基于SVAR模型的气温变化对南京市工业经济的影响研究

目前气象经济学领域探讨气象因子与经济体之间相互动态影响的研究尚不多见,基于此,采用年平均气温序列及工业产值、GDP、劳动力序列,建立多变量的结构向量自回归模型(SVAR模型),通过脉冲响应函数来考察气温对南京市工业经济的动态影响,并用方差分解法揭示其相互影响程度.结果表明总体上气温升高对南京工业有负面影响,但是这种负面作用是趋缓的,平均每年南京工业产值的3.1%受到气温升高带来的负面影响;同时南京工业经济发展对当地气温升高确实存在促进作用,平均每年南京工业经济发展对本地的气温升高的贡献率有4.4%.研究也说明SVAR模型不失为研究气象因子对经济体影响的可行方法.     

地基GPS遥感观测北京地区水汽变化特征

利用2004—2007年SA34（北京大学）站的GPS观测数据,运用GAMIT软件解算反演了间隔30min的连续变化大气水汽总量（PW）。与北京南郊观测场得到的探空结果作比较,均方根误差（RMSE）在2~3mm之间。通过对大气水汽作月平均,得到每月的大气水汽总量口变化曲线,并初步分析了夏季水汽日变化与地面比湿、降水、地面气温以及地面风矢量的关系。结果表明:北京地区夏季7月大气水汽总量最小值出现在08:00（北京时）左右,8月大气水汽总量最小值出现在08:00到12:00左右（各年表现出一定的差异）,夏季大气水汽总量的最大值出现在01:00到03:00;7月和8月的日变化在夜间变化趋势有所不同;大气水汽总量最大值出现时刻与地面小时降水有一定相关性,且大气水汽总量的日变化明显受风矢量日变化的影响。通过对大气水汽总量的时间序列进行小波分析,得到1年大部分时间里,水汽变化存在大约12d的周期。采用前期的大气水汽总量平均值和短时大气水汽总量增量两个条件进行降水的判断,认为夏季降水的出现时刻与差值的高值区有比较好的对应。     

圣帕台风暴雨的非地转湿Q矢量的诊断分析

利用WRF中尺度模式模拟了台风"圣帕"登陆后减弱成热带低压造成湖南省大暴雨的过程,使用模拟输出的高分辨率资料,借助非地转湿Q矢量对这次暴雨过程做了详细的诊断分析.结果表明:非地转湿Q矢量能比较清楚地揭示此次暴雨演变过程,尤其700 hPa的非地转湿Q矢量散度场对降水预报具有较好的指示意义,其散度辐合区域对应着降水的落区,散度辐合强度变化指示着降水强度的变化趋势,并且非地转湿Q矢量散度辐合强度的大小可预示着未来3～6 h降水的强弱,是具有预报价值的;非地转湿Q矢量散度是非地转ω方程的强迫项,并与地形条件共同作用激发了地面中尺度系统的发展与次级环流的形成,是此次暴雨得以发展与维持的机制.     

DIAGNOSTIC ANALYSIS OF ASYMMETRIC SPIRAL RAINBANDS AROUND THE LANDING OF TYPHOON HAITANG (2005)

By using WRF mesoscale model, this paper carries out a numerical simulation and diagnostic analysis of the structural characteristics of the asymmetric spiral rain bands around the landing of Typhoon Haitang during the period of July 19 to 20, 2005. The result indicated that the two rainbands associated with the precipitation centre was mainly located northeast of the typhoon centre. The movement and intensity of the southern rainband corresponded well with the 850-hPa positive vorticity band from 0200 to 1800 UTC July 19, 2005. Under the effect of cyclonic circulation, the positive vorticity band at 850 hPa connected with a southern rain band, leading to the intensification of rainfall in the southern centre of the precipitation. The southward rainband gradually moved toward and then merges with the northward one, strengthening the rainfall in the northern centre of the precipitation. Besides, the relationship between the heavy rainfall and the divergence field of vertical shear wind in the high altitude is analyzed. Finally, the relationship is revealed between the development of the vertical component of convective vorticity vector and the rainfall near the two centres of precipitation in the low altitude.     

