1958～2018年永定河流域蒸发皿蒸发量的变化特征及其影响因子分析

基于永定河流域1958～2018年14个气象站的逐日观测数据,采用气候倾向率、Mann–Kendall趋势检验法分析蒸发皿蒸发量时空变化特征,并通过完全相关系数和多元回归分析识别气候因子与蒸发量的相关程度并定量计算其贡献率。结果表明:在全球气候变暖的背景下,60年来永定河流域气温以0.29°C/10 a的速率上升,而蒸发皿蒸发量不增反减,以－48.88 mm/10 a的速率显著下降（标准化统计量Z=－4.5）,该流域存在明显的“蒸发悖论”现象。流域蒸发量表现出显著的时空分布差异性,在季节上,春、夏季蒸发量分别占全年蒸发量的35%和37%,且春、夏两季蒸发量下降趋势较为显著;在空间上,永定河流域下游平原区下降趋势较上游山区（天镇、蔚县等地区）更为显著。完全相关分析表明,净辐射、平均风速和空气饱和差与蒸发量具有较强的相关性;贡献率分析表明,与基准期 1958～1979 年相比,1980～2018 年平均风速和净辐射减少对蒸发量减少的贡献率分别为77%和66%,空气饱和差的贡献率为－41%,净辐射和平均风速的减少是导致蒸发皿蒸发量下降的主要影响因素。     

全球云水量气候分布及变化趋势特征分析

采用20世纪再分析版本2c数据集的云水量逐月再分析数据,通过数理统计方法,分析了1960～2014年全球、海洋和陆地上空云水量的分布和变化特征及其与水汽通量的关系。结果表明:1）全球云水量空间分布不均,海洋高于陆地且比例约为4﹕3,中低纬海洋、陆地上空云水量变化趋势分别为0.07 g m?2 (10 a)?1和?0.04 g m?2 (10 a)?1,季节性差异主要体现于夏季在热带辐合带和南半球海洋高,冬季在北半球海洋和南半球陆地高。2）对比六大洲发现,云水量最高的南美洲有最快增加趋势,为0.46 g m?2 (10 a)?1,同时云水量最低的非洲有最快降低趋势,为?0.59 g m?2 (10 a)?1。3）中低层整层水汽通量散度场的辐合、辐散区和云水量的高、低值区相对应,云水量与水汽通量散度变化呈负相关（相关系数为?0.44）,负相关关系在赤道附近的低纬地区显著。本文揭示了在全球变暖背景下,大气中云水量分布和变化的时空格局,为模式参数化和未来气候预估提供参考。     

Ka/Ku双波段云雷达反演空气垂直运动速度和雨滴谱方法研究及初步应用

在Ka波段云雷达上升级改造建成的Ka/Ku（Ka和Ku波段波长分别为8.9 mm和2.2 cm）双波段云雷达2019年用于华南云降水垂直结构观测,以改进云内动力和微物理参数探测能力。为了利用该双波段云雷达研究华南降水微物理和动力结构,本文提出了基于双波段云雷达回波强度谱密度（S_Z）数据和最优估计技术的云内空气垂直运动速度（V_air）、雨滴谱（DSD）、含水量（LWC）、雨强（R）的反演方法（DWSZ）,雨区衰减的订正方法。利用2019年在广东龙门观测的一次降水过程数据,对比分析了云雷达反演的微物理参数与雨滴谱直接观测量,并检验了云雷达反演的低层空气垂直运动速度,利用反演结果分析了一次混合云过程的V_air与这些微物理参数的垂直结构和相互关系。结果表明:Ka/Ku双波段云雷达合理反演了微降水微物理和动力参数及其垂直分布,经过衰减订正的Ka和Ku波段回波强度偏差明显减小。该双波段云雷达数据可以用于分析0～30 dBZ回波强度的云降水垂直结构。本次过程为混合云降水,对流单体前部存在明显的上升气流,后部存在下沉气流;从平均垂直结构来看:V_air和粒子平均直径（D_m）在2 km高度层到达最大,粒子数密度（N_w）、LWC和R在2 km以下明显增强,粒子直径却减小,水汽凝结过程、雨滴碰并云滴是本次过程的主要机制。这一工作验证了Ka和Ku波段组合的双波段云雷达的可行性,为Ka/Ku波段云雷达技术的推广,单波段云雷达反演算法进一步改进,云降水精细结构分析等提供了基础。     

微波辐射计在对流天气预报中的应用

选取2016—2018年每年4—9月份RPG-HATPRO型42通道微波辐射计观测的不稳定指数参数(K、SI、CAPE、LI)及水汽参数(IWV、LWP),研究得出各参数触发雷雨大风、短时强降水的阈值条件为K>37℃、SI<-1℃、IWV>60 kg/m~2、LWP>400 g/m~2,而LI、CAPE无法对3种天气类型进行区分。利用费舍判别分析方法,将不稳定指数参数及水汽参数作为预报因子,建立预报方程并进行检验,结果表明:二级判别方程预测对流天气的准确率为76%,可以作为预报对流天气的辅助工具;多级判别方程不能很好地区分3种天气类型,但将其作为修正后的二级判别方程使用,能提高对流天气的测中概率。     

中国区域平均降水量序列构建方法比较研究

由于地面观测台站空间分布不均匀,运用不同区域平均技术方法研究中国降水特征和变化规律的结果存在显著差异,是区域降水变化研究不确定性的重要来源之一。本文以“中国地面与CMORPH（CPC Morphing Technique）融合逐日降水产品”作为参照值,基于中国地面2425 站观测资料,采用5种网格尺寸的经纬度网格面积加权平均方法、省面积加权平均方法、直接平均方法等计算中国区域平均降水量时间序列,比较所得序列统计属性与参照值的偏差,判别方法的优劣。分析表明,1998～2012年,2.5°网格和5.0°网格区域平均方案所得年降水量序列的变化速率和离散程度同参照值最为接近,分别是最优和次优的区域平均方案。省面积加权平均方法对多年平均年降水量的计算准确,但对年降水量变化趋势的估计效果稍差,距平百分率序列很不准确。网格过疏或过密都会使区域平均结果出现较大误差,直接平均方法的误差亦偏大,可靠性较低。     

