The Forecast Skills and Predictability Sources of Marine Heatwaves in the NUIST-CFS1.0 Hindcasts

Using monthly observations and ensemble hindcasts of the Nanjing University of Information Science and Technology Climate Forecast System(NUIST-CFS1.0) for the period 1983–2020, this study investigates the forecast skill of marine heatwaves(MHWs) over the globe and the predictability sources of the MHWs over the tropical oceans. The MHW forecasts are demonstrated to be skillful on seasonal-annual time scales, particularly in tropical oceans. The forecast skill of the MHWs over the tropical Pacific Ocean(TPO) remains high at lead times of 1–24 months, indicating a forecast better than random chance for up to two years. The forecast skill is subject to the spring predictability barrier of El Nino-Southern Oscillation(ENSO). The forecast skills for the MHWs over the tropical Indian Ocean(TIO), tropical Atlantic Ocean(TAO), and tropical Northwest Pacific(NWP) are lower than that in the TPO. A reliable forecast at lead times of up to two years is shown over the TIO, while a shorter reliable forecast window(less than 17 months) occurs for the TAO and NWP.Additionally, the forecast skills for the TIO, TAO, and NWP are seasonally dependent. Higher skills for the TIO and TAO appear in boreal spring, while a greater skill for the NWP emerges in late summer-early autumn. Further analyses suggest that ENSO serves as a critical source of predictability for MHWs over the TIO and TAO in spring and MHWs over the NWP in summer.

The Predictability Limit of Oceanic Mesoscale Eddy Tracks in the South China Sea

Employing the nonlinear local Lyapunov exponent (NLLE) technique, this study assesses the quantitative predictability limit of oceanic mesoscale eddy (OME) tracks utilizing three eddy datasets for both annual and seasonal means. Our findings reveal a discernible predictability limit of approximately 39 days for cyclonic eddies (CEs) and 44 days for anticyclonic eddies (AEs) within the South China Sea (SCS). The predictability limit is related to the OME properties and seasons. The long-lived, large-amplitude, and large-radius OMEs tend to have a higher predictability limit. The predictability limit of AE (CE) tracks is highest in autumn (winter) with 52 (53) days and lowest in spring (summer) with 40 (30) days. The spatial distribution of the predictability limit of OME tracks also has seasonal variations, further finding that the area of higher predictability limits often overlaps with periodic OMEs. Additionally, the predictability limit of periodic OME tracks is about 49 days for both CEs and AEs, which is 5-10 days higher than the mean values. Usually, in the SCS, OMEs characterized by high predictability limit values exhibit more extended and smoother trajectories and often move along the northern slope of the SCS.

Application of the Conditional Nonlinear Local Lyapunov Exponent to Second-Kind Predictability

In order to quantify the influence of external forcings on the predictability limit using observational data,the author introduced an algorithm of the conditional nonlinear local Lyapunov exponent(CNLLE)method.The effectiveness of this algorithm is validated and compared with the nonlinear local Lyapunov exponent(NLLE)and signal-to-noise ratio methods using a coupled Lorenz model.The results show that the CNLLE method is able to capture the slow error growth constrained by external forcings,therefore,it can quantify the predictability limit induced by the external forcings.On this basis,a preliminary attempt was made to apply this method to measure the influence of ENSO on the predictability limit for both atmospheric and oceanic variable fields.The spatial distribution of the predictability limit induced by ENSO is similar to that arising from the initial conditions calculated by the NLLE method.This similarity supports ENSO as the major predictable signal for weather and climate prediction.In addition,a ratio of predictability limit(RPL)calculated by the CNLLE method to that calculated by the NLLE method was proposed.The RPL larger than 1 indicates that the external forcings can significantly benefit the long-term predictability limit.For instance,ENSO can effectively extend the predictability limit arising from the initial conditions of sea surface temperature over the tropical Indian Ocean by approximately four months,as well as the predictability limit of sea level pressure over the eastern and western Pacific Ocean.Moreover,the impact of ENSO on the geopotential height predictability limit is primarily confined to the troposphere.