AN ANALYSIS OF THE ASYMMETRICAL STRUCTURE OF TYPHOON AERE'S PRECIPITATION

The structural characteristics of 2004 typhoon Aere’s precipitation are analyzed using the high-resolution data from the Tropical Rainfall Measuring Mission (TRMM) of the National Aeronautics Space Administration (NASA). It is found that the typhoon’s characteristics vary at different stages of its development. To analyze the asymmetric causation of precipitation distribution, data from the National Center for Environmental Prediction (NCEP) reanalysis are used to calculate the vertical integral of the water vapor flux vector. The results show that because of this process, along with the unique phenomenon of twin-typhoon circulation, the easterly air current of the typhoon’s northern side and the southwesterly air current of its southern side play a joint role in transporting water vapor. Furthermore, its transport effects vary greatly at the different stages of development, showing the peculiarity of the water source for this typhoon process. The distributions of the typhoon convection area—characterized by heavy precipitation and a maximum-value area of the water vapor flux, as well as a strong ascending-motion area—differ at different stages of the typhoon’s development. The non-uniform distribution of water vapor flux and the vertical motion bring about asymmetrical distribution of the typhoon precipitation.     

基于ARM与Linux的姿态角测量系统的研制

选取旋转矢量算法作为姿态角测量系统的姿态矩阵更新算法.针对三子样算法采样周期长的特点,提出了重复使用采样数据的方案以保证系统的实时性.选取陶瓷静电陀螺仪CG-L53为角速率传感器,设计了以AT89S51为MCU的信号采集板,搭建了以ARM芯片为CPU的计算机平台,完成了系统各模块的软件设计.仿真测试结果表明,系统能较稳定地输出姿态角信息.     

基于天气预报的参考作物腾发量LS-SVM预测模型

利用最小二乘支持向量机(LS-SVM)方法,建立了基于天气预报的参考作物腾发量(ET₀)的预测模型.对广利灌区1997~2006年逐日气象信息中的天气类型和风速等级进行量化后,以不同天气预报信息作为输入量,建立10种验证方案,对2007年的逐日ET₀进行预测.经验证,方案1~方案7精度均令人满意,其中方案1精度最高.方案1的输入量为气温、天气类型、风速等级3项的预测值,该方案的模型预测值与计算值的统计参数分别为:均方根偏差E_RMS为0.5182 mm,相对偏差E_R为0.1878,决定系数R2为0.864 8,认同系数I_A为0.966 9,回归系数R_C为0.9867;方案7精度亦较好,且以上指标统计参数依次为0.6576 mm、0.2332、0.986 6、0.774 7及0.986 6,该方案输入量只有气温项,实用性很强.     

Nonlinear spinor field in Bianchi type-I Universe filled with viscous fluid: numerical solutions

We consider a system of nonlinear spinor and a Bianchi type I gravitational fields in presence of viscous fluid. The nonlinear term in the spinor field Lagrangian is chosen to be λ F, with λ being a self-coupling constant and F being a function of the invariants I an J constructed from bilinear spinor forms S and P. Self-consistent solutions to the spinor and BI gravitational field equations are obtained in terms of τ, where τ is the volume scale of BI universe. System of equations for τ and ε, where ε is the energy of the viscous fluid, is deduced. This system is solved numerically for some special cases.      

Response of multilayered transversely isotropic medium due to axisymmetric loads

A novel procedure associated with the precise integration method (PIM) and the technique of dual vector is proposed to effectively calculate the magnitude and distribution of deformations in a homogeneous multilayered transversely isotropic medium. The planes of transverse isotropy are assumed to be parallel to the horizontal surface of the soil system. The linearly elastic medium is subjected to four types of vertically acting axisymmetric loads prescribed either at the external surface or in the interior of the soil medium. There are no limits for the thicknesses and number of soil layers to be considered. By virtue of the governing equations of motion and the constitutive equations of the transversely isotropic elastic body, and based on the Hankel integral transform and a dual vector formulation in a cylindrical coordinate system, the partial differential motion equations can be converted into first‐order ordinary differential matrix equations. Applying the approach of PIM, it is convenient to obtain the solutions of ordinary differential matrix equations for the continuously homogeneous multilayered transversely isotropic elastic soil in the transformed domain. The PIM is a highly accurate algorithm to solve the sets of first‐order ordinary differential equations, which can ensure to achieve any desired accuracy of the solutions. What is more, all calculations are based on the standard method with the corresponding algebraic operations. Computational efforts can be reduced to a great extent. Finally, numerical examples are provided to illustrate the accuracy and effectiveness of the proposed approach. Some more cases are analyzed to evaluate the influences of the elastic parameters of the transversely isotropic media on the load‐displacement responses. Copyright © 2015 John Wiley & Sons, Ltd.     

前向神经网络定解及泛化问题研究

作者借用线性代数方程组定解讨论中的相关概念，类比表述了前人工神经网络的定解和泛化问题，同时说明解决泛化问题仅公给定网络VC维是不够的，还需要研究本集规模、向量维度、相关性以及模拟对象的复杂度，文章出相应的算例，一定程度上讨论了前向神经网络的演进特征。