State Key Laboratory of Water Environment Simulation, School of Environment, Beijing Normal University, Beijing, China.
Int J Environ Res Public Health. 2013 Jun 20;10(6):2578-95. doi: 10.3390/ijerph10062578.
Reducing human vulnerability to chemical hazards in the industrialized city is a matter of great urgency. Vulnerability mapping is an alternative approach for providing vulnerability-reducing interventions in a region. This study presents a method for mapping human vulnerability to chemical hazards by using clustering analysis for effective vulnerability reduction. Taking the city of Shanghai as the study area, we measure human exposure to chemical hazards by using the proximity model with additionally considering the toxicity of hazardous substances, and capture the sensitivity and coping capacity with corresponding indicators. We perform an improved k-means clustering approach on the basis of genetic algorithm by using a 500 m × 500 m geographical grid as basic spatial unit. The sum of squared errors and silhouette coefficient are combined to measure the quality of clustering and to determine the optimal clustering number. Clustering result reveals a set of six typical human vulnerability patterns that show distinct vulnerability dimension combinations. The vulnerability mapping of the study area reflects cluster-specific vulnerability characteristics and their spatial distribution. Finally, we suggest specific points that can provide new insights in rationally allocating the limited funds for the vulnerability reduction of each cluster.
减少工业化城市中人类对化学危害的脆弱性是当务之急。脆弱性绘图是为该地区提供减少脆弱性干预的替代方法。本研究提出了一种通过聚类分析来进行化学危害脆弱性绘图的方法，以实现有效的脆弱性降低。以上海市为研究区域，我们使用接近模型来衡量人类对化学危害的暴露情况，并考虑危险物质的毒性，使用相应的指标来捕捉敏感性和应对能力。我们使用遗传算法对改进的 k-均值聚类方法进行了改进，基本空间单元采用 500 m×500 m 的地理网格。通过平方和误差和轮廓系数的组合来衡量聚类的质量，并确定最佳聚类数量。聚类结果揭示了六组典型的人类脆弱性模式，这些模式表现出明显的脆弱性维度组合。研究区域的脆弱性绘图反映了特定聚类的脆弱性特征及其空间分布。最后，我们建议一些具体的要点，可以为每个聚类的脆弱性降低提供合理分配有限资金的新见解。
