the wave buoy data for the same nominal area, and produced rather different ship performance results. For (b) it was shown that there were also significant differences between the operability calculated for the four different ships at one of the sites. The implications for operability assessment in the ship procurement process are briefly discussed.

Estimating snow depth on Arctic sea ice based on reanalysis reconstruction and particle filter assimilation

The snow depth, an essential metric of snowpacks, can modulate sea ice changes and is a necessary input parameter to obtain altimeter-derived sea ice thickness values. In this study, we propose an innovative snow depth retrieval method with the improved NASA Eulerian Snow on Sea Ice Model (INESOSIM) and the particle filter (PF) approach, namely, INESOSIM-PF. Then, we generate daily snow depth estimates with INESOSIM-PF from 2012 to 2020 at a 50-km resolution. With the use of Operation IceBridge (OIB) data, it can be revealed that compared to the NESOSIM-estimated snow depth, the INESOSIM-PF-estimated snow depth is greatly improved, with a root mean square error (RMSE) decrease of 17.97 % (RMSE: 6.73 cm) and a correlation coefficient increase of 11.85 % (r: 0.71). The INESOSIM-PF-estimated snow depth is close to the satellite-derived snow depth, which is applied in data assimilation. With the use of Multidisciplinary Drifting Observatory for the Study of Arctic Climate (MOSAiC) snow buoy data, it can be verified that INESOSIM-PF performs well in the Central Arctic with an RMSE of 9.23 cm. INESOSIM-PF is robust and the snow depth determined with INESOSIM-PF is less influenced by input parameters with a snow depth uncertainty of 0.74 cm. The variations in the monthly and seasonal snow depth estimates retrieved from INESOSIM-PF agree well with those in the estimates retrieved from two other existing algorithms. Based on the presented snow depth estimates, we can retrieve the sea ice thickness and perform long-term snow depth and sea ice analysis. Snow depth estimates improve the understanding of Arctic environmental change and promote the future development of sea ice models.

Directional wave buoy data measured near Campbell Island, New Zealand

The New Zealand Defence Force (NZDF) has established a permanent wave observation station near Campbell Island, south of New Zealand (52 45.71 S, 169 02.54E). The site was chosen for logistical convenience and its unique location adjacent to the highly energetic Southern Ocean; allowing instrumentation typically deployed on the continental shelf to be used in this rarely observed southern environment. From February 2017, a Triaxys Directional Wave Buoy was moored in 147 m depth, some 17 km to the south of the island, with satellite telemetry of the 2D wave spectra at 3-hourly intervals. To date there have been three deployments on locations, yielding some 784 days of data. Validation of the measured significant wave height against co-located satellite altimeter observations suggests that the predominant wave directions are not attenuated by the island. The data provide a valuable record of the detailed wave spectral characteristics from one of the least-sampled parts of the Global Ocean.