Introduction

These archives contain data and results used in the submitted manuscript titled "Urban light pollution in local climate zones: a comparative study for global cities". There are several (sub)folders with various data. In version 2.03, the published paper was added to the repository (1-s2.0-S026427512501025X-main.pdf).

A. Archive results-LCZs.zip

1) 01_Abs-values: Absolute values of the nighttime radiance (NTR) in delineated local climate zones preceded by an R script in PNG and SVG format (boxplots).

2) 02_Differences: Absolute and relative differences of the nighttime radiance (NTR) between the analysed years and the reference year. An R script produced an output script that generated boxplots in PNG and SVG formats. Note that some relative differences should be misleading (e.g., in the case of minimal values of NTR, the relative increase can be enormous).

3) 03_Cities: Individual results for all the cities, including all boxplots, GIS layers (GeoTiff), maps (PNG), and tables (CSV). Some cities have different names from those used in the manuscript (e.g., Duisburg-Dortmund vs. the Ruhr Region). This folder also contains the LCZ layer.

B. Archive CityAreas.zip

A set of 32 metropolitan areas was selected based on population rankings for each continent. The spatial extent of these areas was based on the Global urban and rural settlement dataset (Liu et al., 2024). The second step was evaluation using the first version of the Global Urban Polygons and Points Dataset (GUPPD)(CEISIN Columbia University and JRC European Commission, 2024). Finally, each selected polygon extent was manually extended to encompass its surroundings using satellite images to account for the effects of (sub-)urbanisation.

C. Archive results-Cities-CorrPlots.zip

The difference between the zones was tested using analysis of variance (ANOVA) and the Tukey HSD test. Evaluating the Tukey HSD test, the difference between individual classes was used as a 'hit' indicator in the correlation matrix ('hit' means statistically significant difference). In most cases, 'hits' were identified in the majority of the LCZ combinations. Note that the number of grid cells for each LCZ needs to be considered.

D. Archive results-table-LP_2012-2023.zip

Aggregated table for all the analysed cities. It contains the following columns: LCZ_code (Integer) with the code of the grid's LCZ, City (String) name, Group (String), and Y2012 (up to Y2023)(Decimal) with NTR values. Note that the first two rows are identical; it is not an error (the surface is practically the same).