Multi-step ahead short-term predictions of storm surge level using CNN and LSTM network

Storm surges pose significant danger and havoc to the coastal residents' safety, property, and lives, particularly at offshore locations with shallow water levels. Predictions of storm surges with hours of warning time are important for evacuation measures in low-lying regions and coastal management plans. In addition to experienced predictions and numerical models, artificial intelligence (AI) techniques are also being used widely for short-term storm surge prediction owing to their merits in good level of prediction accuracy and rapid computations. Convolutional neural network (CNN) and long short-term memory (LSTM) are two of the most important models among AI techniques. However, they have been scarcely utilised for surge level (SL) forecasting, and combinations of the two models are even rarer. This study applied CNN and LSTM both individually and in combination towards multi-step ahead short-term storm surge level prediction using observed SL and wind information. The architectures of the CNN, LSTM, and two sequential techniques of combining the models (LSTM–CNN and CNN–LSTM) were constructed via a trial-and-error approach and knowledge obtained from previous studies. As a case study, 11 a of hourly observed SL and wind data of the Xiuying Station, Hainan Province, China, were organised as inputs for training to verify the feasibility and superiority of the proposed models. The results show that CNN and LSTM had evident advantages over support vector regression (SVR) and multilayer perceptron (MLP), and the combined models outperformed the individual models (CNN and LSTM), mostly by 4％–6％. However, on comparing the model computed predictions during two severe typhoons that resulted in extreme storm surges, the accuracy was found to improve by over 10％ at all forecasting steps.     

桑沟湾不同形态锰的分布、季节变化及影响因素

利用催化动力学分光光度法和两步提取法对2011年4月（春）、8月（夏）、10月（秋）和2012年1月（冬）桑沟湾海域溶解态无机锰(DIMn)和表层沉积物中的锰的含量进行测定。结果表明,桑沟湾4个季节（春季至冬季,后同）DIMn浓度呈现出近岸高、远岸低的分布特点,其平均浓度分别为(60.5±43.1) nmol/L、(42.0±30.5) nmol/L、(23.4±11.2) nmol/L和(18.2±13.5) nmol/L,呈现出明显的季节变化,即春季最高,夏季、秋季次之,冬季最低;与相邻的俚岛湾和爱莲湾相比,桑沟湾春季、夏季DIMn的浓度较高,秋季、冬季则没有显著性差异。桑沟湾表层沉积物中总Mn在4个季节的含量分别为(861±308) mg/kg、(915±322) mg/kg、(589±108) mg/kg、(653±185) mg/kg,表层沉积物中醋酸提取态Mn在4个季节的含量分别为(500±272) mg/kg、(502±232) mg/kg、(322±81) mg/kg、(345±91) mg/kg,两者均表现出近岸高、远岸低的分布特点。醋酸提取态Mn的含量在春季、夏季要显著高于秋季、冬季。悬浮颗粒物的吸附和浮游生物的利用是影响桑沟湾DIMn浓度与分布的重要因素。桑沟湾DIMn的源主要包括河流及地下水输送、大气输送、沉积物−水界面释放;汇主要包括养殖生物的清除、向黄海的输送等。简单箱式模型收支计算结果显示,桑沟湾DIMn的源略大于汇,表明除了养殖生物的清除和向黄海的输送,桑沟湾DIMn还存在其他汇。本研究的结果为桑沟湾DIMn的生物地球化学循环的深入认识提供了基础数据。     

Using triple oxygen isotopes and oxygen-argon ratio to quantify ecosystem production in the mixed layer of northern South China Sea slope region

