Urban vegetation deficits near earthquake-damaged buildings in Kahramanmaraş from 2023 to 2026: Data and Code

This reproducibility archive accompanies the manuscript “Urban vegetation deficits near earthquake-damaged buildings in Kahramanmaraş from 2023 to 2026” The study evaluates post-earthquake urban vegetation trajectories in Kahramanmaraş, Türkiye, using a within-city spatial counterfactual design that compares vegetation near severely damaged buildings with geographically and environmentally comparable donor locations. The archive provides the processed analytical data, scripts, model outputs, validation results, and documentation required to reproduce and independently verify the principal analyses reported in the manuscript and its Supplementary Materials. The core balanced panel contains 946 analysis cells—285 exposed cells and 661 donor cells—observed over 109 retained half-month periods. It includes NDVI and NDMI outcomes, spatial coordinates, exposure and donor classifications, geographic-block identifiers, and 18 pre-event environmental and urban-form covariates. The package includes: - processed Sentinel-2 vegetation-index panels and candidate-cell inventories;- the independently reconstructed 2026 Sentinel-2 extension, including 12 April–September half-month composites and an 11-period season-matched panel;- generalized synthetic-control model inputs and outputs;- joint 2 km geographic-block bootstrap procedures and stored bootstrap summaries;- exposed-, donor-, shared-, and union-block audits;- Landsat land-surface-temperature model outputs, held-out 2022 validation results, alternative-model comparisons, and 2020–2022 pseudo-intervention tests;- canopy-height provenance records for the Meta/WRI High Resolution Canopy Height Map v1 product;- Sentinel-2 acquisition inventories, missing-period documentation, reflectance-scaling rules, and processing-order records;- software-environment specifications, fixed random seeds, run-order instructions, and an automated package verifier; and- a machine-readable manifest, checksum records, and citation metadata. The Sentinel-2 processing workflow documents the acquisition-level reflectance scaling, spectral-index calculation, cloud and shadow filtering, spatial-completeness requirements, and half-month compositing procedures. The 2026 extension was reconstructed independently from public Sentinel-2 observations to verify the partial-season estimates and the season-matched 2026–2023 comparisons. Spatial inference is implemented through a joint block-resampling procedure. A single bootstrap multiplicity is drawn for each block in the union of exposed and donor blocks and is applied simultaneously to both groups. This preserves the dependence created by shared geographic blocks. The archive records 18 exposed blocks, 22 donor blocks, 17 shared block identifiers, and 23 blocks in the union. The thermal-analysis component contains the retained Landsat observation panels, generalized synthetic-control estimates, held-out validation results, pseudo-intervention tests, and block-bootstrap uncertainty estimates for both the full-cell and baseline-vegetation samples. The archive also documents that the vegetation–temperature coupling regressions treat the estimated first-stage vegetation gaps as fixed and therefore do not propagate first-stage estimation uncertainty. Large public-source datasets are not unnecessarily duplicated. In particular, the complete 2018–2025 Sentinel-2 image stacks and the original Meta/WRI canopy-height tile are not included. Source identifiers, acquisition records, public URLs, and processing scripts are supplied so that these inputs can be retrieved or reconstructed from their original providers. The final 2026 verification composites used in the revision audit are included in the archive. To verify the archive after downloading it, extract the ZIP file while preserving its directory structure and follow “ZENODO_USE_INSTRUCTIONS_EN.txt” and “RUN_ORDER.txt.” The included automated verifier checks required files, expected data dimensions, numerical summaries, package integrity, and selected manuscript–supplement consistency conditions. This archive is intended to support transparent assessment, computational verification, and reuse of the study workflow. Users should consult the manuscript and Supplementary Materials for the complete study design, interpretation of the estimates, assumptions, and limitations.