Knowledge about the spatial and temporal pattern of soil water storage under different soil, topographic, and climatic conditions is crucial for agricultural and environmental applications. One of the major challenges in the analysis of the spatial and temporal pattern of soil water data is the presence of systematic trend and noise. Moreover, there is growing evidence that many physical systems determining soil water redistribution have no characteristic length scale and exhibit long-range power-law correlations. The objective of this study was to investigate power-law correlations and the associated cross-over scales of the spatial pattern of soil water storage during dry and wet periods using the detrended fluctuation analysis (DFA) technique. The research was conducted in the semiarid region located at Alvena, SK, Canada. A north-south transect of 624-m length was established on a landscape with a significant slope. Soil water content of the surface 0 to 60 cm depth was measured at 104 locations (6 m intervals) using direct (gravimetric) method, during six occasions between 2002 and 2005. Preliminary results indicate that both the correlation exponents and the associated cross-over scales vary with soil water regimes and topographic indices.