Building damage datasets for the 2023 Türkiye earthquake

This comprehensive dataset was compiled to support research on building damage identification and prediction following the February 2023 Türkiye earthquakes. It integrates pre-seismic building inventories, multi-source post-disaster damage assessments, and key geo-environmental influencing factors. The repository is structured around three core components: (1) GBA_building_footprint: This folder contains a high-resolution Shapefile (Turkey_GBA_building_data) representing the building footprint polygons for the affected regions. (2) Building_damage_data: This folder aggregates building damage evaluations from five distinct sources, facilitating comparative studies and data fusion. It includes ARIA_DPM, damage proxy maps derived from NASA's ARIA project using Sentinel-1 SAR data to indicate ground deformation; UNOSAT, expert-based damage grading (e.g., destroyed, major damage) provided by the UN Operational Satellite Applications Programme in GIS vector format; Microsoft_team, AI-generated or crowdsourced damage assessments from Microsoft's disaster response collaborations; Visual_interpretation, manually annotated damage labels serving as a high-confidence validation dataset; and This_study, containing building damage results obtained using the methodology developed in this study. (3) Influencing_factors: This crucial folder provides raster layers (.tif) of key variables that influence building damage patterns, enabling analysis and prediction of damage distribution. It includes DEM.tif, a digital elevation model essential for assessing topographic effects; Epicenter.tif, representing the distance from each pixel to the main earthquake epicenter as a primary proxy for shaking intensity; Fault.tif, indicating proximity to active faults, a critical factor in ground motion; Lithology.tif, depicting geological or lithological units that strongly affect shaking amplification and liquefaction potential; and PGV.tif, peak ground velocity, a key measured or modeled ground motion parameter directly linked to seismic energy and potential structural damage.本综合数据集旨在为2023年2月土耳其地震后的建筑损伤识别与预测相关研究提供支撑。数据集整合了震前建筑清单、多源灾后损伤评估数据以及关键地质环境影响因子。 本数据集仓库包含三大核心模块： (1) GBA_building_footprint：该文件夹包含高分辨率Shapefile格式文件Turkey_GBA_building_data，用于表征受影响区域的建筑足迹多边形。 (2) Building_damage_data：该文件夹整合了5个不同来源的建筑损伤评估结果，可为对比研究与数据融合提供支持，具体包含以下五类数据： ARIA_DPM：由美国国家航空航天局（NASA）ARIA项目基于Sentinel-1合成孔径雷达（SAR）数据生成的损伤代理地图，用于表征地面形变； UNOSAT：由联合国业务卫星应用方案提供的基于专家经验的损伤分级结果（例如损毁、严重损毁），采用地理信息系统（GIS）矢量格式存储； Microsoft_team：由微软灾害响应合作项目产出的AI生成或众包形式的损伤评估数据； Visual_interpretation：经人工标注的损伤标签数据，可作为高置信度的验证数据集； This_study：包含基于本研究开发的方法所得到的建筑损伤评估结果。 (3) Influencing_factors：该关键文件夹提供了影响建筑损伤分布模式的关键变量栅格图层（.tif格式），可用于损伤分布的分析与预测研究，具体包含以下图层： DEM.tif：数字高程模型，是评估地形效应的核心数据； Epicenter.tif：表征各像素至主震震中的距离，作为地震动强度的主要代理指标； Fault.tif：表征距活动断层的远近，是影响地震动的关键因素； Lithology.tif：展示地质或岩性单元分布，对地震动放大效应与液化潜力具有显著影响； PGV.tif：峰值地面速度，是关键的实测或模拟地震动参数，与地震能量及潜在建筑损伤直接相关。