Dataset, scripts, and figures for the article: "Machine-learning emulation of SBDART spectral transmittance and reflectance under turbid and cloudy skies: physical interpretability and rapid sensitivity analysis"

General description This repository brings together the data, Python scripts, tables, and figures used for the study devoted to the machine-learning emulation of spectral transmittance and reflectance simulated by SBDART under turbid and cloudy skies. The repository includes in particular: • the database generated from large-scale SBDART radiative simulations; • the scripts for data generation, post-processing, statistical analysis, and machine learning; • the descriptive tables used in the methodology; • the main figures of the article; • the SHAP interpretability figures; • the rapid sensitivity figures in the aerosol–cloud parameter space.概述 本仓库汇集了一项研究所用的全部数据、Python脚本、表格与图像，该研究聚焦于浑浊与多云天空下由SBDART模拟的光谱透射率与反射率的机器学习仿真任务。 本仓库具体包含以下内容： • 基于大规模SBDART辐射模拟生成的数据库； • 用于数据生成、后处理、统计分析与机器学习的脚本； • 研究方法章节所用的描述性表格； • 论文主体配图； • SHAP可解释性分析图； • 气溶胶-云参数空间内的快速敏感性分析图。