Neural networks for wave forecasting

The physical process of generation of waves by wind is extremely complex, uncertain and not yet fully understood. Despite a variety of deterministic models presented to predict the heights and periods of waves from the characteristics of the generating wind, a large scope still exists to improve on the existing models or to provide alternatives to them. This paper explores the possibility of employing the relatively recent technique of neural networks for this purpose. A simple 3-layered feed forward type of network is developed to obtain the output of significant wave heights and average wave periods from the input of generating wind speeds. The network is trained with different algorithms and using three sets of data. The results show that an appropriately trained network could provide satisfactory results in open wider areas, in deep water and also when the sampling and prediction interval is large, such as a week. A proper choice of training patterns is found to be crucial in achieving adequate training.     

Real time wave forecasting using neural networks

Forecasting of ocean wave heights, with warning time of a few hours or days, is necessary in planning many operation-related activities in the ocean. Such information is currently derived by numerically solving the differential equation representing wave energy balance. The solution procedure involved is extremely complex and calls for very large amounts of meteorological and oceanographic data. This paper presents a complementary and simple method to make a point forecast of waves in real time sense based on the current observation of waves at a site. It incorporates the technique of neural networks. The network involved is first trained by different algorithms and then used to forecast waves with lead times varying from 3 to 24 h. The results of different training algorithms are compared with each other. The neural output is further compared with the statistical AR models.     

Ocean wave forecasting using recurrent neural networks

The tremendous increase in offshore operational activities demands improved wave forecasting techniques. With the knowledge of accurate wave conditions, it is possible to carry out the marine activities such as offshore drilling, naval operations, merchant vessel routing, nearshore construction, etc. more efficiently and safely. This paper describes an artificial neural network, namely recurrent neural network with rprop update algorithm and is applied for wave forecasting. Measured ocean waves off Marmugao, west coast of India are used for this study. Here, the recurrent neural network of 3, 6 and 12 hourly wave forecasting yields the correlation coefficients of 0.95, 0.90 and 0.87, respectively. This shows that the wave forecasting using recurrent neural network yields better results than the previous neural network application.     

Real-time wave forecasting using genetic programming

The forecasting of ocean waves on real-time or online basis is necessary while carrying out any operational activity in the ocean. In order to obtain forecasts that are station-specific a time-series-based approach like stochastic modeling or artificial neural network was attempted by some investigators in the past. This paper presents an application of a relatively new soft computing tool called genetic programming for this purpose. Genetic programming is an extension of genetic algorithm and it is suited to explore dependency between input and output data sets. The wave rider buoy measurements available at two locations in the Gulf of Mexico are analyzed. The forecasts of significant wave heights are made over lead times of 3, 6, 12 and 24 h. The sample size belonged to a period of 15 years and it included an extensive testing period of 5 years. The forecasts made by the approach of genetic programming indicated that it can be regarded as a promising tool for future applications to ocean predictions.     

Neural network modelling of wave overtopping at coastal structures

A method has been developed to estimate wave overtopping discharges for a wide range of coastal structures. The prediction method is based on Neural Network modelling. For this purpose use is made of a data set obtained from a large number of physical model tests (collected within the framework of the European project CLASH, see e.g. [Steendam, G.J., Van der Meer, J.W., Verhaeghe, H., Besley, P., Franco, L. and Van Gent, M.R.A. (2004). The international database on wave overtopping. World Scientific, Proc. 29th ICCE, vol. 4, pp. 4301–4313, Lisbon, Portugal.]). Moreover, a method was developed to obtain confidence intervals for the overtopping predictions of the neural network.     

Application of artificial neural networks in tide-forecasting

An accurate tidal forecast is an important task in determining constructions and human activities in ocean environments. Conventional tidal forecasting has been based on harmonic analysis using the least squares method to determine harmonic parameters. However, a large number of parameters are required for the prediction of a long-term tidal level with harmonic analysis. Unlike conventional harmonic analysis, this paper presents an artificial neural network (ANN) model for forecasting the tidal-level using the short term measuring data. The ANN model can easily decide the unknown parameters by learning the input–output interrelation of the short-term tidal records. Three field data with three types of tides will be used to test the performance of the proposed ANN model. The numerical results indicate that the hourly tidal levels over a long duration can be predicted using a short-term hourly tidal record.     

