The semiarid region of Brazil undergoes constant changes in land use due to the process of deforestation and high water seasonality, represented among other variables by a decline in the soil moisture content. In this study, the influence of land use on the soil moisture dynamics is analysed. For this purpose, three experimental sites were evaluated, characterised by different types of land use: an agroecosystem of forage cactus, a deforested environment and an area of Caatinga vegetation, located in the semiarid central hinterlands of Brazil. Calibration of the Diviner2000® capacitive probes was carried out using the relationship between the soil volumetric moisture content (θv), obtained by the gravimetric method, and the relative frequency readings of the probes. Weekly measurements of the moisture content were taken at 22 points at depths of 0.00–0.60 m from November 2014 to October 2017, providing 157 sampled days. The mean values and standard deviations of the θv variable for the three surfaces were analysed over twelve periods, and grouped according to the local water regime into rainy, dry and transitional. To analyse the temporal stability of θv, the relative difference and the Spearman correlation matrix were calculated. Finally, multivariate principal component analysis (PCA) was used to identify associations between the mean values of θv and physical properties in the different soil layers. Based on the results, the local calibration curves presented significant coefficients of determination, showing that the equations can be used for reliable estimates of θv. The θv was higher in the Caatinga (0.086 m³ m⁻³), intermediate in the forage cactus (0.064 m³ m⁻³) and lower in the deforested area (0.045 m³ m⁻³). Changes in land use influence the spatial variability of θv in the soil layers, which responded to the local rainfall regime, especially in the upper layers (0.05–0.25 m). Significant differences in θv between soil surface coverage occurred during the dry sub-periods and the transition periods, when the Caatinga was superior to the other areas. The temporal stability analysis identified the forage cactus area as the most representative location for θv estimates, while the Spearman correlation matrix detected the more-unstable areas of moisture, with persistent moisture patterns between the dry sub-periods and transition periods only. Associations between moisture and different soil properties were affected by soil surface coverage and soil layer depth. All the proprieties explained moisture distribution along the profile. It can be concluded that converting the Caatinga vegetation into other soil surface coverages changes soil-moisture distribution patterns, resulting in a reduction in θv of 26% and 47% for forage cactus and deforested areas, respectively. Therefore, the establishment of forage cactus in current deforested areas of the Brazilian semiarid may attenuate the temporal variability and reduction of soil moisture compared to bare soil surfaces.

由于森林砍伐和高水季节性，巴西半干旱地区的土地利用不断变化，除其他变量外，土壤水分含量也有所下降。在这项研究中，分析了土地利用对土壤水分动力学的影响。为此，对三个实验点进行了评估，这些实验点具有不同类型的土地利用：位于巴西半干旱中部腹地的饲用仙人掌农业生态系统，森林砍伐的环境和Caatinga植被区。Diviner2000®电容式探头的校准使用重量法获得的土壤体积含水量（θv）与探头的相对频率读数之间的关系进行。每周在0点深度的22个点进行水分含量测量。2014年11月至2017年10月为00-0.60 m，提供了157个采样日。在十二个时期内分析了三个表面的θv变量的平均值和标准偏差，并根据当地水情将其分为多雨，干旱和过渡时期。为了分析θv的时间稳定性，计算了相对差和Spearman相关矩阵。最后，使用多元主成分分析（PCA）来确定θv平均值与不同土壤层中物理性质之间的关联。根据结果​​，局部校准曲线显示出重要的测定系数，表明该方程式可用于可靠地估计θv。Caatinga的θv较高（0.086 m 在十二个时期内分析了三个表面的θv变量的平均值和标准偏差，并根据当地水情将其分为多雨，干旱和过渡时期。为了分析θv的时间稳定性，计算了相对差和Spearman相关矩阵。最后，使用多元主成分分析（PCA）来确定θv平均值与不同土壤层中物理性质之间的关联。根据结果​​，局部校准曲线显示出重要的测定系数，表明该方程式可用于可靠地估计θv。Caatinga的θv较高（0.086 m 在十二个时期内分析了三个表面的θv变量的平均值和标准偏差，并根据当地水情将其分为多雨，干旱和过渡时期。为了分析θv的时间稳定性，计算了相对差和Spearman相关矩阵。最后，使用多元主成分分析（PCA）来确定θv平均值与不同土壤层中物理性质之间的关联。根据结果​​，局部校准曲线显示出重要的测定系数，表明该方程式可用于可靠地估计θv。Caatinga的θv较高（0.086 m 根据当地水情将其分为多雨，干旱和过渡时期。为了分析θv的时间稳定性，计算了相对差和Spearman相关矩阵。最后，使用多元主成分分析（PCA）来确定θv平均值与不同土壤层中物理性质之间的关联。根据结果​​，局部校准曲线显示出重要的测定系数，表明该方程式可用于可靠地估计θv。Caatinga的θv较高（0.086 m 根据当地水情将其分为多雨，干旱和过渡时期。为了分析θv的时间稳定性，计算了相对差和Spearman相关矩阵。最后，使用多元主成分分析（PCA）来确定θv平均值与不同土壤层中物理性质之间的关联。根据结果​​，局部校准曲线显示出重要的测定系数，表明该方程式可用于可靠地估计θv。Caatinga的θv较高（0.086 m 多元主成分分析（PCA）用于确定θv平均值与不同土壤层中物理性质之间的关联。根据结果​​，局部校准曲线显示出重要的测定系数，表明该方程式可用于可靠地估计θv。Caatinga的θv较高（0.086 m 多元主成分分析（PCA）用于确定θv平均值与不同土壤层中物理性质之间的关联。根据结果​​，局部校准曲线显示出重要的测定系数，表明该方程式可用于可靠地估计θv。Caatinga的θv较高（0.086 m³  m ^-3），在饲用仙人掌中居中（0.064 m ³  m ^-3），在森林砍伐地区中较低（0.045 m ³  m ^-3））。土地利用的变化会影响土壤层θv的空间变异性，这对当地的降雨状况特别是上层（0.05-0.25 m）的降雨状态有响应。当Caatinga优于其他地区时，在干旱子时期和过渡时期，土壤表面覆盖率之间的θv存在显着差异。时间稳定性分析将草食仙人掌区域确定为θv估计值的最有代表性的位置，而Spearman相关矩阵检测到了更不稳定的水分区域，并且仅在干燥子时期和过渡时期之间存在持续的水分模式。水分与不同土壤特性之间的关联受到土壤表面覆盖率和土壤层深度的影响。所有的特性解释了沿着轮廓的水分分布。可以得出的结论是，将Caatinga植被转化为其他土壤表层覆盖物会改变土壤水分的分布方式，从而导致饲用仙人掌和森林砍伐面积的θv分别降低26％和47％。因此，与裸露的土壤表面相比，在巴西半干旱地区目前森林砍伐的地区建立饲用仙人掌可能会减弱时间变化并减少土壤水分。