Population,urbanization and economic scenarios over the Belt and Road region under the Shared Socioeconomic Pathways

The countries throughout the Belt and Road region account for more than 60% of the world's population and half of the global economy. Future changes in this area will have significant influences on the global economic growth, industrial structure and resource allocation. In this study, the proportion of the urban population to the total population and the gross domestic product were used to represent the levels of urbanization and economic development, respectively. The population, urbanization and economic levels of the Belt and Road countries for 2020–2050 were projected under the framework of the IPCC's shared socioeconomic pathways(SSPs), and the following conclusions are drawn.(1) The population, urbanization and economic levels in the Belt and Road region will likely increase under all five pathways. The population will increase by 2%–8%/10 a during 2020–2050 and reach 5.0–6.0 billion in 2050. Meanwhile, the urbanization rate will increase by 1.4%–7.5%/10 a and reach 49%–75%. The GDP will increase by 17%–34%/10 a and reach 134–243 trillion USD.(2) Large differences will appear under different scenarios. The SSP1 and SSP5 pathways demonstrate relatively high urbanization and economic levels, but the population size is comparatively smaller; SSP3 shows the opposite trend. Meanwhile, the economy develops slowly under SSP4, but it has a relatively high urbanization level, while SSP2 exhibits an intermediate trend.(3) In 2050, the population will increase relative to 2016 in most countries, and population size in the fastest growing country in Central Asia and the Middle East countries will be more than double. Urbanization will develop rapidly in South Asia, West Asia and Central Asia, and will increase by more than 150% in the fastest growing countries. The economy will grow fastest in South Asia, Southeast Asia and West Asia, and increase by more than 10 times in some counties with rapid economic development.     

Correspondence between the large volcanic eruptions and ENSO events over AD 1525–2000

ENSO is an interannual mode which may be affected by external forcing, such as volcanic eruptions. Based on the reconstructed volcanic eruptions chronology and ENSO sequences, both 195 large volcanic eruptions(VEI≥4) and 398 ENSO(El Ni?o and La Ni?a) events were extracted from 1525 to 2000. An analysis of the correspondence between the large volcanic eruptions and ENSO events was performed by matching the large volcanic eruptions with the types and magnitudes of ENSO events present in the 0–2 years after the eruptions. The results show the following:(1) The percentages of ENSO events within the 3 years after the large eruptions had increased to 68.3% from 31.7% compared with those with no-eruptions in the previous 0–2 years. In addition, the ratio of El Ni?o to La Ni?a events turned from 2:3 to 1:1, and more El Ni?o events occurred in the 0 year after eruptions in the low-latitudes of the Northern Hemisphere and in the tropics but more La Ni?a events occurred in the 0 year after in the high-latitudes of the Northern Hemisphere and the Southern Hemisphere.(2) After the eruptions, the weak(W) El Ni?o events had increased by 8 percentage points and the very strong(VS) El Ni?o events had decreased by 10 percentage points; conversely, there was a decrease by 15 percentage points of the weak La Ni?a events and an increase by 11.4 percentage points of the very strong La Ni?a events. Specifically, the percentages of strong La Ni?a events increased to a peak at 1(+1) year after the eruptions.(3) The percentage of eruptions followed by single-year ENSO was the greatest. The percentage of ENSO events that occurred in the consecutive 2 years following an eruption was approximately equal to the percentage of events that occurred consecutively 3 years following an eruption, and both sets of ENSO magnitudes showed a decreasing trend.     

Analysis of Spatial–Temporal Evolution of Mining-Induced Fracture Field: A Case Study Using Image Processing in the Shaqu Coal Mine,China

Natural Resources Research - Mining-induced fracture plays a key role in gas drainage for gas burst-prone underground coal mines, especially for closely multilayered coal seams. The layout and...     

空间分析与生物地球化学循环:西藏林周县大骨节病患病村与非患病村的比较研究

西藏林周县是我国大骨节病(KBD)患病较为严重的地区之一,本文将林周县作为研究区,通过使用地理探测器(GeoDetector)量化分析KBD患病率风险因子的影响,并使用环境化学方法验证空间分析结果。通过对10个潜在影响因子的分析以及对当地KBD患病村和非患病村的土壤-水-粮食-人这一生物地球化学循环的环境化学分析,结果表明:(1)林周县KBD由一组多重且交互作用的环境影响因子共同作用影响,其中最重要的控制因子是地层因子;(2)所有环境介质(土壤、水、谷物)及人体组织中的硒元素浓度在KBD患病区均低于非患病区;(3)当地居民对硒和铬的摄入严重不足,尤其是KBD患病村中居民硒元素平均日摄入量(ADD)大约仅为世界卫生组织(WHO)建议的成人基本摄入量下限的4%;(4)我们推测,当地居民患病主要是由于地层这一影响因子,这是由于通过生态系统的迁移转化导致当地人口严重硒缺乏,最终导致地方性生物地球化学硒缺乏。     

基于深度学习理论的无人机影像分类

针对无人机影像深度学习分类方法缺乏现状,本文利用深度学习理论卷积神经网络方法对无人机影像进行了分类。该法首先抽取无人机影像作为训练集和检验集,然后建立一个2个卷积层-池化层的卷积神经网络模型进行深度学习,通过设定参数并运行模型实现无人机影像分类。实验表明,本文提出的方法可完成较复杂地区无人机影像分类,其分类精度与支持向量机方法相当,为无人机遥感影像分类提供了一个崭新的技术视点。     

