| Soil erosion has become one of the most serious environmental problems in the world at present. The direct consequence caused by soil erosion is a thinner soil layer, nutrient loss, the destruction of land resources, exacerbating crisis of food problem. At the same time, sediment may deposit in the waterway, reservoir and lake, and aggravate flood disaster. China is a large country of population and agriculture, and is one of countries with the most serious problem of water and soil erosion. Purple soil is a special soil group, famous for its short generation time and rich in salt. The area of purple soil is 160 thousand km2, and mainly distribute in Sishuan,Chongqing,Yunnan,Guizhou,Hunan,Zhejiang,Jiangxi and Jiangsu province (municipalities) and so on. But purple soil concentrate on Sichuan basin, it takes 68 percent of cultivated land. Abundant in rainfall and concentrate, frequent rainstorm and inferior vegetation cover, All these make purple soil erosion very serious. Research showed that soil erosion degree in purple soil is just less than that in Loess Plateau.Therefore, further analysis of the law of purple soil erosion is the base on protecting soil and water resources, improving soil moisture of farmland and fertilizer management, controling agricultural non-point source pollution, and solving the problem of purple soil erosion and improving the agricultural production.There are many influential factors on soil erosion. Rainfall is the direct motivation factor of soil eerosion. Therefore, the accurate measurement of rainfall erosivity, analysis of rainfall and soil erosion characteristics, the establishment of simulation regression equation are the base of quantitative study and effective control of soil erosion. In this paper, the author take example of theⅡSmall Watershed of Shengzhong Soil and Water Conservation Experiment Station in Sichuan province, collect a vast amount of data of rainfall and runoff, analysis the relationship of rainfall, runoff and soil loss, then explore the characteristics of rainfall and soil erosion in purple soil of northern Sichuan province. The main conclusions are as follows : (1) Rainfall distribution characteristics of the experiment small watershed from 1991 to 2001 are as follows :①Rainfall amount in flood season (from May to October) accounted for 85% or more of annual rainfall. Correlation analysis between Rainfall amount in flood season and annual rainfall gain linear correlation, and the correlation coefficient is 0.9920. Thus the change tendency between rainfall amount in flood season and annual rainfall is the same on the whole. Average rainfall in flood season takes on unimodal distribution, undulatory property is small, and is relatively stable. Comparison annual rainfall with average rainfall, we can get 1991, 1992 and 2001 is normal year, from 1994 to 1997 is drought, and 1993, 1998, 1999 and 2000 is high flow year.②Average rainfall days is 115.4 days from 1991 to 2001, however the average rainfall days in flood season is 70 days. Monthly distribution of rainfall days takes on bimodal type, and the minimum is in July. This shows there will be the possibility of summering drought in this area. But the annual distribution of rainfall days takes on reduced tendency and maintains concordance on the whole. Rainfall days in flood season account from 46.67% to 75.25% of annual rainfall days.③Maximun average rainfall amount in flood season takes on unimodal Distribution. The maximun average rainfall amount exceed 10 mm criterion of erosive rainfall. Therefore, flood season is the main season of soil erosion. The largest rainfall events is 218.2 mm (July, 2001) from 1991 to 2001, and the smallest is 1.1 mm (October, 1998), their difference is 198.36 times.④Largest daily rainfall, largest rainfall events and maximum average rainfall intensity have occurred in May to September from 1991 to 2001. Namely all occurred in flood season, especially in June, July and August. Annual distribution of the largest daily rainfall takes on bimodal type. Annual distribution of rainfall events is more complicated, which the largest rainfall amount occurred On August 18 to 19, 2001. Annual distribution undulatory property of maximum average rainfall intensity is great. Because rainfall intensity is determined by rainfall amount and rainfall duration, which maximum average rainfall intensity occurred on August 19, 1996.(2) In our study R-values of rainfall erosivity are calculated by CREAMS. First, calculate daily rainfall erosivity value in flood season, Sum all daily rainfall erosivity to R_m values. Again cumulative monthly rainfall erosivity to annual rainfall erosivity R_y values. The rainfall erosivity R_m values in flood season (from May to October) were 319.67, 798.96, 993.96, 1185.62, 563.07 and 170.38MJ·mm/hm~2·h respectively. And the largest rainfall erosivity R_m value occurs in August. The sum of rainfall erosivity from June to August accounted for flood season (or annual) 73.87%. Average annual rainfall erosivty R-value is 3777.39MJ·mm/hm~2·h.Distribution of rainfall erosivity value shows erosive rainfall occurred mainly in June, July, August and September, and accounting 82.0% of the totle times. Most of erosive rainfall occurred in September, which is 4.7 times. And the minimun is 1.4 times, which appeared in October. The tendency of times of erosive rainfall occurred is not obvious. The times erosive rainfall is high in 1992, 1998 and 2000, and the times is few in 1994, 1996 and 1997, and the remains is normal. The best algorithm of monthly rainfall erosivity R_m-value with rainfall amount P_m is power function form, which is similar with CREAMS. And the best method to calculate annual rainfall erosivity R_y-value with rainfall amount P_y in flood season is negative exponential function.(3) The change of runoff of theⅡSmall Watershed of Shengzhong Soil and Water Conservation Experiment Station in Sichuan province takes on unimodal type. The largest runoff is in August. The change of runoff from 1991 to 2001 are not obvious, but it generally may be divided into three levels. The changes is high in 1991 and 1993. But the change is low from 1994 to 1998, and the remainings is middle. This shows that the change of runoff not only with rainfall amount, but also with the underlying surface status (e.g. soil, vegetation, farming methods, etc.).Runoff in flood season mainly occurred in June, July and August. The times of runoff in flood season account for 85.8% of the total number. The times trendency of runoff takes on unimodal type, and accord with the change of runoff. The largst times of flow is in August. The times of flow is decreased from 1991 to 2001 on the whole. The largest times of flow is in 1991.(4) The relation between Rainfall and runoff in flood season satisfy negative exponential function in the Small Watershed, The relation between total rainfall amount and runoff is the same in flood season. As erosive rainfall occurred in all the flood season, the total runoff may be expressed as annual runoff.(5) Soil loss is mainly concentrated in the June, July and August (54.742t), which account for 99.09% of total soil loss (55.245t) in flood season in the small watershed. Therefore, soil erosion is mainly in the flood season, especially in June, July and August. Between rainfall amount P_m in flood season and soil loss T_m, totle rainfall amount P_y in flood season and soil loss T_y, and annual rainfall erosivity R_y-value and soil loss L_y meet Negative exponent function regression equation. However rainfall erosivity R_m-value and soil loss L_m meet hyperbolic curve function regression equation.
|
PDF Full Text Request