基于BP神经网络的陕西春季浅层地温预测模型

利用2016—2021年春季(3—5月)陕西98个国家级地面气象观测站逐日观测资料,运用BP神经网络构建了陕西春季不同区域(陕北、关中、陕南)不同月份(3、4、5月)不同土壤深度(5、10、15、20 cm)地温预测模型,并利用2022年数据进行模型预测检验。研究结果表明：模型预测的各深度日平均地温在陕北、关中、陕南地区预测准确率＞95%,且整体表现出土壤深度越深预测准确率越高的趋势,10 cm日平均地温预测准确率＞98%,15、20 cm日平均地温预测准确率＞99%;各深度日平均地温在陕北、关中、陕南地区预测值与实测值的均方根误差≤10 ℃、平均绝对误差≤08 ℃,均方根误差和平均绝对误差整体表现出土壤深度越深误差值越小的趋势;模型预测精度比较理想,可用于陕西春季浅层地温预报业务,为春播及果树花期预报提供技术支持。     

苏联经济区方案定量研究的经验

通常,经济区划的任务包括以下几个主要阶段:1) 确定作为基础区的地域;2)按照基础区选择区划指标系统及其定量评价;3) 选择区域系统模式;4) 阐明课题建立算法,直接把基础区进行分类计算;5)解释已获得的结果,选出最好的区划方案和有关进一步改善工作方法的结论等等。基础区的范围视所要解决的任务性质而定。     

缺氧环境中的成矿作用及构造,岩性和时间对成矿作用的控制

缺氧环境中，金属从海水、古卤水或热水溶液中被吸附和沉淀而发生积聚。在后来的变质和岩浆作用过程中，金属重新活化，形成矿床。这种矿化类型主要受缺氧环境的构造背景类型控制。被动边缘的碳质沈积岩只含有铀、钒和钼的次经济富集；含铜黑色页岩和海底喷流型矿化则限于大陆裂谷和拉拗槽环境或活动边缘的弧后盆地；变质矿术主要与碰掸边缘有关，但也出现于其它的构造环境类型。含铜黑色页岩的形成，受地球演化过程中超大陆裂开时低     

微动勘查技术在地热勘探中的应用

随着能源改革进一步推进,地热能这种中深层清洁能源的勘探开发利用迫在眉睫,目前勘探开发利用重点在中深层位,且主要在城区,限制了常规地震、大地电磁等物探方法的使用,需引进新方法。本文通过微动勘查技术在某城区开展相关实验,对比以往高精度重力、大地电磁测深、地热孔等资料,验证了本次微动实验成果,解释推断的地层划分和构造位置与已知资料基本吻合,值得进一步研究和推广,对实现微动勘查技术在中深层地热能源快速勘探开发具有重要意义。     

博斯腾湖水矿化度趋势面分析初探

本文采用多项式作为趋势面分析数学模型,对博斯腾湖1984年湖水矿化度初步分析表明,博斯腾湖水矿化度分布主要受入湖盐分,入湖水流及湖水循环的影响。同时趋势面显示了入湖盐分及水流的运动路径和扩散情况,从而给出了它们的水文动态特征,取得了比较理想的拟合结果,这是以往传统方法难以实现的。     

Analyses of organic and inorganic contaminants in Salton Sea fish

Chemical contamination of fish from the Salton Sea, a quasi-marine lake in Southern California, could adversely impact millions of birds using the Pacific Flyway and thousands of humans using the lake for recreation. Bairdiella icistia (bairdiella), Cynoscion xanthulus (orangemouth corvina), and Oreochromis spp. (tilapia) were sampled from two river mouths and two nearshore areas of the Salton Sea. Muscle tissues were analyzed for a complete suite of 14 trace metals and 53 pesticides. Fish muscle tissues had concentrations of selenium ranging between 1.89 and 2.73 microg/g wet weight. 4,4'-DDE accounted for 94% of the total DDT metabolites. Total DDTs ranged between 17.1 and 239.0 and total PCBs between 2.5 and 18.6 ng/g wet weight. PCB congeners 132, 138, 153, 168, and 180 comprised over 50% of the total PCBs. Given the potential implementation of a commercial fishing at the Salton Sea in the future, the presence of persistent organic pollutants and selenium warrants further research into the effects of these mixtures on fish populations, and on wildlife and humans consuming fish.     

训练样本对遥感影像分类精度影响研究

以大理SPOT-5 HRG影像和Landsat TM影像作为数据源,在构建标准训练样本数据集的基础上,探索训练样本对遥感影像分类的影响。选取不同训练样本数量组合,分别对监督分类中的平行六面体、最小距离、马氏距离、最大似然法、神经网络和支持向量机6种分类方法进行多次实验,并采用Kappa系数和总体分类精度对实验分类结果的精度进行评价。结果表明:以多次分类结果的平均值作为最终的分类结果能减小随机误差;增加训练样本数量可以减小单次分类引起的随机误差;在不同的训练样本量下,支持向量机的分类精度最高。     

注入诱发地震:利用高应力路径储层和温度诱发应力预处理降低诱发地震风险的策略

如果应力路径足够高,正断层应力机制下的流体注入可以稳定储层。当储层由于注入而显著冷却时,这种稳定性将消失。此外,本文还提出了一种新的激发剪切储层的策略,以减少诱发大震级地震的可能性。本文提出的方法适用于逆断层应力状态,包括初始应力预处理阶段,在该阶段,储层被冷却,压力增加受到限制。这一过程降低了水平总应力,从而也降低了差异应力。然后,储层受到孔隙压力快速增加的刺激,在较低的差异应力下导致剪切破坏,比储层中最初出现的差异应力要小。由于古登堡—里克特b值和差异应力之间的联系,以这种方式激发的储层将表现出更高的b值,从而也减少了发生大地震的可能性。该方法的适用范围包括正断层和走滑断层的应力机制。     

STRUCTURE AND COMPOSITION OF A WATERSHED-SCALE SEDIMENT INFORMATION NETWORK

A "Watershed-Scale Sediment Information Network" (WaSSIN), designed to complement UNESCO’s International Sedimentation Initiative, was endorsed as an initial project by the World Association for Sedimentation and Erosion Research. WaSSIN is to address global fluvial-sediment information needs through a network approach based on consistent protocols for the collection, analysis, and storage of fluvial-sediment and ancillary information at smaller spatial scales than those of the International Sedimentation Initiative. As a second step of implementation, it is proposed herein that the WaSSIN have a general structure of two components, (1) monitoring and data acquisition and (2) research. Monitoring is to be conducted in small watersheds, each of which has an established database for discharge of water and suspended sediment and possibly for bed load, bed material, and bed topography. Ideally, documented protocols have been used for collecting, analyzing, storing, and sharing the derivative data. The research component is to continue the collection and interpretation of data, to compare those data among candidate watersheds, and to determine gradients of fluxes and processes among the selected watersheds. To define gradients and evaluate processes, the initial watersheds will have several common attributes. Watersheds of the first group will be: (1) six to ten in number, (2) less than 1000 km2 in area, (3) generally in mid-latitudes of continents, and (4) of semiarid climate. Potential candidate watersheds presently include the Weany Creek Basin, northeastern Australia, the Zhi Fanggou catchment, northern China, the Eshtemoa Watershed, southern Israel, the Metsemotlhaba River Basin, Botswana, the Aiuaba Experimental Basin, Brazil, and the Walnut Gulch Experimental Watershed, southwestern United States.