Responses of soil microarthropods to warming and increased precipitation in a semiarid temperate steppe

Soil microarthropods are an important component in soil food webs and their responses to climate change could have profound impacts on ecosystem functions. As part of a long-term manipulative experiment, with increased temperature and precipitation in a semiarid temperate steppe in the Mongolian Plateau which started in 2005, this study was conducted to examine effects of climate change on the abundance of soil microarthropods. Experimental warming had slightly negative but insignificant effects on the abundance of mites (−14.6%) and Collembola (−11.7%). Increased precipitation greatly enhanced the abundance of mites and Collembola by 117 and 45.3%, respectively. The response direction and magnitude of mites to warming and increased precipitation varied with suborder, leading to shifts in community structure. The positive relationships of mite abundance with plant cover, plant species richness, and soil microbial biomass nitrogen suggest that the responses of soil microarthropods to climate change are largely regulated by food resource availability. The findings of positive dependence of soil respiration upon mite abundance indicate that the potential contribution of soil fauna to soil CO₂ efflux should be considered when assessing carbon cycling of semiarid grassland ecosystems under climate change scenarios.     

Effects of a one-year rainfall manipulation on soil nematode abundances and community composition

Soil nematodes play a crucial role in the terrestrial nitrogen cycle by accelerating the release of ammonium from microorganisms (bacteria and fungi). As aquatic organisms, nematodes are likely to be affected by predicted changes in precipitation patterns and soil moisture during the 21st century. The objective of this study was to measure the response of soil nematodes to a one-year rainfall manipulation in the sandy, forest soils of the New Jersey Pinelands (USA). We excluded all rain from four replicate field plots and applied double the amount of natural rainfall to four additional plots. We then assessed the impact of these precipitation treatments on nematode abundance and community composition. We found that total nematode abundance increased with more precipitation, and were highly sensitive to annual precipitation amount. This is in contrast to microbial biomass which was previously found to be insensitive to precipitation change. We suggest that any increased microbial growth in high rainfall plots was consumed by microbivorous nematodes. We further suggest that nematodes in the freely draining, sandy soils we studied may be unsuccessful at surviving drought because few water-filled pore spaces remain, as compared to more aggregated soils. All nematode families were sensitive to drought, but the effect was greatest on the Plectidae, while no significant effects were found for the Cephalobidae and Qudsianematidae. While not directly measured, these results provide insight into the relative anhydrobiotic abilities of these families. We found that bacterial-feeding nematodes were most sensitive to drought, suggesting that grazer-induced alterations to the nitrogen cycle are possible if precipitation patterns change in the future.     

Soil erosion from sugar beet in Central Europe in response to climate change induced seasonal precipitation variations

This study estimates the implications of projected seasonal variations in rainfall quantities caused by climate change for water erosion rates by means of a modeling case study on sugar beet cultivation in the Central European region of Upper-Austria. A modified version of the revised Morgan–Morgan–Finney erosion model was used to assess soil losses in one conventional and three conservation tillage systems. The model was employed to a climatic reference scenario (1960–89) and a climate change scenario (2070–99). Data on precipitation changes for the 2070–99 scenario were based on the IPCC SRES A2 emission scenario as simulated by the regional climate model HadRM3H. Weather data in daily time-steps, for both scenarios, were generated by the stochastic weather generator LARS WG 3.0. The HadRM3H climate change simulation did not show any significant differences in annual precipitation totals, but strong seasonal shifts of rainfall amounts between 10 and 14% were apparent. This intra-annual precipitation change resulted in a net-decrease of rainfall amounts in erosion sensitive months and an overall increase of rainfall in a period, in which the considered agricultural area proved to be less prone to erosion. The predicted annual average soil losses under climate change declined in all tillage systems by 11 to 24%, which is inside the margins of uncertainty typically attached to climate change impact studies. Annual soil erosion rates in the conventional tillage system exceeded 10 t ha^− 1 a^− 1 in both climate scenarios. Compared to these unsustainably high soil losses the conservation tillage systems show reduced soil erosion rates by between 49 and 87%. The study highlights the importance of seasonal changes in climatic parameters for the discussion about the impacts of global climate change on future soil erosion rates in Central Europe. The results also indicate the high potential of adaptive land-use management for climate change response strategies in the agricultural sector.     

