Abstract: Soil saturated hydraulic conductivity (Ks), field capacity (FC) and permanent witling percentage (PWP) are the important soil hydraulic parameters. These parameters are also the critical factors in bio-hydrologic models. In order to understand the regional-scale spatial variability of these hydraulic parameters, the Ks, FC and PWP data (0-5-cm soil layer in depth) were obtained from 107 sampling sites in Xinjiang. Both traditional statistics and geostatistics were used to explore the spatial variations of Ks, FC and PWP. Estimations of these parameters were further conducted by multiple stepwise regression (MSR) and pedotransfer functions (PTFs). The results indicated that the mean values of Ks, FC and PWP were 10.999 mm·h^-1, 0.162 g·g^-1 and 0.077 g·g^-1 respectively. These three hydraulic parameters varied moderately with Cvs of 39.88%-96.07%, and had strong spatial correlations in a range of 97-291 km in Xinjiang. MSR performed better than PTFs in estimating soil hydraulic parameters. Based on the land use, bulk density, soil texture, soil organic carbon content and slope aspect, the determination coefficients of the MSR models for Ks, FC and PWP were 0.290, 0.494 and 0.491, and the root mean square errors were 2.540 mm·h^-1，0.039 g·g^-1 and 0.023 g·g^-1, respectively. The development of these equations is beneficial to the rapid estimation of soil hydraulic parameters and can thus provide the key parameters for the agricultural irrigation and eco-hydrological models in arid northwest China.

Key words: soil saturated hydraulic conductivity, field capacity, permanent witling percentage; regional scale, multiple stepwise regression, Xinjiang