Analogue correction method of errors and its application to numerical weather prediction

In this paper, an analogue correction method of errors （ACE） based on a complicated atmospheric model is further developed and applied to numerical weather prediction （NWP）. The analysis shows that the ACE can effectively reduce model errors by combining the statistical analogue method with the dynamical model together in order that the information of plenty of historical data is utilized in the current complicated NWP model, Furthermore, in the ACE, the differences of the similarities between different historical analogues and the current initial state are considered as the weights for estimating model errors. The results of daily, decad and monthly prediction experiments on a complicated T63 atmospheric model show that the performance of the ACE by correcting model errors based on the estimation of the errors of 4 historical analogue predictions is not only better than that of the scheme of only introducing the correction of the errors of every single analogue prediction, but is also better than that of the T63 model.

An Intermediate Reference Datum Static Correction Technique and Its Applications

Static corrections using the conventional method are basically conducted in two steps, the weathering correction followed by the correction from the top of the sub-weathering to the unified datum. However, the conventional method fails to well deal with statics problems in case the top of the sub-weathering is sharply undulated and the lateral velocity of the sub-weathering varies significantly. This brings us to the introduction of a smooth intermediate reference datum （IRD） located under the top of the sub-weathering, which helps to further increase the accuracy of statics based on the weathering corrections, and ensures the imaging quality. Good results based on the IRD technique have been achieved in the complex areas in western China. This paper discusses the IRD functions, its application requirements, and selection of related parameters. Some typical sections for comparison are also given in this paper.

A Deep Learning Method for Bias Correction of ECMWF 24–240 h Forecasts

Correcting the forecast bias of numerical weather prediction models is important for severe weather warnings.The refined grid forecast requires direct correction on gridded forecast products,as opposed to correcting forecast data only at individual weather stations.In this study,a deep learning method called CU-net is proposed to correct the gridded forecasts of four weather variables from the European Centre for Medium-Range Weather Forecast Integrated Forecasting System global model(ECMWF-IFS): 2-m temperature,2-m relative humidity,10-m wind speed,and 10-m wind direction,with a forecast lead time of 24 h to 240 h in North China.First,the forecast correction problem is transformed into an image-toimage translation problem in deep learning under the CU-net architecture,which is based on convolutional neural networks.Second,the ECMWF-IFS forecasts and ECMWF reanalysis data(ERA5) from 2005 to 2018 are used as training,validation,and testing datasets.The predictors and labels(ground truth) of the model are created using the ECMWF-IFS and ERA5,respectively.Finally,the correction performance of CU-net is compared with a conventional method,anomaly numerical correction with observations(ANO).Results show that forecasts from CU-net have lower root mean square error,bias,mean absolute error,and higher correlation coefficient than those from ANO for all forecast lead times from 24 h to 240 h.CU-net improves upon the ECMWF-IFS forecast for all four weather variables in terms of the above evaluation metrics,whereas ANO improves upon ECMWF-IFS performance only for 2-m temperature and relative humidity.For the correction of the 10-m wind direction forecast,which is often difficult to achieve,CU-net also improves the correction performance.

A Limb Correction Method for the Microwave Temperature Sounder 2 and Its Applications

The Microwave Temperature Sounder 2(MWTS-2)is a cross-track radiometer that has 13 channels of sampling radiances emitted from different vertical levels of the atmosphere.Because of the varying scan angles of each field of view within a scan line,observations from the MWTS-2 are subject to strong scan-position-dependent features,i.e.,the limb effect.When examining brightness temperatures(TBs),weather signals observed at every temperature-sounding channel are often concealed by scan-dependent patterns.This study,therefore,proposes a limb correction method to remove scan-dependent features so that the underlying weather signals can be uncovered.Limb-corrected TBs can be used to monitor large-scale patterns over the globe as well as extreme weather events such as typhoons.Limb-corrected TBs are also more correlated with atmospheric physical variables such as temperature and liquid water path.

