Spatial scaling of saturated hydraulic conductivity of soils in a small watershed on the Loess Plateau, China

Purpose

Soil saturated hydraulic conductivity (K _S) is a key variable in hydrologic processes, the parameters of which have strong scale-dependency. Knowing the scaling dependency of K _S is important when designing an appropriate sampling strategy.

Materials and methods

Determinations of K _S were made for 4,865 undisturbed soil samples, collected from a grid with cells of 10?×?10?m in the Daye watershed (50?ha) on the Loess Plateau, China. The dataset was ??re-sampled?? to investigate the effect on K _S of scales that differed by two orders of magnitude in terms of spacing and support, and eight scales of extent. The variance, correlation length, and nugget?Csill ratio derived by analysis of the full dataset were taken to be the true values. Apparent values of variance, correlation length, and nugget?Csill ratio were those calculated for each re-sampled data sub-set.

Results and discussion

Comparing the parameter values at different scales showed that apparent variance increased with increasing extent (p?p?p?=?0.137). Apparent correlation length increased with increasing extent and support (p?p?p?K _S was in the order of extent > support > spacing for all three parameters, with mean coefficient of determination values of 0.96, 0.88, and 0.53, respectively.

Conclusions

The statistical properties investigated for K _S were found to be scaling-dependent, which would benefit sampling strategy design.