【作者】 许明祥；

【导师】 刘国彬；

【作者基本信息】 西北农林科技大学 ， 土壤学， 2003， 博士

【摘要】 土壤质量的恢复、保育是黄土高原植被建设及生态环境可持续发展的关键。本研究针对黄土高原生态环境建设需求及土壤学与水土保持科学研究前沿，以黄土丘陵区侵蚀土壤为研究对象，通过历史资料收集和野外调查相结合、野外试验与室内分析相结合,采用时空互代的方法，从系统分析不同土地利用方式下土壤基本理化及生物质量特征及其相关性入手，拟定了黄土丘陵区侵蚀环境下土壤质量评价指标体系，揭示了生态恢复重建过程中土壤质量演变规律并建立了土壤质量演变的回归模型，在此基础上提出了该区侵蚀土壤质量调控方案，为黄土丘陵区生态恢复重建及侵蚀土壤质量的恢复保育提供了科学依据。主要研究结论如下: 1、建立了黄土丘陵区侵蚀土壤质量评价指标体系及相应的评价模型，拟订了黄土丘陵区侵蚀土壤质量分级评价标准。应用因子分析和判别分析，结合敏感性分析，从10种土地利用类型、208个样点的29项土壤理化及生物指标中筛选出黄土丘陵区侵蚀土壤质量指标为有机质、K10、抗冲性、CEC、蔗糖酶、MWD、速效磷、MICMWD。其中，有机质、K10、抗冲性是表征该区侵蚀土壤质量的关键指标。采用模糊数学中的加权综合法建立了土壤质量综合评价模型。对黄土丘陵区10种土地利用类型的表层土壤质量综合评价结果表明，土壤质量在不同土地利用类型之间差异显著，以天然乔木林地土壤质量最佳，属1级；其次是天然灌木林地和大棚菜地，土壤质量属2级；天然草地土壤质量属于3级，人工乔灌林地土壤质量接近3级，人工草地、撂荒地、农地和果园土壤质量属于4级。研究区土壤质量总体处于中等偏低水平。 2、阐明了不同土地利用方式下土壤化学、物理、生物指标及土壤质量指数随土地利用年限的演变规律，并建立了土壤质量演变的回归模型。撂荒地和人工乔木林地土壤全氮、有机质、活性有机碳、有效氮、速效钾及 C/N、微团聚状况、结构系数、MICMWD、团聚体、MWD、微生物碳和氮、磷酸酶、蔗糖酶及脲酶等指标与利用年限之间有显著或极显著的正相关性；果园和人工灌木林地土壤全氮、有机质与利用年限之间有显著的正相关性；撂荒地土壤渗透系数等物理性质与土地利用年限之间有显著的正相关关系。用线性函数或幂函数对上述相关性进行拟合，建立了黄土丘陵区土壤理化及生物质量指标随土地利用年限演变的回归模型。撂荒地及人工乔灌林地土壤质量指数与利用年限之间呈显著或极显著的正相关性，可以用幂函数表达，建立了黄土丘陵区土壤质量指数随土地利用年限演变的回归模型。据此估算，撂荒地、果园、人工灌木和人工乔木林地土壤质量分别需要 66 年、42 年、25 年和 21 年可达到中等质量水平，而大棚菜地仅需 10年就可达到高等质量水平。 3、在黄土丘陵区，土地利用方式的变化是土壤质量变化的驱动因子。以此为主导的纸坊沟小流域65年来生态系统退化与恢复过程中土壤质量演变可以分为3个阶段。<WP=9>ii 黄土丘陵区生态恢复过程中土壤质量演变及调控 用通用线性回归模型（GLM）计算了土地利用及海拔、坡度、坡向、地形等环境因子在土壤理、化、生物属性及土壤质量指数变异（方差）中所占的百分比。结果表明，土壤理化及生物性质的变异大部分是由土地利用方式的变化引起的，土地利用的变化是土壤质量变异产生的主要来源，可解释97%的土壤质量指数的变异性。 随着流域生态系统退化与恢复进程，31 项土壤理化及生物指标的演变过程可以分为先降后升型、先升后降型及平稳型三种。流域土壤质量指数表现为先降后升型变化。根据纸坊沟小流域土壤质量指数的演变动态，可以将纸坊沟流域 1938 年以来的 65 年间土壤质量演变过程划分为 3 个阶段。1938－1960 年为土壤质量退化期，1960－1990 年为土壤质量稳定期，1990－2002 年为土壤质量恢复期。 4、揭示了土壤有效磷和抗冲性是黄土丘陵区侵蚀土壤质量的限制性因子以及黄土丘陵区侵蚀土壤质量恢复的长时序性和滞后性。提出了优化土地利用结构，增加林草植被覆盖度，建设基本农田，辅以科学有效的水土保持措施和土壤培肥措施的土壤质量调控方案。黄土丘陵区侵蚀土壤质量的恢复具有长时序性和滞后性，在安塞纸坊沟小流域，这种滞后期约为 10 年。不同土壤质量调控措施对土壤质量的影响不同。坡耕地培肥、农耕地撂荒及农耕地转为人工草地对土壤质量的改善作用较缓慢，而天然草地及人工乔灌林地则能有效地提高土壤质量。 5、黄土丘陵区以实现中等土壤质量为目标的人工乔灌和天然草地植被恢复的标准为覆盖度达 65%以上。不论是林草地还是撂荒地，植被盖度与土壤质量间有一定的相关性。其中乔灌林及天然草地、撂荒地植被盖度与土壤质量的相关性达极显著水平。根据林草地植被盖度与土壤质量之间的线性回归模型，人工乔灌和天然草地植被盖度达 65%时土壤质量可达到中等质量水平。更多还原