FY-3D/MERSI-II全球火点监测产品及其应用

FY-3D/MERSI-II全球火点监测产品主要包括全球范围内的火点位置、亚像元火点面积和火点强度等信息,可用于实时监测全球范围的森林草原火灾、秸秆焚烧等生物质燃烧状况。火点判识算法主要根据中红外通道对高温热源的敏感特性,即含有火点的中红外通道像元辐亮度和亮温较远红外通道的辐亮度和亮温偏高,同时较周边非火点的中红外像元偏高,建立合适的阈值可探测含有火点的像元。亚像元火点面积估算主要使用中红外单通道估算,根据亚像元火点面积估算结果对火点强度进行分级,不同的级别表示不同程度的火点辐射强度。基于全球火点自动判识结果,每日生成0.01°分辨率的卫星遥感日全球火点产品,每月生产0.25°×0.25°格点的全球月火点密度图。在利用FY-3D/MERSI-II火点产品开展的全球火点监测应用中,对多起全球重大野火事件进行了监测,为防灾减灾、全球气候变化研究、生态环境保护等方面提供卫星遥感信息支持。     

复杂卫星图像中的小目标船舶识别

船舶作为海上的重要目标,实现对船舶自动识别有重要的意义。针对卫星图像中云雾、海岸背景等复杂海情对船舶识别带来的干扰,以及小目标船舶高漏检率问题,本文提出一种多尺度深度学习模型训练策略,在此基础上构建了一种船舶识别的深度学习网络,该网络可分为多尺度训练、特征提取、生成目标建议区域、船舶分类这4个部分。首先,采用多尺度的训练策略,将多尺度的船舶样本送入网络中进行训练,这样在训练样本中加入了大量小目标船舶的样本,使网络充分提取到小目标船舶的特征;其次,通过卷积神经网络对目标船舶进行特征自适应提取;然后,目标区域建议网络可依据卷积神经网络提取到的特征,在图像中找到感兴趣目标区域,即框定船舶的位置;最后,通过多个全连接层的组合,将高维特征映射到一个4元组中,再运用分类函数输出每一类船舶的概率值,概率值最大的则为该船舶的类别。同时为解决云雾遮挡和海岸背景的干扰,采用了一种负样本增强学习的方法,在样本数据集中加入了大量只含有云雾和海岸背景的图片,进行负样本扩充,增强网络模型对云雾及海岸背景的特征学习能力,以此解决复杂海情的影响。实验结果表明,所提方法有效解决了复杂海情条件下的船舶识别难,以及小目标船舶识别难的问题,实现了复杂海情条件下的船舶识别。同时,与现有成熟的深度学习目标识别算法相比,本文算法的精确度和召回率分别提升了6.98%和18.17%,所训练的模型具有良好的泛化能力和鲁棒性。     

A-FPN算法及其在遥感图像船舶检测中的应用

光学遥感图像船舶检测主要面临两个挑战:光学遥感图像背景复杂,船舶检测易受海浪、云雾及陆地建筑等多方面干扰;遥感图像分辨率低,船舶目标小,对于其分类与定位带来很大困难;针对上述问题,在FPN的基础上,提出一种融入显著性特征的卷积神经网络模型A-FPN (Attention-Based Feature Pyramid Networks)。首先,利用卷积提取图像特征金字塔;然后,利用顶层金字塔逐级构建显著特征层,抑制背景信息,通过金字塔顶层的细粒度特征提高浅层特征的表达能力,构建自上而下的多级显著特征映射结构;最后利用Softmax分类器进行多层级船舶检测。A-FPN模型利用显著性机制引导不同感受下的特征进行融合,提高了模型的分辨能力,对遥感图像处理领域具有重要应用价值。实验阶段,利用公开的遥感目标检测数据集NWPU VHR-10中的船舶样本进行测试,准确率为92.8%,表明A-FPN模型适用于遥感图像船舶检测。     

O2-O2云反演算法及其在TROPOMI上的应用

利用遥感技术对大气环境污染进行监测时,云是影响痕量气体反演精度的重要因素,因此在痕量气体反演中需要对云的影响进行校正,通常使用的云参数主要是有效云量和云压。本文基于O₂-O₂ 477 nm吸收波段构建了O₂-O₂云反演算法：首先,根据有效云量和云高与连续反射率和O₂-O₂斜柱浓度之间的对应关系,结合假定的云模型利用VLIDORT辐射传输模型建立关于有效云量和云压的查找表;然后,通过差分吸收光谱技术拟合卫星载荷观测的大气层顶辐射,获得O₂-O₂斜柱浓度并计算连续反射率;最后,结合辅助数据,根据查找表进行插值反演获得有效云量和云压。通过将算法应用到OMI观测数据,将反演结果与OMCLDO2产品进行对比验证,有效云量和云压空间分布一致,相关系数R均超过0.97;并还将该算法应用于下一代大气成分监测仪器TROPOMI,与FRESCO+产品对比,有效云量和云压空间分布基本一致,当地表类型为海洋时,有效云量相关系数R大于0.97,云压相关系数R大于0.94,云压反演结果存在一定的区别;通过将O₂-O₂云反演算法和FRESCO+云压反演结果与CALIOP Cloud Layer产品进行对比,结果表明,在低云情况下,O₂-O₂云反演算法线性回归方程斜率为0.782,截距为198.0 hPa,相关系数R为0.850,算法表现优于FRESCO+。而在高云情况下,FRESCO+反演结果更接近CALIOP云压结果。在OMI和TROPOMI上的应用表明O₂-O₂云反演算法在大气云反演中具有较高的准确性和可行性,可以为大气痕量气体反演的校正提供云参数,并为中国同类型卫星载荷的云反演算法提供算法参考。