论文标题
人为夜灯的时空网络演变1992-2015
Spatiotemporal Network Evolution of Anthropogenic Night Light 1992-2015
论文作者
论文摘要
夜灯的卫星成像提供了从1992年到现在的全球开发记录。并非所有定居点都可以用夜灯检测到，而是建筑环境的连续性，在这些环境中，大多数人口通常都可以使用。将连续场分割为离散的像素的空间连续子集产生一个空间网络，其中每个连续段代表一个不同的网络组件。将照明开发的连续体作为不同空间连接的空间网络，可以在空间环境中应用电源法网络结构的生成条件，以提供对在定居点和其他土地覆盖类型中观察到的类似规模的一般说明。这项研究介绍了一种新的方法，以结合卫星来源的夜间光观测的互补来源，以量化从1992年至2015年的全球空间网络结构的演变。网络组件的面积 - 外膜分布显示多帧量表与较大的组件变得越来越多。网络组件的等级大小分布是线性的，并且由幂律的幂定律很好地描述了-108的0.08，所有27个地理，年度和连接程度的子集都表明稳定性的缩放属性。网络组件内的亮度区域分布显示，从连续低相位的较小组件到不连续的，几乎均匀的亮度分布的突然过渡，对于较大的组件，表明尽管大小和形状的延续，网络结构的根本变化。这些结果表明，对于行政定义的城市人口而言，城市规模的规模不一致，对于物理定义的结算网络区域更加一致，可以通过空间网络增长过程更简单地解释。
Satellite imaging of night light provides a global record of lighted development from 1992 to present. Not all settlements can be detected with night light, but the continuum of built environments where a rapidly growing majority of the population lives generally can. Segmenting a continuous field into discrete spatially contiguous subsets of pixels produces a spatial network in which each contiguous segment represents a distinct network component. Representing the continuum of lighted development as spatial networks of varying spatial connectivity allows the generative conditions for power law network structure to be applied in a spatial context to provide a general explanation for similar scaling observed in settlements and other land cover types. This study introduces a novel methodology to combine complementary sources of satellite-derived night light observations to quantify the evolution of the global spatial network structure of lighted development from 1992 to 2015. Area-perimeter distributions of network components show multifractal scaling with larger components becoming increasingly tortuous. Rank-size distributions of network components are linear and well described by power laws with exponents within 0.08 of -1 for all 27 subsets of geography, year and degree of connectivity - indicating robust scaling properties. Area distributions of luminance within network components show an abrupt transition from continuously low-skewed for smaller components to discontinuous, nearly uniform luminance distributions for larger components, suggesting a fundamental change in network structure despite a continuum of size and shape. These results suggest that city size scaling, observed inconsistently for administratively-defined city populations, is more consistent for physically-defined settlement network area and can be explained more simply by a spatial network growth process.
