A probabilistic approach for predicting rainfall soil erosion losses in semiarid areas

The implementation of soil and water conservation structures in semiarid areas, usually poses a difficult design problem. This is, in large part, due to the high variability of rainfall and the huge potential impact of extreme hydrologic events on structures and on the landscape in general. Magnitudes of runoff and soil loss or sedimentation rates in those environments are better not assessed by conventional modelling techniques, which tend to average out event magnitude and recurrence variability in time and space. A probability-based approach is proposed here to analyse and predict rainfall erosion losses. The maximum annual storm and its associated erosivity is used as a core element in the assessment of annual interrill and rill erosion rates. Frequency and cumulative soil loss distributions are obtained by combining verified annual and maximum daily rainfall frequency distributions with a proposed erosion algorithm. This stochastic representation of erosion permits to evaluate soil losses for the maximum annual storm, as well as annual erosion rates as a function of recurrence interval. The proposed method was verified with a short series of measured soil loss data in Cape Verde. The physical basis underlying the prediction algorithm and method in general, could be sustained by experimental data and field survey evidence. The method seems applicable to arid and semiarid ecosystems with a high seasonal concentration of precipitation and with rainfall limited to only a few major storm events.     

垦植方式对山地果园水土流失的影响

应用径流小区法,长期（1996-2006）定位研究福建中部尤溪县垦植方式对山地果园水土流失的影响,结果表明：1）该地区11年平均年降雨量1575.5mm,平均年蒸发量1461.1mm,5月、6月、8月是强降雨高发期,也是水土流失防治的关键期;2）在山地果园开发中,平台开垦的径流系数为0.0688,明显小于顺坡开垦（0.2865）;3）植草能使果园0-30cm土层的土壤含水量提高0.8％-0.9％;4）顺坡清耕区侵蚀模数为2388.2t／（km^2·a）,梯台清耕区土壤侵蚀量达10.3kg/m,土壤侵蚀主要源于台壁受冲刷,顺坡植草区、梯台植草区基本不发生土壤侵蚀。     

A simple approach to soil loss prediction: a revised Morgan–Morgan–Finney model

A revised version of the Morgan–Morgan–Finney model for prediction of annual soil loss by water is presented. Changes have been made to the way soil particle detachment by raindrop impact is simulated, which now takes account of plant canopy height and leaf drainage, and a component has been added for soil particle detachment by flow. When tested against the same data set used to validate the original version at the erosion plot scale, predictions made with the revised model gave slopes of a reduced major-axis regression line closer to 1.0 when compared with measured values. The coefficient of efficiency, for sites with measured runoff and soil loss, increased from 0.54 to 0.65. When applied to a new data set for erosion plots in Denmark, Spain, Greece and Nepal, very high coefficients of efficiency of 0.94 for runoff and 0.84 for soil loss were obtained. The revised version was applied to two small catchments by dividing them into land elements and routing annual runoff and sediment production over the land surface from one element to another. The results indicate that, when used in this way, the model provides useful information on the source areas of sediment, sediment delivery to streams and annual sediment yield.     

Throughfall,runoff and soil erosion after prescribed burning in gorse shrubland in Galicia (NW Spain)

The first‐year effect of two different prescribed burning treatments on throughfall, runoff and soil erosion was evaluated in gorse shrubland (Ulex europaeus L.) in Galicia (NW Spain). The treatments compared were: intense burn, light burn and control (no burn). Accumulated annual throughfall represented between the 81 and 87 per cent of total rainfall in intensely burned and lightly burned areas, respectively, whereas in the unburnt areas it was 60 per cent. No significant differences between burning treatments were found for the annual throughfall. However, runoff was significantly greater in intensely burned plots (1·5‐times) than in lightly burned plots. Burning also resulted in a significant increase in runoff (between 2·5 and 1·7‐times, respectively) compared with controls. Total soil losses were small in all treatments, but the intense burn caused significantly greater soil erosion (5·8‐times) compared with the unburned areas. Soil losses after the light burn did not significantly differ from the control although they were higher (2·3‐times). The relationships obtained between erosion and several rainfall parameters were significantly different in burned areas compared to the control. The same response was observed for runoff. Annual erosion losses showed a strong dependence on percentage of bare soil even for small values of this variable. Litter thickness was also a very important variable influencing on erosion rates. This study indicated that by combining ignition techniques and high litter moisture content to maintain the percentage of bare soil below 85 per cent, soil erosion was low. Nevertheless, this result was constrained by the low rainfall that occurred during the study. Copyright © 2005 John Wiley & Sons, Ltd.     