Understanding the Low Predictability of the 2015/16 El Niño Event Based on a Deep Learning Model

The 2015/16 El Niño event ranks among the top three of the last 100 years in terms of intensity,but most dynamical models had a relatively low prediction skill for this event before the summer months.Therefore,the attribution of this particular event can help us to understand the cause of super El Niño–Southern Oscillation events and how to forecast them skillfully.The present study applies attribute methods based on a deep learning model to study the key factors related to the formation of this event.A deep learning model is trained using historical simulations from 21 CMIP6 models to predict the Niño-3.4 index.The integrated gradient method is then used to identify the key signals in the North Pacific that determine the evolution of the Niño-3.4 index.These crucial signals are then masked in the initial conditions to verify their roles in the prediction.In addition to confirming the key signals inducing the super El Niño event revealed in previous attribution studies,we identify the combined contribution of the tropical North Atlantic and the South Pacific oceans to the evolution and intensity of this event,emphasizing the crucial role of the interactions among them and the North Pacific.This approach is also applied to other El Niño events,revealing several new precursor signals.This study suggests that the deep learning method is useful in attributing the key factors inducing extreme tropical climate events.

Predictability of the upper ocean heat content in a Community Earth System Model ensemble prediction system

Upper ocean heat content(OHC)has been widely recognized as a crucial precursor to high-impact climate variability,especially for that being indispensable to the long-term memory of the ocean.Assessing the predictability of OHC using state-of-the-art climate models is invaluable for improving and advancing climate forecasts.Recently developed retrospective forecast experiments,based on a Community Earth System Model ensemble prediction system,offer a great opportunity to comprehensively explore OHC predictability.Our results indicate that the skill of actual OHC predictions varies across different oceans and diminishes as the lead time of prediction extends.The spatial distribution of the actual prediction skill closely resembles the corresponding persistence skill,indicating that the persistence of OHC serves as the primary predictive signal for its predictability.The decline in actual prediction skill is more pronounced in the Indian and Atlantic oceans than in the Pacific Ocean,particularly within tropical regions.Additionally,notable seasonal variations in the actual prediction skills across different oceans align well with the phase-locking features of OHC variability.The potential predictability of OHC generally surpasses the actual prediction skill at all lead times,highlighting significant room for improvement in current OHC predictions,especially for the North Indian Ocean and the Atlantic Ocean.Achieving such improvements necessitates a collaborative effort to enhance the quality of ocean observations,develop effective data assimilation methods,and reduce model bias.

Analysis of Predictability of a Large-scale Short-duration Heavy Precipitation Process in Nanchang City

Based on the observation data of automatic stations and sounding data,the circulation characteristics and physical quantities of a large-scale short-duration heavy precipitation process in Nanchang City on July 7,2020 were diagnosed and analyzed,and the ability of several numerical forecasting products to predict this process was tested.The results show that the short-duration heavy precipitation process was triggered in the process of the subtropical high changing from lifting to the north to retreating to the south under the weather background of the confrontation between the northerly flow behind the trough and the strong southwest warm and wet flow on the north side of the subtropical high.The strong southwest warm and wet flow provided abundant water vapor,and the southern pressing of the lower energy front and the invasion of the cold air near the surface layer provided unstable energy and dynamic conditions for the heavy precipitation.The changing trend of the subtropical high from lifting to the north to retreating to the south during 08:00 to 20:00 on July 7 was not predicted by numerical forecast,and there was a large deviation in the forecast of the time and intensity of the southern pressing of the northerly flow behind the trough,so the guidance of numerical forecast for heavy precipitation was not strong,which was not conducive to the prediction of the short-duration heavy precipitation.It was predicted that the subtropical high would move slightly to the south on July 6 compared with the previous day,and the forecast adjustment of the high-level weather system can be used as a sign of the forecast change,which needs to be paid certain attention in the daily forecast.

Recent Advances in China on the Predictability of Weather and Climate

This article summarizes the progress made in predictability studies of weather and climate in recent years in China,with a main focus on advances in methods to study error growth dynamics and reduce uncertainties in the forecasting of weather and climate.Specifically,it covers(a)advances in methods to study weather and climate predictability dynamics,especially those in nonlinear optimal perturbation methods associated with initial errors and model errors and their applications to ensemble forecasting and target observations,(b)new data assimilation algorithms for initialization of predictions and novel assimilation approaches to neutralize the combined effects of initial and model errors for weather and climate,(c)applications of new statistical approaches to climate predictions,and(d)studies on meso-to small-scale weather system predictability dynamics.Some of the major frontiers and challenges remaining in predictability studies are addressed in this context.

