Mesogeos
收藏资源简介:
Mesogeos是一个大规模的多用途数据集,专为地中海地区的野火建模设计。该数据集整合了代表野火驱动因素(气象、植被、人类活动)的变量以及17年(2006-2022)的历史野火点火和烧毁区域记录。Mesogeos以云友好的时空数据集形式,即数据立方体,将所有变量统一在一个1km x 1km x 1天分辨率的网格中。数据立方体结构为评估机器学习在各种野火建模任务中的应用提供了机会。数据集的创建旨在通过机器学习技术,提高对地中海地区野火发生和发展的理解和预测能力,从而增强野火预防和管理,适应不断变化的气候条件。
Mesogeos is a large-scale, multi-purpose dataset specifically designed for wildfire modeling in the Mediterranean region. This dataset integrates variables representing wildfire driving factors (meteorology, vegetation, human activities) alongside 17-year (2006–2022) historical records of wildfire ignitions and burned areas. Mesogeos unifies all variables within a grid with a resolution of 1km × 1km × 1 day, presented as a cloud-friendly spatiotemporal dataset, i.e., a data cube. The data cube structure provides opportunities to evaluate machine learning applications across diverse wildfire modeling tasks. This dataset was developed to improve understanding and predictive capabilities of wildfire occurrence and progression in the Mediterranean region using machine learning techniques, thereby enhancing wildfire prevention and management to adapt to changing climatic conditions.

- 1Mesogeos: A multi-purpose dataset for data-driven wildfire modeling in the Mediterranean国家天文台,天文、空间应用和遥感研究所,雅典国家天文台 · 2023年



