The VIIRS nighttime lights dataset constitutes progress in the measurement of night lights radiance, with monthly data at a pixel of roughly 0.5km × 0.5km. We identify a downward bias in the reported radiance when the number of cloud-free images in a month is low. This bias often takes on large values from -10% to -30%. We develop a cautious bias-correction scheme which partially addresses this problem. This scheme is applied upon the pixel-level dataset to create an improved dataset. The bias-corrected data hews closer to the ground truth as seen in household survey data.
Citation: But clouds got in my way: Bias and bias correction of VIIRS nighttime lights data in the presence of clouds, Ayush Patnaik, Ajay Shah, Anshul Tayal, Susan Thomas, XKDR Working Paper 7, October 2021.