Forecast Error and Predictability for the Warm-sector and the Frontal Rainstorm in South China

In south China, warm-sector rainstorms are significantly different from the traditional frontal rainstorms due to complex mechanism, which brings great challenges to their forecast. In this study, based on ensemble forecasting, the high-resolution mesoscale numerical forecast model WRF was used to investigate the effect of initial errors on a warmsector rainstorm and a frontal rainstorm under the same circulation in south China, respectively. We analyzed the sensitivity of forecast errors to the initial errors and their evolution characteristics for the warm-sector and the frontal rainstorm. Additionally, the difference of the predictability was compared via adjusting the initial values of the GOOD member and the BAD member. Compared with the frontal rainstorm, the warm-sector rainstorm was more sensitive to initial error, which increased faster in the warm-sector. Furthermore, the magnitude of error in the warm-sector rainstorm was obviously larger than that of the frontal rainstorm, while the spatial scale of the error was smaller. Similarly, both types of the rainstorm were limited by practical predictability and inherent predictability, while the nonlinear increase characteristics occurred to be more distinct in the warm-sector rainstorm, resulting in the lower inherent predictability.The comparison between the warm-sector rainstorm and the frontal rainstorm revealed that the forecast field was closer to the real situation derived from more accurate initial errors, but only the increase rate in the frontal rainstorm was restrained evidently.

Robust monitoring machine:a machine learning solution for out‑of‑sample R_(2)‑hacking in return predictability monitoring

The out-of-sample R^(2) is designed to measure forecasting performance without look-ahead bias.However,researchers can hack this performance metric even without multiple tests by constructing a prediction model using the intuition derived from empirical properties that appear only in the test sample.Using ensemble machine learning techniques,we create a virtual environment that prevents researchers from peeking into the intuition in advance when performing out-of-sample prediction simulations.We apply this approach to robust monitoring,exploiting a dynamic shrink-age effect by switching between a proposed forecast and a benchmark.Considering stock return forecasting as an example,we show that the resulting robust monitoring forecast improves the average performance of the proposed forecast by 15%(in terms of mean-squared-error)and reduces the variance of its relative performance by 46%while avoiding the out-of-sample R^(2)-hacking problem.Our approach,as a final touch,can further enhance the performance and stability of forecasts from any models and methods.

Bitcoin:Exploring Price Predictability and the Impact of Investor Sentiment

This article addresses the predictability of Bitcoin’s price by examining relationships between Bitcoin and financial and emotional variables such as the Fear and Greed Index(FGI),the American Interest Rate(FED),and the Stock Market Index(NASDAQ).Through the use of statistical techniques such as the Johansen Cointegration Test and Granger Causality,as well as forecasting models,the study reveals that,despite the notorious volatility of the cryptocurrency market,it is possible to identify consistent behavioral patterns that can be successfully used to predict Bitcoin returns.The approach that combines VAR models and neural networks stands out as an effective tool to assist investors and analysts in making informed decisions in an ever-changing market environment.

基于融合注意力机制LSTM网络的地下水位自适应鲁棒预测

地下水水位是旱天污水管网地下水入渗量的重要影响因素,快速精准地预测地下水水位能有效提升旱天污水管网地下水入渗量估算准确度,辅助优化管网病害治理与维护策略。针对目前城市复杂水文预测存在的准确度低、灵敏度低、泛化能力弱等问题,本文提出了一种新的鲁棒自适应水位预测算法。首先,对水文数据进行预处理,解决了数据时间跨度大、噪声多、缺失及异常、非平稳等问题。其次,针对不同输入特征对预测指标的影响,在模型训练阶段提出一种新的空间变量注意机制,可快速识别与水位关联的关键变量,并对输入特征赋予不同的影响权重。然后,针对不同序列长度对预测效果的影响,还设计了自适应时间注意力机制,帮助网络自适应地找出与不同时间序列长度预测指标相关的编码器隐藏状态,以更好地捕捉时间上的依赖关系。在此基础上,以上下文向量作为输入,提出一种融合注意力机制的长短时记忆网络水文预测算法。最后,通过意大利Petrignano水文数据验证了所提算法的有效性,并与GRU、Elman、LSTM、VA–LSTM和S–LSTM等方法进行预测性能比较。结果表明,基于融合注意力机制的LSTM网络在面临大规模、噪点多的复杂数据时有优于其它几种算法的预测效果,表明该算法具有强自适应性和鲁棒性。本文研究结果可以为市政排水策略合理调整、及时控制提供参考。