Using plot soil loss distribution for soil conservation design

Soil conservation design is generally based on the estimation of average annual soil loss but it should be developed taking into account storms of a given return period. However, use of frequency analysis in soil erosion studies is relatively limited. In this paper, an investigation on statistical distribution of soil loss measurements was firstly carried out using a relatively high number of simultaneously operating plots of different lengths, λ (11, 22, 33 and 44 m) at the experimental station of Sparacia (southern Italy). Using a simple normalization technique, the analysis showed that the probability distribution of the normalized soil loss is independent of both the scale length λ and the temporal scale, which are completely represented by the mean soil loss calculated for a given event using all replicated data collected in plots having the same length. Then, a comparison between the frequency distribution of soil loss and rainfall erosivity index of the USLE was carried out. An estimating criterion of the annual soil loss of a given return period was also developed. By this criterion, the frequency distribution of the rainfall erosivity factor can be used to design soil conservation practices.     

Concentrated flow erosion rates as affected by rock fragment cover and initial soil moisture content

Concentrated flow experiments using a small hydraulic flume and a constant flow discharge and bed slope have been conducted in order to investigate the effects of rock fragment cover (Rc) on sediment yield for an initially wet and an initially air-dry loamy topsoil. The experimental results indicate that Rc reduces concentrated flow erosion rates (E) in an exponential way (i.e., E=e^−bRc), which is similar to previously reported relations for other water erosion processes such as interrill erosion and sheet-rill erosion measured on runoff plots. The decay rate (b) of this exponential relationship increased throughout the experiments because of scour-hole development and bed armouring. The concentrated flow erosion rates and b-values also depend on the initial moisture content of the topsoil. Depending on Rc, mean concentrated flow erosion rates were 20% to 65% less on initially wet compared to initially air-dry topsoils. The mean value for b was 0.032 for the initially wet, but only 0.017 for the initially air-dry topsoil, indicating that a rock fragment cover is less efficient in reducing concentrated flow erosion rates when the topsoil is initially air-dry than when it is initially wet. The results help explain the data scatter in reported relationships between Rc and interrill–rill erosion rates. They also indicate that a given surface rock fragment cover will offer more protection to wet topsoils than to dry topsoils, which are very common in Mediterranean environments. Event-based water erosion models should incorporate effects of antecedent soil moisture content as well as those of Rc on concentrated flow erosion rates.     

基于机器学习理论的土壤侵蚀模型构建

土壤侵蚀一直是环境问题中的重点和难点。由于影响土壤侵蚀的因素众多,传统的预测模型存在数据获取困难、适用范围小、研究周期长等不足,使得对土壤侵蚀的预测无法做到快速、便捷。支持向量机(Support Vector Machine,SVM)是机器学习中的一个重要模型,具有非线性映射、自我学习能力、全局最小值、对输入数据变化不敏感等优点,在建立土壤侵蚀量相关性预测模型方面较传统预测模型具有更强的优势。本研究应用浙江省诸暨市浦阳江水文站的降雨数据,利用ArcGIS地理信息系统确定水文站上游流域为研究区域。以降雨量、研究区域地理数据维度(包括坡度数据、坡长数据、土壤信息、土地利用类型)作为影响因子,输入支持向量机模型,进行流域内土壤侵蚀量预测。将水文站土壤侵蚀量实测数据作为对照值,用模型输出值检验,从而在取值范围内选择出模型最优的参数组。用影响因子数据和土壤侵蚀量数据对使用最优参数的模型进行检验,模型的预测准确率最高达到75%。其中,降雨量对土壤侵蚀量的影响最大,降雨量单因子预测准确率在70%以上,其余因子预测准确率在3.5%左右。最终得到一个土壤侵蚀量相关性预测模型,通过水文站降雨数据以及地理信息,即可预测当地土壤侵蚀量,准确率达到75%。     

