基于灰色关联分析与RBF神经网络的水面蒸发量预测

利用灰色关联度法分析了影响水面蒸发量的主要气象因素,并通过RBF神经网络建立了水面蒸发量的经验预测模型.计算结果表明,所建立的模型简单易行,精度较高,可以应用于生产实践中.     

基于随机森林模型的农业电子商务浅析——洛川苹果网络销售问题与对策

随着互联网产业的飞速发展,电子商务开始进入农业领域。以电子商务起步较早的"洛川苹果"作为研究对象进行调研。基于随机森林模型的决策树集成算法,对农业网络销售体系整体进行数据分析,模型构建,从问题表象出发挖掘其在不同部分的影响因子,最终基于影响因素解决问题,提出合理化建议:加强农村基础设施建设、健全农村公共服务体系以及完善农村电子商务培训制度等,因地制宜,推进农业电子商务的健康发展。     

基于BP神经网络与数学形态学的彩色地质图面要素信息智能提取

地质图是一个区域地质研究的重要成果,也是前人留下的宝贵资料,更融合了地质专家的丰富知识。本研究的目的是通过新的思路将彩色地质图信息提取出来,使其结果能直接进行数据分析,并用于决策和分析。以机器学习为指导,在分析半结构化（栅格）地质图特征的基础上,根据图例信息,提出一种彩色地质图信息提取新思路,对彩色地质图进行分层信息提取,并结合数学形态学和多层前向反馈式神经网络,探索半结构化数据转换为结构数据的有效技术方法。利用图像信息提取技术将半结构化地质图转化为结构化数据,可用于成矿预测等研究。这一变化将改变传统地质数据的结构,地质研究的信息基础和来源将会增加,对于获取更多的数据源和信息源,进一步开展地质分析研究具有重要意义。     

新生代构造运动对黄土高原地貌分异与黄河形成的控制作用

黄土高原面积广大,不同区域的新生代构造演化有很大差别,根据新生代构造运动特征与沉积过程的差异,将黄土高原划分为鄂尔多斯稳定地块、陇西地区和汾渭新生代裂谷3个新生代构造运动分区。新生代构造运动控制黄土高原新生代沉积与侵蚀过程,从而产生明显地貌分异,并可划为3个一级地貌分区: 1)鄂尔多斯黄土地貌区。第四纪黄土覆盖在新近纪红土梁、峁及红土盆地等古地形之上,后被侵蚀,形成复杂多变的黄土地貌景观; 2)陇西黄土地貌区。主体为新近纪红土地貌,上覆少许第四纪黄土,实际为红土地貌区; 3)汾渭裂谷黄土地貌区。裂谷盆地的地貌主体为河流两侧的阶地和台地,其上堆积第四纪黄土,形成不同时期的黄土台塬。鄂尔多斯周边新生代裂谷的出现控制了黄河的发育。新近纪末银呼裂谷与汾渭裂谷进入快速发展期,吕梁山相应隆起; 鄂尔多斯地块西北部也轻微抬升,使古黄河初步形成并流经晋陕峡谷一带,进入三门湖并东流入海。第四纪时期早期,黄河曾长期处于河湖并存期,中更新世末,黄河溯源侵蚀使湖泊消失,并呈现目前的黄河地貌景观。     

The Synergism between Methanogens and Methanotrophs and the Nature of their Contributions to the Seasonal Variation of Methane Fluxes in a Wetland: The Case of Dajiuhu Subalpine Peatland

Wetland ecosystems are the most important natural methane (CH₄) sources, whose fluxes periodically fluctuate. Methanogens (methane producers) and methanotrophs (methane consumers) are considered key factors affecting CH₄ fluxes in wetlands. However, the symbiotic relationship between methanogens and methanotrophs remains unclear. To help close this research gap, we collected and analyzed samples from four soil depths in the Dajiuhu subalpine peatland in January, April, July, and October 2019 and acquired seasonal methane flux data from an eddy covariance (EC) system, and investigated relationships. A phylogenetic molecular ecological networks (pMENs) analysis was used to identify keystone species and the seasonal variations of the co-occurrence patterns of methanogenic and methanotrophic communities. The results indicate that the seasonal variations of the interactions between methanogenic and methanotrophic communities contributed to CH₄ emissions in wetlands. The keystone species discerned by the network analysis also showed their importance in mediating CH₄ fluxes. Methane (CH₄) emissions in wetlands were lowest in spring; during this period, the most complex interactions between microbes were observed, with intense competition among methanogens while methanotrophs demonstrated better cooperation. Reverse patterns manifested themselves in summer when the highest CH₄ flux was observed. Methanoregula formicica was negatively correlated with CH₄ fluxes and occupied the largest ecological niches in the spring network. In contrast, both Methanocella arvoryzae and Methylocystaceae demonstrated positive correlations with CH₄ fluxes and were better adapted to the microbial community in the summer. In addition, soil temperature and nitrogen were regarded as significant environmental factors to CH₄ fluxes. This study was successful in explaining the seasonal patterns and microbial driving mechanisms of CH₄ emissions in wetlands.     