三维成矿预测关键问题

三维成矿预测是当前深部找矿预测和勘查的重要方法和手段,其方法体系和实践应用均已取得大量成果,但同时存在若干关键科学技术问题,导致其进一步发展受到制约。本文从多尺度三维成矿预测方法体系不完善、不确定性分析与优化研究薄弱、三维成矿预测要素挖掘存在瓶颈、缺少针对三维成矿预测的三维深度学习模型和方法等关键问题出发,对目前三维成矿预测领域相关方面的研究进展进行综合分析,并提出针对上述关键问题可能的解决方案和研究方向。预期未来三维成矿预测领域的研究工作将创新发展出多种方法,实现对三维预测信息的深度挖掘;构建形成适用的三维深度学习模型和训练方法,有效增强三维成矿预测结果的预测能力;通过系统性地开展三维成矿预测不确定性研究,进一步优化预测过程和结果,有效提高三维成矿预测方法的可靠性和准确性;形成面向多尺度三维成矿预测的方法体系,更有效地指导矿集区-矿田-勘查区块(矿床)等不同级别的深部矿产资源找矿勘查工作。相关关键问题的解决将进一步深化和完善三维成矿预测理论和方法体系,促进三维成矿预测理论方法的实践应用,显著提升深部找矿预测和勘查工作的效率与水平,助力深部找矿突破。

Subseasonal predictability of the transition of the stratospheric polar vortex:A case study in winter 1987/88

平流层极涡作为冬季次季节尺度上一个重要的可预测性来源,其强度在1987/88年冬季表现为1979-2019年最显著的转折,即在前(后)冬极端偏弱(强),因此在本文中选取这一个例研究了该年冬季平流层极涡在次季节尺度上的可预测性,结果表明弱极涡和强极涡事件的预测与模式能否准确预测上传行星波的强度紧密相关,同时,发现前期对流层欧亚遥相关波列可能是弱极涡事件发生的关键预兆信号.此外,模式对平流层极涡强度和北大西洋涛动预测误差之间存在显著正相关关系,表明模式减少平流层极涡的预测误差可能可以提高北大西洋涛动及相关对流层气候预测.

干旱形成机制与预测理论方法及其灾害风险特征研究进展与展望

在全球变暖背景下,干旱事件发生的频率和强度不断增大、影响不断加重,干旱发生规律的异常性和机制的复杂性也更为突出,对干旱形成机制、预测理论方法及灾害风险变化规律等方面都提出了新的挑战,也制约了当前干旱预测、预警及其灾害防控能力的提高。近年来,在国家重点基础研究发展计划(973计划)课题等多个国家级项目支持下,已在干旱灾害形成机制与预测理论方法及其风险特征方面取得了一系列新成果。通过动力诊断、数值模拟和田间试验等方法,开展了干旱形成的多因子协同作用和多尺度叠加机制、干旱致灾过程的逐阶递进特征,以及干旱灾害风险分布演化的主控因素等方面的研究。对如下几方面的新进展进行了系统总结归纳:(1)厘清了全球变暖背景下青藏高原热力、海温、夏季风、遥相关等多因子对干旱形成的作用机制。(2)发现了降水亏缺时间尺度和农作物不同生长阶段的干旱敏感性规律。(3)揭示了变暖背景下典型区域干旱灾害风险分布及其变异的新特征;构建了干旱灾害风险新概念模型。(4)研发了东亚季风区的季节和次季节干旱集成预测系统。在总结归纳已取得研究成果的基础上,对未来干旱形成机制及其灾害风险科学研究进行了展望,提出了5个重点研究方向:(1)多因子联动及其多尺度叠加效应对干旱形成的影响;(2)系统整合人类活动和决策以及相关反馈的气候模式研究;(3)揭示陆-气耦合和大气环流协同作用对干旱的影响;(4)认识干旱灾害对粮食安全和生态安全影响的关键过程;(5)提高不同气候情景下干旱预估的准确度。

基于Adaboost-PSO-SVM的铝电解槽健康状态诊断方法研究

针对铝电解槽在铝电解生产过程中故障频发的问题,提出了一种基于支持向量机(support vector machine,SVM)的铝电解槽健康状态诊断模型,考虑传统的支持向量机只能适用于二分类问题,采用自适应推进算法(adaptive boosting,Adaboost)将支持向量机的二分类问题转化为多分类问题用于求解铝电解槽健康状态诊断问题,充分考虑了子模型的权重,强化了模型的适用性。并利用粒子群优化算法(particle swarm optimization,PSO)对其超参数寻优,提高模型的预测精度。实验结果表明,提出的铝电解槽健康状态诊断模型的准确率和Macro-F1分数分别达到94.70%和0.9453,相较于其他传统模型均有显著提升。

