A proposed directional function and wind-wave directional spectrum

Ａｐｒｏｐｏｓｅｄｄｉｒｅｃｔｉｏｎａｌｆｕｎｃｔｉｏｎａｎｄｗｉｎｄ－ｗａｖｅｄｉｒｅｃｔｉｏｎａｌｓｐｅｃｔｒｕｍ￥ＷｅｎＳｈｅｎｇｃｈａｎｇ；ＷｕＫｅｊｉａｎ；ＧｕａｎＣｈａｎｇｌｏｎｇ；ＳｕｎＳｈｉｃａｉａｎｄＺｈａｎｇＤａｃｕｏ（Ｒｅｃｅｉ...     

Computation of wind-wave spectra in coastal waters with SWAN on unstructured grids

An unstructured-grid procedure for SWAN is presented. It is a vertex-based, fully implicit, finite difference method which can accommodate unstructured meshes with a high variability in geographic resolution suitable for representing complicated bottom topography in shallow areas and irregular shoreline. The numerical solution is found by means of a point-to-point multi-directional Gauss–Seidel iteration method requiring a number of sweeps through the grid. The approach is stable for any time step while permitting local mesh refinements in areas of interest. A number of applications are shown to verify the correctness and numerical accuracy of the unstructured version of SWAN.     

基于潮汐表数据同化的天文潮数值预报模型及其模拟预报效果

潮汐表是利用长期潮汐观测结果经调和分析实现的主要港湾潮汐预报结果,具有较高的预报精度,而通常的天文潮数值预报目前还难以达到潮汐表的预报精度.本研究在建立常规天文潮数值预报模型的基础上,建立了基于潮汐表数据同化的天文潮数值预报模型,并分别采用这2种模型预报福建沿岸海域的天文潮.其结果表明同化模型的预报结果无论是在潮时还是在潮高均明显优于常规模型;同化模型能显著地改善所研究的沿岸海域90个水位点中至少45个水位点的潮汐预报结果,而其他水位点的预报结果也有不同程度地改善.     

基于多层感知器的风涌浪分离方法

风涌浪分离是研究风浪、涌浪各自特性的基础,但受限于海浪谱数据的匮乏,基于海浪谱的风涌浪分离方法难以普及应用,有效的解决办法是采用波浪观测中容易获取的基本波要素进行风涌浪分离。现有方法无法利用基本波要素全面计算出风浪、涌浪的比例及其特征参数,为此本文将机器学习引入到风涌浪分离中,以多层感知器模型为基础,提出了一种利用基本波要素、风要素准确计算出风涌浪参数的方法。该方法需要每个测站提供至少466笔、建议766笔及以上的实测波浪数据作为训练样本,适用于台湾海峡3个测站,在计算精度上显著优于基于海浪频谱的传统风涌浪分离方法,可为本海域缺乏海浪谱的测站提供替代性的风涌浪计算方案,有助于扩大实测风涌浪资料的来源,进而加强风涌浪分布特性以及预警预报研究。     

Wave direction analysis from data buoys

This paper describes a new procedure of directional wave analysis from pitch-roll buoy measurements. The two previous procedures adopted by the National Data Buoy Center (NDBC) [Steele, K. E., Lau, J. C. K. and Hsu, Y.-H. L. (1985) Theory and application of calibration techniques for an NDBC directional wave measurement buoy. IEEE Journal of Oceanic Engineering OE-10(4), 382-396; Steele, K. E., Teng, C.-C. and Wang, D. W. C. (1992) Wave direction measurements using pitch-roll buoys. Ocean Engineering 19(4), 349-375] are relevant to our formulations. In these two studies, an estimate for the total phase shift of the sea surface displacement/slope spectra from the measured buoy heave/pitch and heave/roll spectra was calculated either by a weighted average method or a maximum heave/pitch quad-spectrum method. These two formulations were based on a fundamental assumption of symmetric hull-mooring effect on pitch and roll motions, which will never be true in the oceans. In the present study we essentially incorporate the basic formulations of NDBC, but calculate two estimates for this total phase shift.Examples of directional wave analysis from data measured by a 3 m diameter discus buoy during Typhoon Herb are presented in this paper. This data set was also analyzed by the weighted average method of Steele et al. (1985) which yielded unsatisfactory results of wave directions during severe wave climates.     