黄土区退耕草地合理放牧可减少土壤CO2排放和土壤侵蚀

【目的】在退耕草地实施合理放牧,有助于减少土壤CO2排放、 减缓土壤侵蚀。为验证此假设,本研究选择黄土高原渭北旱原坡地,建立退耕草地放牧、 退耕草地不放牧和传统农业耕作三种处理的对比试验小区,定量研究了退耕草地合理放牧相对于退耕草地在减少土壤CO2排放和土壤侵蚀的作用及其影响因素,为探寻在我国西部退耕还草区实施畜牧业生产与环境保护的协调发展模式提供科学依据。【方法】在建立的退耕草地放牧、 退耕草地不放牧和传统农业耕作3种处理的试验小区,利用LI-8100 碳通量自动测量仪原位监测植物生长期(4~8月)和放牧前后土壤CO2排放速率的变化,同时利用时域反射仪(TDR)测定表层0—10 cm土壤含水量,用地温表测定土壤表层2 cm和5 cm的温度。利用环境放射性核素 7Be示踪技术监测较大降雨事件引起的土壤侵蚀速率,同时取样测定侵蚀区土壤有机碳含量,比较不同处理小区侵蚀导致的土壤有机碳流失量。【结果】观测期间,3种处理CO2平均排放速率大小顺序为退耕草地[3.69±0.39 μmol/(m2·s)]退耕草地放牧[3.00±0.44 μmol/(m2·s)]传统农耕地[1.99±0.22 μmol/(m2·s)],坡耕地退耕还草后土壤CO2排放增加了85%,而合理放牧使退耕草地土壤CO2排放量减少了19%。放牧后退耕草地土壤CO2排放速率平均减少了11%,减少值在2% ~ 41%之间。观测期内,退耕草地放牧后土壤侵蚀速率比农耕地和退耕草地分别减少了93% 和77%。坡耕地退耕还草后土壤CO2排放增加主要由于草被植物引起土壤有机碳储量增加和土壤侵蚀强度减小,放牧后退耕草地土壤CO2排放减少主要与动物踩踏引起土壤容重明显增加及草类植被地上部分向土壤中输入的有机碳的减少有关。水分、 温度影响因子无法解释3种处理间土壤CO2排放差异。【结论】合理放牧不仅能显著减少退耕草地土壤CO2排放,而且可以有效控制退耕草地土壤和有机碳侵蚀流失。放牧期间动物的踩踏作用引起草地土壤容重显著增加是退耕草地土壤CO2排放量和土壤侵蚀速率减少的主要原因。本研究结果揭示,在我国黄土高原和类似的退耕还草地区实施合理放牧既可以促进当地畜牧业生产,又能控制土壤侵蚀和减少CO2的排放,是一种值得探究的草地可持续发展管理模式。     

概念性土壤侵蚀模型的建立及在紫色土小流域的应用

概念性土壤侵蚀模型在描述流域基本过程的同时对参数要求较低,现有概念性模型中计算产沙所需参数主要靠律定或借鉴经验值。该文介绍了1个无需流域出口产沙量长序列观测资料来律定且能适用于研究流域的概念性土壤侵蚀模型。模型中引入了分布式水文模型WetSpa Extension作为水文模块,结合流域内试验小区上建立的流量—产沙量经验关系计算侵蚀量,再结合泥沙输移比构建起产沙模块。通过在紫色土地区小流域的应用表明,模型能够得到较合理的流域出口产流量、产沙量以及侵蚀率的空间分布等模拟结果,在没有试验小区的邻近流域地区也具有推广性,且能作为评价水保措施效益的有力工具。该概念性模型对于流域出口泥沙资料稀少的地区具有很高的应用价值和协助流域管理的 潜力。     

Impact of expected increase in precipitation intensities on soil loss—results of comparative model simulations

