Abstract:As a direct consequence of greater landscape fragmentation worldwide, it is becoming increasingly common for agricultural landscapes to be dissected by boundaries between crop types. However, our understanding of processes associated with these boundaries remains comparatively undeveloped. Cropland-grassland mosaics resulting from intensification of farming activity on extensive grassland are now common in north of China. These land use changes not only influence the original ecosystem processes but also generate new ecological processes that impact on the regional environment at a larger scale. Recently the topic of spatial heterogeneity of soil moisture has received more attention from ecologists because it plays an important role in the hydrological cycle through effects on the interaction between land and atmosphere, thus climate and plant growth in different ecological scales. Despite this emerging focus few studies have examined the spatial heterogeneity of soil moisture in cropland-grassland mosaics in north of China. Here we report data investigating the spatial pattern for soil moisture variation in cropland (C), the cropland-grassland boundary (B) and grassland (G) of a cropland-grassland mosaic. The scale of heterogeneity for soil moisture was assessed by evaluating sampling units of differing sizes (0.5 m×0.5 m, 1 m×1 m, 2 m×2 m) using classical statistical and geostatistical methodologies. Soil water content of C was significantly higher than that of G (P < 0.05). Soil water content of B was intermediate between C and G and differed significantly from both C and G (P < 0.05). The B zone displayed a higher coefficient of variation (CV) for soil moisture than the adjacent systems (P < 0.05), The geostatistical analysis determined that values of the magnitude of spatial heterogeneity (MSH) for soil moisture for B were 0.814, 0.763 and 0.883, respectively, in the three sampling grain sizes (0.5 m×0.5 m, 1 m×1 m, 2 m×2 m). There was also a strong autocorrelation with the ranges A0 15.44, 27.24 and 19.09 m, respectively. The MSH values of soil moisture for G were 0.537, 0.837 and 0.650, and there was a moderate to strong autocorrelation with the ranges A0 6.009, 12.74 and 30.99 m, respectively, in the three different sampling grain sizes. The MSH of soil moisture for C was 0.706, which had a moderate autocorrelation with range 27.28 m in grain size 2 m×2 m, whereas there were nugget effects at the other finer sampling grain sizes. The MSH of the cropland-grassland mosaic displayed a nonlinear relationship across different sampling scales in the field. We found soil moisture for B displayed the highest spatial heterogeneity (i.e. significant patchiness), whereas the soil moisture for C was more homogeneous in comparison with soils under grassland. With respect to the characteristics of the B transition zone, there was an abrupt change of increase then rapidly decreases for soil moisture from cropland to grassland in the cropland-grassland mosaic. The data therefore support the ‘ecological boundary effects’ hypothesis that the ecological boundary is more heterogeneous than that of the adjacent systems and also show that abrupt ecological boundary effects occur. Our results suggest that the historical shift from perennial herb to monocrop culture in grasslands ecosystem led to changes in the small-scale spatial structure of soil moisture and related processes. These changes in land utilization may therefore alter ecosystem function and further impact the regional climate.