中分辨率遥感像元尺度生物土壤结皮覆盖与植被及土壤间的交互关系

生物土壤结皮（Biological soil crust，BSC）作为干旱、半干旱地区健康生态系统的重要组成部分，具有固碳、固氮、降低土壤水分蒸发、减少土壤径流等多种生态功能，这些功能与BSC的发展阶段及覆盖度密切相关，使其在维系荒漠自然生态系统的稳定性中扮演重要角色。因此，结合遥感数据尺度定量研究影响BSC分布的环境因素是评估沙区生态系统稳定性并以此为依据进行沙化土地治理的重要基础性研究。影响BSC覆盖度的环境因子较为复杂，现有研究方法存在两方面局限，其一是多为小尺度定性分析，其二是多侧重于孤立地分析BSC与环境因子间的单向关系。利用结构方程模型（Structural Equation Modeling，SEM）中的非递归路径分析（Nonrecursive Path Analysis，NPA）对遥感30 m分辨率像元尺度BSC覆盖度、植被覆盖度、土壤pH值、盐度、有机质与粒度进行路径分析，旨在使用综合性方法从整体上阐明大尺度BSC分布与植被、土壤间的交互关系。结果表明：（1）在毛乌素沙地，BSC覆盖度受各环境因子综合影响，无法用单一变量说明。BSC覆盖度与植被覆盖度、土壤有机质、平均粒径和细颗粒占比呈极显著正相关（P<0.01），与粗颗粒占比呈极显著负相关（P<0.01）。（2） BSC覆盖度与植被覆盖度通过有机质相互影响，BSC覆盖度对植被覆盖度有较大的正向直接影响，路径系数（Path Coefficient，PC）=0.43（P>0.05），植被覆盖度对BSC覆盖度有交大的负向影响（PC=－0.22；P>0.05）。（3）土壤平均粒径和细颗粒占比均正向影响BSC的覆盖度，其中，平均粒径对BSC覆盖度的总体影响较大（PC=0.67；总效果值=0.590），细颗粒占比对BSC覆盖度的间接影响较大（间接效果值=0.052）。（4）盐度对BSC覆盖度呈显著负向直接影响（PC=－0.41；P<0.05；总效果值=－0.398），pH值对BSC覆盖度有极小的正向影响（总效果值=0.072）。上述研究结果可为遥感探测BSC、制定有效的荒漠生态系统保护与修复政策措施提供科学依据。

Interactions among biological soil crust, vegetation, and soil on medium resolution pixel scale of remote sensing

As an important part of healthy ecosystem in arid and semi-arid areas, biological soil crust (BSC) has many ecological functions, such as carbon and nitrogen fixation, reducing soil water evaporation and soil runoff. These functions are closely related to the species and coverage of BSC. In view of this, BSC plays an important role in maintaining the stability of desert ecosystem. Therefore, taking the scale of remote sensing data into account and quantifying the environmental factors that affect the distribution patterns of BSC are the key to ecological restoration and governance of desertification land. The environmental factors that affect the BSC coverage are complex. There were two limitations in the existing research methods. On the one hand, most of the researches were small-scale analysis between BSC and environmental factors in the field. On the other hand, Numerous studies focused on analyzing the one-way relationships between BSC and environmental factors in isolation. In our research, to systematically sort the interactions among the distribution of BSC, vegetation, and soil, the structural equation modeling (SEM) which is a comprehensive method was used. The paths among BSC coverage, vegetation coverage, soil pH, salinity, organic matter, and particle size on a 30-meter pixel scale were analyzed. The results showed that:(1) in the Mu Us Sandy Land, BSC coverage is affected by various environmental factors and cannot be explained by a single variable. The BSC coverage was greatly and positively correlated with vegetation coverage, soil organic matter, average particle size, and the ratio of fine particles (P< 0.01), and negatively correlated with the ratio of coarse particles (P< 0.01). (2) BSC coverage and vegetation coverage interacted with each other through organic matter. BSC coverage had a positive effect on vegetation coverage (path coefficient=0.43, P>0.05). Vegetation coverage had a negative effect on BSC coverage (PC=-0.22; P >0.05). (3) the average particle size and the proportion of fine particles were all positively affected on BSC coverage. Among them, the average particle size had a great impact on the coverage of BSC (PC=0.67; total effect value=0.590). The proportion of fine particles had an indirect impact on BSC coverage (indirect effect value=0.052). (4) the salinity significantly and negatively impacted BSC coverage (PC=-0.41, P< 0.05). The value of soil pH had a little positive effect on BSC coverage (total effect value=0.072). The results provided scientific evidence for BSC detection, protection, and restoration, as well as the basis for policymaking to restore the desert ecosystem.