运用分形模型研究了川南天然常绿阔叶林及其人工更新成檫木(Sassafras tzumu)林、柳杉(Cryptomeria fortunei)林和水杉(Metasequoia glyptostroboides)林后土壤团粒结构,探讨了分形维数与林地土壤水源涵养功能、肥力特征和微生物数量之间的关系。结果表明:天然常绿阔叶林人工更新后土壤团粒结构的分形维数和结构体破坏率增大、土壤物理性质变差、养分含量和微生物数量降低,3种人工林中,檫木林较好、水杉林次之、柳杉林最差;土壤团聚体、水稳性团聚体和水稳性大团聚体含量越高分形维数越小;在湿筛条件下,土壤结构体破坏率随分形维数的降低而减小;土壤团粒结构的分形维数与土壤物理性质、养分含量和微生物数量之间存在显著的回归关系。这表明天然常绿阔叶林人工更新后由于不同林分对林地土壤组成结构的维护效果不同,导致更新后林地土壤物理、化学和生物性质变化,林地土壤团粒结构的变化,进而影响其分形维数的大小。因此,分形维数可作为天然常绿阔叶林及其人工更新后林地土壤水源涵养功能、肥力特征和微生物活动情况的一项综合性定量化评价指标。同时,为保护天然常绿阔叶林、选择适宜的更新树种和天然常绿阔叶林人工更新后林地土壤的科学管理提供依据,也为退耕还林中树种的选择提供参考。

Aims Fractal theory, a study tool popular in recent years, offers a new means to quantitatively investigate soil structure. Soil structure is the basis of soil fertility, which is the basic property of soil. It can be comprehensively reflected by soil physical, chemical and organism properties, and change of soil structure will result in changes in other soil properties. Fractal features of soil aggregate structure under natural evergreen broadleaved forest and regeneration of artificial systems is rarely studied. We chose to study natural evergreen broadleaved forest and three artificial plantations of Sassafras tzumu, Cryptomeria fortunei and Metasequoia glyptostroboides in southern Sichuan Province. Our objective was to determine a) effects of artificial regeneration on fractal features of soil aggregate structure, b) effects of different plantations on fractal dimension of soil aggregate structure, c) relationships between fractal dimension and soil physical properties, nutrient content and microbe number, and d) use of fractal dimension of soil aggregate structure for evaluating the water conservation, fertility and microbe activity of soil.

Methods Soils were collected from each forest to determine a) fractal dimension of soil aggregate structure using Yang Peiling's approach and b) soil physical and chemical properties and soil microbe number. The relationship between fractal dimension and soil physical properties, nutrient content and microbe number was analyzed with regression analysis.

Important findings Natural evergreen broadleaved forest and artificial regeneration resulted in increased fractal dimension of soil aggregate structure and percent of construction damage, poorer soil physical properties and reduced nutrient content and number of microbes. The higher the content of aggregates, water-stable aggregates and water-stable big aggregates in soil, the smaller the fractal dimension of soil aggregate structure. With wet sieving condition and decreased fractal dimension, the percent of construction damage decreased. There were close relationships between fractal dimension of soil aggregate structure and soil natural water content, bulk density, capillary porosity, non- capillary porosity, original infiltration coefficient, stable infiltration coefficient, content of organic matter, total-N, hydrolysis-N, total-P, available-P, total-K, available-K and the number of bacteria, fungi and actinomyces. Different stands had different effects for maintenance of soil structure, which resulted in changes of soil physical, chemical and organism properties under natural evergreen broadleaved forest and artificial regeneration, and changes of soil aggregate structure had an effect on the value of fractal dimension. This indicates that fractal dimension can be used as a comprehensive quantitative index to evaluate water conservation function, fertility states and microbe activity of soil for natural evergreen broadleaved forest and artificial regeneration. It provides a solid foundation for protecting natural evergreen broadleaved forest, choosing appropriate trees for its artificial regeneration, managing soil after artificial regeneration and choosing trees for converting farmland to forest.