利用卫星云图资料制作热带气旋预报路径的一种算法

根据台风生成后在云场中移动的环境条件,引入一个环境作用于台风中心的热力梯度力方向,用该热力梯度力方向与当前台风的移动方向相结合,研究台风移动过程的变化规律,寻找一种简易可行的台风路径预报方法．经过多年实践,本文揭示出台风中心未来沿着当前移动方向与环境作用于台风中心的热力方向合成移动的基本规律：当前台风中心移向与前方晴空区中轴线相交时,台风沿着当前移向前进到与阻挡轴线相交点相距4个纬距的位置时便发生偏转,逐渐与阻挡轴线走向趋于一致;若当前台风中心距相交点的距离小于或等于4个纬距时,则从当前位置发生偏转．文中利用以上规律研制出台风移向变化方程和移动轨迹方程．     

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.     

计算结构可靠度的RBF神经网络响应面法

对功能函数不能明确表达的问题进行可靠度分析,常采用响应面法。其中二次多项式响应面法应用较为广泛,采用与此方法相同的思路,提出了RBF神经网络响应面法,并通过算例与常用的BP神经网络响应面法进行了对比分析,该方法在学习速度、迭代次数等方面均优于BP神经网络响应面法。该方法用于大型复杂结构的可靠性分析,可相应提高工作效率和解题质量,具有一定实际应用价值。     

声速预测方程在浅地层剖面资料处理中的应用

海底沉积层的声速足浅地层剖面资料采集和处理的关键参数之一,通常的做法是将地层的声速设定为一个经验值,而实际上声速并非定值.通过对国内外主要沉积物声速颅测方程的比较,利用卢博等建立的适用于中国东南近海的声速经验公式,在某人工岛构造调查中,根据地质钻孔获取的孔隙度参数计算各沉积层的平均声速,建立相应的声速结构剖面,应用于浅...     

基于观测和再分析数据的LSTM深度神经网络沿海风速预报应用研究

基于海洋气象历史观测资料和再分析数据等,利用LSTM深度神经网络方法,开展在有监督学习情况下的海面风场短时预报应用研究。以中国近海5个代表站为研究区域,通过气象台站观测数据和ERA-Interim 6 h再分析数据构建数据集。选取21个变量作为预报因子,分别构建两个LSTM深度神经网络框架(OBS_LSTM和ALL_LSTM)。经与2017年WRF模式6 h预报结果对比分析,得出如下结论:构建的两个LSTM风速预报模型可以大幅降低风速预报误差,RMSE分别降低了41.3%和38.8%,MAE平均降低了43.0%和40.0%;风速误差统计和极端大风分析发现,LSTM模型能够抓住地形、短时大风和台风等敏感信息,对于大风过程预报结果明显优于WRF模式;两种LSTM模型对比发现,ALL_LSTM模型风速预报误差最小,具有很好的稳定性和鲁棒性,OBS_LSTM模型应用范围更广泛。     

小样本机器学习算法的特性分析与应用

基于经典统计学的机器学习算法,在解决小样本学习问题时表现得不能令人满意。在总结分析小样本机器学习算法特点的基础上,以支持向量机(SVM)学习算法为例,定量分析了影响其泛化性能、学习性能的几个因素,实验结果与理论分析结论取得了良好的一致性;SVM用于解决KTH-TIPS纹理图像分类问题,取得了很好的实验结果。     

Weak and strong constraint data assimilation in the inverse Regional Ocean Modeling System (ROMS): Development and application for a baroclinic coastal upwelling system

