SVM方法在武汉区域夏季暴雨预报业务中的应用

以2003～2006年ECMWF数值预报产品、Japan和T213降水预报场资料为基础,应用支持向量机（SVM）方法和CMSVM应用软件平台,通过对训练样本进行交叉验证和模型核参数的逐渐逼近,分区建立了武汉区域16,个区SVM24h暴雨预报模型,并在武汉区域5～7月进行实时业务应用。通过预报结果检验,5～7月武汉区域暴雨预报TS评分为33．59％,其中湖北省平均成绩为34．69％、湖南省平均成绩为34．15％、河南省平均成绩为31．71％。检验结果较好,表明SVM方法在区域性暴雨预报中具有一定的预报能力和参考价值。     

分块DPCM与5/3整数小波变换结合的天文图像无损压缩

针对海量的天文图像数据与有限的存储空间和带宽资源之间矛盾日益突出这一问题,提出一种无损压缩方法,首先将超大天文图像分块,再使用差分脉冲编码调制和5/3整数小波变换,最后使用霍夫曼算法编码。对该方法的原理和具体实现做了详细的分析与介绍,通过实验验证该方法比天文中常用的tar、PKZip、WinZip、WinRar软件在压缩比上分别提高了30%、29%、26%、2%,压缩速度远大于WinZip和WinRar;且该算法实现简单,适合硬件实现和利于并行处理。     

任一力学量■的矩阵元Fmn的微扰近似计算

本文以能量和波函数的各级微扰近似的基态展开式为基础,推导出任一力学量的矩阵元F_(mn)的各级微扰近似表示式。并且指出只要构成能量展式的级数收敛很快,满足能够应用微扰的条件,那么对于任一力学量F的微扰计算,只要计算矩阵元F_(mn)的前二级近似就已足够精确。     

LC元件参数的数字测量方法与设计

本文着重说明如何将LC元件的諸参数转换为相应的数字量,以适应数字測量的要求。本文采用的方法与设计,具有电路结构简单,易于实现的特点。     

Impacts of the Wenchuan Earthquake on the giant panda nature reserves in China

The Wenchuan Earthquake that occurred in May of 2008 caused damages to large areas of Sichuan, Gansu, and Shaanxi provinces in China. Reports from local governments and related management agencies show that the giant panda nature reserves in the earthquake-hit areas were heavily damaged. Our estimates in this paper of the impacts of the earthquake on the giant panda in the earthquake-hit areas were made based on the interpretation of remote sensing images and information collected by field survey. A rapid assessment method was designed to estimate the damages of the earthquake on giant panda habitats. By using visual interpretation methods, we decoded the remote sensing images of the disaster area in the 49 giant panda nature reserves. Research results showed that the Wenchuan Earthquake and the succeeding secondary geological disasters caused great damages to the giant panda nature reserves and disturbed the normal life of the giant pandas there (e.g., landscape fragmentation increased significantly). Undoubtedly, the life of the giant pandas there was affected. However, although the earthquake caused certain impacts on the giant pandas, it did not really threat their survival. Even so, we still strongly advocate for protection of the giant pandas, and have prioritized a couple of measures to be taken to restore the giant panda nature reserves in the earthquake-hit areas.     

Solar-like and Mira-like oscillations of stars - A uniform excitation mechanism

Using a non-local and time-dependent theory of convection, we have cal- culated the linear non-adiabatic oscillations of the radial and low-degree F-p39 modes for evolutionary models from the main sequence to the asymptotic giant branch for stars with solar abundance （X = 0.70, Z = 0.02） in the mass range of 0.6-3.0 3//o. The results show that iow luminosity cool stars tend to be solar-like oscillators, whose low-order modes are stable, but intermediate and high order p-modes are pulsationally unstable; their unstable modes have a wide range in frequency and small values for amplitude growth rates. For stars with increasing luminosity and therefore lower tem- perature, the unstable modes shift towards lower orders, the corresponding range of frequency decreases, and the amplitude growth rate increases. High luminosity red gi- ant stars behave like typical Mira-like oscillators. The effects of the coupling between convection and oscillations on pulsational instability have been carefully analyzed in this work. Our research shows that convection does not simply act as a damping mechanism for oscillations, and the complex nature of the coupling between convec- tion and oscillations makes turbulent convection sometimes behave as damping, and sometimes as excitation. Such a picture can not only naturally account for the red edge of the instability strip, but also the solar-like oscillations in low luminosity red stars and Mira-like ones in high luminosity red giants.     

Are There Pulsationally Unstable Low-degree p1 Modes in the Sun?

By using a non-local time-dependent theory of stellar convection, the solar non-adiabatic pulsations of the low- and intermediate-degree (l < 25) modes are calculated. The results show that the non-radial p1 modes of l = 1–5 are pulsationally unstable. However, the adjacent g, f, p2-p5 modes and the p1 modes of l > 5 are stable. From the analysis of the diagram of integrated work it is discovered that the excitation of oscillations comes from the radiation zone beneath the convective region. Whether the sun possesses unstable low-degree p1 modes is of signi?cant importance for clarifying the excitation mechanism of solar ?ve-minute oscillations.     

参数调整对气象预测的B-P网络模型性能的影响

为了探索ＢＰ网络的参数调整特性,进行了参数α、β的选取对ＢＰ算法的收敛速度和模型的稳定性的影响研究。通过ＢＰ网络用于气象预测建模的参数调整个例分析表明：参数α、β的取值对ＢＰ模型的稳定性无显著影响,但参数值的调整尤其是β值的调整对建模的收敛速度有明显的影响。     

影响流域土壤流失的降水量的PPAR(K)预测模型

介绍了投影寻踪自回归（ＰＰＡＲ（Ｋ））的基本思想及其算法。用ＰＰＡＲ（Ｋ）和均生函数（ＭＧＦ）两种方法对泸州市和北碚市降水量的时间序列进行了建模预测。结果表明ＰＰＡＲ（Ｋ）模型具有较高的稳定性和预测精度。     

四川旱涝震灾害的人工神经网络外推预测

选取旱、涝、震灾害之间的一步转移概率作为Ｂ－Ｐ人工神经网络训练样本的输入信息，建立了四川旱、涝、震的人工神经网络灾型预测模型。Ｂ－Ｐ网络模型应用于实例预测结果与用主分量分析法的趋势预测结果精度接近。