LAUSR

The available statistics on electric consumption (EC) can hardly show the spatial heterogeneity within political boundaries, and the previous studies paid few attention on the relations between EC and urban planning. The present study developed models between nighttime light values and EC using nighttime light images (NLI) from the Suomi National Polar-Orbiting Partnership-Visible Infrared Imaging Radiometer Suite (NPP-VIIRS) and statistical EC data from local government of Xi’an. A geographic information system (GIS) was utilized to map the distribution of EC on a grid scale, and the spatiotemporal dynamic of EC was obtained. In addition, a slope model and the EC grading threshold were conducted to determine the urban functional districts (UFD). The results were shown as follows: (1) the relation between TNL and EC is significant positive, and the performance of linear regression model ( $\text{R}^{2}=0.76$ , Root Mean Square Error (RMSE) $=9.44\times 10^{4}$ , Mean Absolute Error (MAE) $=6.73\times 10^{3}$ , Mean Percent Error (MPE) = −3.71%) is better than logarithmic model ( $\text{R}^{2}=0.68$ , RMSE $=1.09\times 10^{5}$ , MAE $=7.77\times 10^{4}$ , MPE = −5.44%). (2) In Xi’an, the EC continuously increased during 2013-2017, and the change of EC demonstrated obvious features that increased slowly in the city center, and spread outward to the surrounding region especially to the south of the study area. (3) The UFD have been divided into five sections including tourist district, commercial district, industrial zone, residential community, and ecological zone, respectively. This study will thus provide a reference and scientific basis for urban planning and rational allocation of electric power.