GC23E-1183
Spatiotemporal dynamics of human settlement patterns in the Southeast U.S. from DMSP/OLS nightlight time series, 1992-2013

Tuesday, 15 December 2015
Poster Hall (Moscone South)
Cuizhen (Susan) Wang, University of South Carolina Columbia, Columbia, SC, United States and Linlin Lu, Institute of Remote Sensing and Digital Earth, Chinese Academy of Sciences, Key Laboratory of Digital Earth Science, Beijing, China
Abstract:
The Southeast U.S. is listed one of the fastest growing regions by the Census Bureau, covering two of the eleven megaregions of the United States (Florida and Piedmont Atlantic). The Defense Meteorological Satellite Program (DMSP)’s Operational Line-scan System (OLS) nighttime light (NTL) imagery offers a good opportunity for characterizing the extent and dynamics of urban development at global and regional scales. However, the commonly used thresholding technique for NTL-based urban land mapping often underestimates the suburban and rural areas and overestimates urban extents. In this study we developed a novel approach to estimating impervious surface area (ISA) by integrating the NTL and optical reflectance data. A geographically weighted regression model was built to extract ISA from the Vegetation-Adjusted NTL Urban Index (VANUI). The ISA was estimated each year from 1992 to 2013 to generate the ISA time series for the U.S. Southeast region. Using the National Land Cover Database (NLCD) products of percent imperviousness (2001, 2006, and 2010) as our reference data, accuracy assessment indicated that our approach made considerable improvement of the ISA estimation, especially in suburban areas. With the ISA time series, a nonparametric Mann-Kendall trend analysis was performed to detect hotspots of human settlement expansion, followed by the exploration of decennial U.S. census data to link these patterns to migration flows in these hotspots. Our results provided significant insights to human settlement of the U.S. Southeast in the past decades. The proposed approach has great potential for mapping ISA at broad scales with nightlight data such as DMSP/OLS and the new-generation VIIRS products. The ISA time series generated in this study can be used to assess the anthropogenic impacts on regional climate, environment and ecosystem services in the U.S. Southeast.