Data archive for "Duplicate-informed allocation of stations and subsamples for petroleum hydrocarbon monitoring in the northern South China Sea" (v1.0)

This deposit is the data part of a paired Data and Code archive supporting the manuscript "Duplicate-informed allocation of stations and subsamples for petroleum hydrocarbon monitoring in the northern South China Sea" (v1.0, prepared 28 September 2026). How to use it Download this record together with the companion Software record (https://doi.org/10.5281/zenodo.23054555), which is linked from this record under Related works. Extract both ZIPs into the same parent directory, keeping the shared top-level folder Northern_SCS_Oil_v1.0; their contents merge into analysis/ and figures/. Do not add an extra Data/ or Code/ level. Run commands from that shared top-level folder. Contents (214 files) - figures/data/ : reconstructed laboratory records, the 76-location analytical population, the processed satellite slick catalogue, covariates and prediction grids (16 files, plus 2 pre-correction copies under _pre_correction_backup/) - figures/tables/ : analytical results, uncertainty parameters, network-design comparisons and sensitivity outputs (88 files) - figures/source_data/ : plot-level values for each figure panel (34 files) - figures/figures/ : archived rendered panels plus figure_map.json (69 files) - root : COLUMN_INVENTORY.csv , Data_FILE_MANIFEST.csv , Data_README.md , LICENSE.txt , SOURCE_MAPPING.md Units and conventions Concentrations are mg/L; log concentrations and precision parameters use log10. CSV index columns are identifiers, not measurements. Missing values are not zero. The 76-location population and the 43 duplicate pairs are not interchangeable sample sizes: duplicate subsamples are separately processed portions of the same collected water, not independent field replicates. Scope The archive starts from extracted laboratory records; it is not an instrument-raw-data deposit. External products and published comparison data retain their source attribution and provider conditions. Provenance manifests preserve historical processing details, including local paths that may not exist on another machine; portable fallback inputs are included for laboratory reconstruction and replication mapping. Verification boundary The replication map, the bias diagnostics and the new allocation sensitivity analysis were rerun for this version; independent simulation checks were made for the principal m=1/m=4 classification results, and Python syntax was checked for all 58 scripts. Heavy network/model fits and the figure scripts were not rerun in this packaging pass: their archived outputs are supplied and are not represented as newly regenerated. The new sensitivity results supplement, rather than overwrite, the original Monte Carlo tables. Corrections figures/data/CORRECTIONS.md records each value changed after the analytical files were first assembled, the evidence for the change and everything downstream affected; pre-correction copies are kept in figures/data/_pre_correction_backup/. Geographic coverage Northern South China Sea, 109.05-113.78 E, 15.67-22.76 N. Licence CC BY 4.0. Please cite this record and the companion Software record when reusing the data.