Predicting urban expansion in Moscow based on night lights

Abstract

Night-lights obtained by the DMSP/OLS sensor offer a unique opportunity to measure urban expansion in the past two decades. We apply a method to project the existing night-lights time series in the future in order to forecast urban expansion. The rapidly expanding Moscow city in the Russian Federation is selected as the case study. Night-lights are projected up to the year 2025 by exponential smoothing. It is demonstrated by the results that the method can be used to obtain both spatially explicit, i.e. actual maps, as well as synoptic forecasts, e.g. by means of forecast variables such as the Sum of Lights (SoL). These forecasts are accompanied by estimation of confidence intervals, providing upper and lower bounds for future values. The method presented can be applied globally. © 2017 SPIE.