The Impact of the Numbers of Monitoring Stations on the National and Regional Air Quality Assessment in China During 2013–18

China national air quality monitoring network has become the core data source for air quality assessment and management in China. However, during network construction, the significant change in numbers of monitoring sites with time is easily ignored, which brings uncertainty to air quality assessments. This study aims to analyze the impact of change in numbers of stations on national and regional air quality assessments in China during 2013–18. The results indicate that the change in numbers of stations has different impacts on fine particulate matter (PM_2.5) and ozone concentration assessments. The increasing number of sites makes the estimated national and regional PM_2.5 concentration slightly lower by 0.6?2.2 μg m^?3 and 1.4?6.0 μg m^?3 respectively from 2013 to 2018. The main reason is that over time, the monitoring network expands from the urban centers to the suburban areas with low population densities and pollutant emissions. For ozone, the increasing number of stations affects the long-term trends of the estimated concentration, especially the national trends, which changed from a slight upward trend to a downward trend in 2014?15. Besides, the impact of the increasing number of sites on ozone assessment exhibits a seasonal difference at the 0.05 significance level in that the added sites make the estimated concentration higher in winter and lower in summer. These results suggest that the change in numbers of monitoring sites is an important uncertainty factor in national and regional air quality assessments, that needs to be considered in long-term concentration assessment, trend analysis, and trend driving force analysis.     

基于广义回归神经网络的砂土液化综合判别方法

基于厦门地区大量钻孔试验数据,分别采用规范法和Seed法对该区饱和砂土进行液化判别。然后选取二者判别结果相同的数据作为训练和测试样本,运用广义回归神经网络,对二者判别结果分歧的钻孔数据进行二次判别。结果表明:广义回归神经网络性能良好,预测准确度高。此外,这种综合判别方法也提高了饱和砂土液化判别的准确度,并为其他地区饱和砂土的液化判别研究提供借鉴和参考。     

青藏高原困难地区开展区域重力调查工作遇到的问题与对策

青藏高原的大部分地区自然条件恶劣、人口稀少、交通不便、基础测绘程度低、通讯条件极其落后,属于典型的开展区域重力调查工作的困难地区.在这些地区开展区域重力调查工作,在重力仪选择及标定、GPS控制点的建立及其控制范围、重力测点水准高程求取,以及近区地形改正、物性标本采集等方面都会遇到很多与内地一般地区工作时不一样的问题.笔者通过对青藏高原地区开展的多个1∶20万区域重力调查项目的野外工作过程进行总结,分析了困难地区区域重力调查项目实施过程中遇到的一些技术问题,认为可通过尽量选择高精度、稳定性好的重力仪,合理延长闭合时间,采用CQG2000似大地水准面模型求取测点水准高程,扩大GPS基准站的控制半径等办法解决.这些措施可作为以后在青藏高原地区及国内其他困难地区开展区域重力调查工作时的参考.     

多目标区域地球化学编图数据整理方法—以海河流域编图为例

多目标区域地球化学编图是科学表达区域土壤地球化学分布和特征的图示方法,大区域的编图一般可能涉及到对不同的工作区、不同实验室和不同时间的测试数据,以及不同景观条件和网度的表层、深层土壤元素数据或不同分度带的网格数据分别进行拼接合并,因此区域编图对原始数据的整理和网格化处理不可避免.而不同的数据整理方法及其计算参数所形成的结果是非常不同的,数据本身的特点也对编图具有很大的影响.为了真实客观地反映区域内元素的地球化学分布特征,数据整理和处理的方法及技术参数的选择是关键技术问题.作者就化探数据处理中常用的几种插值方法进行了对比评价,并分别从网格间距及搜索半径大小等方面做比较,提出了自己在编图中采用的数据处理方法及选用的技术参数.     

基于CALIPSO卫星产品的中国区域异质核化冰云比例差异

利用2012年1月～2015年12月CALIPSO冰云3级月平均产品(CAL_LID_L3_Ice_Cloud)对中国区域异质核化冰云比例进行分析,研究冰云的比例分布与季节变化特征,探究中国区域冰云分布情况和影响冰云分布的因素。结果表明:中国南北部冰云的比例差异与地面2 m温度、地表长波净辐射通量和来自地面的冰核有关。-40～0℃之间的冰云比例水平分布,中国大陆北部比中国大陆南部高约30%,季节性变化规律明显,冬季南北差异大,夏季南北差异小;云箱温度在-30～10℃之间的冰云,中国大陆北部比例比大陆南部平均高出约5%,春季、秋季、冬季北部冰云比例均高于南部,夏季温度在-8℃以下,南部的冰云比例超过了北部的冰云比例;在对流层(10 km以下),冰云比例分布中国大陆北部比南部高,南北部之间冰云比例差异在6 km左右显示出峰值,约50%,四季冰云比例随高度增加均呈现增大的趋势,且北部冰云比例均高于南部。