Urbanization in European regions based on night lights Demetris Stathakis n , Vassilis Tselios, Ioannis Faraslis University of Thessaly, Department of Urban and Regional Planning Engineers, Volos 38334, Greece article info Article history: Received 6 August 2015 Received in revised form 6 October 2015 Accepted 8 October 2015 Available online 24 October 2015 Keywords: Urbanization Night lights DMSP/OLS NUTS Europe abstract Defense Meteorological Satellite Program/Operational Linescan System (DMSP/OLS) night lights are used as a source to capture urbanization trends in European regions. The data are intercalibrated to form a comparable time series. The Sum of Lights (SoL) index is then calculated per region. To eliminate the problem of annual fluctuation, the overall and the decadal slope of the SoL trend are also calculated. The analysis of the results shows that there exist significantly diverse urbanization patterns within European regions that can be effectively captured by DMSP/OLS data. It is also shown that GDP and population are positively associated with urbanization in Europe, however less strongly at regional than national levels. Overall, the interpretation of SoL changes in the European regions is more complex compared to those in the developing parts of the globe, primarily due to dif- ferences in energy saving polices. & 2015 Elsevier B.V. All rights reserved. 1. Introduction Cities occupy approximately 3% of the earth's surface but their “ecological footprints” are tens to hundreds of times their area, resulting to significant land use changes (Grimm et al., 2008). Despite the importance of cities, the design of earth observation sensors has been focused on natural sys- tems (Elvidge et al., 1999). Recently, efforts to map urban areas at continental and global scales have been successful in exploiting the capacity of optical sensors, primarily by MODIS and LANDSAT data (Schneider et al., 2010; World Bank, 2015). These efforts resulted for the first time to a relatively accurate and independent estimate of urban extent and form. The coverage and spatial resolution (250–500 m) of MODIS is suitable for mapping urban areas at regional and national scales. The main drawback is however that MODIS data do not go back much in time. The sensor was launched in 1999. LANDSAT's archive, covering more than four decades, is also not suitable in this case. Its small footprint requires laborious mosaicking to cover continental scales. The high associated costs render repetition infeasible. In terms of content, the main barrier in processing optical data is the fact that urban areas are spectrally mixed with barren land, especially in parts of the globe where cities are not surrounded by vegetation which has a conveniently distinct spectral signature (Stathakis and Faraslis, 2014; Stathakis et al., 2012). One appealing alternative to bridge the data gap for monitoring urban areas is the Defense Meteorological Satellite Program–Operational Linescan System (DMSP/OLS) sensors. The OLS instrument has a broad 0.5–0.9 μm VNIR band (Elvidge et al., 1999), approximately corresponding to LAND- SAT's panchromatic band (TM bands 1–4 combined). DMSP/ OLS sensors capture images during the night, typically between 8:30 and 9:30 p.m. local time (Amaral et al., 2005). However, overpass times can differ by as much as two hours (Elvidge et al., 2014). Night lights observed from space are a quite straightforward indication of human presence. Inhabited areas are clearly outlined (Croft, 1973; Imhoff et al., 1997a, 1997b). Human presence in night lights is not purely related to nighttime (census) population (Elvidge et al., 1999). Fea- tures of day-time population (a.k.a. ambient population) are also evident (industry, businesses, traffic etc.). The radiometric resolution of OLS is relatively coarse. Only six bits are available Contents lists available at ScienceDirect journal homepage: www.elsevier.com/locate/rsase Remote Sensing Applications: Society and Environment http://dx.doi.org/10.1016/j.rsase.2015.10.001 2352-9385/& 2015 Elsevier B.V. All rights reserved. n Corresponding author. Tel.: þ30 24210 74489. E-mail addresses: dstath@uth.gr (D. Stathakis), vtselios@uth.gr (V. Tselios), faraslis@uth.gr (I. Faraslis). Remote Sensing Applications: Society and Environment 2 (2015) 26–34