Evaluation of Precipitation Datasets from TRMM Satellite and Down-scaled Reanalysis Products with Bias-correction in Middle Qilian Mountain,China

Accurate estimates of precipitation are fundamental for hydrometeorological and ecohydrological studies,but are more difficult in high mountainous areas because of the high elevation and complex terrain.This study compares and evaluates two kinds of precipitation datasets,the reanalysis product downscaled by the Weather Research and Forecasting(WRF)output,and the satellite product,the Tropical Rainfall Measuring Mission(TRMM)Multisatellite Precipitation Analysis(TMPA)product,as well as their bias-corrected datasets in the Middle Qilian Mountain in Northwest China.Results show that the WRF output with finer resolution perfonns well in both estimating precipitation and hydrological simulation,while the TMPA product is unreliable in high mountainous areas.Moreover,bias-corrected WRF output also performs better than bias-corrected TMPA product.Combined with the previous studies,atmospheric reanalysis datasets are more suitable than the satellite products in high mountainous areas.Climate is more important than altitude for the\falseAlarms'events of the TRMM product.Designed to focus on the tropical areas,the TMPA product mistakes certain meteorological situations for precipitation in subhumid and semiarid areas,thus causing significant"falseAlarms"events and leading to significant overestimations and unreliable performance.Simple linear bias correction method,only removing systematical errors,can significantly improves the accuracy of both the WRF output and the TMPA product in arid high mountainous areas with data scarcity.Evaluated by hydrological simulations,the bias-corrected WRF output is more reliable than the gauge dataset.Thus,data merging of the WRF output and gauge observations would provide more reliable precipitation estimations in arid high mountainous areas.

Research on angle-dependent to mographic static correction

Reflection data in CMP has influenced seriously in static calculations,especially in some highly weathered and structurally altered circumstances. Because of static correction in the existed problems and requests,the authors studied the angle dependent tomographic static correction,and discussed its basic theory, including the establishment of forward model,the calculation theory of tomography and tomographic static correction. The usage of theoretical models and practical information on the method has been validated. The results show that using these methods to calculate static correction in a complex area,the quality of static correction is greatly improved.

融合DEM与FY-4A数据的ECMWF预报产品深度学习订正方法

精准的数值天气预报是精细化气象公共服务和商业服务的重要前提。欧洲中期天气预报中心(European Center forMedium Weather Forecasting,ECMWF)预报产品在全球被广泛采用,但始终存在系统预报误差。针对数值天气预报中的误差和多源数据融合中的非线性映射等问题,设计了一个ECMWF数值预报产品的深度学习订正模型(Numerical Forecast CorrectionNetwork,NFC-Net)。NFC-Net引入了FY-4A卫星观测数据、数字高程模型数据(Digital Elevation Model,DEM)和ERA5历史实况数据订正预报结果,利用多源数据空间分辨率对齐模块、时空特征提取模块解决多源异构数据特征的提取与融合问题,并通过UNet网络实现ECMWF预报产品的订正。为了评估所提算法的性能,利用NFC-Net对ECMWF产品中的2 m气温和10 m风速两个天气要素开展订正试验,并将试验结果与ECMWF预报结果、ANO方法订正结果、Convlstm方法订正结果、Fuse-CUnet方法订正结果和ERA5实况进行对比。结果显示,NFC-Net模型订正的2 m气温和10 m风速的均方根误差(Root Mean Squared Error,RMSE)分别较ECMWF预报产品下降49.71%和50.86%。表明NFC-Net模型可以充分利用多源数据有效改善复杂地形条件下的订正结果。NFC-Net模型可用于订正ECMWF预报结果,显著提升数值天气预报的精度。

基于ECOC平衡随机森林的雷达降水粒子分类

针对数据不平衡情况下的降水粒子分类问题,提出了一种基于纠错输出码(error correcting output code,ECOC)平衡随机森林的双偏振气象雷达降水粒子分类方法。首先,将多类别降水粒子数据集编码为多个二分类数据集;然后,对二分类数据集进行有放回的平衡重采样,构建多棵分类回归树;最后,利用所有的分类回归树联合进行降水粒子分类。对实测数据的处理结果表明,所提方法能够在保证总体准确率较高的情况下,大幅提高少数类的分类效果。

