-
摘要:
受全球气候变化影响，台风风暴潮造成的损失显著增加，准确构建高效、合理的损失评估模型对海洋灾害防灾减灾工程具有重大现实意义。使用4组指标构建台风风暴潮指标体系，并通过主成分分析筛选出输入因子。采用麻雀搜索算法优化支持向量机模型对台风风暴潮损失分级和直接经济损失进行评估，与其他优化算法进行比较分析，发现该模型具有更好的预测精确性。对指标体系中的4组指标分别进行评估，得出指标的有效性大小为危险性指标>气候变化指标>易损性指标>防灾减灾能力指标，表明了该实验的合理性，为防灾减灾事业提供了有效的评估方式。
Abstract:Affected by global climate change, the losses caused by typhoon storm surge are increasing gradually. Building an accurate, efficient and reasonable loss assessment model is highly demanded for marine disaster prevention and mitigation projects. Four sets of indexes were used to construct the index system of typhoon storm surge, and the input factors were selected by principal component analysis. The sparrow search algorithm (SSA) was used to optimize the support vector machine model for loss classification and direct economic loss assessment of typhoon storm surge. Compared with other optimization algorithms, the SSA model showed better prediction accuracy. In addition, the four sets of indicators in the index system were evaluated individually, from which the order of effectiveness of them is: danger level > climate change > vulnerability > disaster prevention and mitigation capability. This study showed the rationality of the experiment and provided an effective assessment method for disaster prevention and mitigation.
-
-
表 1 台风风暴潮灾害损失等级划分标准
Table 1. Classification standard of disaster loss by typhoon storm surge
指标 等级 5（重灾） 4（大灾） 3（中灾） 2（小灾） 1（轻灾） X01受灾人口/万人 （+∞, 800] （800, 600] （600, 300] （300, 100] （100, 0] X02死亡（含失踪）人数/人 （+∞, 200] （200, 100] （100, 30] （30, 10] （10, 0] X03直接经济损失/亿元 （+∞, 100] （100, 50] （50, 20] （20, 5] （5, 0] X04农田受灾面积/千公顷 （+∞, 500] （500, 200] （200, 100] （100, 1] （1, 0] X05海水养殖受灾面积/千公顷 （+∞, 50] （50, 20] （20, 10] （10, 1] （1, 0] X06海岸工程损毁/千米 （+∞, 600] （600, 400] （400, 100] （100, 30] （30, 0] X07倒塌房屋/万间 （+∞, 20] （20, 10] （10, 5] （5, 1] （1, 0] X08船只损毁/只 （+∞, 4 000] （4 000, 2 000] （2 000, 1 000] （1 000, 300] （300, 0] 综合分级标准 （+∞, 6.09] （6.09, 2.49] （2.49, 0.17] （0.17, −1.07] （−1.07, −1.48] 
下载: 导出CSV
表 2 不同模型效果对比
Table 2. Comparison in effects of different models
算法类型 训练集正确数 测试集正确数 训练集
正确率/%测试集
正确率/%SVM 37 8 97.37 66.67 GA-SVM 37 9 97.37 75 PSO-SVM 38 8 100 66.67 SSA-SVM 38 9 100 75
下载: 导出CSV
表 3 不同模型的测试集评估效果比较
Table 3. Comparison in assessment effect of test sets of different models
算法类型 MAE MAPE NRMSE CC SVM 15.624 9 1.786 6 0.109 9 0.679 3 GA-SVM 15.185 9 0.293 9 0.087 3 0.765 7 PSO-SVM 14.776 5 1.217 0 0.094 9 0.722 1 SSA-SVM 14.153 6 0.063 3 0.018 0 0.834 9
下载: 导出CSV
表 4 不同指标类型效果比较
Table 4. Comparison in the effect of different index types
序号 指标类型 MAE MAPE NRMSE CC 1+2+3+4 14.153 6 0.063 3 0.018 0 0.834 9 一 1+2+3 11.905 2 0.062 3 0.009 1 0.881 3 二 1+2+4 18.852 0 0.103 3 0.043 1 0.737 5 三 1+3+4 19.792 8 1.503 7 0.112 0 0.047 8 四 2+3+4 15.816 0 0.417 6 0.059 6 0.644 0 五 1 19.828 9 1.443 8 0.110 3 0.113 0 六 2 14.862 3 1.178 4 0.088 6 0.696 8 七 3 19.829 2 1.453 6 0.110 4 0.133 2 八 4 19.830 4 1.449 4 0.123 3 0.110 2
下载: 导出CSV
-
[1] CHURCH J A,WHITE N J. A 20th century acceleration in global sea-level rise[J]. Geophysical Research Letters,2006,33(1):313-324.
[2] 郭腾蛟,李国胜. 基于验证性因素分析的台风风暴潮灾害经济损失影响因子优化分析[J]. 自然灾害学报,2020,29(1):121-131.
[3] BERKE P,LARSEN T,RUCH C. A computer system for hurricane hazard assessment[J]. Computers Environment and Urban Systems,1984,9(4):259-269.
[4] MARK P,GEORGE S,STEVE C,et al. State of Florida hurricane loss projection model:atmospheric science component[J]. Journal of Wind Engineering and Industrial Aerodynamics,2005,93(8):651-674.
