遇见数据集

EVAonline: An open-source web platform for global reference evapotranspiration estimation via multi-source data fusion

收藏
Zenodo2026-04-16 更新2026-05-26 收录
官方服务:

资源简介:

This dataset provides the complete validation results and reproducibility package for EVAonline, an open-source system designed for high-accuracy reference evapotranspiration (ETo) estimation using multi-source data fusion. 📦 What's Included 186,287 daily observations spanning 30 years (1991–2020) across 17 Brazilian cities (MATOPIBA region + Piracicaba/SP). Four data sources: Reference (ground truth): Xavier et al. BR‑DWGD (0.1° resolution). Reanalysis A: NASA POWER (MERRA-2, 0.5° resolution). Reanalysis B: Open‑Meteo Archive (ERA5‑Land, ~9 km resolution). Fused product: EVAonline Adaptive Kalman Fusion results. 📂 Repository Structure 📁 data/ - Raw inputs, intermediate files, and final figures. 📁 notebooks/ - Demonstration notebooks for each provider (NASA, Open‑Meteo, MET Norway, NWS) and utility examples 📁 scripts/ - API adapters, Kalman fusion core, FAO‑56 Penman‑Monteith engine, and sequential pipeline (scripts 1 to 7) to reproduce all manuscript results. 📄 Root Files - complete_validation_analysis.ipynb (main entry point), tutorial_full_pipeline.ipynb (step‑by‑step guide using Piracicaba/SP), environment.yml and requirements.txt (Python 3.12 environment). 🔬Methodology Notes Input vs. output distinction: open_meteo_raw and nasa_power_raw contain raw meteorological variables (temperature, humidity, wind, radiation) used as inputs to EVAonline’s FAO‑56 PM engine. eto_open_meteo contains ETo values pre‑calculated by the Open‑Meteo service, used for benchmarking. Validation Scope: Focuses on the mathematical core (Kalman filter, weighted fusion, FAO‑56) – not the full web platform (FastAPI, Celery, Redis). Climatological rigor: 30‑year period (1991–2020) follows WMO guidelines for climate normals. 📈Scientific Performance EVAonline significantly outperforms individual global sources: Metric EVAonline Nasa Power Open-Meteo MAE (mm/d) 0.42 0.84 0.86 PBIAS (%) +0.71 +15.78 +13.02 KGE 0.81 0.41 0.43 MAE reduction: 38.7–50.8% Bias reduction: 91.4–95.5% 📄 License and Citation Data license: Creative Commons Attribution 4.0 International (CC BY 4.0). Software license: GNU AGPL‑3.0 (see GitHub repository). Recommended citation: Cunha Soares, Â. S. M., Ribeiro, V., Duarte, S., PERRELLA BALESTIERI, J. A., Padovani, C. R., Bordignon, Á. J. Z., Maciel, C., & Marques, P. (2026). EVAonline: An open-source web platform for global reference evapotranspiration estimation via multi-source data fusion (1.0.1) [Data set]. Zenodo. https://doi.org/10.5281/zenodo.19339254 🔗 Links Live repository: https://github.com/angela-cunha-soares/EVAONLINE Associated article: Submitted to Computers and Electronics in Agriculture (2026)

本数据集为EVAonline提供完整的验证结果与可复现性套件。EVAonline是一款基于多源数据融合实现高精度参考蒸散量(reference evapotranspiration, ETo)估算的开源系统。 📦 数据集包含内容 涵盖巴西17座城市(MATOPIBA区域及巴西圣保罗州皮拉西卡巴市)1991年至2020年共30年的186287条日度观测数据。 四类数据源: 1. 参考真值(ground truth):Xavier等人提出的BR-DWGD数据集,空间分辨率为0.1°。 2. 再分析数据A:NASA POWER(MERRA-2),空间分辨率为0.5°。 3. 再分析数据B:Open-Meteo档案馆数据(ERA5-Land),空间分辨率约9 km。 4. 融合产品:EVAonline自适应卡尔曼融合结果。 📂 仓库结构 📁 data/ 目录:存放原始输入文件、中间产物与最终可视化图表。 📁 notebooks/ 目录:包含各数据源(NASA、Open-Meteo、MET Norway、美国国家气象局NWS)的演示脚本与实用工具示例。 📁 scripts/ 目录:包含API适配器、卡尔曼融合核心模块、FAO-56彭曼-蒙特斯(Penman-Monteith)计算引擎,以及用于复现论文全部结果的1至7号顺序执行流水线脚本。 根目录文件:complete_validation_analysis.ipynb(主入口脚本)、tutorial_full_pipeline.ipynb(基于皮拉西卡巴市的全流程分步指南)、environment.yml与requirements.txt(适配Python 3.12的环境配置文件)。 🔬 方法学说明 ### 输入与输出区分 open_meteo_raw与nasa_power_raw包含原始气象变量(气温、湿度、风速、辐射),作为EVAonline的FAO-56彭曼-蒙特斯引擎的输入数据;eto_open_meteo包含由Open-Meteo服务预先计算得到的ETo值,用于基准对比测试。 ### 验证范围 本次验证聚焦于系统的数学核心模块(卡尔曼滤波、加权融合、FAO-56算法),而非完整的Web平台(FastAPI、Celery、Redis组件)。 ### 气候学严谨性 采用1991年至2020年的30年周期,符合世界气象组织(World Meteorological Organization, WMO)关于气候基准期的规范要求。 📈 科学性能表现 | 评估指标 | EVAonline | NASA POWER | Open-Meteo | | :---: | :---: | :---: | :---: | | 平均绝对误差(mm/d,Mean Absolute Error, MAE) | 0.42 | 0.84 | 0.86 | | 相对偏差百分比(Percent Bias, PBIAS) | +0.71 | +15.78 | +13.02 | | Kling-Gupta效率系数(KGE) | 0.81 | 0.41 | 0.43 | 平均绝对误差降低幅度:38.7%~50.8%;相对偏差降低幅度:91.4%~95.5%。 📄 许可与引用规范 数据许可协议:知识共享署名4.0国际许可协议(Creative Commons Attribution 4.0 International, CC BY 4.0)。 软件许可协议:GNU AGPL-3.0协议(详见GitHub仓库)。 推荐引用格式:Cunha Soares, Â. S. M., Ribeiro, V., Duarte, S., PERRELLA BALESTIERI, J. A., Padovani, C. R., Bordignon, Á. J. Z., Maciel, C., & Marques, P. (2026). EVAonline: An open-source web platform for global reference evapotranspiration estimation via multi-source data fusion (1.0.1) [数据集]. Zenodo. https://doi.org/10.5281/zenodo.19339254 🔗 相关链接 在线仓库:https://github.com/angela-cunha-soares/EVAONLINE 关联论文:已提交至《Computers and Electronics in Agriculture》(2026年)

提供机构:
Zenodo
创建时间:
2025-12-02
二维码
社区交流群
二维码
科研交流群
商业服务