Multiple environmental factors analysis of flash flood risk in Upper Hanjiang River, southern China
2020
Zhong, Ming | Jiang, Tao | Li, Kun | Lu, Qiaoqian | Wang, Jiao | Zhu, Jiajun
Identifying the environmental factors and analyzing the causal mechanism of flash floods help to manage the risk. Maximum 24-h precipitation (MP), digital elevation (DE), slope degree (SD), soil type (ST), drainage density (DD), and vegetation cover (VC) are selected as the risk factors of flash floods in this study. Precipitation is the important meteorological components in flash floods; thus spatial characteristics of precipitation trend have been analyzed by using Mann-Kendall tests, and a positive trend of precipitation in Upper Hanjiang River is detected. Then, association rule mining approach is proposed to investigate the multiple environmental factors of flash floods, in which both single and multiple dimension data mining have been conducted by Apriori algorithm. Considering the high rate of 5-year return period floods in the flash flood inventory, further association rule mining after sampling has been conducted in order to deeply mine the causal patterns of flash floods in different risk magnitudes. Results show that soil type, slope degree, and digital elevation are the dominant environmental factors of flash floods in the study area, and precipitation is one of the important causal factors in severe flash flood hazards. It is also highlighted that flash floods might easily occur even with a slight rainfall present due to the instability of sand clay and saturated soil moisture. The proposed novel use of field data and data mining has the potential for providing procedures and solutions for an effective interpretation of flash flood mechanism. The results are expected to be applicable for decision-making and sustainable management in flooding risk.
显示更多 [+] 显示较少 [-]