rotcore-doi v1.3.9: Dataset for "Geometric Consistency Limits of Single-Core Vortex Descriptions in Turbulent Flows"
收藏资源简介：
Description-----------This dataset (rotcore-doi v1.3.9) accompanies the manuscript“Geometric Consistency Limits of Single-Core Vortex Descriptions in Turbulent Flows”.It provides all sample-level radii, robust error summaries, and identity-residualstatistics used in the paper, together with minimal code to regenerate the inlinetables and figures. Compared to earlier versions (v1.3.6–v1.3.8), v1.3.9 recomputes all per-case/per-methodsummary metrics (rel_med, MAE, MAPE, Spearman ρ, etc.) directly from the currentsample-level table `data/metrics_long_v1.3.9.csv`. The summary tables used in thefinal manuscript (Tables 2–3, S1–S2, S5–S6) are now consistent with this file.Previous single-snapshot audit tables from v1.3.6 and intermediate v1.3.8 summariesare retained under `legacy/` inside the archive for historical reference anddiagnostic checks, but they are no longer used to define the main performance metrics. Content overview----------------- Four internal incompressible Navier–Stokes regimes: (i) a high-resolution tropical-cyclone (Typhoon) simulation, (ii) mesoscale ocean eddies (Ocean), (iii) a strong multi-vortex tropical-cyclone surrogate (TC MultiVortex), (iv) forced–dissipative two-dimensional turbulence (2D Turbulence). - Three independent vortex segmenters embedded in the rotational core capacity identity (RCCI): λ_ci (swirling strength), a Γ₂-based circulation proxy (“Γ₂-proxy”), and a PH-like multi-level 0D persistence method. - A-pipeline core summaries (TRIAD tuple L_core, φ, s, N_*), B-pipeline radii, and the capacity-based predictions r_pred for every sample. - Sample-level CSV file: `data/metrics_long_v1.3.9.csv` containing (r_pred, r_meas), regime/method labels, and metadata for all samples. Legacy single-snapshot tables from v1.3.6 and v1.3.8 are stored under `legacy/` (e.g. `legacy/metrics_long_v1.3.8.csv`, `legacy/triad_v136.csv`) and are provided only for reproducibility of earlier audit runs. - Aggregated CSV tables for summary metrics: `data/triad_v139.csv`, `data/Aprime_ext_v139.csv`, `data/metrics_aux_v139.csv`, `data/metrics_robust_v139.csv`, `data/coverage_v139.csv`, `data/identity_resid_v139.csv`, `data/shape_stats_v139.csv`, together with corresponding legacy versions under `legacy/`. - A minimal Python environment file and small scripts that reconstruct all tables and the identity scatter plot directly from the CSV inputs. - A machine-readable metrics specification `docs/METRICS_SPEC_v1.3.9.json` documenting the exact definitions of all error metrics, coverage measures, and identity-residual diagnostics. Reproducibility---------------Using the provided environment and scripts, a reader can reproduce every number inthe tables of the manuscript (including Supplementary Tables S1–S6) and regeneratethe RCCI identity plot without access to the original DNS solver or raw velocity fields.The dataset is intended as a stable, append-only archive of the numerical audit: newversions add files or recomputed tables but do not remove earlier snapshots, which arekept under `legacy/` for transparency. Recommended citation--------------------If you use this dataset, please cite it as: Y. J. Lee, “rotcore-doi v1.3.9: Dataset for ‘Geometric Consistency Limits ofSingle-Core Vortex Descriptions in Turbulent Flows’”, Zenodo,doi:10.5281/zenodo.17690547.
数据集说明 本数据集（rotcore-doi v1.3.9）配套于学术论文《Geometric Consistency Limits of Single-Core Vortex Descriptions in Turbulent Flows》（湍流中单核心涡旋描述的几何一致性极限）。本数据集提供了论文中使用的全部样本级半径、稳健误差汇总结果与身份残差统计量，同时附带用于复现文中表格与插图的极简代码。 相较于此前版本（v1.3.6~v1.3.8），v1.3.9版本直接基于当前样本级数据表`data/metrics_long_v1.3.9.csv`重新计算了所有按案例/按方法的汇总指标（rel_med、平均绝对误差（MAE）、平均绝对百分比误差（MAPE）、斯皮尔曼ρ（Spearman ρ）等）。最终论文中使用的汇总表（表2~表3、补充表S1~S2、S5~S6）现已与该文件保持一致。v1.3.6版本的单快照审核表与v1.3.8版本的中间汇总结果现已保留在归档的`legacy/`目录中，仅用于历史参考与诊断校验，不再用于定义核心性能指标。 内容概览 - 四种内部不可压缩纳维-斯托克斯（Navier–Stokes）流场工况： (i) 高分辨率热带气旋（台风）模拟数据集 (ii) 中尺度海洋涡旋（海洋）数据集 (iii) 强多涡旋热带气旋替代模拟（多涡旋热带气旋）数据集 (iv) 强迫-耗散二维湍流（二维湍流）数据集 - 三种嵌入旋转核心容量识别（Rotational Core Capacity Identity, RCCI）框架的独立涡旋分割方法：λ_ci（旋度强度）、基于Γ₂的环流代理指标（“Γ₂代理”），以及类PH（PH-like）的多级别0维持续方法。 - A流程核心汇总结果（TRIAD元组L_core、φ、s、N_*）、B流程半径，以及所有样本基于容量的预测半径r_pred。 - 样本级CSV文件：`data/metrics_long_v1.3.9.csv`，包含（r_pred、r_meas）、工况/方法标签与所有样本的元数据。v1.3.6与v1.3.8版本的遗留单快照表现存储于`legacy/`目录下（例如`legacy/metrics_long_v1.3.8.csv`、`legacy/triad_v136.csv`），仅用于复现早期审核运行结果。 - 用于汇总指标的聚合CSV数据表：`data/triad_v139.csv`、`data/Aprime_ext_v139.csv`、`data/metrics_aux_v139.csv`、`data/metrics_robust_v139.csv`、`data/coverage_v139.csv`、`data/identity_resid_v139.csv`、`data/shape_stats_v139.csv`，同时在`legacy/`目录下提供对应的遗留版本。 - 一份极简Python环境配置文件与小型脚本，可直接基于CSV输入文件重构所有表格与身份散点图。 - 一份机器可读的指标规范文档`docs/METRICS_SPEC_v1.3.9.json`，其中详细说明了所有误差指标、覆盖度度量与身份残差诊断方法的准确定义。 可复现性 借助本数据集提供的环境与脚本，研究人员无需访问原始直接数值模拟（DNS）求解器或原始速度场，即可复现论文表格（含补充表S1~S6）中的全部数据，并重新生成RCCI身份识别图。本数据集旨在作为数值审核的稳定、仅追加式归档：新版本仅添加文件或重新计算汇总表，不会移除早期快照，这些快照均保留在`legacy/`目录中以保证透明度。 推荐引用方式 若使用本数据集，请按以下方式引用： Y. J. Lee, “rotcore-doi v1.3.9: Dataset for ‘Geometric Consistency Limits of Single-Core Vortex Descriptions in Turbulent Flows’”, Zenodo, doi:10.5281/zenodo.17690547.
