EFFICIENCY OF THE SPAW MODEL IN ESTIMATION OF SATURATED HYDRAULIC CONDUCTIVITY IN CALCAREOUS SOILS

Abstract

The measurement of soil hydraulic properties is tedious, time-consuming, and costly. An alternative
approach is to formulate models that utilize the physical and chemical properties of the soil as input
variables to predict soil saturated hydraulic conductivity (Ks). However, the previous studies have not
paid attention to the calcium carbonate content in their models; it can lead to reducing the size and
number of the pores in the soil which, in turn, can lead to reduction Ks
. Here we evaluated the ability of
the Soil, Plant, Atmosphere, and Water (SPAW) model to predict Ks under different states of compaction
for calcareous soils with wide-ranging textures sampled along a precipitation gradient in northwestern
Iraqi Kurdistan. The results revealed that the best match occurred under loose to normal state of
compaction for these soils. Among the soil properties, sand content was high significantly correlated with
Ks followed by CaCO3, clay, organic matter content, silt and Electrical conductivity. A pedotransfer
function (PTF) was proposed using these data and its results were compared to these from the SPAW
model. Root mean square error (RMSE) and coefficient of variation (CV) for the comparison between
measured Ks values and those predicted by the SPAW model were very high 2.7×10-4
cm s−1
and 166%
respectively, that due to the values of Ks predicted by the SPAW model are overestimated for calcareous
soils, for these reasons the accuracy of the SPAW model was improved via calibration. The RMSE and
CV of the calibrated SPAW model were dropped to 9.8×10-5
cm s−1
and 61.2%, respectively. Additionally,
the accuracy of our best PTF that constructed from sand, clay, and CaCO3 was slightly higher than the
calibrated SPAW model. Therefore, it is recommended to use the calibrated SPAW model for predicting
Ks
in calcareous soils.