PTFs for predicting LLWR from various soil attributes including cementing agents

Abstract

Least limiting water range (LLWR), the range of soil water content at which plant growth is least limited by water potential, soil aeration or soil mechanical resistance is routinely calculated from water release curve (WRC) and soil resistance curve (SRC). There is no enough information about the effect of various soil attributes including cementing agents (metal oxides, carbonates and organic carbon) on LLWR. The present study evaluates the effect of several soil characters, including cementing agents, texture, sodium adsorption ratio (SAR), bulk density (Db) and cation exchange capacity (CEC) on LLWR and develops proper pedotransfer functions (PTF) for its prediction. Disturbed and undisturbed samples of 32 soils with wide range of properties were collected from Ahar, Horand and Tabriz regions, northwest of Iran. Undisturbed soil samples were equilibrated to matric pressures of 0.001, 0.004MPa in hanging columns and of 0.01, 0.03, 0.1, 0.5, 1.5MPa in pressure plate and the equivalent water contents were measured gravimetrically. Penetration resistance of each sample at the mentioned matric pressures was measured by a hand cone penetrometer. Soil water contents at field moisture capacity (θfc) and permanent wilting point (θwp) (matric pressures of 0.01 and 1.5MPa, respectively) were predicted using WRC; water content at 2MPa penetration resistance (θsr) was estimated from SRC. Water content at 10% air filled porosity (θafp) was taken as θs-0.1 and LLWR computed from the above moisture coefficients. The relative influence of soil characters, as independent variables on the moisture coefficients (θafp, θfc, θsr, θwp) and on LLWR was evaluated separately using multiple linear stepwise regression and then appropriate pedotransfer functions were developed to predict LLWR. The relative influences of soil attributes on the moisture coefficients and on LLWR were not similar. Clay content, bulk density (Db) and ammonium oxalate extractable iron produced considerable effects on LLWR. Grouping the examined soil samples according to Db or clay content led to more accurate prediction of LLWR (R2=0.86 and 0.76, respectively) than forcing all samples in a single group (R2=0.31). In samples with Db≥1.4Mg/m3, citrate-bicarbonate-dithionate extractable aluminum, as cementing agent, turned to be the second most influential soil attribute (after clay) on LLWR while the latter was not affected by calcium carbonate equivalent. Even though SAR significantly (P<0.01) affected both θwp and θfc, its net effect on LLWR was insignificant.