The impact of the expected climate change on the frequency and extent of soil erosion processes is hardly assessable so far. This is mainly because available models of climate change reliably produce at best mean daily precipitation data, whereas erosion is the result of extreme but short time rainfall and runoff events, normally lasting no longer than a few hours. The frequency and intensity of these extreme rainfall events are expected to increase in some regions, which could lead to increased erosion rates. Mathematical models are able to describe erosion rates under conditions of these extreme events, however, so far prognostic meteorological data necessary for the application of these models are not available.The use of a new method for the projection of meteorological time series and their extremes using global climate simulations [Enke and Spekat, 1997, Enke, 2000, Enke, 2003, Enke et al., 2005 and Enke et al., in press] permits for the first time an approximation of future soil loss.This research is based on simulated, high resolution data for extreme rainfall events in the period of 2031–2050, which reproduces the mean frequency, intensity and duration of future events with high precipitation intensities relevant to erosion within the investigated seasonal period from June to August. The simulations are performed for two exemplary sites in Saxony, based on the EROSION 2D model (Schmidt, J., 1990. A mathematical model to simulate rainfall erosion, Catena, Suppl. 19), which is a process-based soil erosion model for simulating soil erosion and deposition by water on single slopes. Simulated precipitation for the 2031–2050 time period is used to model soil loss, and results are compared to soil loss based on 20 years of measured precipitation from 1981 to 2000.The simulation results allow the impacts of climate change on erosion rates to be quantified by comparing current climate with predicted, future climate. However, expected changes in land use due to changed economic conditions are not taken into account in this analysis.     

The development, calibration and field testing of a soil loss and a runoff model derived from a small-scale physical simulation of the erosion environment on arable land in Zimbabwe

Soil loss and runoff prediction models were developed for a clay soil from a five-factor small-scale physical simulation of the field environment by following a central composite rota table experimental design. The uncalibrated models efficiently ranked observed annual soil losses and runoff over a 4–year period between nine field treatments consisting of two bare fallows, two weed fallows and five plots cropped to soyabeans. When calibrated against the field data, the soil loss model predicted the 4–year mean losses from cropped and bare fallow treatments to within 6% for two of the plots, to within 12% for five of the plots and to within 14% for all seven treatments. Over the same period, the runoff model predicted mean annual runoff for the cropped and bare fallow treatments to within 4% for four of the plots and to within 16% for all seven treatments. Percentage vegetal cover proved to be an adequate parameter for describing the role of the soyabean crop in runoff and soil loss processes for a wide range of planting densities of the crop. However, it did not prove to be an efficient index for weed fallows and it was apparent that factors other than simple above-ground cover became important soon after germination.     

基于抽样调查资料估算区域土壤侵蚀量

区域土壤侵蚀量估算是土壤侵蚀调查的重点和难点,为了利用抽样调查数据定量估算区域土壤侵蚀状况,以2011年第1次全国水利普查中的水力侵蚀抽样调查资料为基础,利用中国土壤流失方程(CSLE),通过地理信息系统估算2011年辽宁省的降雨侵蚀力、土壤可蚀性、地形、水土保持措施等侵蚀因子及侵蚀量,在此基础上对全省的水力侵蚀强度进行分级和面积统计,并与抽样调查结果进行对比.结果显示:全省各土壤侵蚀因子的估算结果与抽样调查结果较为接近;土壤水蚀以微度和轻度侵蚀为主,每年全省约85％的土地处于微度侵蚀(＜1000t/km2),而强度侵蚀(＞5 000 t/km2)面积仅占2％;抽样调查得到的独立工矿用地、果园、旱地和其他土地的侵蚀模数较估算得出的侵蚀模数大,而在草地、其他林地等土地利用类型上抽样调查得到的侵蚀模数却较估算得出的侵蚀模数小,主要原因可能是省域和抽样单元计算地形因子的DEM数据精度不同.总体上,该研究为如何利用抽样调查数据估算全国土壤侵蚀状况提供参考.     

Interactions between soil properties and moisture content in crust formation, runoff and interrill erosion from tilled loess soils