We describe the development and preliminary application of the inverse Regional Ocean Modeling System (ROMS), a four dimensional variational (4DVAR) data assimilation system for high-resolution basin-wide and coastal oceanic flows. Inverse ROMS makes use of the recently developed perturbation tangent linear (TL), representer tangent linear (RP) and adjoint (AD) models to implement an indirect representer-based generalized inverse modeling system. This modeling framework is modular. The TL, RP and AD models are used as stand-alone sub-models within the Inverse Ocean Modeling (IOM) system described in [Chua, B.S., Bennett, A.F., 2001. An inverse ocean modeling system. Ocean Modell. 3, 137–165.]. The system allows the assimilation of a wide range of observation types and uses an iterative algorithm to solve nonlinear assimilation problems. The assimilation is performed either under the perfect model assumption (strong constraint) or by also allowing for errors in the model dynamics (weak constraints). For the weak constraint case the TL and RP models are modified to include additional forcing terms on the right hand side of the model equations. These terms are needed to account for errors in the model dynamics.Inverse ROMS is tested in a realistic 3D baroclinic upwelling system with complex bottom topography, characterized by strong mesoscale eddy variability. We assimilate synthetic data for upper ocean (0–450 m) temperatures and currents over a period of 10 days using both a high resolution and a spatially and temporally aliased sampling array. During the assimilation period the flow field undergoes substantial changes from the initial state. This allows the inverse solution to extract the dynamically active information from the synthetic observations and improve the trajectory of the model state beyond the assimilation window. Both the strong and weak constraint assimilation experiments show forecast skill greater than persistence and climatology during the 10–20 days after the last observation is assimilated.Further investigation in the functional form of the model error covariance and in the use of the representer tangent linear model may lead to improvement in the forecast skill.     

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.     

Near-shore swell estimation from a global wind-wave model: Spectral process,linear, and artificial neural network models

Estimation of swell conditions in coastal regions is important for a variety of public, government, and research applications. Driving a model of the near-shore wave transformation from an offshore global swell model such as NOAA WaveWatch3 is an economical means to arrive at swell size estimates at particular locations of interest. Recently, some work (e.g. Browne et al. [Browne, M., Strauss, D., Castelle, B., Blumenstein, M., Tomlinson, R., 2006. Local swell estimation and prediction from a global wind-wave model. IEEE Geoscience and Remote Sensing Letters 3 (4), 462–466.]) has examined an artificial neural network (ANN) based, empirical approach to wave estimation. Here, we provide a comprehensive evaluation of two data driven approaches to estimating waves near-shore (linear and ANN), and also contrast these with a more traditional spectral wave simulation model (SWAN). Performance was assessed on data gathered from a total of 17 near-shore locations, with heterogenous geography and bathymetry, around the continent of Australia over a 7 month period. It was found that the ANNs out-performed SWAN and the non-linear architecture consistently out-performed the linear method. Variability in performance and differential performance with regard to geographical location could largely be explained in terms of the underlying complexity of the local wave transformation.     

BP网络学习参数模糊自适应算法的实现

前馈神经网络BP算法的改进方案中,对网络训练(学习)过程中学习率和惯性系数进行模糊自适应调节,以提高收敛速度,是一项很有效的措施。文中具体分析了如何根据设计者的先验知识确定模糊规则和隶属函数,并以三比特异或函数(或称奇偶分类)的实现为例,验证了这种算法的改进、加速了BP网络的学习过程。     

A computer model for forecasting ocean wave parameters at specific sites around the the New Zealand coast

An applications model for forecasting frequency‐directional wave spectra at any appropriately specified site is described. There are two stages to the calculations. Firstly, a spectrum is calculated based on results at nearby gridpoints from an ocean wave prediction model. This is then adjusted to make the spectrum consistent with the local wind history. Verifications of the model are made at sites off Cape Egmont and Great Barrier Island, North Island, New Zealand. These give encouraging results for the shape of the frequency spectrum, with reasonable skill evident in the energetic parts of the spectrum. The significant wave heights also agree well, with the model estimates explaining two thirds of the measured variance.     

Neural network for wave forecasting among multi-stations

Unlike in the open sea, the use of wind information for forecasting waves may encounter more ambiguous uncertainties in the coastal or harbor area due to the influence of complicated geometric configurations. Thus this paper attempts to forecast the waves based on learning the characteristics of observed waves, rather than the use of the wind information. This is reported in this paper by the application of the artificial neural network (ANN), in which the back-propagation algorithm is employed in the learning process for obtaining the desired results. This model evaluated the interconnection weights among multi-stations based on the previous short-term data, from which a time series of waves at a station can be generated for forecasting or data supplement based on using the neighbor stations data. Field data are used for testing the applicability of the ANN model. The results show that the ANN model performs well for both wave forecasting and data supplement when using a short-term observed wave data.     

Methods for automatic cloud classification from MODIS data

Izvestiya, Atmospheric and Oceanic Physics - In this paper, different texture-analysis methods are used to describe different cloud types in MODIS satellite images. A universal technique is...