Initialization and Simulation of a Typhoon Using 4-Dimensional Variational Data Assimilation——Research on Typhoon Herb (1996)

The MM5 and its four dimensional variational data assimilation (4D-Var) system are used in this paper. Based on the National Centers for Environmental Prediction/National Center for Atmospheric Research (NCEP/NCAR) reanalysis data, the authors generate an optimal initial condition for a typhoon by using the bogus data assimilation (BDA) scheme. BDA is able to recover many of the structural features of typhoons including a warm-core vertex, the correct center position, and the strong circulation. As a result of BDA using a bogus surface low, dramatic improvement is achieved in the 72 h prediction of typhoon Herb. Through several cases, the initialization by BDA effectively generates the harmonious inner structure of the typhoon, but which is lacking in the original analysis field. Therefore the intensity forecast is improved greatly. Some improvements are made in the track forecast, but more work still needs to be done.

A STUDY ON SATELLITE DATA ASSIMILATION WITH DIFFERENT ATOVS IN TYPHOON NUMERICAL EXPERIMENTS

Based on the newly developed Weather Research and Forecasting model(WRF)and its three-dimensional variational data assimilation(3DVAR)system,this study constructed twelve experiments to explore the impact of direct assimilation of different ATOVS radiance on the intensity and track simulation of super-typhoon Fanapi(2010)using a data assimilation cycle method.The result indicates that the assimilation of ATOVS radiance could improve typhoon intensity effectively.The average bias of the central sea level pressure(CSLP)drops to 18 hPa,compared to 42 hPa in the experiment without data assimilation.However,the influence due to different radiance data is not significant,which is less than 6hPa on average,implying limited improvement from sole assimilation of ATOVS radiance.The track issue is studied in the following steps.First,the radiance from the same sensor of different satellites could produce different effect.For the AMSU-A,NOAA-15 and NOAA-18,they produce equivalent improvement,whereas NOAA-16 produces slightly poor effect.And for the AMSU-B,NOAA-15 and NOAA-16,they produce equivalent and more positive effect than that provided by the AMSU-A.Second,the assimilation radiance from different sensors of the identical satellites could also produce different effect.The assimilation of AMSU-B produces the largest improvement,while the ameliorating effect of HIRS/3assimilation is inferior to that of AMSU-B assimilation,while the AMSU-A assimilation exhibits the poorest improvement.Moreover,the simultaneous assimilation of different radiance could not produce further improvement.Finally,the experiments of simultaneous assimilation radiance from multiple satellites indicate that such assimilation may lead to negative effect due to accumulative bias when adding various radiance data into the data assimilation system.Thus the assimilation of ATOVS radiance from a single satellite may perform better than that from two or three satellites.

Aeolus水平径向风速产品资料同化在台风预报中的应用研究

为评价星载激光测风雷达产品资料同化对台风预报的影响,以2020年台风“黑格比”和2019年台风“利奇马”为例,选用Aeolus卫星的水平径向风速产品,开展星载激光测风雷达资料同化及其对台风预报影响研究。首先基于ERA5再分析资料对Aeolus产品进行统计检验和偏差订正,并统计各高度层观测背景差的均方差作为观测误差用于同化试验。然后通过WRFDA(Weather Research and Forecast Model Data Assimilation)模式同化Aeolus产品并利用WRF(Weather Research and Forecast Model)模式开展预报试验。试验结果表明,Aeolus产品精度较高,统计所得各层观测误差在3~5 m s^(-1)范围且随高度的升高而增大,订正后资料满足三维变分同化无偏假设;在台风“黑格比”和“利奇马”期间同化该产品能使模式具备更加合理的环流形势,能够有效提高对台风“黑格比”和“利奇马”路径及强度的预报效果。

初始场水汽通量在一次登陆台风降水模拟中的作用

用三维变分同化方法,通过同化加密地面观测资料和X波段雷达观测资料,研究初始场陆地侧水汽通量对台风“狮子山”登陆期间降水模拟效果的影响.结果表明,同化加密地面资料能够明显改进初始场陆地侧大气边界层的热动力场,通过减弱海-陆热力对比及陆地向登陆台风的大气边界层东北的水汽通量,改善登陆台风外围雨带影响下的沿岸陆地-海洋两侧地区降水的模拟效果. X波段雷达资料通过增强初始场中雷达探测范围内的低层大气水汽通量改善登陆台风外围雨带降水的模拟效果.