Soil-surface seals and crusts resulting from aggregate breakdown reduce the soil infiltration rate and may induce erosion by increasing runoff. The cultivated loess areas of northwestern Europe are particularly prone to these processes.Surface samples of ten tilled silty loamy loess soils, ranging in clay content from 120 to 350 g kg⁻¹ and in organic carbon from 10 to 20 g kg⁻¹, were packed into 0.5 m² plots with 5% slopes and subjected to simulated rainfall applied at 30 mm h⁻¹. The 120 minutes rainfall events were applied to initially field-moist soil, air-dried soil and rewetted soil to investigate the effect of soil moisture content prior to rainfall. Runoff and eroded sediments were collected at 5 minutes intervals. Aggregate stability of the soils was assessed by measuring particle-size distribution after different treatments.All soils formed seals. Runoff rates were between 70 and 90% by the end of the rainfall event for field-moist plots. There were large differences between soil runoff rates for the air-dried and rewetted plots. Interrill erosion was associated with runoff, and sediment concentration in runoff readily reached a steady-state value. Measurements of aggregate stability for various treatments were in good agreement with sealing, runoff and erosion responses to rainfall. Runoff and erosion were lower for air-dried plots than for field-moist plots, and were either intermediate or lowest for rewetted plots, depending on soil characteristics. Soils with a high clay content had the lowest erosion rate when they were rewetted, whereas the soil with a high organic-carbon content had the lowest erosion rate in air-dry conditions. The results indicate the complexity of the effect of initial moisture content, and the interactions between soil properties and climate.     

Analysis and modeling of soil conservation measures in the Three Gorges Reservoir Area in China

Experimental plots were constructed in the Zhangjiachong Watershed of the Three Gorges Reservoir Area to evaluate soil erosion of traditional slope land farming and effects of soil conservation measures. Surface runoff and sediment from the watershed and each plot were collected and measured during 2004–2007. Field investigations indicated that hedgerows were the best for soil erosion control, followed by stone dike terraces and soil dike terraces. The Water Erosion Prediction Project (WEPP) model was used to simulate erosion of annual and rainfall events both at the watershed and plot levels. The low deviation, high coefficient of determination and model efficiency values for the simulations indicated that the WEPP model was a suitable model. The soil erosion rate distribution was modeled to determine where serious erosion would occur during rainfall events in the Zhangjiachong Watershed and so control measures can be taken.     

Comparison of EUROSEM,WEPP, and MEDRUSH model calculations with measured runoff and soil‐loss data from rainfall simulations in Hungary

A useful method to evaluate the effectiveness of soil‐erosion models is to compare the models' soil‐loss and runoff calculations with measured data from experimental plots subjected to artificial rainfall. This study was conducted to develop a set of statistics to compare the performance of the soil‐erosion models EUROSEM, WEPP, and MEDRUSH. Rainfall (six rainfall intensities, two replicates), runoff, and soil‐loss data from artificial plots at two locations in Hungary were used to assess the accuracy of the different models. The soil types within the plots represented a wide range of soil properties and are soil types that are commonly used for agriculture. The results showed that the three soil‐erosion models performed with varying effectiveness dependent on basic soil properties. However, statistical analysis showed the EUROSEM model to be the best for estimating soil loss in Hungary.     

CLIGEN降水要素在黄土塬区的适应性评估

 CLIGEN是目前较全面产生降水要素(降水量、历时、达到最大降水强度的时间与降水总历时的比率、最大降水强度与平均降水强度的比率)的天气发生器,其生成降水要素的质量直接影响水文和农业响应模型的输出结果。利用黄土高原长武1957—2001年的日气象观测数据、王东沟流域1988—2001年的降水要素数据和CLIGEN生成的100年日气象数据,对CLIGEN模型产生日、月、年降水量的均值和方差、概率分布、降水极端值和降水历时、强度进行评估。结果表明:CLIGEN对日、月和年降水量均值的模拟效果较好,相对误差都不大于1.0%;对标准差的模拟结果偏低,相对误差的绝对值小于6.6%;没有模拟出日降水量的概率分布,但是较好地模拟出了月和年降水量的概率分布;对日、月和年最大降水量的模拟误差较大,表明CLIGEN对极值的模拟精度有待提高。CLIGEN很好地模拟出连续降水的频率,但是连续干旱天数在20d以内的累积频率的平均相对误差为8.9%;CLIGEN产生的最大降水强度与平均降水强度的比率高于实测数据;相对于实测数据,CLIGEN模拟的降水历时和降水量具有相同的趋势,对小降水量或短历时的模拟结果偏高,对大降水量或长历时的模拟结果偏低。     

Testing of the hydromechanical prediction model of soil erosion under the conditions of Georgia