Simulating site-specific impacts of climate change on soil erosion and surface hydrology in southern Loess Plateau of China

Proper spatial and temporal treatments of climate change scenarios projected by General Circulation Models (GCMs) are critical to accurate assessment of climatic impacts on natural resources and ecosystems. The objective of this study was to evaluate the site-specific impacts of climate change on soil erosion and surface hydrology at the Changwu station of Shaanxi, China using a new spatiotemporal downscaling method. The Water Erosion Prediction Project (WEPP) model and climate change scenarios projected by the U.K. Hadley Centre's GCM (HadCM3) under the A2, B2, and GGa emissions scenarios were used in this study. The monthly precipitation and temperature projections were downloaded for the periods of 1900–1999 and 2010–2039 for the grid box containing the Changwu station. Univariate transfer functions were derived by matching probability distributions between station-measured and GCM-projected monthly precipitation and temperature for the 1950–1999 period. The derived functions were used to spatially downscale the GCM monthly projections of 2010–2039 in the grid box to the Changwu station. The downscaled monthly data were further disaggregated to daily weather series using a stochastic weather generator (CLIGEN). The HadCM3 projected that average annual precipitation during 2010–2039 would increase by 4 to 18% at Changwu and that frequency and intensity of large storms would also increase. Under the conventional tillage, simulated percent increases during 2010–2039, compared with the present climate, would be 49–112% for runoff and 31–167% for soil loss. However, simulated soil losses under the conservation tillage during 2010–2039 would be reduced by 39–51% compared with those under the conventional tillage in the present climate. The considerable reduction in soil loss in the conservation tillage indicates the importance of adopting conservation tillage in the region to control soil erosion under climate change.     

Adjustment of CLIGEN parameters to generate precipitation change scenarios in southeastern Australia

Global climate model predictions are often downscaled with stochastic weather generators to produce suitable climate change scenarios for impact analysis. Proportional adjustment to generated daily precipitation and direct adjustment to parameter values for weather generators have been used for assessing the impact of climate change on runoff and soil loss. Little is known of how these parameter values should be realistically adjusted, the amount of adjustment, and whether the adjustments are correlated among different parameters. Rainfall in southeastern Australia has significantly increased since the late 1940s. Rainfall records in Sydney show a similar trend. Long term daily and 6-min intensity data from Sydney have made it possible to examine how CLIGEN parameter values have changed in relation to the underlying significant increase in rainfall. This study shows that for Sydney, most of the increase in rainfall is a result of the increase in wet day precipitation. The increase in the standard deviation of wet-day precipitation is greater than that in the mean, implying a greater rainfall variability during wetter periods. The wet-following-wet transition probability, and maximum 30-min rainfall intensity are all positively and significantly correlated with the change in wet-day precipitation. The change in peak intensity is about half the change in rainfall. No significant relationship can be established between the changes in mean monthly rainfall and those in the skewness coefficient for wet day precipitation and wet following dry transition probability for the site. Simultaneous adjustment of all these parameters is needed for generation of precipitation change scenarios for the region. Using simple proportional adjustment to generated precipitation sequences would lead to maximum impacts on runoff and soil loss predicted with WEPP, while attributing precipitation change equally to the change in wet day precipitation and the number of wet days would under-estimate the magnitude of the impacts considerably for the site.     

Impact of climate change on soil erosion, runoff, and wheat productivity in central Oklahoma