Quantifying the gross and net production is an essential component of carbon cycling and marine ecosystem studies.Triple oxygen isotope measurements and the O_2/Ar ratio are powerful indices in quantifying the gross primary production and net community production of the mixed layer zone,respectively.Although there is a substantial advantage in refining the gas exchange term and water column vertical mixing calibration,application of mixed layer depth history to the gas exchange term and its contribution to reducing indices error are unclear.Therefore,two cruises were conducted in the slope regions of the northern South China Sea in October 2014(autumn) and June 2015(spring).Discrete water samples at Station L07 in the upper 150 m depth were collected for the determination of δ~(17)0,δ~(18)O,and the O_2/Ar ratio of dissolved gases.Gross oxygen production(GOP) was estimated using the triple oxygen isotopes of the dissolved O_2,and net oxygen production(NOP) was calculated using O_2/Ar ratio and O_2 concentration.The vertical mixing effect in NOP was calibrated via a N_2O based approach.GOP for autumn and spring was(169±23) mmol/(m~2·d)(by O_2) and(189±26) mmol/(m~2·d)(by O_2),respectively.While NOP was 1.5 mmol/(m~2·d)(by O_2) in autumn and 8.2 mmol/(m~2·d)(by O_2) in spring.Application of mixed layer depth history in the gas flux parametrization reduced up to 9.5% error in the GOP and NOP estimations.A comparison with an independent O_2 budget calculation in the diel observation indicated a26% overestimation in the current GOP,likely due to the vertical mixing effect.Both GOP and NOP in June were higher than those in October.Potential explanations for this include the occurrence of an eddy process in June,which may have exerted a submesoscale upwelling at the sampling station,and also the markedly higher terrestrial impact in June.     

TOPEX Constrained Solution of Mean Sea Surface by Joint Processing of TOPEX,ERS-1 and GEOSAT Altimetric Measurements

We present an improved crossover adjustment procedure to determine mean sea surface height using TOPEX, 35-day repeat phase ERS-1, Geosat, and 168-day repeat phase ERS-1 satellite altimeter data. The mean sea surface frame defined by the TOPEX data is imposed as certain constraints in our crossover adjustment procedure rather than held fixed as in some other procedures. The new procedure is discussed in detail. Equations are developed to incorporate the a priori information of Topex data as well as other satellite altimeter data. The numerical computation result shows that the rms crossover discrepancies are reduced by an order of 1 cm when the Topex data is not fixed. Furthermore, the computed mean sea surface is less noisy and more realistic than that computed by the traditional procedure.     

Integrating Uncertainty Theories with GIS for Modeling Coastal Hazards of Climate Change

Prediction of coastal hazards due to climate change is fraught with uncertainty that stems from complexity of coastal systems, estimation of sea level rise, and limitation of available data. In-depth research on coastal modeling is hampered by lack of techniques for handling uncertainty, and the available commercial geographical information systems (GIS) packages have only limited capability of handling uncertain information. Therefore, integrating uncertainty theory with GIS is of practical and theoretical significance. This article presents a GIS-based model that integrates an existing predictive model using a differential approach, random simulation, and fuzzy set theory for predicting geomorphic hazards subject to uncertainty. Coastal hazard is modeled as the combined effects of sea-level induced recession and storm erosion, using grid modeling techniques. The method is described with a case study of Fingal Bay Beach, SE Australia, for which predicted responses to an IPCC standard sea-level rise of 0.86 m and superimposed storm erosion averaged 12 m and 90 m, respectively, with analysis of uncertainty yielding maximum of 52 m and 120 m, respectively. Paradoxically, output uncertainty reduces slightly with simulated increase in random error in the digital elevation model (DEM). This trend implies that the magnitude of modeled uncertainty is not necessarily increased with the uncertainties in the input parameters. Built as a generic tool, the model can be used not only to predict different scenarios of coastal hazard under uncertainties for coastal management, but is also applicable to other fields that involve predictive modeling under uncertainty.     

