喀斯特小流域土壤含水率空间异质性及其影响因素

该文基于网格取样(80 m×80 m),利用地统计学和经典统计学方法,研究了典型喀斯特小流域旱季表层(0~10 cm)土壤含水率(soil moisture content,SMC)的空间变异特征,并分析了其与容重(bulk density,BD)、毛管孔隙度(capillary porosity,CP)、非毛管孔隙度(non-capillary porosity,NCP)、土壤有机碳(soil organic carbon,SOC)、碎石含量(rock fragment content,RC)等土壤理化性质以及坡度(slope gradient,SG)、坡向(slope aspect,SA)、裸岩率(bare rock,BR)等地形因子的关系。结果显示,SMC半方差函数的最优拟合模型为指数模型,变程为381.00 m,块基比为0.382,属于中等程度的空间相关性。普通克里格插值结果显示,SMC呈现出随海拔升高而降低的分布规律。Pearson相关分析表明,除SOC外,其他土壤理化指标均对SMC有显著影响(p0.05);各地形因子中仅SG对SMC有显著影响。协方差分析表明,RC、CP和NCP对SMC的方差解释达到显著水平(p0.05),地形部位(上坡、中坡、下坡、洼地)、土地利用类型(乔木林、灌木林、灌草丛、耕地)及二者的交互作用均未达到显著水平。这说明土壤理化性质是影响SMC的直接因素,地形部位和土地利用类型均通过改变土壤理化性质来影响SMC。该结果有利于辨别喀斯特小流域旱季SMC的主要影响因素,从而为该地区土壤水源涵养功能的提高及流域水文过程研究提供科学依据。

Spatial variability of soil moisture content and its influencing factors in small Karst catchment during dry period

Abstract: Surface soil moisture content (SMC) has a crucial effect on rainfall infiltration, runoff generation, evaporation and other soil hydrological processes. However, knowledge of SMC distribution was limited in Karst areas with discontinuous thin soils containing high content of rock fragments. In this paper, based on 80 m×80 m sampling grid, 162 undisturbed and disturbed soils was sampled to measure SMC and other soil properties, including bulk density (BD), capillary capacity (CP), non-capillary capacity (NCP), soil organic carbon (SOC), and rock fragment content (RC). Environmental factors including topographical position (upper-slope, middle-slope, lower-slope and depression), land use type (forestland, shrubland, shrub-grassland and farmland), slope gradient (SG), slope aspect (SA) and bare rock (BR) were investigated around the sampling points. Spatial variability of SMC and its influencing factors were analyzed by both geostatistical and classical analysis methods. The results showed that the mean value of SMC was 34.43%. The SMC had a moderate variation with the coefficient of variation of 0.33. The geostatistical results showed that the semivariance of SMC was best fitted by exponential model with a higher determination coefficient of 0.910. The range was 381.00 m and the nugget/sill value was 0.382, indicating a moderate spatial correlation of SMC. When the lag distance was smaller than 200 m, the variation in 120° direction was higher than that of 30° direction. However, when the lag distance was exceeded 200 m, no anisotropy was found in 120° and 30° directions with the anisotropy ratio fluctuated around 1. The Kriging map showed that SMC generally decreased with the increasing altitude. Most of the slope had a lower SMC value less than 35%. The highest SMC value always appeared in the east of the depression which was higher than 50%. The Pearson correlation analysis showed that CP had a significant (p < 0.01) positive correlation with SMC but RC and BD had a significant negative correlation with SMC. NCP had a significantly negatives effect on SMC with the significant value of 0.022. However, SOC only significantly affected SMC at 0.10 level. This suggested that all the soil properties had important influence on SMC. However, for the topographical factors, only SG had a significant influence on SMC with the correlation coefficient of -0.435 (p < 0.001). The analysis of covariance showed that the interpretation of RC, CP, and NCP was significant (p < 0.05). However, topographical location, land use type and their interaction effect were not significant. This indicated that the soil properties were the direct influencing factors. Topographical locations and land use types influenced SMC mainly by changing the soil properties. The explained variation of topographical location was higher than that of land use type. This indicated that topographical locations had more important effect on SMC than land use types. These results will help to understand the spatial distribution of SMC and to distinguish the main influential factors of SMC in Karst catchment. They also provide knowledge on ecological conservation and restoration in Karst regions.