一种基于属性值变化程度定权的聚类算法

针对经典K-means聚类算法以欧氏距离作为相似度判断法则进行聚类划分,而未考虑聚类对象的各属性值对聚类划分的影响程度存在差异的问题,该文提出了一种基于属性值变化程度定权的聚类算法。通过采用Iris dataset数据进行实验,该算法相对于其他聚类算法获得了更好的聚类效果,且该算法适用于生物物种分类、遥感影像识别等工作领域,能提高聚类运算的精准度。     

GPS高层建筑物动态特性的小波分析提取

针对高层建筑物GPS动态监测数据特点,该文将小波分析技术应用于高层建筑物动态特性提取,研究了数据处理时4种阈值选取准则的好坏,针对观测数据不同阶段振幅和频率不同,提出分段、分层处理数据的思想,为了获得更全面的观测信号,将小波包分析应用于数据处理。结果表明:自适应阈值准则更适用于高层建筑物动态数据处理,采用分段分层的处理方法,既能满足去噪要求又能提取丰富的变形信息,小波包分析能有效提取高层建筑物动态特性,保留更多的动态信息。     

Network-based triple-frequency carrier phase ambiguity resolution between reference stations using BDS data for long baselines

Network-based ambiguity resolution (AR) between reference stations is the prerequisite to realize a precise real-time kinematic positioning service. With the help of BDS triple-frequency signals, we can efficiently deal with the ionospheric delay and tropospheric delay, and achieve rapid and reliable AR. To overcome the inaccurate ionospheric delay estimated by the geometry-free three carrier ambiguity resolution (GF TCAR) technique, which leads to failure in the original ambiguity resolution, we propose an ionospheric-free (IF) TCAR method to resolve the ambiguity between the reference stations over long baselines. Taking full advantage of the known positions of the reference stations, the easily resolved extra-wide-lane (EWL) ambiguity, and the IF phase combinations, we can reliably fix the wide-lane (WL) ambiguity. A Kalman filter is applied to estimate precise IF ambiguities and the original ambiguity is resolved with the fixed WL ambiguity. A numerical analysis with triple-frequency BDS data from three long baselines of a CORS network is provided to compare the AR performance of GF TCAR with that of IF TCAR. The results show that both methods can reliably resolve the WL ambiguity with a remarkable correctly-fixed rate of higher than 99%, and the reliably-fixed rates of the IF TCAR slightly increase from 92.19, 94.67 and 94.61–98.26, 99.54 and 97.51% for the three baselines. Herein “correctly-fixed” and “reliably-fixed” mean the difference between the float ambiguity and the true one are less than ± 0.5 and ± 0.25 cycles, respectively. On the other hand, the AR performance of the original signals with the IF TCAR method is much better than that with the GF TCAR method attaining a 100% correctly-fixed rate, while the GF TCAR method can hardly fix the original ambiguity with the largest bias being as much as 4 cycles because of the amplified systematic bias.     

A new multipath mitigation method based on adaptive thresholding wavelet denoising and double reference shift strategy

Multipath disturbance is one of the major error sources in high-accuracy positioning for global navigation satellite system (GNSS). Although various methods based on software and hardware strategies have been developed to mitigate this error, they are still limited by different kinds of factors and the effect is not ideal. After analyzing the existing methods, a new single-difference sidereal filtering method, based on adaptive thresholding wavelet denoising and double reference shift strategy (ATDR), is proposed to mitigate multipath effects for static short-baseline GNSS applications. The key idea of the proposed method is the use of both the adaptive thresholding wavelet denoising to extract an accurate multipath correction model from the reference Day and the double reference shift strategy to mitigate multipath for subsequent Day 2 more accurately and efficiently. By applying the introduced adaptive thresholding method, the average improvement rate of the RMS values of the single-difference residuals can reach about 15.82% compared with the constant thresholding method. Moreover, after applying the proposed ATDR method, the 3D positioning precision is improved by about 37.73% for the single epoch mode with 30 s data sampling rate and about 31.22% for the continuous mode with 1 s high sampling rate compared with the original results. Even compared with the constant thresholding single orbital reference (CTSR) method, the improvement percentage is about 33.94% in single epoch mode and about 25.40% in continuous mode for 3D positioning precision, respectively. In conclusion, the results of the two experiments indicate that the proposed ATDR method performs much better than the CTSR method in mitigating multipath for different sampling rates and different processing modes in the measurement domain for GNSS static short-baseline postprocessing applications.     

Modelling coastal land use change by incorporating spatial autocorrelation into cellular automata models

This paper presents a spatial autoregressive (SAR) method-based cellular automata (termed SAR-CA) model to simulate coastal land use change, by incorporating spatial autocorrelation into transition rules. The model captures the spatial relationships between explained and explanatory variables and then integrates them into CA transition rules. A conventional CA model (LogCA) based on logistic regression (LR) was studied as a comparison. These two CA models were applied to simulate urban land use change of coastal regions in Ningbo of China from 2000 to 2015. Compared to the LR method, the SAR model yielded smaller accumulated residuals that showed a random distribution in fitting the CA transition rules. The better-fitting SAR model performed well in simulating urban land use change and scored an overall accuracy of 85.3%, improving on the LogCA model by 3.6%. Landscape metrics showed that the pattern generated by the SAR-CA model has less difference with the observed pattern.