风云三号C星微波湿度计资料全天候同化对台风玛莉亚预报的影响

随着风云三号系列卫星的成功发射,越来越多的卫星微波直接观测资料应用于数值天气预报的资料同化系统。并且由于卫星微波全天候同化技术可以充分利用晴天及云雨区微波观测资料,在增加同化使用的观测数据的基础上,有效提高数值天气预报准确率,该技术在卫星资料同化领域也颇受瞩目。本研究选取2018年7月的台风玛莉亚,利用WRF(Weather Research and Forecasting)模式及其同化系统WRFDA(WRF Data Assimilation)中三维变分方法,探讨风云三号C星微波湿度计观测资料的全天候同化技术在区域模式中的适用性,以及其在不同模式驱动场中的预报表现。通过对比仅同化晴空区域卫星资料的试验和全天候同化的试验结果发现,全天候条件下更多的云雨区域观测资料被有效利用,能够更好地模拟出台风玛莉亚核心区域的暖心和对称风速结构,有效改善湿度场的预报,对台风路径的预报误差平均降低了大约34%~62%,且这种正面影响均能在不同模式驱动场中得到体现。

基于监测数据的桥梁台风期结构响应研究

针对沿海地区台风过境后桥梁结构安全状况的评价难题,以浙江沿海地区一座无背索独塔斜拉桥为背景工程,提取2022年台风“梅花”登陆浙江沿海期间该桥健康监测系统获得的全过程数据,通过数据分析研究台风历经前后环境状况的变化情况,以及桥梁关键结构响应(主梁振动、拉索振动、应变和转角数据)的演化过程。结果表明:台风过境过程中,环境温度在台风中降低、湿度增大;结构响应特征中的主梁竖向振动、主梁应变等参数未发生显著变化,拉索振动加速度变化最为显著。台风过境后,结构响应恢复到过境前水平,表明台风并未对结构整体产生显著影响。基于监测数据分析台风过境前后状态,可以显示台风历经前后外部环境和结构内部响应变化趋势,揭示控制桥梁安全的关键响应,可作为台风过后桥梁安全性评价的重要方法。

多源数据驱动的台风动态风险评估

灾害动态风险评估是灾前风险管理的重要依据。文章基于2009—2022影响中国东南5省的108个台风案例近4000个县级灾情,融合30类多源风险要素指标,建立台风动态风险评估样本库,并利用随机森林算法建立了6个台风灾害风险评估模型,用于灾害事件发生前对受灾人口、紧急转移安置人口、农作物受灾面积、倒塌和严重损坏房屋、直接经济损失等风险等级以及综合风险等级进行评估。通过实际灾情与模型结果进行验证,灾害风险评估结果准确率整体达到80%以上,表明该模型具有较好的泛化能力,可用于实际灾害评估工作。实验对比发现,训练样本量提高1~2个数量级能使模型评估准确率提升3%~14%,表明灾害风险大数据的积累对灾害风险评估研究具有重要意义。