The potential for global climate changes to increase the risk of soil erosion is clear, but the actual damage is not. The objectives of this study were to evaluate the potential impacts of climate change on soil erosion, surface runoff, and wheat productivity in central Oklahoma. Monthly projections were used from the Hadley Centre's general circulation model, HadCM3, using scenarios A2a, B2a, and GGa1 for the periods of 1950–1999 and 2070–2099. Projected changes in monthly precipitation and temperature distributions between the two periods were incorporated into daily weather series by means of a stochastic weather generator (CLIGEN) with its input parameters adjusted to each scenario. The Water Erosion Prediction Project (WEPP) model was run for four climate scenarios including a recent historical climate and three tillage systems (conventional tillage, conservation tillage, and no-till). HadCM3-projected mean annual precipitation during 2070–2099 at El Reno, Oklahoma decreased by 13.6%, 7.2%, and 6.2% for A2a, B2a, and GGa1, respectively; and mean annual temperature increased by 5.7, 4.0, and 4.7 °C, respectively. Predicted average annual soil loss in the tillage systems other than no-till, compared with historical climate (1950–1999), increased by 18–30% for A2a, remained similar for B2a, and increased by 67–82% for GGa1. Predicted soil loss in no-till did not increase in the three scenarios. Predicted mean annual runoff in all three tillage systems increased by 16–25% for A2a, remained similar for B2a, and increased by 6–19% for GGa1. The greater increases in soil loss and runoff in GGa1 were attributed to greater variability in monthly precipitation as projected by HadCM3. The increased variability led to increased frequency of large storms. Small changes in wheat yield, which ranged from a 5% decrease in B2a to a 5% increase in GGa1, were because the adverse effects of the temperature increase on winter wheat growth were largely offset by CO₂ rise as well as the bulky decrease in precipitation occurred outside the growing season. The overall results indicate that no-till and conservation tillage systems will be effective in combating soil erosion under projected climates in central Oklahoma.     

Changes in variability of soil moisture alter microbial community C and N resource use

Grassland ecosystems contain ∼12% of global soil organic carbon (C) stocks and are located in regions where global climate change will likely alter the timing and size of precipitation events, increasing soil moisture variability. In response to increased soil moisture variability and other forms of stress, microorganisms can induce ecosystem-scale alterations in C and N cycling processes through alterations in their function. We explored the influence of physiological stress on microbial communities by manipulating moisture variability in soils from four grassland sites in the Great Plains, representing a precipitation gradient of 485-1003 mm y⁻¹. Keeping water totals constant, we manipulated the frequency and size of water additions and dry down periods in these soils by applying water in two different, two-week long wetting-drying cycles in a 72-day laboratory incubation. To assess the effects of the treatments on microbial community function, we measured C mineralization, N dynamics, extracellular enzyme activities (EEA) and a proxy for substrate use efficiency. In soils from all four sites undergoing a long interval (LI) treatment for which added water was applied once at the beginning of each two-week cycle, 1.4-2.0 times more C was mineralized compared to soils undergoing a short interval (SI) treatment, for which four wetting events were evenly distributed over each two-week cycle. A proxy for carbon use efficiency (CUE) suggests declines in this parameter with the greater soil moisture stress imposed in LI soils from all four different native soil moisture regimes. A decline in CUE in LI soils may have been related to an increased effort by microbes to obtain N-rich organic substrates for use as protection against osmotic shock, consistent with EEA data. These results contrast with similar in situ studies of response to increased soil moisture variability and may indicate divergent autotrophic vs. heterotrophic responses to increased moisture variability. Increases in microbial N demand and decreases in microbial CUE with increased moisture variability observed in this study, regardless of the soils’ site of origin, imply that these systems may experience enhanced heterotrophic CO₂ release and declines in plant-available N with climate change. This has particularly important implications for C budgets in these grasslands when coupled with the declines in net primary productivity reported in other studies as a result of increases in precipitation variability across the region.     

Climate change impacts on soil erosion in Midwest United States with changes in crop management

This study investigates potential changes in erosion rates in the Midwestern United States under climate change, including the adaptation of crop management to climate change. Previous studies of erosion under climate change have not taken into account farmer choices of crop rotations or planting dates, which will adjust to compensate for climate change. In this study, changes in management were assigned based on previous studies of crop yield, optimal planting date, and most profitable rotations under climate change in the Midwestern United States. Those studies predicted future shifts from maize and wheat to soybeans based on price and yield advantages to soybeans. In the results of our simulations, for 10 of 11 regions of the study area runoff increased from + 10% to + 310%, and soil loss increased from + 33% to + 274%, in 2040–2059 relative to 1990–1999. Soil loss changes were more variable compared to studies that did not take into account changes in management. Increased precipitation and decreasing cover from temperature-stressed maize were important factors in the results. The soil erosion model appeared to underestimate the impact of change in crop type, particularly to soybeans, meaning that erosion increases could be even higher than simulated. This research shows that future crop management changes due to climate and economics can affect the magnitude of erosional impacts beyond that which would be predicted from direct climate change alone. Prediction of future soil erosion can help in the management of valuable cropland and suggest the need for continually changing soil conservation strategies.     

