Multivariate Models for Predicting the Maximal Diameter of the Wetted Area under Surface Drip Irrigation System

Abstract

Water shortage has been and will continue to be a key global-scale threat to agricultural production. One approach to mitigate the intensity this problem is the efficient use of water and this necessitates introduction of high efficient irrigation systems like drip irrigation. Reliable information about the dimensions of wetted soil under drip irrigation enables designers to find out optimal emitter flow rates and spacing to offer efficient use of irrigation water. Accordingly, the current study was initiated and the main objectives were to predict the ultimate diameter of wetting area under emitters from dripper discharge and other properties of the dominant soils in Erbil plain. To achieve the above objective 24 sites were selected over the indicated plain keeping in mind covering a wide spectrum of soil properties. At each site the soil moisture distribution in horizontal and vertical directions were monitored under three drip discharges of 1.2, 2.5 and 3.5 l hr^-1 such that each line represented a discharge level. The results indicated that among a host of input variable, emitter discharge, soil clay content and saturated hydraulic conductivity were the most influential factor affecting the maximal diameter of the wetted area (D).  A linear and a nonlinear model were also derived for predicting the maximal diameter of the wetted area (D). The mean absolute percentage errors were 10.37 and 8.57 % respectively. Similarly, a linear model was proposed for predicting the wetting depth with a reasonable accuracy. Additionally, the results also confirmed  that the model proposed by Schwartzman and Zur, (1986) had poor predictability for estimating D in the area under study.