基于多种算法的小安溪流域降雨侵蚀力时空演变特征

Spatiotemporal characteristics of rainfall erosivity in Xiaoanxi Basin using multiple algorithms

降雨侵蚀力是土壤侵蚀预报模型中的重要指标,精确计算降雨侵蚀力能够提高区域水土流失预报精度。该研究以长江上游末端支流小安溪为研究区域,基于流域内国家基本站点近10年逐分钟降雨数据,采用K均值聚类法进行侵蚀性降雨雨情分类;以通用土壤流失方程计算降雨侵蚀力的结果为标准,在6种不同时间尺度的计算模型中优选简易算法,应用推荐模型计算长序列逐日降雨侵蚀力并分析其时空演变特征。结果表明:1)研究区侵蚀性降雨可分为3类,I类为主要的降雨类型,产生的降雨侵蚀力最小,仅33.90 MJ·mm/(hm^(2)·h);Ⅲ类最剧烈,达1176.86 MJ·mm/(hm^(2)·h);2)优选日降雨量模型B计算得到1960—2018年的平均降雨侵蚀力为2037.14~2464.71 MJ·mm/(hm^(2)·h);铜梁站点在研究时段内增加3.66%,其余站点呈下降趋势,永川变化幅度最大,减少12.39%;年内变化呈双峰特征,高值期集中于5—9月,春冬和夏秋季节降雨侵蚀力的空间差异性明显;3)空间变化特征方面,多年平均降雨侵蚀力从上游向下游依次递增,小安溪流域汇入干流涪江处水力侵蚀潜在危险最大。该研究结果可为小安溪流域水土保持提供理论和方法借鉴。

Rainfall erosivity is characterized by the potential capability of rainfall to cause soil loss in the regional soil and water conservation.It is a high demand to collect the rainfall data with high time resolution for the better evaluation of rainfall erosivity.In this study,multiple simple algorithms were extended to calculate the rainfall erosivity in the time series under certain accuracy requirements,further to meet the research needs of the temporal and spatial evolution of erosivity.The Xiaoanxi Basin in the Yangtze River was taken as the study area with frequent human activities and serious soil disturbance.First,the K-means clustering was used to categorize all the erosive rainfalls in the recent 10 years using the minute-by-minute rainfall data from Yongchuan,Dazu,Tongliang and Hechuan meteorological stations from 2009 to 2018.Second,an optimal simple algorithm was determined from six selected algorithms,according to the standard calculation on the classic rainfall erosivity.A long-sequence calculation of rainfall erosivity was then carried out for the daily rainfall data from 1960 to 2018.Third,a Mann-Kendall test was applied to analyze the rainfall erosivity characteristics from spatial and temporal aspects,using rescales range and inverse-distance-weighted spatial interpolation.The results showed that:1)The percentage of annual mean rainfall,duration,intensity,frequency,and erosivity of erosive rainfall in the total rainfall were 71.45%,38.06%,52.93%,15.11%and 96.30%,respectively.The erosive rainfall over the basin was divided into three categories as follows.Category I was characterized by the low rainfall,medium duration,low intensity with the least erosivity of 33.90 MJ·mm/(hm^(2)·h),and the highest percentage of occurrences of the total erosive rainfall up to 87.43%.The categoryⅡof the erosive rainfall presented a higher erosivity of 330.12 MJ·mm/(hm^(2)·h)and a frequency of 10.01%.The categoryⅢbelonged to the highly intensive rainfall with the most erosivity of 1176.86 MJ·mm/(hm^(2)·h),accounting for only 2.56%of the total occurrence times.2)The simple algorithm B was illustrated as the optimized calculation for the long-sequence rainfall erosivity in the study area from 1960 to 2018.The multi-year average rainfall erosivity was among 2037.14 to 2464.71 MJ·mm/(hm^(2)·h)in the study basin.The rainfall erosivity showed a decreasing trend in all selected stations,except Tongliang with an increase of about 3.66%.The Yongchuan station performed the best with the change range of 12.39%in terms of change magnitude.The annual variation of rainfall erosivity showed a bimodal distribution with the highest value in the first half of July.The period of high rainfall erosivity was from May to September,where the cumulative value accounted for 84%-88%of the annual rainfall erosivity.Therefore,it is necessary to strengthen the prevention and control of water and soil loss in this period.3)In spatial distribution,the rainfall erosivity showed the increasing feature from the upstream to downstream.There was an outstanding spatial discrepancy of rainfall erosivity in different seasons.The potential risk of hydraulic erosion most likely appeared at downstream,where the Xiaoanxi River merges into the Fujiang River Basin.