聊古一井数字化氡观测资料的分析与应用

对聊古一井SD－3A数字化氡观测资料的变化特征进行了分析，与该井原有的模拟观测资料进行了对比研究，并对数字化氡观测资料的应用与分析方法进行了探讨。认为尽管SD－3A自动测氡仪所测为逸出氡与模拟观测溶解氡的物理量有所不同，但两具有基本一致的变化趋势，且数字化氡具有较强的映震能力。     

核废物处置场岩体断裂的遥感分析

本文利用ＴＭ和ＳＩＲ－Ａ遥感图像对甘肃北山区４个预选的核废物处置场址的岩体断裂进行了解释和编图。并运用线性构造统计分析方法，对岩体断裂发育程度进行了统计分析。     

白云石的微化反应与有序度

试验证实白云石的微化反应特征明显受有序度的影响，白云石粉末在稀盐酸中的起泡程度随有序度的降低而加剧，在茜素红溶液中的染色程度和速度随有序度的降低而增强。低有序度白云石或完全无序的原白云石，其微化反应与方解石极为相似，反之，高有序度白云石与菱镁矿相似。另外还讨论白云石微化反应的机理。     

NUMERICAL STUDY ON THE INTERANNUAL OSCILLATION OF SEA SURFACE TEMPERATURE IN THE SOUTH CHINA SEA

A two and a half layer oceanic model of wind-driven, thermodynamical general circulation is appliedto study the interannual oscillation of sea surface temperature (SST) in the South China Sea (SCS). Themodel consists of two active layers: the upper mixed layer (UML) and the seasonal thermocline, with themotionless abyss beneath them. The governing equations which include momentum, continuity and sea.temperature for each active layer, can describe the physics of Boussinseq approximation, reduced gravityand equatorial β-plane. The formulas for the heat flux at the surface and at the interface between twoactive layers are designed on the Haney scheme. The entrainment and detrainment at the bottom of theUML induces vertical transport of mass,momentum and heat, and couples of dynamic andthermodynamic effect.Using leap-frog integrating scheme and the Arakawa-C grid the model is forced bya time-dependent wind anomaly stress pattern obtained from category analysis of COADS. The numerical results indicate that t     

Performance of the stereoscopic system of the HEGRA imaging air erenkov telescopes: Monte Carlo simulations and observations

Based on the Monte Carlo simulations we have studied the performance of the HEGRA system of imaging air erenkov telescopes (IACTs) in its present configuration of 4 IACTs as well as in its future final configuration of 5 IACTs. Here we present the results on the basic characteristics of the IACT system which are used in the standard data analysis procedure, i.e., the collection areas, the detection rates, the angular resolution, the energy resolution, and the γ/hadron-separation efficiency. By comparing several key Monte Carlo predictions with experimental results it is possible to check the accuracy of the simulations. The Monte Carlo results concerning hadron-nuclear showers are tested with the recorded cosmic ray events and the results concerning photon-induced showers are tested with a large data sample of γ-rays observed from BL Lac object Mkn 501 during its high flaring activity in 1997. Summarizing the simulations and current observations we give the basic recommendations of using the instrument and the major values of its sensitivity.     

基于Sentinel-1A数据的多种机器学习算法识别冰山的比较

冰山识别对于海洋环境监测和船只安全运行等具有重要的意义,是北极航道开通和北极开发过程中的重要内容。采用合成孔径雷达(SAR)影像进行冰山识别具有独特的优势,多种机器学习算法均可用于SAR影像的冰山识别中。为了最大限度地发挥机器学习算法的性能,有必要对不同机器学习算法及其搭配使用的特征与特征标准化方法进行评估,从而进行最优冰山识别方法的选择。因此,本文基于Sentinel-1A SAR影像,采用多种机器学习方法、多种特征组合及多种特征标准化方法进行冰山识别,并比较各流程方法的识别性能差异。采用的机器学习算法包括贝叶斯分类器(Bayes)、反向神经网络(BPNN)、线性判别分析(LDA)、随机森林(RF)以及支持向量机(SVM);特征标准化方法包括Min-max标准化、Z-score标准化及log函数标准化;数据集是含有12个SAR影像特征的969个冰山与非冰山样本,样本主要位于格陵兰岛东海岸。分类效果采用接收者操作特性(ROC)曲线下的面积(AUC)进行衡量。结果显示,最佳搭配下的RF的AUC值最高,达到了0.945,比最差的Bayes高出0.09。从识别率上来看,RF在冰山查全率为80%的情况下非冰山查全率达到92.6%,效果最好,比第2位的BPNN高出1.4%,比最差的Bayes高出2.6%;BPNN在冰山查全率为90%的情况下非冰山查全率达到87.4%,比第2位的RF高出0.8%,比最差的Bayes高出2.7%。上述结果表明,对冰山识别而言,选择最优的机器学习算法和最佳的特征与特征标准化方法都是十分重要的。     

