Abstract: By the methods of geostatistics, this paper studied the spatial heterogeneity of surface soil moisture in a broad-leaved Korean pine forest in Changbai Mountains. Three sampling plots sized 50 m×50 m were installed, and the moisture content in 0-7.5 cm soil layer was measured with a time domain reflectometry. The coefficient of variation of the surface soil moisture was 24.32% in plot 1, 24.11% in plot 2, and 23.60% in plot 3, suggesting a moderate variability. The variogram of the surface soil moisture could be well fitted by spherical model, and had a higher spatial heterogeneity, which mainly came from spatial autocorrelation. The structural ratio of surface soil moisture was 57.9% in plot 1, 83.3% in plot 2, and 90% in plot 3, and the regressed range (distance beyond which samples were spatial independent, an indicator of patch size) was 5.5-13.1 m, being consistent with the results from Bayesian method. Through Kriging interpolation, the surface soil moisture content was 49.3% in plot 1, 52.8% in plot 2, and 42.6% in plot 3.

Key words: broad-leaved Korean pine forest, soil moisture, geostatistics, spatial heterogeneity, climatic mean value, rice, delayed growth-type cold damage, evaluation, Liaoning Province.