基于TLE数据的大气密度反演与校正

用于航天器轨道预报的热层密度模型普遍存在30%左右的误差,影响LEO卫星的精密轨道确定和载荷控制。基于低轨航天器平运动变化与大气密度的关系,使用GRACE(gravity recovery and climate experiment)卫星TLE数据反演2003、2007年沿轨大气密度,通过比较反演值、模型值和实测值的关系分析误差产生原因,使用对数正态分布拟合密度比值。通过分析太阳辐射、地磁指数对大气密度变化的影响,提出一种基于空间环境指数的热层大气密度模型校正与预报方式。使用该方法对2003、2004、2007、2008年的MSIS86模型计算密度进行修正,将模型平均相对误差从33.33%~59.62%降低到11.55%~15.13%,太阳活动低年改进量是高年的1.5~2倍。对2009年经验模型结果进行预报校正,将预报误差降低36.49%,提高了模型精度。

一种融合多要素时空特征的数值预报风场订正模型

目前,数值天气预报产品已广泛应用于气象预报业务中。针对数值预报模式本身不足导致的预报误差,为进一步提高其预报准确率,提出一种融合多要素时空特征的基于U-net和3D注意力机制的风场订正模型。采用该模型对中国气象局研发的国家气象中心GRAPES-3KM模式预报的近地面10 m风场进行偏差订正,以RMSE、MAE作为评价指标,与原数值预报产品、传统订正方法以及U-net、CU-net模型进行比较。实验结果表明,经过所提模型订正后的10 m纬向风RMSE相较原预报数据、LASSO回归、U-net、CU-net模型降低了4.19%~42.67%,MAE降低了6.06%~45.29%;10 m经向风RMSE指标降低了8.55%~41.35%,MAE降低了6.54%~40.82%;10 m全风速RMSE降低了6.14%~29.41%,MAE降低了1.5%~21.08%。所提模型相较对照模型有更好的订正效果,同时未出现订正结果过于平滑的情况。

基于多目标协同训练的风电功率预测提升算法

“双碳”目标下,电力系统加速转型,风电预测技术对构建高比例新能源的新型电力系统具有重要意义。为提升风电功率预测的准确性和鲁棒性,文中提出一种基于多目标协同训练的数值天气预报(numerical weather predicition,NWP)隐式校正算法。首先,分析了NWP校正的必要性和基于NWP显式校正的二步预测法存在的问题;然后,针对二步预测法存在的问题,基于多目标协同训练的优化方式利用神经网络进行NWP隐式校正,以端到端的方式训练模型,同时实现NWP隐式校正和风电功率预测的功能。结合某风电场实测数据开展具体算例分析,证明了所提算法对短期及中长期风电功率预测均有提升作用。此外,该算法仅需1个网络且避免了二次计算,节省了计算存储成本。

DIN 75220标准阳光模拟试验与海南户外暴露试验对比研究

为了对比DIN 75220标准阳光模拟试验与我国海南户外暴露试验,以某型汽车整车和PC、PS标准板为研究对象,对自然曝晒试验和基于DIN 75220标准的整车阳光模拟试验中汽车典型部位微环境、汽车典型部件失效行为以及温度校正太阳辐射量的结果进行了对比研究。分析了2种试验方法中各统计量间的相关关系,进一步结合2种方法下的油漆件色差、内外饰件间缝隙变化结果以及关键部件典型失效问题,提出了适合用于模拟我国海南典型湿热自然环境的定制化阳光模拟试验循环。研究结果表明,基于DIN 75220标准的1个周期整车阳光模拟试验可以反映约80%的湿热自然环境整车曝晒试验3~6个月出现的失效问题。DIN 75220标准中,潮湿气候循环试验可以较好地模拟我国海南湿热自然环境,35个潮湿气候循环可以模拟海南整车自然曝晒1 a。

