We found a match
Your institution may have access to this item. Find your institution then sign in to continue.
- Title
Multi-period early-warning precipitation identification method for the easily waterlogged districts in Jiangxi province, China.
- Authors
Bai, Hua; Wang, Bingxiang; Li, Bin; Liu, Zhangjun; Wen, Zhenyu; Zhang, Yang; Xiao, Feng; Xu, Xinfa; Huang, Yongfeng
- Abstract
Urban easily waterlogged districts need more systematic monitoring as the key disaster-forming and managing spatial scale, hard to identify the early-warning precipitation (EP). We proposed new algorithms to extract the urban easily waterlogged hilly and plain districts' polygons in Jiangxi Province, China. The districts' disaster risk level prediction model was built through eleven intelligent algorithms' optimization based on the field survey data. The multi-period EP values were identified using the scenario simulation by the model and further verified compared to the government-issued ones. It indicated that 532 easily waterlogged districts possessed sufficient sample representativeness, represented by the significant coefficient of variation ≥ 0.66 for slope, waterlogging disaster area, and depth. The fine Gaussian support vector machine and fine K-Nearest Neighbor algorithms performed higher prediction accuracies of more than 86% and 83% in the model construction and verification districts, respectively. The 1-, 3-, and 6-h severe waterlogging disaster's EP medians were averaged to be 16.5–20.8, 31.5–35.8, and 35–39.3 mm differed among the optimum algorithms respectively, smaller than the government-issued ones. In comparison, the 1-, 3-, and 6-h moderate waterlogging disaster's EP medians were averaged to be 10.5–16.1, 25.5–31.1, and 29–34.6 mm respectively, larger than the issued ones.
- Subjects
JIANGXI Sheng (China); CHINA; PRECIPITATION (Chemistry); NAIVE Bayes classification; SUPPORT vector machines; WATERLOGGING (Soils); K-nearest neighbor classification; FIELD research
- Publication
Theoretical & Applied Climatology, 2024, Vol 155, Issue 4, p2705
- ISSN
0177-798X
- Publication type
Article
- DOI
10.1007/s00704-023-04774-w