The Annual Visible Night Light (VNL) V2 (VIIRS) images at 500-m spatial resolution for the period 2012 to 2024 (Elvidge et al., 2021) have been used to extrapolate the values backwards for years 2000–2011. This was done by fitting a logistic regression (per pixel) and then predicting the values for the previous years (see nightlights_stack_500m.R). After consistent time-series have been produced, I also derived the difference between year 2024 and year 2000 (nightlights.difference_viirs.v21_m_500m_s_2000_2024_go_epsg4326_v20230318.tif): this shows average rate of change for the 25 years period. Use with caution: extrapolation of values can lead to artifacts. For most of the land surface, however, it appears that the growth of night lights follows exponential growth function and hence nights in the past can be represented accurately by fitting decay / logistic regression function. Original values from the Annual VNL V2 product have been converted from 0–200 to 0–2000 scale and are available as Cloud-Optimized GeoTIFFs. Principal components (PC1, PC2, PC3, PC4) were derived using SAGA GIS (sums-of-squares-and-cross-products matrix) method. The first PC1 usually matches the long-term mean value, PC2 matches the 1st derivation in values. File "nightlights_dmsp.v10_m_1km_s_19920101_20241231_go_epsg4326_v20251006.tif" contains 33 years 1992 to 2024, but at 1 km resolution. To cite the Annual VNL V2, please use: Elvidge, C. D., Zhizhin, M., Ghosh, T., Hsu, F. C., & Taneja, J. (2021). Annual time series of global VIIRS nighttime lights derived from monthly averages: 2012 to 2019. Remote Sensing, 13(5), 922. https://doi.org/10.3390/rs13050922 Historic night light images (1 km resolution) are also available from Figshare: Li, X., Zhou, Y., Zhao, M., & Zhao, X. (2020). A harmonized global nighttime light dataset 1992–2018. Scientific data, 7(1), 168. https://doi.org/10.1038/s41597-020-0510-y