Scale dependency of spatial variability of soil moisture in gravel-mulched field

Knowledge of the scale dependency of the spatial variability of soil moisture is of paramount importance in the study of soil–atmosphere interactions and hydrological processes. We present a case study of the effects of sampling area and spacing on the spatial variability of soil moisture in a 32 × 32 m gravel-mulched field in an arid area of northwestern China. The coefficient of variation, correlation length and Moran’s index, which are commonly used in spatial analysis, were applied to each scenario to characterise the degree of spatial variability of soil moisture. The spatial variability was weak at all sampling scales. All indices increased to various degrees with an increase in sampling area. The correlation length decreased with increasing sampling spacing, and neither the coefficient of variation nor Moran’s index were significantly correlated with sampling spacing, indicating that sampling spacing had little effect on soil-moisture variability. The spatial distributions of soil moisture were irregular, with peaks and valleys at different sampling spacings but tending to gradually stabilise as sampling spacing increased. A sampling spacing of 8 m was reasonable, because it best-characterised the spatial distribution of soil moisture. This study also provides a theoretical reference for the establishment of optimal sampling schemes, which allow a considerable savings in both time and cost, and provides theoretical support for ecological agricultural production in gravel-mulched fields.