应用随机辐射传输模型反演云南松林分郁闭度

随机辐射传输模型可用于模拟水平分布不均一森林的辐射传输过程。本文以云南松林分为研究对象，提出一种应用随机辐射传输模型的郁闭度反演方法。该方法以随机辐射传输模型中参数与林分郁闭度的定量关系为基础，提出了针对云南松的冠型等效模型，构建了郁闭度和卫星反射率（GF-1和Landsat 8卫星影像）的查找表，并实施了反演。基于野外实测的30个样地进行了郁闭度数据验证，并和基于NDVI回归模型的反演方法进行对比。结果表明，反演结果能够较准确反映云南松林分郁闭状况（R²=0.8345，RMSE=0.0688），通过冠型修正能够降低反演误差，冠型等效模型是合理的。反演方法机理清晰且适用范围广，研究成果可为大面积森林郁闭度反演提供模型和方法支持。     

A SIMPLE PROGNOSTIC CLOSURE ASSUMPTION TO DEEP CONVECTIVE PARAMETERIZATION:I

In this work,the problem of dependency of the predicted rainfall upon the grid-size in mesoscale numerical weatherprediction models is addressed.We argue that this problem is due to (i) the violation of the quasi-equilibrium assump-tion,which is underlying most existing convective parameterization schemes,and states that the convective activity maybe considered in instantaneous equilibrium with the larger-scale forcing;and (ii) the violation of the hydrostatic approx-imation,made in most mesoscale models,which would induce too large-scale circulation in occurrence of strong con-vection.On the contrary,meso-β and meso-α scale models,i.e.models with horizontal grid size ranging from 10 to 100km,have a capacity to resolve motions with characteristic scales close to the ones of the convective motions.Wehypothesize that a possible way to eliminate this problem is (i) to take a prognostic approach to the parameterization ofdeep convection,whereby the quantities that describe the activity of convection are no longer diagnosed from the instan-taneous value of the large-scale forcing,but predicted by time-dependent equations,that integrate the large-scale forc-ing over time;(ii)to introduce a mesoscale parameter which varies systematically with the grid size of the numericalmodel in order to damp large-scale circulation usually too induced when the grid size becomes smaller (from 100 km to10 kin).We propose an implementation of this idea in the frame of one existing scheme,already tested and used for along time at the French Weather Service.The results of the test through one-dimensional experiments with the Phase Ⅲof GATE data are reported in this paper;and the ones on its implementation in the three-dimensional model with theOSCAR data will be reported in a companion paper.     

裂隙各向异性介质横波双折射特性研究

借助于有限元方法,对裂隙各向异性介质一雏四分量弹性波场中横波双折射的有关问题进行了研究。它们是裂隙方位与相速度,双折射与震源主频,双折射与裂隙丰度,双折射与能量分配,双折射与延迟时,双折射及其振幅谱等。这些问题对深入认识横波双折射特性,对理论和实际工作都有一定帮助。     

A SIMPLE PROGNOSTIC CLOSURE ASSUMPTION TO DEEP CONVECTIVE PARAMETERIZATION:Ⅱ

A series of 3D predictions,dealing with the development of a heavy storm observed during the OSCAR experiment,were carried out by utilizing the PERIDOT model,and introducing alternatively the cumulus parameterization scheme of Bougeault (1985) and the prognostic one (Chen,1989;Chen and Bougeault,1993),with three different grid sizes:160 km,80 km,40 km.The feasibility of the new prognostic scheme and its improvement on the problem of dependency of the predicted rainfall upon the grid size of the numerical model were verified by comparison of the rainfall observed and those predicted.The results demonstrate that,in general,the predicted rainfall increases when the grid size decreases for both diagnostic and prognostic schemes.However,with the new prognostic scheme,the numerical model is capable,on the one hand,for the larger grid sizes,to increase the rainfall,which is under-estimated with the scheme of Bougeault (1985);on the another hand,for the smaller grid sizes,to reduce the rainfall,which is usually over-estimated.In other word,there is an obvious improvement on the problem under study.