A hydromechanical model for predicting water (rain-induced) soil erosion was tested on the experimental plots of the Research Institute of Tea and Subtropical Crops in Zendidi village (the Ajara Autonomous Republic) and the Sabashvili Institute of Soil Science, Agrochemistry, and Melioration in Khevi and Kitskhi villages (Upper Imeretia, Western Georgia). A comparison of factual and predicted values of rain-induced erosion for the plots with permanent black fallow showed that the model overestimated the average annual soil loss for the yellow-brown strongly eroded soil in Zendidi village by 23.22 t/ha (133%). This value ranged in different years from 18 to 1052%. For the plots with corn, the predicted value of annual erosion was by 16.94 t/ha higher than the factual value (overestimation of 488%). A comparison of factual and predicted values of rainfall erosion for the plots under sprinkling irrigation also showed that the predicted soil loss was higher than the factual one by 4.14–30.40 t/ha for corn, 6.76–11.14 t/ha for winter wheat, and 15.75–24.12 t/ha for the plots with stubble of winter wheat and barley. Thus, the hydromechanical model for predicting water erosion inadequately describes it under the conditions of Western Georgia and has to be refined.     

Assessment of soil erosion on arable land using Cs measurements: a case study from Jaslovske Bohunice, Slovakia

This study was carried out to obtain a representative set of data on long-term erosion rates from a pilot area located close to the Jaslovske Bohunice village, in western Slovakia using the ¹³⁷Cs-method. The study area chosen was representative of the hilly loess cultivated areas of Slovakia. The sampling strategy was based on a multiple transect approach. Analyses of the samples for ¹³⁷Cs activity were made at the Nuclear Power Plant Research Institute, Jaslovske Bohunice. The ¹³⁷Cs-method was used to obtain long-term estimates of soil erosion in the Jaslovske Bohunice site, a representative hilly loess cultivated area of Slovakia. The estimated reference ¹³⁷Cs inventory was 2910 Bq m⁻², with a coefficient of variation of 4.3%.Examination of the ¹³⁷Cs redistribution in relation to the topography of the study area revealed that, within individual transects the ¹³⁷Cs inventories were closely related to major landforms. The ¹³⁷Cs inventories were considerably lower on the slopes than on the plateau and they were highest in the valley. However, when plotted against a selection of individual quantitative slope parameters, i.e. the S and the LS factors of the USLE or slope inclination, the correlations obtained were weak.Three conversion models, i.e. the proportional model (PM), the simplified mass balance model (MBM1) and the standard mass balance model (MBM2), from the set of models developed at Exeter University, Great Britain were selected to interpret the resulting ¹³⁷Cs measurements into soil erosion/deposition rates. The mean erosion rates estimated with the PM were 22.4, 35.6 with MBM1 and 17.3 t ha⁻¹ per year with MBM2. There was a good agreement between the average of these mean erosion rates (25.1 t ha⁻¹ per year) for the Jaslovske Bohunice site and the estimated mean soil erosion rate obtained for small erosion plots (15 t ha⁻¹ per year) for conditions similar to the study site. Nevertheless, further research on the application of the ¹³⁷Cs-method, in particular the independent validation of the results obtained, is needed. Several issues requiring further study have been highlighted.     

Nutrient losses from a vineyard soil in Northeastern Spain caused by an extraordinary rainfall event

Vineyards are one of the lands that incur the highest soil losses in Mediterranean environments. Most of the studies that report about this problem only focus on soil losses and few investigations have addressed the nutrient losses associated with erosion processes during the storms. The present research evaluates the loss of nitrogen, phosphorus and potassium in vineyard soils located in a Mediterranean area (NE Spain), after an extreme rainfall event recorded on 10 June 2000. The total rainfall of this event was 215 mm, 205 mm of which fell in 2 h 15 min. The maximum intensity in 30-min periods reached 170 mm h⁻¹. This rainfall produced a large amount of sediments both inside and outside the plots, with the consequent soil mobilisation and loss of nutrients. The estimate of soil loss was based on the subtraction of two very accurate digital elevation models (DEMs) of different dates in GIS, and measures of the nutrient content of sediment collected in the plot. Soil loss in the study plot reached 207 mg ha⁻¹. Most sediment was produced by concentrated surface runoff. Nutrient losses amounted as 108.5 kg ha⁻¹ of N, 108.6 kg ha⁻¹ of P and 35.6 kg ha⁻¹ of K. The proposed method allowed mapping the sediment contribution and deposition areas and the distribution of the nutrient load and losses within the plot.