基于特征变量扩展的含气饱和度随机森林预测方法

采用数据驱动的方式,提出了一种基于随机森林机器学习算法训练出含气饱和度地震预测方法,并将该方法应用于中国西部复杂天然气藏中,分别对单井资料和二维地震资料进行了含气饱和度预测与分析。研究结果表明:(1)抽取井旁道纵波速度、横波速度和密度3个弹性参数叠前地震反演结果作为基本特征变量样本,引入边界合成少数类过采样技术对基本特征变量样本和对应的含气饱和度样本进行平衡化处理;利用扩展弹性阻抗结合数学变换自动生成一系列的扩展变量;再利用随机森林对特征变量进行含气饱和度预测重要性排名,并优选重要性较高的特征变量进行含气饱和度随机森林训练。(2)该方法大幅减少了特征变量提取和优选的人工工作量,且有效减少了信息冗余以及因含气饱和度样本不平衡导致的训练偏倚问题,有效增强了随机森林算法在含气饱和度地震预测方面的能力。(3)实际单井应用中预测的含气饱和度与测井解释的含气饱和度的相关系数可达0.9855;在二维地震资料应用中,该方法比基于常规未平衡化的11个弹性参数作为随机森林输入预测出的含气饱和度精度更高。

渗出砂岩型铀矿成矿预测与找矿标志

文章基于渗出砂岩型铀矿成矿机理和模式,认为渗出砂岩型铀成矿作用不同于渗入砂岩型铀矿成矿作用,其成矿预测和找矿标志也不同;提出了渗出砂岩型铀矿成矿作用两大基本关键条件识别标志:红杂色含矿建造原生成因和其中控矿灰色砂体的后生成因识别;在提出的红杂色沉积建造中渗出砂岩型铀矿“上红下黑、上下连通、红中找灰、灰中找矿”总体找矿新思路基础上,阐明渗出砂岩型铀成矿区域预测评价条件和标志:深部富铀富有机质沉积岩建造、区域构造、区域建造、放射性异常信息和综合预测标志等,提出并阐述“小凹陷成大矿”条件;系统建立渗出砂岩型铀矿床预测定位标志体系,特别是野外可识别的宏观标志,包括控矿构造、沉积建造、蚀变改造、铀矿化砂岩颜色、外来有机质特征等,并对比了渗入和渗出砂岩型铀成矿预测标志。提出的渗出砂岩型铀矿识别标志体系不仅对区分“渗入”和“渗出”两种矿化成因,而且对厘清控矿要素、指导成矿预测和找矿工程部署具有重要意义和价值。

基于混合分解和PCG-BiLSTM的风速短期预测

为降低风速的随机性对风力发电的影响,提高风速短期预测的精准度,提出一种基于混合分解、双通道输入、多分支PCG-BiLSTM深度学习模型的短期风速预测方法。首先,将全年风速数据分为春、夏、秋、冬4个季度,选取春季作为主要实验对象;其次,利用奇异谱分解(SSD)和变分模态分解(VMD)以降低原始春季风速数据复杂度,生成具有不同模态且复杂度低的子分量,两种不同模式子分量组合为混合分量,实现不同模式分解算法的优势互补;最后,将混合分量以双通道的形式输入到多分支PCG-BiLSTM深度学习模型中,其模型的每个分支由卷积神经网络(CNN)与门控循环单元(GRU)并联组成时空特征提取模块,用于提取两种分解分量组合的混合分量的时空特征,各分支提取对应混合分量的时空特征经聚合后再由双向长短期记忆网络(BiLSTM)进一步提取风速信号的正向和反向双向波动规律,进而得到最终的风速预测结果。多组实验结果表明:提出的组合预测方法在短期风速预测中具有较高的精度和泛化能力,优于其他传统预测方法。

基于雷达回波进行降水场预测的无监督学习模型训练策略

为了提高降水场预测模型的学习效率与预测性能,在预测模型的训练阶段提出一个改善的训练策略,使其可以充分学习物体运动轨迹以及物体运动时的外观变化。通过在一个雷达回波数据集和一个公开数据集上进行对应实验,可以显示出该方法在两项指标的性能表现上具有明显提高,证明了该方法的有效性。

基于地质三维建模的矿床蚀变带深部资源预测方法

随着矿产资源需求的持续增加,加之浅表矿产资源的逐渐匮乏,如何有效开发深部矿产资源成为国家重点研究课题之一。以往二维矿床信息无法确定精准的深部资源储量,阻碍矿产资源的开发进度,本文提出基于地质三维建模的矿床蚀变带深部资源预测方法研究。应用地质三维建模软件(GOCAD)与计算机技术,构建矿床蚀变带深部地质三维模型,以此为基础,确定最佳单元尺度,划分矿床蚀变带深部地质统计单元,提取矿床蚀变带深部地质变量,计算地质变量权重数值,探究深部地质统计单元之间的联系度,基于三维证据权法预测矿床蚀变带深部资源。实验数据显示:应用该方法获得的矿床蚀变带深部资源预测误差最小值为1.01%,充分证实了该方法的深部资源预测精度更高。