基于浮标站和ERA5再分析资料的浙江沿海台风浪特征分析

为了分析台风影响下浙江沿海风和浪的演变特点,利用浙江省海洋浮标站监测数据和欧洲中期天气预报中心第五代全球气候大气再分析数据(European Centre for Medium-Range Weather Forecasts Reanalysis v5,ERA5),选取2010年以来严重影响浙江的7次台风个例,对台风作用下浙江沿海海面风和浪的演变特点进行分析。结果表明:在台风影响过程中,海浪波型多数呈现混合浪-风浪-混合浪的演变规律;涌浪波型的出现与台风强度及其与浮标站的距离和方位有关,也与海洋潮汐现象紧密相关。台风影响期间,浙江沿海浪高的变化受风速和风向共同作用影响。在风向不变的情况下,持续风速增大对浪高的增大有明显作用;风向的变化也会对浪高变化产生影响,向岸风和离岸风的转变会造成浪高出现剧烈变化。ERA5再分析资料有效波高在台风浪较大时会呈现偏小的趋势,分析订正后的ERA5有效波高发现,台风浪有效波高大值区与台风中心位置相关。研究结果可为严重影响浙江沿海的台风浪预报服务提供参考。

A nowcasting model for the prediction of typhoon tracks based on a long short term memory neural network

It is of vital importance to reduce injuries and economic losses by accurate forecasts of typhoon tracks. A huge amount of typhoon observations have been accumulated by the meteorological department, however, they are yet to be adequately utilized. It is an effective method to employ machine learning to perform forecasts. A long short term memory（LSTM） neural network is trained based on the typhoon observations during 1949–2011 in China＇s Mainland, combined with big data and data mining technologies, and a forecast model based on machine learning for the prediction of typhoon tracks is developed. The results show that the employed algorithm produces desirable 6–24 h nowcasting of typhoon tracks with an improved precision.

CORRECTION OF ASYMMETRIC STRENGTHENING OF QUIKSCAT WIND FIELD AND ASSIMILATION APPLICATION IN TYPHOON SIMULATION

As an approach to the technological problem that the wind data of QuikSCAT scatterometer cannot accurately describe the zone of typhoon-level strong wind speed, some objective factors such as the typhoon moving speed, direction and friction are introduced in this study to construct the asymmetric strengthening of the QuikSCAT wind field. Then by adopting a technology of four-dimensional data assimilation, an experiment that includes both the assimilation and forecasting phases is designed to simulate Typhoon Rananim numerically. The results show that with model constraints and adjustment, this technology can incorporate the QuikSCAT wind data to the entire column of the model atmosphere, improve greatly the simulating effects of the whole-column wind, pressure field and the track as well as the simulated typhoon intensity covered by the forecast phase, and work positively for the forecasting of landfall locations.

SATELLITE-BASED ANALYSIS ON THE CONCENTRIC EYEWALL REPLACEMENT CYCLES OF SUPER TYPHOON MUIFA (1109)

Multisatellite data is used to analyze the characteristics of three eyewall replacement cycles(ERCs) during the lifetime of Typhoon Muifa(1109).Spiral rainbands evolutions,concentric eyewall(CE) structure modes,CE durations,and intensity changes are discussed in detail.In addition,an ERC evolution model of Typhoon Muifa is given.There are four main findings.(1) The outer spiral rainband joins end to end to form the outer eyewall after it disconnects from the original(inner) eyewall.The inner eyewall weakens as the outer eyewall becomes axisymmetric and is intensified.The contraction of the outer eyewall causes the inner eyewall to dissipate rapidly.Finally,the ERC ends with an annular eyewall or spiral rainbands.(2) Although the CE duration times of Typhoon Muifa's three ERCs covered a large range,the CE structures were all maintained for approximately 5 h from the formation of the axisymmetric outer eyewall to the end of the cycle.(3) There is no obvious precipitation reflectivity in the eye or moat region for the subsidence flow.The convection within the two eyewalls is organized as a radially outward slope with increasing height.(4) Typhoon intensity estimation results based on ADT may not explain the intensity variations associated with ERC correctly,while the typhoon's warm core data retrieved from AMSU-A works well.

A baroclinic typhoon model with a moving multi-nested grid andvariational adjustment initializationII. Forecast experiments

A 3-dimensional baroclinic typhoon model with a moving multi-nested grid and its initialization are described first. Prediction results are improved by using a simple but effective data assimilation method in which the initial field is adjusted by the sixth hour's typhoon report and the weak-constraint variational principle. Finally someforecast examples made by this typhoon model are given.

