占位，持续更新。

第1篇

RichardFlorida, Charlotta Mellander and Tim Gulden. Global Metropolis: AssesingEconomic Activity in Urban Centres Based on Nighttime Satellite Images. TheProfessional Geographer, 2012, 64:2 178-187.

看点：

1.使用全球卫星遥感图片，识别都市区。

2.避免不同国家城市化区域有不同定义，从而不具可比较性的问题。

这篇文章只有方法的部分值得看，其他的没必要看。

Methods

These arecompletely contiguously lighted areas as seen from space at night.

These data provide a measure of light intensity for each 30 arc second cell between -65。 and 65。 latitude. These cells cover approximately 1 km2 at the equator and become somewhat smaller at higher and lower latitudes.

This data limitations is that the relationship between light emissions and economic activity breaks down as higher levels of urbanization expand vertically rather than horizontally. We would be forced to estimate central cities differently from their surroundings in any case. To deal with the problem of saturation of urban cores, we break the process of estimating economic activity from light emissions into two stages. We estimate activity levels for low-light areas, including urban peripheries, as a direct function of light level. We separately estimate urban cores as a function of both area and shape.

We then set a light threshold that captures the essence of the U.S. metropolitan areas. We thus calibrate the model by using detailed published data for gross regional product (GRP) for 363 U.S. metropolitan areas (Global Insight, 2006). This calibration yields an objective measure of the precision of this process. We are able to reproduce the GRP estimates with a standard error of 34 percent.

数据来源

We utilizelight emission data for the year 2000. The base data are from the EarthObservation Program of the National Oceanic and Atmospheric Administration’sNational Geophysical Data Center (NGDC).

没看懂的地方

1. 数据。Although the data capture low light levels that are consistent with low-density suburban and electrified semirural areas, the measured emission level saturates far from the most economically intense center of a major city due to the design of the sensors and the processing algorithms used by NGDC. The fall-off in brightness gradient as this threshold is quite steep and occurs in the inner suburbs of large U.S. urban areas (Weier, 2000).

2.校准。We thus calibrate the model by using detailed published data for gross regional product (GRP) for 363 U.S. metropolitan areas (Global Insight, 2006).

3. These are completely contiguously lighted areas as seen from space at night. 怎么操作？

---------------------------------------------------------------------------------------------------------------------------------
第2篇
Elvidge C.D.,Baugh K. E., Anderson S.J., Sutton P.C., and Ghosh T. The Night LightDevelopment Index: a spatially explicit measure of human development fromsatellite data. Social Geography, 2012, 7, 23-35.

研究目的

We havedeveloped the Night Light Development Index (NLDI), applying the Lorenz curveanalysis to characterize the co-distribution of nocturnal lighting and people.

探讨人口与灯光之间的关系。

数据，看看理解的快。
http://ngdc.noaa.gov/eog/

补充

Henderson Vernon, Storeygard Adam, and Weil David. A bright idea for measuring economic growth. The American Economic Review, 2011, 101(3), 194-199.

卫星遥感影片的两大劣势
（1）经济活动与卫星图片的灯光之间的关系在不同的地方、不同时间都是变化的。
（2）湿度，折射，阳光、月光和云彩覆盖的时间等，在不同地方都不同。

卫星遥感影片的两大优势
（1）测度误差。发展中国家的数据误差很大，用卫星遥感影片可以一定程度上解决。
（2）补充数据。很多经济发展指标缺失，可以用卫星遥感影片数据补充。

数据
Light emanation is measured by a digital number (DN) on a linear scale between 0 and 63, which corresponds closely but not exactly to the quantity of light (as measured by true radiance) reaching the satellite.  

几个问题
1.灯光下的城市边界是怎么画出来的？操作层面上回答。
2.怎么建立灯光强度与经济增长的关系？因为二者之间的关系因时间、区域不同而不同。
我的一个想法是：使用灯光的亮度、范围、高度等信息，不知道有没有反映灯光高度的信息，比如100m处有灯光，还是仅仅20m高处有灯光。换句话说，要充分挖掘灯光的信息。
3.怎么使用灯光数据模拟小城市的发展？恐怕很难。

强度似乎就反映高度。问题是：经济发展水平不同的区域，可能灯光强度一样。
还是第2个问题：怎么建立灯光强度与经济增长的关系？因为二者之间的关系因时间、区域不同而不同。