SDGSAT-1 nighttime light data improve village-scale built-up delineation

Abstract

While previous urban delineation studies have been based mainly on Defense Meteorological Satellite Program (DMSP) and Visible Infrared Imaging Radiometer Suite (VIIRS) nighttime light (NTL) data at a 1-km or 500-m spatial resolution, the higher 10-m-spatial-resolution Sustainable Development Science Satellite-1 (SDGSAT-1) NTL data offer the possibility of sophisticated village-scale analysis. In this study, we developed the quantile method with downward heuristic rules for built-up delineation based on SDGSAT-1 NTL data and proposed a new threshold method, named the trend method framework, for urban dynamic delineation in combination with SDGSAT-1 NTL data and DMSP-VIIRS NTL data. Then, the methods developed in this study were applied to five urban agglomerations (the Beijing-Tianjin-Hebei Urban Agglomeration (BTH), Yangtze River Delta (YRD), Guangdong–Hong Kong–Macao Greater Bay Area (GHM), Chengdu–Chongqing Economic Circle (CCC), and Hainan Free Trade Port (HTP)) in China. The results showed that both the downward quantile method and the trend method can be used to delineate village-scale built-up areas. In the delineation of built-up areas based on SDGSAT-1 NTL data, the trend method provides clearer information on the built-up areas within lighted areas, leading to higher accuracy than the downward quantile method. In the built-up area delineation based on DMSP and VIIRS NTL data at 1-km resolution, the trend method combines SDGSAT-1 NTL data and accounts for information on time-series changes in urbanization, thus allowing for the more accurate extraction of built-up areas. In 2020, the trend method had higher delineation accuracies than the quantile method, with effectiveness accuracies (EAs) of 0.91, 0.77, 0.67, 0.94, and 0.65 for the YRD, BTH, CCC, GHM, and HTP, respectively, compared with the European Space Agency (ESA) land use and land cover (LULC) product. The trend method has very good application prospects for delineating built-up areas using the high-resolution SDGSAT-1 NTL dataset.