地磁变化场的MEEMD-样本熵-LSSVM预测模型

目的 针对地磁变化场时间序列的混沌特性,提出了一种改进的集成经验模态分解(modified ensemble em-pirical mode decomposition,MEEMD)-样本熵-最小二乘支持向量机(least square support vector machine,LSSVM)的地磁变化场预测模型。首先,利用MEEMD-样本熵将非平稳的地磁变化场时间序列分解为一系列复杂度差异明显的地磁变化场子序列;然后,针对每一个子序列分别建立LSSVM模型,选择各自适合的最优模型参数;最后,以地磁台站实测的地磁变化场数据为例进行实验,并与基于单一LSSVM以及RBF径向基神经网络的两种预测模型进行比较。实验结果表明,MEEMD-样本熵-LSSVM模型的预测值能紧跟地磁变化场的变化趋势,相比另外两种模型,体现出更好的预测效果,在地磁Kp指数小于3时,预测3h平均绝对误差为1.63nT。     

基于MVO优化神经网络的GNSS高程异常拟合方法

针对普通神经网络的梯度消失和易陷入局部极值的问题,提出一种基于多元宇宙优化算法(multi-verse optimizer, MVO)的BP神经网络优化方法(MVO-BP),利用MVO全局寻优的特性求取BP神经网络各层之间可靠的神经元阈值与连接权,从而使神经网络预测模型具备更高的预测精度。建立基于MVO-BP算法的GNSS高程异常拟合预测模型,并采用实际工程中少量高程异常数据进行算法可行性检验。结果表明,相较于常规的BP神经网络法及多面函数法,MVO-BP法精度更高、适用性更强,可为实际工程测量中正常高的求取提供参考。     

成都平原粘性土动力学参数统计分析

成都平原地震构造环境主要受近场中强地震和外围大地震的影响。成都平原内的第四系分布广泛,主要为河流相的砂卵石层夹粘土、粉土层。这种地层在土层地震反应计算时往往会产生一个峰值,具有显著的放大作用。本文共收集了107组土动力学参数的实验资料,统计分析了粉质粘土和粘土两种粘性土的实测土动力学参数,给出了它们在不同深度的动剪切模量比和阻尼比的统计值。然后,选取1个典型钻孔,建立了土层地震反应分析模型,分别运用本文"统计值"、"十五结果"、"规范值"、"推荐值"进行地震土层反应计算,从反应谱形状、地表峰值加速度和反应谱特征周期等方面,验证了本文"统计值"的适用性和针对性,结果表明在盆地内使用"规范值"和"推荐值"时应谨慎,不然可能会对工程的抗震设防产生不利的影响。本文的"统计值"比"规范值"和"推荐值"更适用于成都平原地区,对各类工程建设的场地地震安全性评价具有一定的借鉴和参考价值。     

水分胁迫对绿洲影响的数值模拟

运用耦合了包含土壤-植被-水文参数化陆面过程的非静力平衡中尺度大气模式MM5,将水平分辨率提高到1km,通过数值模拟的方式,分别从地气间水热交换、对绿洲小气候影响、土壤湿度变化以及径流量的角度,研究了2002年7月下旬我国西北干旱区绿洲环境在不同水量滴灌条件下的变化,从理论上找到了较适宜绿洲小气候维持的滴灌水量。结果表明：1、绿洲下垫面为农田时,从土壤保湿和绿洲小气候维持角度考虑,500m^3／hm^2／10天的滴灌水量对7月下旬黑河流域中游绿洲是一个比较合适的灌溉量。2、500m^3/hm^2／10天的灌溉水量仅为同期大水漫灌定额下限的一半,上限的1／3,间接证明了滴灌是一种高效节水的灌溉方式,可以为干旱区节约大量水资源。     

喀喇昆仑山厦呈、吕莫慕士塔格山段现代冰川分布及变化特征

在全球气候普遍变暖、冰川大面积退缩的大背景下,为揭示山岳冰川十分发育的喀喇昆仑山脉现代冰川分布及变化特征,以中等空间分辨率(ETM和Landsat 8)卫星数据为主要数据源,采用人机交互式解译方法完成了厦呈慕士塔格山、吕莫慕士塔格山段冰川变化遥感调查,基本查明了区内冰川面积、分布特征和变化状况,共调查1 373处冰川,其中中国境内有554处,境外克什米尔印控区有819处; 同时对区内典型冰川变化进行了2期对比分析,得出冰舌退缩、前进或稳定现象同时存在的结论,为喀喇昆仑山脉冰川变化研究提供了参考依据,为我国西部边海防地区水资源的开发利用和冰雪灾害的防治等提供了科学依据。     

