Reproduction package and derived data for "Rooftop photovoltaic potential and carbon offsets for building-stock carbon peaking: A building-by-building assessment of Qingdao, China"

Overview. This deposit contains the documentation, supplementary tables and reproduction package supporting the manuscript "Rooftop photovoltaic potential and carbon offsets for building-stock carbon peaking: A building-by-building assessment of Qingdao, China" (submitted to Renewable Energy). The study assesses rooftop photovoltaic (PV) potential for 2,229,005 height-attributed buildings in Qingdao (534.3 km² of roof footprint) and couples the resulting hourly generation to a building-stock turnover carbon-peaking model. Contents (104 files, 5.6 MB). The archive is organised in seven numbered groups. – Root (7 files) — SI1_Supplementary_Documentation.pdf and 00_Supplementary_Material_Index.md give the packaging map, the recomputation specification for the 168-row hourly absorption matrix (SI-B), the data dictionary, host-model input notes and script notes; also included are the SI label mapping, the package README, an upload note and restore_script_inputs.py. – 01_Supplementary_A_Methods (1 file) — SI-A: model equations and parameterisation. – 02_Supplementary_B_Hourly_Matrix (3 files) — the 168-row SI-B absorption and avoided-emission matrix (3 scenarios × 7 load shapes × 8 absorption settings), its recomputation specification and stage35_si_b_168.py. – 03_Supplementary_C_Coupling_Scenarios (5 files) — SI-C1 coupling sensitivity grid (216 rows), SI-C4 annual trajectories 2015–2050 and peaking summaries, SI-C5 carbon-factor and host-baseline sensitivity. – 04_Supplementary_D_Code_and_Data (68 files) — Code_and_Metadata/ (24 files: citywide summary, historical distributed-PV capacity model, host-model inputs, hourly NASA POWER inputs for three representative points, the citywide hourly generation series 2023–2025, the machine-readable manifest, the OpenStreetMap ingestion path and a requirements lock); Reproduction_Package/ (39 files: 24 script files under scripts/, of which 22 are numbered stage scripts covering the chain from building footprints to available roof area, installed capacity, hourly generation, integration correction and carbon-peaking coupling, plus the OSM deviation pathway, reference_pipeline.py, the verification scripts, data dictionaries, a de-identified per-building scalar table for the Shibei District pilot (42,259 buildings) and a 200-building synthetic validation set); SI-D1 district-level capacity and shaded generation; SI-D5 OSM-vs-AI footprint deviation. – 05_Supplementary_E_Validation (6 files) — SI-E1 stratified footprint-match sample, SI-E2 match summary, SI-E3 Sobol indices with convergence series (D = 10) and the 216-row structural scenarios, SI-E4 output distribution from N = 1024 Saltelli sampling (12,288 evaluations), and the validation note. – 07_Figure_S1_Block_Shading_Validation (3 files) — Figure S1, the block-scale validation of the 12-sector solar-trajectory shading method against a 300 m raster DSM, with its per-building and summary tables. – 08_Raw_Data_and_Sensitivity_Tests (11 files) — height-error propagation (7 tables), load-shape sensitivity, LCOE screening and storage/weather-year sensitivity. Reproducibility. The scripts read only the bundled inputs; no proprietary geometry and no environment variables are required. Verification entry points: verify_bundled_inputs.py (7 checks), verify_si_b.py (11 checks) and reference_pipeline.py --validate (13 checks). Recomputation of the 168-row SI-B matrix from the shipped inputs reproduces the shipped curtailment rates to within 0.1 percentage point on average (162 of 168 rows identical cell by cell; maximum deviation of deliverable energy 10.9 GWh, ≤ 0.03%). Restrictions. The complete building geometry (2,229,005 footprints with heights) is not included: it originates from a licensed commercial remote-sensing product that permits publication of aggregate results but prohibits redistribution of the raw geometry. It is available only under a signed agreement (tier L3 of the data-availability statement in the manuscript). An open-source alternative pathway (OpenStreetMap + NASA POWER) is provided so that the method chain can be reproduced in any city. Licence and reuse. Documentation and derived data: CC BY 4.0. Code: MIT licence, with the full reproduction package included in this deposit under 04_Supplementary_D_Code_and_Data/. Please cite this deposit and the associated article.