吐哈盆地台北凹陷桃东沟群烃源岩风化校正及生排烃潜力

吐哈盆地台北凹陷桃东沟群地层钻井较少,而野外露头往往因风化作用导致地化参数失真。为探明桃东沟群烃源岩受风化影响程度和生排烃潜力,基于岩石热解、总有机碳、镜质体反射率等实验数据,应用风化校正法、生烃潜力法对烃源岩进行了风化校正和生排烃潜力计算。结果表明:风化作用对照壁山、坎尔其和恰勒尔剖面的样品影响较大,对塔尔朗沟剖面样品的影响较小;风化作用对生烃潜量的影响最大,其次是w (氯仿沥青“A”)的影响,对w (TOC)的影响最小;桃东沟群烃源岩有机质类型主要为Ⅲ型和Ⅱ_(2)型,部分为Ⅰ型和Ⅱ1型,有机质丰度较高,为中等—很好烃源岩,Ro为0.53%~1.49%,平均为1.03%,处于成熟阶段;台北凹陷桃东沟群烃源岩在R_(o)为0.62%时达到排烃门限,最大生烃潜力指数为615.88 mg/g,生烃中心位于连木沁次洼北部;桃东沟群烃源岩常规油气地质资源量为2.46×10^(8) t,具有良好的勘探前景。研究成果可为吐哈油田中二叠统桃东沟群油气资源的勘探部署提供理论指导。

考虑光伏机理与数据驱动结合的短期功率预测

精确的光伏功率预测对电网的可靠与稳定运行至关重要。现有研究大多数都是将天气条件直接作为数据驱动的输入,未深入分析天气条件与光伏输出功率直接耦合关系,因此将机理模型与数据驱动方法相结合,提出一种新型的光伏功率预测方法。首先,建立光伏系统物理模型,依据建立的物理模型得到不同的辐照度分量以及光伏电池温度。其次,将这些关键的天气特征重新构建数据驱动的输入,实现光伏机理与数据驱动结合的短期功率预测。最后,进行误差修正然后得到最终的光伏功率预测结果。根据光伏系统实测数据集进行仿真分析,结果表明:因为从物理模型得到了关键天气特征,考虑了天气条件与天气因素的耦合关系,预测精度有了明显提升,验证了所提方法的有效性。

相控阵天气雷达强度距离订正方法与试验验证

相控阵天气雷达具有波束灵活、扫描速度快等优势,已成为天气雷达领域的重要发展趋势。但是,相控阵天线的波束宽度和增益随扫描角度变化,并且经常使用宽波束发射、多个窄波束同时接收的工作模式,导致其强度距离订正比传统天气雷达更为复杂。文中在传统天气雷达强度距离订正方法的基础上,分析了发射增益、接收增益和波束宽度随扫描角度和波束形状的变化情况,给出了相控阵天气雷达在不同发射波束形状情况下的订正方法。并利用C波段相控阵天气雷达在江苏的观测试验数据,与南京新一代天气雷达观测数据进行了对比分析,验证了该方法的正确性。

计及多误差场景集划分的超短期NWP风速修正方法

超短期风电功率预测对于机组运行控制和能源调度有着指导性的作用。为了削弱数值天气预报(numerical weather prediction,NWP)风速对超短期预测精度的影响,提出了一种计及多误差场景集划分的超短期NWP风速修正方法。采用双向长短期记忆网络(bidirectional long-short term memory,BILSTM)对NWP风速未来4 h的预报误差进行预测,对风速误差预测值进行误差场景集划分,根据误差场景集训练不同的BILSTM网络进行误差匹配和风速预报误差预测对风速进行修正,再根据修正结果采用多模型进行风电功率超短期预测。将所提方法应用于中国内蒙古某风电场进行算例验证。结果表明,该方法有效降低了NWP风速预报误差,在原有数据基础上,相较于未修正NWP的风速,RMSE值降低了1.859,MAE值降低了1.464,MAPE值降低了26.01%。其中,BP神经网络超短期功率预测精度提高了7.5%,GRU深度网络提高了8.7%,多元线性回归模型提高了9.6%,证明了该方法的有效性。