【Abstract】 Soil quality restoration and soil management is the key for vegetation construction andeco-environment sustainable development in the Loess Plateau. Aimed at the requirement ofeco-environment construction and the hotspot in soil and environment sciences in the LoessPlateau, taking on the soil in the erosion environment on the hilly Loess Plateau as researchobject, through field experiment and lab test, combining with historical data collection andfield investigation, soil quality evolvement mechanism in the process of ecosystem restorationwas studied in this project. Soil quality indicators in the erosion environment were identified.Regression models for soil quality assessment were also established, and limiting soil qualityfactors were identified. Finally recommendation for vegetation restoration and soil qualityimprovement was proposed. The goal was to provide scientific reference for vegetationrehabilitation and soil quality care in the hilly Loess Plateau, and enrich the theory andmethod for soil quality management. The main results were as follows: 1. Soil quality assessment indicators and corresponding assessment model for theerosion environment in the hilly gully region of Loess Plateau were established, thecriterion for soil quality assessment in the region were also set up. Through factor analysisand discriminant analysis, combining with sensitivity analysis, 8 soil quality indicatorsincluding organic matter, K10、anti-scrubility、CEC、invertase、MWD、available phosphorousand MICMWD, were identified from 29 soil chemical , physical and biological indicators of208 soil sampling sites in 10 landuse types. Integrated soil quality evaluation model wasestablished by applying weighted integrating method in fuzzy mathematics. The integratedevaluation results to the topsoil in 10 landuse types on the hilly-gully region of Loess Plateaushowed that there were significant differences in soil quality among landuse types. Soilquality in natural forestland belonged to grade 1 with the best soil quality; soil quality innatural shrub land and vegetable land was in the next place and belonged to grade 2; soilquality in natural grassland belong to grade 3 and in planted woodland and shrub land wasnear to grade 3; soil quality in planted grassland, revegetated grassland, cropland and orchard<WP=11>iv 黄土丘陵区生态恢复过程中土壤质量演变及调控belonged to grade 4. As a whole, soil quality in the research area was in lower level. 2. The evolvement rule of soil chemical, physical, biological indicators and soilquality index was clarified with landuse years in different landuse types. The regressionmodels for soil quality evolvement were also established. There were significant positivecorrelations between landuse years and soil total nitrogen, organic matter, labile carbon,available nitrogen, available potassium, C/N, situation of micro-aggregate, structurecoefficient, MICMWD, MWD, aggregate stability, SMBC, SMBN, phosphatase, invertaseand urease, in revegetated grassland and planted woodland. Significant positive correlationswere also found between landuse years and soil total nitrogen and organic matter in orchardand planted shrub land. Significant positive correlation was showed between landuse yearsand K10 in revegetated grassland. Linear function or power function was suitable for fittingthe correlations mentioned above and was used to establish the regression models for soilchemical, physical and biological quality evolvement with landuse years. Based on the soilquality evolvement models established, we estimated that the time needed for soil quality toreach to middle level was 66, 42, 25 and 21 years respectively in revegetated grassland,orchard, planted shrub land and planted woodland. It’s only 10 years for soil quality invegetable land to reach to high level. 3. In the hilly-gully region of Loess Plateau, land use change was the driving forcefor soil quality change. Effected by land use change, th更多还原