Assimilation and Simulation of Typhoon Rusa (2002) Using the WRF System

Using the recently developed Weather Research and Forecasting (WRF) 3DVAR and the WRF model, numerical experiments are conducted for the initialization and simulation of typhoon Rusa (2002). The observational data used in the WRF 3DVAR are conventional Global Telecommunications System (GTS) data and Korean Automatic Weather Station (AWS) surface observations. The Background Error Statistics (BES) via the National Meteorological Center (NMC) method has two different resolutions, that is, a 210-km horizontal grid space from the NCEP global model and a 10-km horizontal resolution from Korean operational forecasts. To improve the performance of the WRF simulation initialized from the WRF 3DVAR analyses, the scale-lengths used in the horizontal background error covariances via recursive filter are tuned in terms of the WRF 3DVAR control variables, streamfunction, velocity potential, unbalanced pressure and specific humidity. The experiments with respect to different background error statistics and different observational data indicate that the subsequent 24-h the WRF model forecasts of typhoon Rusa's track and precipitation are significantly impacted upon the initial fields. Assimilation of the AWS data with the tuned background error statistics obtains improved predictions of the typhoon track and its precipitation.

THE EFFECT OF SAMPLE OPTIMIZATION ON THE ENSEMBLE KALMAN FILTER IN FORECASTING TYPHOON RAMMASUN(2014)

In a limited number of ensembles, some samples do not adequately reflect the true atmospheric state and can in turn affect forecast performance. This study explored the feasibility of sample optimization using the ensemble Kalman filter(EnKF) for a simulation of the 2014 Super Typhoon Rammasun, which made landfall in southern China in July 2014. Under the premise of sufficient ensemble spread, keeping samples with a good fit to observations and eliminating those with poor fit can affect the performance of En KF. In the sample optimization, states were selected based on the sample spatial correlation between the ensemble state and observations. The method discarded ensemble states that were less representative and, to maintain the overall ensemble size, generated new ensemble states by reproducing them from ensemble states with a good fit by adding random noise. Sample selection was performed based on radar echo data. Results showed that applying En KF with optimized samples improved the estimated track, intensity,precipitation distribution, and inner-core structure of Typhoon Rammasun. Therefore, the authors proposed that distinguishing between samples with good and poor fits is vital for ensemble prediction, suggesting that sample optimization is necessary to the effective use of En KF.

改进残差网络及时序气象卫星云图的台风等级分类研究

结合了传统残差网络在数据样本小环境下分支卷积层特征的浪费问题,且考虑台风云图时空关联性强、特征复杂因素,参考日本国立情报学研究所在西北太平洋上空通过数个气象卫星拍摄的8000多景高分辨率台风云图,编制了适应于残差神经网络的时序台风云图分类训练集和测试集。为满足数据集及台风特征,有效优化了原始残差网络的残差块,并得到了恒等映射残差块。经由增加卷积输出来促进分支通路更好的被利用,保留台风图像时序特性,提高网络性能。实验显示,W-ResNets模型在自建台风数据集上的训练精度达到99.60%,测试精度达到76.19%,相较于浅层卷积神经网络的测试精度高出8.48%,相比于使用传统的残差神经网络提高了2.87%,为进一步验证模型的泛化性能,采用MNIST通用数据集进行网络对比实验,宽残差网络得到98.786%的测试精度,优于传统残差网络。文中的W-ResNets预报模型可在小样本台风数据集推广使用。