Implications of climate change scenarios for soil erosion potential in the USA

Atmospheric general circulation models (GCMs) project that increasing atmospheric concentrations of CO2 and other greenhouse gases May, result in global changes in temperature and precipitation over the next 40-100 years. Equilibrium climate scenarios from four GCMs run under doubled CO2 conditions were examined for their effect on the climatic potential for sheet and rill erosion in the conterminous USA. Changes in the mean annual rainfall factor (R) in the Universal Soil Loss Equation (USLE) were calculated for each cropland, pastureland and rangeland sample point in the 1987 National Resources Inventory. Projected annual precipitation changes were assumed to be from differences in either storm frequency or storm intensity. With all other USLE factors held constant these changes in R translated to changes in the sheet and rill erosion national average of +2 to +16 per cent in croplands, -2 to +10 per cent in pasturelands and -5 to +22 per cent in rangelands under the eight scenarios. Land with erosion rates above the soil loss tolerance (T) level and land classified as highly erodible (eredibility index >8) also increased slightly. the results varied from model to model, region to region and depended on the assumption of frequency versus intensity changes. These results show the range of sensitivity of soil erosion potential by water under projected climate change scenarios. However, actual changes in soil erosion could be mitigated by alterations in cropping patterns and other management practices, or possibly by increased crop growth and residue production under higher atmospheric CO2 concentrations.     

肃北高寒草原不同放牧强度土壤养分变化特征

研究了肃北高寒草原不同放牧强度下不同土层土壤养分及 5 种微量元素有效态含量变化特征。结果表明：①高寒草原土壤物理性质的变化对土壤养分及微量元素具有重要的调控作用;②随着放牧强度的提高,0 ~ 10、10 ~ 20 cm 土层土壤体积质量均呈不同程度的增加,土壤孔隙度和土壤含水量则呈显著的递减趋势;③轻度放牧草地土壤有机质、土壤全 N 含量高于中度放牧和重度放牧草地;20 ~ 30 cm 土层有机质随放牧强度的增大呈明显下降趋势,即随放牧强度的增大深层土壤肥力呈退化趋势;肃北高寒草原的速效养分以多 N 少 P 富 K 为特点,土壤速效 N、P、K 含量在总体上随放牧强度的增加呈下降趋势;④肃北高寒草原 5 种微量元素的高低顺序依次是：Na＞Fe＞Mn＞Cu＞Zn,不同放牧强度下各微量元素的变化一致,顺序依次是：轻度放牧＞对照＞中度放牧＞重度放牧;⑤放牧强度对 10 ~ 20 cm 的土层影响最大,随放牧强度的增大,地表植物营养吸收层土壤营养成分、微量元素呈降低,导致地表植被生长能力降低,最终导致地表土壤沙化,最后使草地大面积退化。     

Implications of decadal changes in precipitation and land use policy to soil erosion in Basilicata,Italy

This paper examines the implications of changes in precipitation and land use to soil erosion from 1955 to 2002 in Basilicata, a hilly portion of southern Italy. Analysis of daily precipitation records reveals statistically significant trends using both non-parametric and parametric approaches. The inter-annual variability of precipitation increases in intensity; primarily between October and January. From 1955 to 2000, the length of dry spells greatly increased, while wet days decreased. A land use change map was produced for the three study areas using aerial photos (1955) and orthophotos (1997 and 2002), integrated with field surveys. Results show that land use is highly dynamic in Basilicata, especially due to the application of the European Union's Common Agricultural Policy (CAP) measures. The EU policies resulted in reclamation of badlands and degraded grasslands for agriculture, principally the cultivation of durum wheat. This farming practice and the abandonment of some of the remodeled areas have increased the risk of soil erosion and desertification processes, and is manifest in land degradation by rill networks and gullying.     

