工业部门污染物治理协同控制温室气体效应评价——基于重庆市的实证分析

以高能耗为主要特征的工业部门是大气污染物和温室气体的重要排放源.为推动协同管控,文中结合生态环境部在重庆市组织开展的试点工作,对工业企业NOx污染治理协同控制温室气体的效应进行了量化分析.结果表明,以末端治理为手段的NOx治理措施协同控制温室气体的效果为负,即工业企业去除1 t NOx会直接或间接增加CO2排放1.81...     

水库对河流营养盐滞留效应研究进展

综述了国内外水库对河流营养盐滞留效应的主要进展,尽管在大型水库是否显著改变原有河流输送营养盐的问题上存在诸多争议,反映在滞留效率的估算上,存在前后矛盾的结果,但大量的文献仍支持筑坝显著改变了河流原有输送营养盐的特性,并对下游、河口产生深远的影响.本文将"滞留"的主要过程分为泥沙过滤作用和生物过滤作用,并列举了若干经典的估算滞留效率的方法和部分水库对营养盐的滞留效率.     

武汉市日供水量与气象要素的相关分析

利用武汉市逐日供水量资料与温度、降水、日照等气象要素进行同步相关分析,选取相关性较高的气温作为预报因子,建立了日供水量的简易预测模型。     

内伶仃岛及其附近海面低层大气的平均结构

本文综合分析了内伶仃岛低空温、风探测资料和海岛及其附近海面大气边界层的平均结构,讨论了海陆环流,并把实测风速廓线与理论计算结果进行了比较。      

面向对象的高分遥感影像信息提取技术研究

以World View-2高分遥感影像数据为例,采用规则集开发的模式进行地图要素提取,对主要制图要素图像进行了分割、分类试验,实现了多尺度分割的层次构建和制图地物要素分级分类。     

Precise positions of Phoebe determined with CCD image-overlapping calibration

This paper presents 210 positions of Phoebe, the ninth satellite of Saturn, observed with the 1-m telescope at the Yunnan Observatory during the years 2003–2005, using a CCD image-overlapping calibration method proposed recently by Peng et al. After the observed positions of Phoebe are compared with its theoretical positions computed by the new JPL ephemerides DE405 and SAT199, the mean residuals (observed minus computed) are 0.21 and −0.05 arcsec in right ascension and declination, respectively, with a standard deviation of 0.06 arcsec.     

导航电子地图增量更新服务及关键技术研究

以移动通讯数据传输技术、网络技术和时态数据库技术为基础,构建了导航地理信息的变化发现、分析甄别、采集、加工、处理、存储和发布服务一体化的增量式更新服务模式、技术、软件、质检和标准体系,论述了多源数据融合、时空模型及空间关系的动态实时维护、多尺度增量式导航电子地图综合处理、路径规划和质检方法等关键性技术,实现了导航电子地图近实时更新服务。     

Thickness distribution and extent of sea ice and snow in Prydz Bay and surrounding waters observed in 2012/2013 austral summer

Three ship-based observational campaigns were conducted to survey sea ice and snow in Prydz Bay and the surrounding waters(64.40°S–69.40°S, 76.11°E–81.29°E) from 28 November 2012 to 3 February 2013. In this paper, we present the sea ice extent and its variation, and the ice and snow thickness distributions and their variations with time in the observed zone. In the pack ice zone, the southern edge of the pack ice changed little, whereas the northern edge retreated significantly during the two earlier observation periods. Compared with the pack ice, the fast ice exhibited a significantly slower variation in extent with its northernmost edge retreating southwards by 6.7 km at a rate of 0.37 km?d-1. Generally, ice showed an increment in thickness with increasing latitude from the end of November to the middle of December. Ice and snow thickness followed an approximate normal distribution during the two earlier observations(79.7±28.9 cm, 79.1±19.1 cm for ice thickness, and 11.6±6.1 cm, 9.6±3.4 cm for snow thickness, respectively), and the distribution tended to be more concentrated in mid-December than in late November. The expected value of ice thickness decreased by 0.6 cm, whereas that of snow thickness decreased by 2 cm from 28 November to 18 December 2012. Ice thickness distribution showed no obvious regularity between 31 January and 3 February, 2013.     

不同方法对冬小麦地块级估产的适用性研究

为了探究不同估产模型对冬小麦估产的精度和适用性,提高像元级和地块级冬小麦估产精度。利用机器学习方法支持向量机（SVM）模型、随机森林（RF）模型和深度学习方法长短期记忆（LSTM）模型,对冬小麦进行产量预估。结果表明：在像元级上,LSTM模型、RF模型估产精度高于SVM模型,LSTM模型不仅能够表征作物在生育期的生长变化,还能降低人为因素干扰,估产结果更加客观真实,而SVM模型、RF模型易受到特征共线性和过拟合的影响,因此SVM模型、RF模型更适用于区域性小范围估产。在地块级上,SVM模型、RF模型和LSTM模型地块级估产精度比像元级估产精度均有提高,且LSTM模型估产精度明显高于SVM模型和RF模型,因此LSTM模型更适用于地块级估产,且LSTM模型泛化能力强、普适性高,适用于大范围农作物估产。