Application of artificial neural networks in typhoon surge forecasting

A typhoon-surge forecasting model was developed with a back-propagation neural network (BPN) in the present paper. The typhoon's characteristics, local meteorological conditions and typhoon surges at a considered tidal station at time t−1 and t were used as input data of the model to forecast typhoon surges at the following time. For the selection of a better forecasting model, four models (Models A–D) were tested and compared under the different composition of the above-mentioned input factors. A general evaluation index that is a composition of four performance indexes was proposed to evaluate the model's overall performance. The result of typhoon-surge forecasting was classified into five grades: A (excellent), B (good), C (fair), D (poor) and E (bad), according to the value of the general evaluation index. Sixteen typhoon events and their corresponding typhoon surges and local meteorological conditions at Ken–fang Tidal Station in the coast of north-eastern Taiwan between 1993 and 2000 were collected, 12 of them were used in model's calibration while the other four were used in model's verification. The analysis of typhoon-surge forecasting results at Ken–fang tidal station show that the Model D composing 18 input factors has better performance, and that it is a suitable BPN-based model in typhoon-surge forecasting. The Model D was also applied to typhoon-surge forecasting at Cheng-kung Tidal Station in south-eastern coast of Taiwan and at Tung-shih Tidal Station in the coast of south-western Taiwan. Results show that the application of Model D in typhoon-surge forecasting at Cheng-kung Tidal Station has better performance than that at Tung-shih Tidal Station.     

Wave height forecasting in Dayyer, the Persian Gulf

Forecasting of wave parameters is necessary for many marine and coastal operations. Different forecasting methodologies have been developed using the wind and wave characteristics. In this paper, artificial neural network (ANN) as a robust data learning method is used to forecast the wave height for the next 3, 6, 12 and 24 h in the Persian Gulf. To determine the effective parameters, different models with various combinations of input parameters were considered. Parameters such as wind speed, direction and wave height of the previous 3 h, were found to be the best inputs. Furthermore, using the difference between wave and wind directions showed better performance. The results also indicated that if only the wind parameters are used as model inputs the accuracy of the forecasting increases as the time horizon increases up to 6 h. This can be due to the lower influence of previous wave heights on larger lead time forecasting and the existing lag between the wind and wave growth. It was also found that in short lead times, the forecasted wave heights primarily depend on the previous wave heights, while in larger lead times there is a greater dependence on previous wind speeds.     

Prediction model for occurrence of impact wave force

This study investigates the applicability of neural networks to predict whether impact wave force will act on the upright section of a composite breakwater. We employ a three-layered neural network whose units of input layer are h/L, H/h, d/h and B_M/h (h: the total water depth; L: the wavelength; H: the wave height; d: the water depth above the mound; B_M: the horizontal distance from the shoulder of mound to the caisson). Teach signals are 0.99 and 0.01 according to the cases of occurrence and absence of impact wave force, respectively. The neural network whose parameters are determined through self-learning can accurately predict whether impact wave force occurs.     

Neural network prediction of a storm surge

The occurrence of storm surge does not only destroy the resident's lives, but also cause the severe flooding in coastal areas. Therefore, accurate prediction of storm surge is an important task during the coming typhoon. Conventional numerical methods and experienced methods for storm surge prediction have been developed in the past, but it is still a complex ocean engineering problem which many factors, including the central pressure of typhoon, the speed of the typhoon, the heavy rainfall, coastal topography and local features influence the variation of storm surge. In fact, this problem is still a complex nonlinear relationship that can not solved efficiently by these two methods. Therefore, this paper presents an application of the neural network for forecasting the storm surge. The original data of Jiangjyun station in Taiwan will be used to test the performance of the present model. The results indicate that the neural network can be efficiently forecasted storm surge using the four input factors, including the wind velocity, wind direction, pressure and harmonic analysis tidal level.     

Derivation of coastal wind and wave parameters from offshore measurements of TOPEX satellite using ANN

The paper discusses an artificial neural network (ANN) approach to project information on wind speed and waves collected by the TOPEX satellite at deeper locations to a specified coastal site. The observations of significant wave heights, average wave period and wind speed at a number of locations over a satellite track parallel to a coastline are used to estimate corresponding values of these three parameters at the coastal site of interest. A combined network involving an input and output of all the three parameters, viz., wave height, period and wind speed instead of separate networks for each one of these variables was found to be necessary in order to train the network with sufficient flexibility. It was also found that network training based on statistical homogeneity of data sets is essential to obtain accurate results. The problem of modeling wind speeds that are always associated with very high variations in their magnitudes was tackled in this study by imparting training in an innovated manner.     

Wave prediction using wave rider position measurements and NARX network in wave energy conversion

Several control methods of wave energy converters (WECs) need prediction in the future of wave surface elevation. Prediction of wave surface elevation can be performed using measurements of surface elevation at a location ahead of the controlled WEC in the upcoming wave. Artificial neural network (ANN) is a robust data-learning tool, and is proposed in this study to predict the surface elevation at the WEC location using measurements of wave elevation at ahead located sensor (a wave rider buoy). The nonlinear autoregressive with exogenous input network (NARX NN) is utilized in this study as the prediction method. Simulations show promising results for predicting the wave surface elevation. Challenges of using real measurements data are also discussed in this paper.     