Twenty years of tillage research in subarctic Alaska: I. Impact on soil strength, aggregation, roughness, and residue cover

Soil properties and surface characteristics affecting wind erosion can be manipulated through tillage and crop residue management. Little information exists, however, that describes the impact of long term tillage and residue management on soil properties in the subarctic region of the United States. This study examines the impact of 20 years of tillage and residue management on a broad range of physical properties that govern wind erosion processes on a silt loam in interior Alaska. A strip plot experimental design was established in 1983 and included intensive tillage (autumn and spring disk), spring disk, autumn chisel plow, and no tillage with straw either retained on or removed from the soil surface. Soil and residue properties measured after sowing barley (Hordeum vulgare L.) in May 2004 included penetration resistance, soil water content, shear stress, bulk density, random roughness, aggregate size distribution, and residue cover and biomass. No tillage was characterized by larger aggregates, greater soil strength (penetration resistance and shear stress), wetter soil, and greater residue cover compared to all other tillage treatments. Despite crop failures the previous 2 years, crop residue management influenced residue biomass and cover, but not soil properties. Autumn chisel and spring disk appeared to be viable minimum tillage options to intensive tillage in controlling erosion. Autumn chisel and spring disk promoted greater roughness, aggregation, and residue cover as compared with intensive tillage. Although no tillage appeared to be the most effective management strategy for mitigating wind erosion, no tillage was not a sustainable practice due to lack of weed control. No tillage also resulted in the formation of an organic layer on the soil surface over the past 20 years, which has important ramifications for long term crop production in the subarctic where the mean annual temperature is <0 °C.     

气候变化对新疆地区棉花生产的影响

经分析发现新疆棉区近50年来气侯变暖变湿并且日照时数明显减少。利用COPRAS动力评估模型研究得出：气侯变化对新疆棉区不同区域的棉花生长发育影响是不同的，北疆棉区、南疆盆地西缘区和南疆盆地东缘区棉花的开花期和吐絮期明显提前，棉花停止生长期明显延后，70年代、80年代和90年代比60年代全生育期分别平均延长8．2d、2．4d和5．2d；近50年来新疆地区棉花模拟产量明显增加，平均增产为15．7kg/hm^2/10年，尤其在北疆棉区。棉花模拟产量波动性明显加强，说明受气侯变化的影响棉花生产风险也在加大。     

红壤丘陵区坡长对作物覆盖坡耕地土壤侵蚀的影响

坡长对坡耕地土壤侵蚀的影响随雨强的不同而变化,为解决南方红壤丘陵区坡耕地水土流失问题,该文采用野外人工模拟降雨的方法,研究了南方红壤丘陵区作物覆盖坡耕地上不同雨强下坡长对其土壤侵蚀的影响,并探讨了侵蚀增强的临界雨强和设置水土保持措施的合理坡长,结果表明:产沙量随坡长延长整体呈增大趋势,但存在一定的波动,二者的关系可用幂函数(决定系数0.84)表示。坡长延长相同长度时,产沙量不呈比例增加,但每隔4 m产沙量增量有减少的趋势,且径流侵蚀产生的泥沙中主要为粒径0.002~0.02 mm的粉粒及粒径0.002 mm的黏粒,加剧了耕地土壤粗化,因此,可每隔4 m设置水土保持措施,有效减少坡耕地水土流失。坡面径流侵蚀产沙量随着雨强的增大而增加,坡长越长,产沙量随雨强增加速度越快,二者呈幂函数关系(决定系数0.76),60 mm/h是红壤丘陵区侵蚀增强的临界雨强;雨强、坡长与产沙量均呈正相关关系,且雨强对坡耕地产沙量的影响较坡长大。对不同雨强下坡长对作物覆盖坡耕地土壤侵蚀的影响研究,可以为南方红壤丘陵区坡耕地水土流失的治理提供一定的理论依据。     

Runoff and soil erosion from areas of burnt scrub: comparison of experimental results with those predicted by the WEPP model

The effects of burning on runoff and soil erosion from scrub-bearing hillslopes in northwest Spain were investigated by monitoring of experimental plots over a 4-year period. At the beginning of this period, two plots (BP1 and BP2) were subjected to low-intensity controlled burns, and two plots left as controls; in the year following the controlled burns, however, one of the control plots (plot WF) was burnt in a wildfire of higher intensity than the controlled burns. Runoff and erosion losses from the BP plots were only slightly higher than from the control plot, with the loss in no case exceeding 300 g m⁻² yr⁻¹. Erosion losses from plot WF were considerably higher (1314 g m⁻² over the 1st year post-burning, 8.5 times higher than from the control plot). These results were compared with those predicted using the Water Erosion Prediction Project (WEPP) hillslope model (Version 95.7). In general, WEPP predictions of total runoff volume over the study period were acceptable: coefficients of determination for the regression of predicted on observed values were 0.41, 0.68, 0.66 and 0.57 for the control plot and plots BP1, BP2 and WF, respectively. Erosion losses were likewise predicted with reasonable accuracy, though the model showed a consistent tendency to under-estimate, particularly with plot WF.     

