Regional-scale variation and distribution patterns of soil saturated hydraulic conductivities in surface and subsurface layers in the loessial soils of China

SummarySaturated hydraulic conductivity (Ks) is an important soil property that shows a high degree of spatial heterogeneity. There is a lack of research that investigates and determines Ks at a regional scale, due to the challenges associated with the required intensive sampling. To determine the closely correlated factors affecting Ks at a regional scale and to then generate a regional distribution map of Ks, we selected 382 sampling sites across the Loess Plateau of China (620,000 km2) and collected undisturbed and disturbed soil samples from two soil layers (0–5 and 20–25 cm). We found that both surface Ks and subsurface Ks had log(base 10)-normal distributions, and demonstrated strong spatial variability (CV = 206% and 135%, respectively). Surface LogKs was most closely correlated with LogSand, LogSilt, LogSG (slope gradient), LogSSWC (saturated soil water content), vegetation coverage and land use; while subsurface LogKs was correlated with LogClay, SSWC, LogSG, LogAltitude, LogGY (growth year) and land use. Geostatistical analysis indicated that semivariograms of surface and subsurface Log Ks could be best fitted by an isotropic exponential model, with effective ranges of 204 km and 428 km, respectively. Distribution maps of Ks produced by kriging indicated a pronounced spatial pattern and demonstrated an obvious spatial depth gradient. The spatial distribution patterns of Ks at a regional scale in the loessial soils of China comprehensively reflected soil hydraulic properties and the combined effects of soil texture, vegetation, topography and human activities.

► Surface and subsurface Ks demonstrated a strong spatial variability at a regional scale.
► Surface LogKs was dominated by LogSand, LogSilt, LogSG, LogSSWC, VC and LD.
► Subsurface LogKs was controlled by LogClay, SSWC, LogSG, LogAL, LogGY and LD.
► Isotropic exponential model was the best-fitted model of LogKs at the two layers.
► The magnitudes and patterns of Ks demonstrated an pronounced spatial depth gradient.