缺失风浪数据补足的神经网络模型

基于波浪数据的完备性对于海岸海洋工程设计而言非常关键,详细阐述了风浪观测数据补足神经网络模型的建立方法,构建了两个网络模型,以已有观测资料为样本进行了验证.结果表明,两个网络的训练效果均很好,且单输出目标的分层模拟要优于多输出目标的单层模拟.表明了利用人工神经网络推导缺失波浪条件的可行性.     

遗传算法与神经网络相结合的热带气旋强度预报方法试验

以1960~2001年共41 a的7月和8月西行进入南海海域的热带气旋样本为基础,采用遗传算法与神经网络相结合的方法,进行了热带气旋强度预报模型的预报建模研究.并根据相同的热带气旋个例,将这种遗传-神经网络热带气旋强度预报模型与气候持续法热带气旋强度预报方法进行对比分析,试验预报结果表明,遗传-神经网络方法具有更好的预报能力.     

Temporal significant wave height estimation from wind speed by perceptron Kalman filtering

Significant wave height estimates are necessary for many applications in coastal and offshore engineering and therefore various estimation models are proposed in the literature for this purpose. Unfortunately, most of these models provide simultaneous wave height estimations from wind speed measurements. However, in practical studies, the prediction of significant wave height is necessary from previous time interval measurements. This paper presents a dynamic significant wave height prediction procedure based on the perceptron Kalman filtering concepts. Past measurements of significant wave height and wind speed variables are used for training the adaptive model and it is then employed to predict the significant wave height amounts for future time intervals from the wind speed measurements only. The verification of the proposed model is achieved through the dynamic significant wave height and wind speed time series plots, observed versus predicted values scatter diagram and the classical linear significant wave height models. The application of the proposed model is presented for a station in USA.     

Spectral wave data assimilation for the prediction of waves in the North Sea

Spectral observations from pitch-and-roll buoys have been assimilated in a North Sea wave model, in order to study their impact on the wave analysis and forecast. The assimilation is based on Optimal Interpolation (OI) of a limited number of characteristic spectral parameters. In a case study, the propagation of the corrections through the model domain is followed, and it is clarified for which wave conditions the data assimilation has the largest influence on the forecast: this is especially the case for swell waves with long travel times between the assimilation site and the location where validation is carried out. A 1-year test has been carried out in which an analysis and subsequent forecast were produced four times a day. From a statistical analysis of the results a modest but systematic improvement of the 12-h forecast is found. When only swell cases are selected, the impact is more pronounced. It is argued that for shelf seas like the North Sea, more progress is to be expected from extension of the ‘conventional' observations network (buoys and wave radars) than from satellite measurements.     

HF radar data quality requirements for wave measurement

HF radar wave measurements are presented focussing on theoretical limitations, and thus radar operating parameters, and quality control requirements to ensure robust measurements across a range of sea states. Data from three radar deployments, off the west coast of Norway, Celtic Sea and Liverpool Bay using two different radar systems, WERA and Pisces, and different radio frequency ranges, are used to demonstrate the wave measurement capability of HF radar and to illustrate the points made. Aspects of the measurements that require further improvements are identified. These include modifications to the underlying theory particularly in high sea states, identification and removal of ships and interference from the radar signals before wave processing and/or intelligent partitioning to remove these from the wave spectrum. The need to match the radio frequency to the expected wave peak frequency and waveheight range, with lower radio frequencies performing better at higher waveheights and lower peak frequencies and vice versa, is demonstrated. For operations across a wide range of oceanographic conditions a radar able to operate at more than one frequency is recommended for robust wave measurement. Careful quality control is needed to ensure accurate wave measurements.     

A forecasting model for wave heights based on a long short-term memory neural network

To explore new operational forecasting methods of waves, a forecasting model for wave heights at three stations in the Bohai Sea has been developed. This model is based on long short-term memory(LSTM) neural network with sea surface wind and wave heights as training samples. The prediction performance of the model is evaluated,and the error analysis shows that when using the same set of numerically predicted sea surface wind as input, the prediction error produced by the proposed LSTM model at S...     

Data assimilation of partitioned HF radar wave data into Wavewatch III

In this study the assimilation of HF radar data into a high resolution, coastal Wavewatch III model is investigated. An optimal interpolation scheme is used to assimilate the data and the design of a background error covariance matrix which reflects the local conditions and difficulties associated with a coastal domain is discussed. Two assimilation schemes are trialled; a scheme which assimilates mean parameters from the HF radar data and a scheme which assimilates partitioned spectral HF radar data. This study demonstrates the feasibility of assimilating partitioned wave data into a coastal domain. The results show that the assimilation schemes provide satisfactory improvements to significant wave heights but more mixed results for mean periods. The best improvements are seen during a stormy period with turning winds. During this period the model is deficient at capturing the change in wave directions and the peak in the waveheights, while the high sea state ensures good quality HF radar data for assimilation. The study also suggests that there are both physical and practical advantages to assimilating partitioned wave data compared to assimilating mean parameters for the whole spectrum.