Climate change and soil erosion on agricultural land in england and wales

To assess the effects of climate change on soil erosion we need to model changes in rate, frequency and extent of erosion. Present day rates of soil erosion for agricultural land in England and Wales are known from a national monitoring scheme and also from a local one. The latter, for the South Downs, covers a seven-year period and includes climatic data. This shows a strong correlation between total erosion and a Rainfall Index. The availability of these databases allows us to use existing models such as EPIC and an Expert System to predict erosion rates for postulated warmer and wetter (winter) conditions. EPIC is particularly suitable for specific sites where detailed data exists and crop yield implications can also be modelled. A rule-based Expert System approach allows us to examine erosion rates at a different scale across the landscape. We postulate that water erosion rates on arable land in the lowlands will increase markedly in severity, frequency and extent especially if land use changes. In the uplands predicted climatic warming suggests a longer growing season and fewer frosts: these may lead to a decrease in erosion of overgrazed eroding slopes. Increases in erosion rates are not inevitable if policy decisions are taken and implemented in good time.     

降雨强度对优先流特征的影响及其数值模拟

为了研究优先通道发育程度对不同降雨强度下优先流特征的影响,该文选取不同深度的田间原状土样,在不同降雨强度下进行土壤优先流的室内物理模拟试验,并采用双渗透介质模型Hydrus-1D对优先流发育过程进行数值模拟。结果表明:优先通道越发育,其优先流程度越高。降雨强度对优先流速率及其增加幅度的影响随土壤优先通道发育程度的增加而增加。优先通道发育的土柱,降雨强度越大时,其土柱下端初始出流时间越短,而优先通道一般发育的土柱,其土柱下端初始出流时间均随降雨强度增加呈先减小而后增加的规律。当降雨强度小于土柱最大导水能力(原状土柱表面出现积水时的导水能力)时,优先流速率随降雨强度增加而显著增加,而当降雨强度接近或大于其最大优先导水能力时,优先流速率增加缓慢或趋于稳定。土壤优先通道对降雨入渗的导水能力(即优先流速率与降雨强度之比)随降雨强度增加而变小。双渗透介质模型能较理想地模拟优先流发育过程,对土壤基质向大孔隙排水过程的模拟较差,但模型不适用于裂隙发育的土壤。各土柱的优先流程度为93.6%~99.9%。在研究降雨强度对优先流的影响规律时,该研究考虑了优先通道发育程度对此规律的影响,丰富了优先流理论,同时对农业节水、土壤污染迁移、地下水污染风险评价以及滑坡机理等研究方面具有重要的科学意义。     

基于土地利用结构特征的土壤侵蚀强度预测研究——以四川省为例

研究计算四川省土地利用结构特征指数和土壤侵蚀强度指数结果表明,该省土地利用结构特征指数与土壤侵蚀强度指数的变化具有同步性,用土地利用结构特征指数预测其土壤侵蚀强度变化是合理可行的。     

Effects of soil erosion on long-term soil productivity in the black soil region of northeastern China

China's northeastern Black Soil Region, one of the country's most important crop production areas, has been seriously affected by soil erosion. This study evaluated the effects of soil erosion on the long-term productivity of this region. We used a modified productivity index (MPI) model (MPI is a number between 0 and 1, with 1 indicating highest productivity) to assess the current effects of soil erosion on soil productivity, as well as to predict long-term change in productivity. Samples from 21 black soil profiles yielded varying MPI values, although most MPI values were indicative of moderate productivity. Organic matter content and available water capacity impact MPI values in the region, whereas soil clay content and pH were less important. Overall, organic matter content and available water capacity of soil profiles decreased consistently as depth of erosion increased. Modeling indicated that MPI in the region will decrease by 0.0052 for each centimeter of topsoil eroded; this rate represents 1% of the current average MPI for the study area. The model predicts a 9.6% productivity reduction over 100 years and a 48.3% reduction over 500 years.     

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%。