不同降雨重现期下山洪汇流过程关键因子识别研究——以赣江上游为例

Identification of key factors in the process of mountain flood confluence during different rainfall return periods: Upper Ganjiang River as an example

森林植被与山洪过程耦合关系是生态水文学研究的热点问题,而给定降雨条件下不同重现期山洪汇流过程关键影响因子识别研究不足严重制约着山洪灾害的预警预报。该研究选取江西赣江上游易灾地区为研究案例,将洪峰模数、汇流时间和1h立即转移临界雨量这3个参数作为山洪汇流过程表征因子,以生态参数、人类扰动、森林起源及林龄4个因子集共计14个因子作为环境解释变量,采用10 min不同重现期降雨实现暴雨区划,在乡镇尺度上利用Canoco5.0软件中RDA模型并进行关键因子识别。研究结果表明:NDVI在1%、2%、5%、20%和50%不同降雨重现期下对山洪汇流过程影响贡献最大且变化幅度相对稳定,其贡献率在37.8%~44.2%之间,而植被覆盖度指数(VC)在80%时贡献率最大;天然林、中林龄、250以上坡耕地及人口密度对山洪汇流过程的影响作用不容忽视;不同重现期下解释变量和响应变量因子的空间分布特征及其相关性变化不大;植被指数与山洪汇流过程耦合作用显著,可作为山洪预警参考指标之一。

The coupling relationship between forest vegetation and flash floods is a hot issue in ecological hydrology research.The insufficient research on the identification of key influencing factors in the process of mountain flooding during different rainfall return periods under given rainfall conditions seriously restricts the ability to produce early warning systems and forecast mountain flood disasters.We selected the flood-prone area in the upper reaches of the Jiangxi Ganjiang River as the study case.Then a total of 14 factors,including ecological parameters,human disturbance,forest origin,and forest age,were used as the environmental explanatory variables.At the same time,the rainstorm regionalization was realized using 10 min of rainfall during different recurrence periods.All the factors are counted to the township level,and the key factors are identified by the Redundancy Analysis(RDA)model in the Canoco5.0 software.The results showed the contribution rate of NDVI in 1%,2%,5%,20%,and 50%rainfall reappearing periods to be the largest and relatively stable,the contribution rate was between 37.8%and 44.2%,and the maximum contribution rate of the vegetation coverage(VC)index was estimated at 80%.The effects of natural forest,mid stand age,over 250 of slope farmland and population density on the process of flash flood and confluence should not be disregarded.The spatial distribution characteristics,the correlation of the explanatory variable and the response variable factors across different recurrence periods showed little change.The coupling effect between the vegetation index and mountain flood convergence process is significant,and it can be used as a reference indicator for mountain flood warning systems.