A self-adjusting method to generate daily consistent nighttime light data for the detection of short-term rapid human activities

Nighttime light (NTL) has become an emerging indicator of the magnitude and changes in human settlement and activities. A recently released Lunar-BRDF-corrected NTL product, Black Marble VNP46A2, with enhanced temporal (daily) and spatial (15 arc-second) resolutions, has the potential to monitor human reactions to shortterm events with rapid light intensity variations. However, the VNP46A2 NTL exhibits unexpected large daily variations resulting from two residual errors (spatial observational coverage mismatch and angular effect), leading to difficulty in extracting true change signals of NTL time series. In this study, we developed a Selfadjusting method featuring Filter and Angular effect Correction (SFAC) to eliminate the two errors without any auxiliary data. The proposed adjusted-average filtering effectively reduced the mismatch error and enhanced data consistency with less "blooming effect" than other filtering methods. The normalized standard deviation of daily NTL was largely reduced in the selected airport areas after filtering. The variation caused by the angular effect has also been corrected by a relative calibration method using the 16 days' cycle of satellite view angles, which apparently decreases the dependence of the light intensity on the view angles. We further explored and found that the detectability of short-term rapid events (i.e., festival/event celebration with increased NTL and power outage with reduced NTL) using SFAC-corrected VNP46A2 outperformed the original product. More detailed information on rapid events and evolutions in small spaces is expected to be digested with the highquality, SFAC-corrected daily consistent VNP46A2 NTL data.