浮游植物生物量研究 Ⅰ.浮游植物生物量细胞体积转化法

在海洋生态动力学研究过程中,采用浮游植物细胞数量来估算浮游植物丰度可以说是不够精确的,因为不同种的浮游植物细胞大小差别很大,只有浮游植物的生物量才能正确反映海洋生态系中的能量分布．本文以拟合浮游植物细胞相似体积方法,基于胶州湾生态动力学研究所获资料,计算了87种中国近海常见浮游植物的细胞体积、鲜重、碳含量、氮含量．     

免疫多糖对栉孔扇贝酸性磷酸酶、碱性磷酸酶和超氧化物歧化酶活性的影响

在栉孔扇贝的血清、血细胞和肝脏提取液中均发现有酸性磷酸酶（ＡＣＰ）、碱性磷酸酶（ＡＬＰ）和超氧化物歧化酶（ＳＯＤ）活性,其中,ＡＣＰ和ＡＬＰ在肝脏中的活性最高。采用１．０％的虫草多糖或海藻多糖作为免疫药物,对栉孔扇贝进行注射刺激后,发现血清中的ＡＣＰ、ＡＬＰ和ＳＯＤ活性有显著提高,血细胞中的ＡＣＰ和ＳＯＤ活性也有所增加,而肝脏提取液的这三种酶活性的变化不很明显。研究认为这两种多糖具有增强栉孔扇贝免疫活性的作用,在水产养殖业中有一定的开发价值。     

山东栖霞桃村地区中酸性火山-侵入杂岩岩石地球化学及其成因研究

山东栖霞桃村地区出露中生代中酸性火山-侵入杂岩,对采自桃村地区青山组下段的中酸性火山-侵入岩以进行了岩石学和地球化学研究。桃村地区的中酸性火成岩的地球化学特征为:高钾低钠,高铝低镁;Cr,Co,Ni含量较低,Sr/Y及(La/Yb)N比值很高;微量元素蛛网图显示富集不相容元素K,Rb,Ba和Th,以及不相容元素Ce,Hf和Ti,亏损高场强元素Ta,Nb以及P;轻重稀土元素分馏程度高,具有类似埃达克岩的性质。综合分析其地球化学特征显示,青山组下段中酸性杂岩为增厚的下地壳熔融,这种地壳的增厚作用可能由古太平洋板块俯冲作用引起,而在俯冲作用过程中青山组中酸性杂岩的源岩被俯冲沉积物混染。     

Three-dimensional liquid sloshing in a tank with baffles

A numerical model has been developed to study three-dimensional (3D) liquid sloshing in a tank with baffles. The numerical model solves the spatially averaged Navier-Stokes equations, which are constructed on a non-inertial reference frame having six degree-of-freedom (DOF) of motions. The large-eddy-simulation (LES) approach is employed to model turbulence by using the Smagorinsky sub-grid scale (SGS) closure model. The two-step projection method is employed in the numerical solutions, aided by the Bi-CGSTAB technique to solve the pressure Poisson equation for the filtered pressure field. The second-order accurate volume-of-fluid (VOF) method is used to track the distorted and broken free surface. The baffles in the tank are modeled by the concept of virtual boundary force (VBF) method. The numerical model is first validated against the available analytical solution and experimental data for two-dimensional (2D) liquid sloshing in a tank without baffles. The 2D liquid sloshing in tanks with baffles is then investigated. The numerical results are compared with other results from available literatures. Good agreement is obtained. Finally, the model is used to study 3D liquid sloshing in a tank with vertical baffles. The effect of the baffle is investigated and discussed.     

Mechanism and enhancement of photoluminescence from silicon nanocrystals implanted in SiO2 matrix

Photoluminescence (PL) spectra of Si nanocrystals (NCs) prepared by 130 keV Si ions implantation onto SiO₂ matrix were investigated as a function of annealing temperature and implanted ion dose. PL spectra consist of two PL peaks, originated from smaller Si NCs due to quantum confinement effect (QCE) and the interface states located at the surface of larger Si NCs. The evolution of number of dangling bonds (DBs) on Si NCs was also investigated. For hydrogen-passivated samples, a monotonic increase in PL peak intensity with the dose of implanted Si ions up to 3× 10¹⁷ ions /cm² is observed. The number of DBs on individual Si NC, the interaction between DBs at the surface of neighbouring Si NCs and their effects on the efficiency of PL are discussed.