Measurement of blooming effect of DMSP-OLS nighttime light data based on NPP-VIIRS data

Abstract

Nighttime light (NTL) imagery offers important data for socio-economic research since it depicts human activities at night. Blooming effect is one of the main problems of DMSP-OLS NTL data, but the problem has been largely mitigated in NPP-VIIRS data. Blooming effect smooths light intensity through diffusing the brightness from the bright area to the dark areas. This effect could be quantified by the alteration of spatial heterogeneity in light intensity which can be represented by spatial texture indices. Taking NPP-VIIRS data as the standard, this study calculates relative difference of spatial texture indices between NPP-VIIRS and DMSP-OLS images to measure the blooming effect of DMSP-OLS data at pixel scale in mainland China. Results show that within the real urban area, spatial texture indices of DMSP-OLS pixels are generally smaller than that of NPP-VIIRS, suggesting that blooming effect underestimates spatial heterogeneity and it is more serious towards to urban centres where nighttime light is stronger. DMSP-OLS underestimates 90% spatial heterogeneity when NPP-VIIRS value >10. On the contrary, in the region outside real urban area, spatial texture indices of DMSP-OLS data are greater than that of NPP-VIIRS, suggesting that blooming effect introduces fake spatial heterogeneity outside the real urban area. Abbreviations: NTL: Nighttime light; DMSP: Defense Meteorological Satellite Program; OLS: Operational Linescan System; NPP: National Polar-orbiting Partnership; VIIRS: Visible Infrared Imaging Radiometer Suite; NOAA: The National Oceanic and Atmospheric Administration; NGDC: The National Geophysical Data Center; NCV: Normalized coefficient of variation; DN: Digital number.