基于复contourlet域隐马尔科夫树模型的海面溢油合成孔径雷达图像相干斑抑制

海面溢油SAR图像中的相干斑噪声严重影响了后续的图像分割、特征提取和分类.为了更有效地抑制海面溢油SAR图像相干斑,文中提出了一种基于复contourlet域隐马尔科夫树模型的海面溢油SAR图像相干斑抑制方法.首先对观测图像取对数并进行复contourlet变换;然后在复contourlet域中用隐马尔科夫树模型对相邻尺度间的带通方向子带系数进行建模,并依据贝叶斯最小均方误差准则估计无噪系数;最后进行逆复contourlet变换和指数变换,得到相干斑抑制后的图像.大量实验结果表明,与Lee、Kuan、Frost及Gamma Map等4种经典滤波方法以及小波域和contourlet域隐马尔科夫树模型方法相比,文中方法从主观视觉和客观定量评价两方面来看综合性能更为优越,是一种行之有效的SAR遥感图像海面溢油检测的预处理方法.     

Adaptive evolution of the rural human–environment system in farming and pastoral areas of northern China from 1952–2017

The theory on the cyclic adaptation between society and ecosystems sheds new light on the evolution and internal structure of human–environment systems. This paper introduces the risk index(RI) and adaptation capacity index(ACI) to evaluate the rural human–environment system. An evaluation index system for the adaptability of rural human–environment systems is configured in the context of climate change and policy implementation. On this basis, the stages, features, dominant control factors, and evolution mechanism were examined vis-à-vis the adaptability of the rural human–environment system in Darhan Muminggan Joint Banner from 1952 to 2017. The main results are as follows:(1) The evolution of the rural human–environment system can be divided into three stages, namely, the reorganization and rapid development stage(1952–2002) with population, cultivated land, livestock and degraded grassland increasing by 260%, 13%, 134% and 16.33%, respectively. The rapid to stable development stage(2003–2010) with population increasing by 2.8%; cultivated land, livestock and degraded grassland decreasing by 2.3%, 13.6% and 10.7%, respectively. The stable to release stage(2011–2017) with population, cultivated land, livestock and degraded grassland decreasing by 2.6%, 0.2%, 10.6% and 3.8%, respectively.(2) With the passage of time, the ACI of the rural human–environment system first increased slightly(–0.016–0.031), followed by a slight decline(0.031–0.003), and culminating in a rapid increase(0.003–0.088). In terms of spatial patterns, adaptability is high in the middle, moderate in the north, and low in the south.(3) The evolution of adaptability in the rural human–environment system was mainly controlled by the per capita effective irrigation area(22.31%) and the per capita number of livestock(23.47%) from 1990 to 2000, the desertified area(25.06%) and the land use intensity(21.27%) from 2000 to 2005, and the per capita income of farmers and herdsmen(20.08%) and the per capita number of livestock(18.52%) from 2010 to 2007.(4) Under the effects of climate change and policy interventions, the cyclic adaptation of the rural human–environment system was propelled by the interactions between two kinds of subjects: farmers and herdsmen on the one hand and rural communities on the other hand. The interaction affects the adaptive behavior of the two kinds of subjects, which in turn drives the cyclic evolution of the system. As a result, the system structure and functions developed alternatively between coordinated and uncoordinated states. Small-scale adaptive behaviors of farmers and herdsmen have a profound impact on the evolution of the rural human–environment system.     

Review of Rapid Transformation of Floodplain Wetlands in Northeast China: Roles of Human Development and Global Environmental Change

Northeast China is the region with the largest area of wetlands in China. The Sanjiang Plain and the Songnen Plain are large freshwater marsh distribution regions that are affected by climate warming and by the increasing frequency and density of extreme weather and are the regions most subject to disturbances by human activities in Northeast China. The wetlands of the Sanjiang Plain and the Songnen Plain have shrunk severely in the past 60 years, and wetland functions have been reduced substantially because of climate change, unreasonable land use, fire episodes, engineering and construction works and urbanization. Large-scale agricultural development started in the 1950s has been the most important driving factor for wetland loss and degradation in the Sanjiang Plain. Water shortage has been the most important factor for degradation and fragmentation of wetlands in the Songnen Plain. To mitigate wetland degradation and better protect wetlands, special regulations, long-term mechanisms and technical support of wetland protection should be established. A wetland compensation program should be implemented, and technologies for increasing the adaptive capacity of wetlands should be developed. Moreover, it is most important to find the balanced threshold between agricultural development and wetland protection.