【作者】 王彬；

【导师】 郑粉莉；

【作者基本信息】 西北农林科技大学 ， 水土保持与荒漠化防治， 2013， 博士

【摘要】 土壤可蚀性是土壤侵蚀研究的前沿科学问题，也是土壤侵蚀预报及其环境效应评价模型的重要参数。本论文针对我国土壤侵蚀预报的需要，尤其是土壤退化问题凸显的东北黑土区土壤侵蚀预警的需求，在集成已有研究成果的基础上，采用野外调查、田间动态监测与原位测定、室内理化分析和模拟降雨试验等相结合的研究方法，确定了表征土壤可蚀性的关键因子及其主要指标，阐明了各关键因子的动态变化特征和内在机制，评价了EPIC、USLE、RUSLE2和Dg等现有土壤可蚀性估算模型在我国（流域尺度和区域尺度）的适用性，最终建立了适用于我国主要水蚀区土壤可蚀性估算模型。主要结论如下：1)确定了影响土壤可蚀性的关键因子及其主要表征指标。采用敏感性分析、相关分析、通径分析和因子降维联合分析的方法，确定了土壤可蚀性关键影响因子，包括土壤质地类关键因子、土壤结构类关键因子、土壤力学和团聚作用类关键因子，以及土壤有机质类关键因子；其中，土壤质地类关键因子对土壤可蚀性的确定具有决定性作用，为核心关键因子；土壤有机质类关键因子则主要通过对其它3个关键因子的作用而对土壤可蚀性进行影响，为土壤可蚀性辅助关键因子。同时，提出了4类关键因子的主要表征指标为土壤黏粒含量（CLA）、土壤颗粒几何平均直径（Dg）、>0.25mm水稳性团聚体含量（WSA>0.25）、机械破碎作用下的土壤团聚体平均重量直径(MWD_SW)、土壤抗剪强度（τ）、土壤容重（ρ）、消散作用下的土壤团聚体平均重量直径(MWD_SW)和土壤有机质（SOM）；建议在实际应用中可根据指标易获取性选择不同的指标组合。2)明晰了土壤质地类关键因子的变化特征并提出了一种优化的土壤质地信息表达方法。通过对Shirazi方法进行修正，提出了适用于任意粒径划分标准的修正颗粒几何平均直径和几何标准差的计算方法。修正算法改善了原方法对不同质地分类制外延性差的缺点，将任意粒径分布条件下所得到的粉粒含量和粘粒含量平均转换相对误差降低至-4.83%和19.31%。同时发现，土壤粘粒含量在流域尺度上随土层厚度的增加呈显著的波动增加趋势；坡面尺度上，坡顶部位或至坡肩部位为主要的物质输出区，而坡脚部位为主要的物质输入区（沉积部位）。而在时间尺度上，土壤质地关键因子在长历时农耕作用下，粘粒含量随开垦历史的增加呈显著的减小趋势，但未达到改变土壤质地的水平。3)阐明了土壤有机质关键因子的时空变化特征。土壤有机质层厚度存在明显的空间变异性，流域尺度上土壤有机质层平均厚度由流域上游至下游逐渐增加，变化范围为23⁵1cm；坡面尺度上，有机质层厚度由坡顶至坡脚部位逐渐加厚，在坡面宣布或坡脚部位由于发生沉积，导致埋藏土壤现象出现。土壤有机质含量在时间尺度上存在变化，对于次降雨和年际间时间尺度，土壤有机质含量存在不显著的下降趋势；而对于多年时间尺度，土壤有机质含量呈显著减小的趋势。同时发现，在黑土区短坡长、直型坡条件下，表层土壤主要以<0.25mm微团聚体的形式迁出，次降雨过程的单位面积微团聚体有机质流失量为13.2³5.6g/m2。4)揭示了土壤团聚体破碎机制。发现随粘粒含量的增高土壤团聚体稳定性增强；而随初始含水量的增加土壤团聚体受消散作用的破坏程度明显减弱，其作用强度可分三个阶段：“气爆”作用阶段（初始含水量<10%）、消散作用迅速减小阶段（初始含水量介于10%²0%）和消散作用极弱阶段（初始含水量>20%）。研究区土壤的团聚体破碎机制主要为消散作用和粘粒膨胀作用，团聚体破坏作用按作用程度的排序为：消散作用（“气爆”现象）>粘粒膨胀作用>机械破坏作用。5)研究了冻融循环作用对土壤团聚体的作用机制。冻融循环作用通过初始含水量和冻融循环次数对团聚体稳定性进行作用；随着土壤初始含水量的增加，3⁵和1²mm两粒级的土壤团聚体稳定性下降。冻融循环次数的作用则表现为，在冻融作用初期（循环次数<3），冻融作用对各粒级各初始含水量的土壤团聚体表现出破坏的作用；而随着冻融次数的增加，冻融循环则表现出促进团聚体形成的作用。冻结温度对土壤团聚体稳定性的影响不显著。6)分析了干湿交替作用下土壤团聚体粒径再分布的过程及相应的阈值现象。干湿交替过程初期对团聚体稳定性造成的影响最大，且其破坏作用对大团聚体更为明显；干湿交替作用对各粒级初始粒径团聚体主要表现为促进<0.2mm微团聚体向0.2¹mm粒径团聚体转化的作用；同时发现，该粒径转化过程是在干湿交替作用力的累积作用下完成的，存在阈值现象（3次），一旦达到形成新生团聚体的阈值后其聚合作用随之下降。7)建立了适用于我国主要水蚀区的土壤可蚀性估算模型。系统评价了USLE、RUSLE2、EPIC和Dg模型在我国（流域尺度和区域尺度）的适用性，发现RUSLE2适用于宾州河流域，而USLE模型和Dg模型分别在东北黑土区和黄土高原地区的区域尺度上表现出较好的估算结果。基于中国土壤可蚀性基础数据库，建立了包含Dg和SOM两个因子且具有明显物理意义的土壤可蚀性估算模型“DG-OM”，模型具有较高的预报精度，能够满足我国土壤可蚀性估算的需要。同时，发现了我国四个主要水蚀区的土壤可蚀性分布情况为（以加权平均值为依据）：黄土高原地区>东北黑土区>西南紫色土区≈南方红壤区（第四纪红粘土）。更多还原

【Abstract】 Soil erodibility is a key indicator to evaluate soil susceptibility to erosion and crucial forpredicting soil loss and evaluating its environmental effects. Mechanism study on soilerodibility, which plays an important role in the domain of soil erosion, provides theoreticalfoundation for soil loss quantification and prediction. To meet needs of implementing soilconservation practices in China, especially for one of the serious eroded area--the black soilregion in NE China, a systematic study on the dynamic mechanism of soil erodibility and itscalculation was conducted. The experimental investigations on dynamic mechanism of soilerodibility and its indicators were studied by field investigating, in-situ monitoring, rainfallsimulation experiments, combining with laboratory physical-chemical analysis and statisticalanalysis. The key soil erodibility factors and their suitable indicators were proposed; theintrinsic mechanisms and variation characteristics of soil erodibility key indicators (e.g., soiltexture, soil organic matter content and soil aggregate stability), which caused the variationsof soil erodibility, were exposited. Moreover, the applicability of EPIC (Erosion ProductivityImpact Calculator), USLE (Universal Soil Loss Equation), RUSLE2(Revised Universal SoilLoss Equation), and Dg models (soil erodibility estimator based on geometric mean diameterof the soil particles) were assessed, at the watershed scale and regional scale; revised modelsfor the aforementioned models were also given. A