首页|期刊导航|中国城市林业|从平面到立体:景观格局对风灾的影响

从平面到立体:景观格局对风灾的影响OA

From Planar to Three-Dimensional:Analyzing the Impact of Landscape Patterns on Wind Disasters via Machine Learning

中文摘要英文摘要

[目的]比较二维与三维景观格局对城市大风灾情的解释能力差异,识别关键影响因子及其尺度效应,为沿海高密度城市风灾防控提供量化依据.[方法]以上海市为研究区,构建涵盖二维和三维特征的景观格局指标体系;采用多尺度对比分析,结合皮尔逊相关分析与随机森林模型,系统识别影响大风灾情的关键景观格局指标,阐明其尺度依赖性与空间分异规律.[结果]三维指标对大风灾情的整体解释力优于二维指标,其中占空度(VD)贡献最高,二维指标中建筑密度(BD)贡献最突出;"高占空度+高建筑密度"组合对应更高的边际贡献,高程对VD效应具有调节作用;识别出4 km为最优分析尺度,确定VD>0.45、BD>0.35 为关键风险阈值.[结论]大风灾情的空间分异主要受城市三维景观格局影响,且其影响具有尺度效应、非线性特征及因子协同作用.三维景观格局对风灾风险的主要影响作用有助于推动城市防灾规划从平面向立体转型,为沿海高密度城市的韧性空间规划提供科学依据.未来可引入风场模拟并开展多城市比较研究,进一步验证研究的普适性,以深化对景观格局影响风灾风险机制的理解.

[Objective]To compare the explanatory differences between two-dimensional and three-dimensional landscape patterns for urban windstorm damage,identify key drivers and scale effects,and provide quantitative evidence for wind-disaster mitigation in dense coastal cities.[Method]Taking Shanghai as the study area,this study developed a landscape pattern indicator system integrating two-dimensional and three-dimensional characteristics.Multi-scale comparative analysis,Pearson correlation analysis,and random forest models were used to identify the key landscape pattern indicators affecting windstorm damage and clarify their scale dependence and spatial differentiation.[Result]Three-dimensional indicators showed stronger overall explanatory power for windstorm damage than two-dimensional indicators.Volumetric density(VD)contributed the most,while building density(BD)was the most important two-dimensional indicator.High VD combined with high BD corresponded to higher marginal contributions,and elevation moderated the effect of VD.The optimal analysis scale was 4 km,and VD>0.45 and BD>0.35 were identified as key risk thresholds.[Conclusion]The spatial differentiation of windstorm damage is mainly shaped by the urban three-dimensional landscape pattern,with effects showing scale dependence,nonlinear characteristics,and synergistic interactions among key factors.This study clarifies the main role of three-dimensional landscape patterns in windstorm risk,supports the shift of urban disaster-prevention planning from a planar to a volumetric perspective,and provides direct scientific support for resilience-oriented spatial planning in high-density coastal cities.

戴代新;刘欣妍

同济大学建筑与城市规划学院,上海 200092同济大学建筑与城市规划学院,上海 200092

三维景观格局机器学习大风灾情尺度效应

three-dimensional landscape patternmachine learningwindstorm damagescale effect

《中国城市林业》 2026 (1)

165-175,11

2025年度国家外国专家个人类项目(面向生态防灾减灾的城市绿色基础设施灾害韧性评价方法国际经验借鉴)

10.12169/zgcsly.2026.01.04.0001

评论