Impact of nighttime light data saturation correction on the application of carbon emissions spatialization: A comparative study of the correction effect and application effect based on five methods in China

Nighttime light (NTL) remote sensing data have attracted wide attention in monitoring carbon dioxide (CO2) emissions of fossil fuel consumption in recent years due to its unique perspective of observing human activity and low cost. However, saturation errors in NTL data are believed to limit the accuracy of the application. Therefore, this study compared the desaturation ability and the effect of spatializing CO2 emissions in China based on five saturation correction methods of Letu + Tong, SARMRC (the saturation correction method based on a regression model and radiance -calibrated NTL data), VANUI (index -based vegetation adjusted NTL urban index), EANTLI (enhanced vegetation index (EVI)-adjusted NTL index) and LERNCI (land surface temperature (LST)-and-EVIregulated-NTL-city index). Based on this, we explored the influence of saturation correction on the spatialization of CO2 emissions. The main findings of this study include: (1) The saturation correction effect is limited by the bias of the NTL image itself in some regions, the auxiliary data, and the regional differences of the fitting models, which fail to reach the best effect in spatializing CO2 emissions at the provincial level. However, as downscaling to the prefectural level, the Letu + Tong image with the advantage of desaturation has a better effect on spatializing CO2 emissions. (2) The saturation correction effect of Letu + Tong and EANTLI images is relatively better with advantages in three evaluation criteria than other images. In most provinces, the original NTL and SARMRC images are better at spatializing CO2 emissions. (3) Implementing inter -calibration was found to be generally beneficial in enhancing the effects of saturation correction. This study provides an important basis for improving the quality of NTL data and helps to reflect the spatial pattern of CO2 emissions more accurately and efficiently.