soil erodibility estimator with highpredicting precision for China was also established and validated. Main results of this studywere as follows:1) A set of soil erodibility key factors and corresponding indicators were proposed byusing a combined analysis of sensitivity analysis, correlation analysis, path analysis, andfactor dimension reduction. The key factors reflecting soil erodibility behaviors included: soiltexture key factor, soil structure key factor, soil shear-strength key factor, and soil organicmatter key factor. Among them, soil texture key factor was the fundamental index for soilerodibility, and was decisive for quantifying the K value. While, it’s also found that soilorganic matter key factor played an assisting role for the other three soil erodibility factors,which was the same path for the other key factors to affecting soil erodibility. Moreover, thecharacterized indicators for the soil erodibility key factors were indicated, which containedCLA (clay content), Dg (geometric mean diameter of the soil particles), WSA>0.25(>0.25mm water-stable aggregate content), MWDWS(mean weight diameter of soil aggregate by shakingtreatment in LB method), soil shear-strength (τ), soil bulk density (ρ), MWDSW(mean weightdiameter of soil aggregate by slow wetting treatment in LB method), and SOM (soil organicmatter content). An adjustable combination of the aforementioned characterized indicatorswas suggested, by considering with the specific research purpose and the ease of indicatorobtaining.2) Dynamic characteristics of soil texture key factor were analyzed, and a more efficientsoil texture expression method was suggested. Based on Shirazi’s method, a revised soiltexture expression method containing Dg and δg was proposed, which has the feature tonormalizing and conversing soil particle distribution information. The revised Shirazi’smethod was improved by extending its extensionality for different soil texture taxonomies,which make sure that the revised method can be applied for Chinese texture taxonomy andother optional classification. The related error was reduced to-4.83%and19.34%for siltcontent and clay content respectively, when converted from the optional classification to theUSDA soil texture taxonomy. Results also showed that clay content represent a significantfluctuations increase trend along the increase of soil thickness (A horizon) at a watershedscale; while, for the slope scale, it showed that the upper slope and slope shoulder area areeroded area, and the slope toe is found as a materials imported area (i.e., depositional area).Moreover, for the time scale, clay content showed a significant decrease trend along the longreclamation time and it did not reach the level to change soil texture.3) Spatial and temporal variations of soil organic matter key factor were clarified. Asignificant spatial varies of A horizon thickness was found for both watershed scale and slopescale. At watershed scale, it showed that A horizon thickness increased from the upper reachto the down reach of Binzhouhe Basin, and the thickness changed from23to51cm; at aslope scale, it indicated that the thickness of A horizon was thicker at slope toe than the toparea of a slope, and a few soil profiles with berried soil phenomenon were also been found atthe slope toe area. We also exposited temporal variations of soil organic matter at differenttime scale. Under a single rainstorm or few years time scale (less than five years in thisresearch), soil organic matter content represent a slight decrease trend, and for a long timeseries (30to100years) a significant decrease was found. Moreover, we found that <0.25mmmicro-aggregate was the main transport particle size under a short slope-length and straightblack