Association Between Urbanization and Observed Warming Trends across the Korean Peninsula: A Multi-Station Analysis Controlling for Latitude and Elevation (1980–2010)

This deposit contains the analysis code and the station-level derived datasets supporting everynumber, table and figure reported in the accompanying manuscript. The study combines 1980-2010 monthly mean temperature records from 105 stations of the KoreaMeteorological Administration's ASOS network with four epochs of the Ministry of Environment's30 m national land-cover map, and tests whether the change in built-up land-cover ratio within1 km and 5 km buffers around each station is associated with that station's long-term warming rate(Theil-Sen slope, Mann-Kendall test), controlling for latitude and elevation. Contents:- scripts/01_core_pipeline (19 scripts): tile indexing, station-coordinate verification, buffer zonal statistics, Mann-Kendall / Theil-Sen trend estimation, simple and multiple regression, group comparison, Benjamini-Hochberg FDR correction, spatial-autocorrelation diagnostics, Ordinary Kriging maps, and figure generation.- scripts/02_revision_analyses (20 scripts): sensitivity analyses (station relocation, seasonal completeness, buffer radius, absolute urban/rural thresholds, fixed common-station panel), breakpoint homogeneity testing (SNHT, Pettitt), relative homogeneity testing against neighbouring stations, coordinate-perturbation Monte Carlo analysis, the regional fixed-panel anomaly series, and the 2010-2025 extension.- data/derived: regression inputs and outputs at station level.- data/figure_data: plotting data for each manuscript figure, named by the submitted figure number and split into main_text/, supplementary/ and model_output/.- data/sensitivity: complete results of all sensitivity and robustness analyses, including TableS1_station_summary_66stations.csv and TableS2_relative_homogeneity_by_station.csv, which are Supplementary Tables S1 and S2 of the manuscript.- data/tables: Moran's I, cluster-robust standard errors, BH-FDR adjusted p-values, kriging cross-validation and land-cover QC checks.- environment/requirements.txt: the fixed package versions used to produce the reported results. The two source datasets (KMA ASOS observations and MOE land-cover maps) are not redistributedhere, as they are subject to the issuing agencies' own terms; README.md documents how to obtainthem and which analyses can be reproduced from the derived data alone. All reported statistics were computed with conventional, non-AI statistical software; themanuscript's "Use of Artificial Intelligence Tools" section documents where generative AI toolswere used in the project.