[5] RONALD L I,MARK E J,TOM A S. Assessing hurricane effects. Part 1. Sensitivity Analysis[J]. Reliability Engineering and System Safety,2002,78(2):131-145. doi: 10.1016/S0951-8320(02)00133-3
[6] CHARLES S,PAUL F,NEIL B,et al. HAZUS-MH flood loss estimation methodology. II. Damage and loss assessment[J]. Natural Hazards Review,2006,7(2):72-81. doi: 10.1061/(ASCE)1527-6988(2006)7:2(72)
[7] GAURAV T,SUSHIL K,ASHISH R,et al. A high-resolution mesoscale model approach to reproduce super typhoon Maysak (2015) over northwestern Pacific Ocean[J]. Earth Systems and Environment,2019,3(1):101-112. doi: 10.1007/s41748-019-00086-0
[8] 孙丰霖. 基于证据理论的风暴潮灾害损失评估[J]. 热带海洋学报,2022,41(1):1-81. doi: 10.11978/2020047
[9] 江斯琦,刘强. 基于改进神经网络及地理信息系统空间分析的风暴潮经济损失评估[J]. 科学技术与工程,2020,20(22):9243-9247. doi: 10.3969/j.issn.1671-1815.2020.22.058
[10] MAHAPATRA M,RATHEESH R,RAJAWAT A S. Storm surge vulnerability assessment of Saurashtra coast,Gujarat,using GIS techniques[J]. Natural Hazards,2017,86(2):821-831. doi: 10.1007/s11069-016-2718-x
[11] BARADARANSHORAKA M,PINELLI J P,GURLEY K,et al. Hurricane wind versus storm surge damage in the context of a risk prediction model[J]. Journal of Structural Engineering,2017,143(9):04017103.1-04017103.10.
[12] YANG S,LIU X,LIU Q,et al. A study of storm surge disasters based on extreme value distribution theory[J]. Journal of Coastal Research,2017,33(6):1423-1435.
[13] YANG S,LIU X,LIU Q. A storm surge projection and disaster risk assessment model for China coastal areas[J]. Natural Hazards,2016,84(1):649-667. doi: 10.1007/s11069-016-2447-1
[14] SUN H,WANG J,YE W. A data augmentation-based evaluation system for regional direct economic losses of storm surge disasters.[J]. International Journal of Environmental Research and Public Health,2021,18(6):2918. doi: 10.3390/ijerph18062918
[15] 陈仕鸿,唐丹玲,隋广军. 基于SVR的广东省台风灾害损失评估[J]. 海洋环境科学,2013,32(6):939-943.
[16] 冯倩,刘强. 基于SVM-BP神经网络的风暴潮灾害损失预评估[J]. 海洋环境科学,2017,36(4):615-621.
[17] 王甜甜,刘强. 基于BAS-BP模型的风暴潮灾害损失预测[J]. 海洋环境科学,2018,37(3):457-463. doi: 10.12111/j.cnki.mes20180323
[18] 于福江, 董剑希, 叶琳. 中国风暴潮灾害史料集[M]. 北京: 海洋出版社, 2015.
[19] 叶金玉,林广发,张明锋. 自然灾害风险评估研究进展[J]. 防灾科技学院学报,2010,12(3):20-25. doi: 10.3969/j.issn.1673-8047.2010.03.004
[20] 冯爱青,高江波,吴绍洪,等. 气候变化背景下中国风暴潮灾害风险及适应对策研究进展[J]. 地理科学进展,2016,35(11):1411-1419.
[21] 张继权, 李宁. 主要气象灾害风险评价与管理的数量化方法及其应用[M]. 北京: 北京师范大学出版社, 2007.
[22] 毕业亮, 王华彩, 夏兵, 等. 雨源型城市河流水污染特征及水质联合评价: 以深圳龙岗河为例[J]. 环境科学, 2011（8）: 1-20.
[23] 谭丽荣,陈珂,王军,等. 近20年来沿海地区风暴潮灾害脆弱性评价[J]. 地理科学,2011,31(9):1111-1117.
[24] KIM S,MATSUMI Y,PAN S et al. A real-time forecast model using artificial neural network for after-runner storm surges on the Tottori Coast,Japan[J]. Ocean Engineering,2016,122:44-53. doi: 10.1016/j.oceaneng.2016.06.017
[25] VAPNIK V. The nature of statistical learning theory[M]. New York: Springer-Verlag New York, 1995.
[26] 郁磊, 史峰, 王辉, 等. MATLAB智能算法30个案例分析[M]. 北京: 北京航空航天大学出版社, 2015.
[27] XUE J,SHEN B. A novel swarm intelligence optimization approach:sparrow search algorithm[J]. Systems Science and Control Engineering,2020,8(1):22-34.
[28] CHANG C C,LIN C J. LIBSVM:a library for support vector machines[J]. ACM Transactions on Intelligent Systems and Technology,2011,2(3):27.
[29] LIU H,TIAN H,LIANG X,et al. New wind speed forecasting approaches using fast ensemble empirical model decomposition,genetic algorithm,Mind Evolutionary Algorithm and Artificial Neural Networks[J]. Renewable Energy,2015,83:1066-1075. doi: 10.1016/j.renene.2015.06.004
[30] 李素,袁志高,王聪,等. 群智能算法优化支持向量机参数综述[J]. 智能系统学报,2018,13(1):70-84.
-
图(5)
表(4)
计量
- 文章访问数: 1614
- PDF下载数: 48
- 施引文献: 0