soil slope. Soil organic matter was eroded and transport combined with themicro-aggregate, and the loss rate was13.2to35.6g/m2for a single rainfall event.4) The breakdown mechanism of soil aggregate was indicated by applying the threetreatments (i.e., slow wetting treatment, fast wetting treatment, and stirring after pre-wetting treatment) of LB method and Yoder method. We found that clay content has a significanteffect on soil aggregate stability, aggregate stability increased with the increase of claycontent. Impact of initial soil moisture on aggregate stability was also found; it showed thatthe slaking effect was weakened along the increase of soil initial moisture. The decreaseprocess of slacking effect showed obviously periodic behaviors, three stages were divided to I)high-intensity slaking stage, with the initial soil moisture less than10%, II) sharplydecreasing stage, with initial soil water content between10%and20%, and III) stable stage,whose slacking effect nearly die out, with initial moisture larger than20%. The dominant soilaggregate breakdown mechanism were slaking and swelling for the research area, and theorder of their affected degree on breaking soil aggregates was as follow: slaking> swelling>mechanical breakdown effect.5) We investigated the mechanism of soil aggregate stability’s variation by exploringfreeze-thaw cycles’ impact on soil aggregate stability and micro-aggregate distribution.Results showed that the role of freeze-thaw cycles on aggregate stability was realized by theinfluences of initial aggregate water content and numbers of freeze-thaw cycles. It indicatedthat stabilities for3-5mm and1-2mm soil aggregates decreased with the increase of initialwater content. Moreover, freeze-thaw serious destroyed soil aggregate stability at verybeginning (i.e., freeze-thaw cycles less than3times); on the contrary, freeze-thaw cyclesturned to increasing soil aggregate stability along the increase times of freeze-thaw cycles.Different frozen temperature (-10℃and-25℃) did not show significant difference tofreeze-thaw cycles impact.6) We explored the dry-wetting cycles’ impact on soil aggregate stability and itsmicro-aggregate distribution, and found threshold value for the dry-wetting effect. Resultsshowed that serious damage on aggregate stability happened at the beginning of dry-wettingcycles, and the damage degree was more obvious on the large soil aggregate (i.e.,3-5mmaggregate). It’s also found that dry-wetting cycles promoted <0.2mm micro-aggregateconvert to0.2-1mm aggregate for all initial soil aggregate size. And the promotion wascomplicated by the accumulation of dry-wetting effect; there was a threshold as3timesdry-wetting cycles, when it reached the threshold the degree of dry-wetting impact onaggregate stability reducing sharply.7) A Chinese soil erodibility estimator was established, based on a Chinese soilerodibility database. A comprehensive assessment for USLE, RUSLE2, EPIC, and Dg models’applicability was conducted at watershed scale and regional scale. Results showed thatRUSLE2was suitable for the research watershed, and the USLE and Dg models can beapplied directly for the black soil region and the Loess Plateau, respectively, without calibration. Based on the Chinese soil erodibility database, a multiple regression, obtained bythe nonlinear best fitting techniques, yielded a significant relationship (DG-OM model),explaining K values with a combination of Dg (geometric mean diameter) and OM (soilorganic matter). Moreover, soil erodibility values for the four main water erosion areas inChina were in the order as: the Loess Plateau> the black soil region (NE China)> the purplesoil region (SE China)> the red soil region (Quaternary red clay, South China).更多还原