基于相似卡尔曼滤波的安徽省WRF模式风速预报订正

应用改进的相似卡尔曼滤波方法对2020年4—12月安徽省19个基本气象站WRF模式预报的10 m风速进行误差订正。结果表明:订正后的风速预报平均偏差从1.35 m·s^(-1)降低至0.08 m·s^(-1),基本消除了模式的系统误差;均方根误差从1.77 m·s^(-1)减小至0.81 m·s^(-1)。平均风速为3 m·s^(-1)以上的较大风,风速预报的均方根误差从2.01 m·s^(-1)降低至1.19 m·s^(-1),表明该方法不仅可以有效减小模式的系统误差,还可以大幅减小模式的随机误差。相似卡尔曼滤波可以对无法精确模拟物理过程的数值模式进行误差订正,提高模式在天气系统剧烈变化时的预报准确率,适用于气象要素24~72 h的连续预报。

一种X波段相控阵天气雷达反射率因子订正算法

不同雷达观测相同目标的反射率因子差异会影响雷达组网应用效果,对无波束阻挡的X波段雷达而言,这种差异主要由雷达定标偏差和信号衰减引起,其中衰减包含降水导致的衰减以及天线罩水膜导致的衰减。为了订正定标偏差、天线罩水膜衰减以及降水衰减,设计了一个基于相控阵雷达组网的订正方法。首先,将天线罩水膜导致的衰减看作定标偏差的一部分,使用组网衰减订正算法做初次衰减订正;然后,通过构建一个雷达网观测偏差函数,使用梯度下降法求解各雷达之间的定标偏差;最后,将订正了定标偏差的原始反射率因子再次使用组网衰减订正算法进行二次衰减订正。对上述订正方法,用于广东省佛山市7部X波段相控阵天气雷达探测的一次冷空气降水过程进行检验。结果显示,传统的PIA(Path-integrated attenuation)算法订正结果与广州S波段天气雷达观测结果相关系数为0.53,均方根误差为9.0 dB,而该算法这两项数值分别为0.64和8.4 dB,优于PIA;在台风外围的局地强降水过程中也得到了相同的结论。

强风事件识别及预报订正方法研究

以如东海上风电场升压站激光雷达测风资料为基础,提出了一种强风事件识别方法,设计并比较了三种预报强风事件识别方案。基于决策树和一元线性回归方法,分别开展了针对强风事件的订正方法研究。结果发现:三种预报强风事件识别方案中,等分位阈值方案明显更优,事件命中率达到76.1%,匹配时长命中率达到87.6%;采用消偏阈值方案和等分位阈值方案预报的强风事件时长会更接近观测强风事件时长;等分位阈值方案识别的事件基本可以覆盖到各次观测强风事件的全程;两个订正模型相对于模式预报都有一定提升与改进,其中决策树比一元线性回归模型更优,其平均绝对误差、相对误差和均方根误差明显更小。

联合气象驱动生成和风速修正辅助的风电功率日前预测

随着风电渗透率的不断提高,准确预测风电功率对于构建高比例新能源的新型电力系统具有重要意义。针对该问题,文章提出一种联合气象驱动生成和风速修正辅助的风电功率日前预测方法,通过分析不同类型风电出力的气象特征,以气象特征作为驱动条件设计生成对抗网络,通过修正的数值天气预报(numerical weather prediction,NWP)信息指导风电功率的生成,根据以修正风速计算的参考功率和历史风电出力的连续性从多组生成风电功率中确定最终预测功率。使用实际风电场数据对所提出的方法进行了测试,并与几种典型的预测方法进行对比验证,结果表明该方法具有更高的准确度。