FY-3A SATELLITE MICROWAVE DATA ASSIMILATION EXPERIMENTS IN TROPICAL CYCLONE FORECAST

China's new generation of polar-orbiting meteorological satellite FY-3A was successfully launched on May 26,2008,carrying microwave sounding devices which had similar performance to ATOVS of NOAA series.In order to study the application of microwave sounding data in numerical prediction of typhoons and to improve typhoon forecasting,we assimilated data directly for numerical forecasting of the track and intensity of the 2009 typhoon Morakot(0908)based on the WRF-3DVar system.Results showed that the initial fields of the numerical model due to direct assimilation of FY-3A microwave sounding data was improved much more than that due to assimilation of conventional observations alone,and the improvement was especially significant over the ocean,which is always without conventional observations.The model initial fields were more reasonable in reflecting the initial situation of typhoon circulation as well as temperature and humidity conditions,and typhoon central position at sea was also adjusted.Through direct 3DVar assimilation of FY-3A microwave data,the regional mesoscale model improves the forecasting of typhoon track.Therefore,the FY-3A microwave data could efficiently improve the numerical prediction of typhoons.

基于网格化多源异构数据的台风灾害下杆塔受损数量预测

台风灾害下受灾区域杆塔受损数量预测对提高电网公司防灾减灾能力具有重要意义。为此从全新的统计学习角度出发,基于气象数据、电网数据、地理数据等多源异构数据原始样本,对数据进行分类变量处理、特征筛选、标准化等预处理,并以网格内的杆塔受损数量为响应变量,基于梯度提升决策树算法构建台风灾害下杆塔受损数量预测模型,对台风灾害下配网10 kV杆塔受损数量进行预测。利用广东省徐闻县的数据样本对模型进行训练测试与验证,仿真结果表明所提方法可较为准确地预测台风灾害下配网杆塔受损数量,为电网公司的灾前风险量化评估、制定应急方案等提供辅助决策参考。

扩展水凝物控制变量的雷达资料同化对台风“天鸽”数值模拟的影响

利用WRF(Weather Research and Forecast)模式及WRFDA(WRF model data assimilation system)系统,针对2017年台风“天鸽”个例通过同化雷达径向速度(Vr)和反射率因子(RF),研究水凝物控制变量的雷达资料同化对台风分析预报的影响。研究表明:雷达径向速度的直接同化有效地改进了模式初始场中台风涡旋区的中小尺度信息,分析场中产生了气旋性的风场增量,对模式背景场中的台风有显著增强作用。通过在传统控制变量中扩展针对水凝物的控制变量可有效地同化雷达反射率因子资料,对初始场的水物质进行调整,并对随后确定性预报的台风路径和强度都有一定的正效果。此外,相比没有水凝物控制变量的雷达同化试验,加入了水凝物控制变量的雷达资料同化试验降水预报效果更好。这为将我国近海的地基多普勒天气雷达用于台风初始化分析和预报提供了一定的技术支撑和保障。

“印-太海洋环境变异与海气相互作用”国际合作项目成果介绍--海洋与台风相互作用方向

本文介绍了我国开展的“印-太海洋环境变异与海气相互作用”国际合作项目取得的研究进展。从海洋与台风相互作用研究方向出发,阐述了项目中在台风组网观测、台风与海洋中尺度涡相互作用机制、海洋对台风的低频响应与调制等方面取得的重要成果。基于多元海洋数据同化技术与海气耦合模式的改进,我国在台风预报技术方面取得了重大突破,大大降低了台风强度的预报偏差。依托本项目增进了对印-太海域海气相互作用过程和机理的科学认识,提升了我国海洋国际合作地位和影响力。

基于Leaflet与Vue的台风风暴潮预报可视化系统组件化设计与实现

分析了海洋动力过程常用的可视化分析方法,并对台风风暴潮预报业务进行了开源WebGIS引擎的对比与选型。为了满足台风风暴潮预报可视化的多元化和轻量化需求,采用了Leaflet作为WebGIS框架引擎,基于组件化开发思想,阐述了采用Leaflet与Vue的组件化页面设计及相应实现方法。此方法可以描述台风风暴潮预报可视化系统的总体交互逻辑与可视化效果。