Ali Javadi - Water Engineering Department, College of Agriculture, Isfahan University of Technology, Isfahan ۸۴۱۵۶-۸۳۱۱۱, Iran
Mohammad Shayannejad - Water Engineering Department, College of Agriculture, Isfahan University of Technology, Isfahan ۸۴۱۵۶-۸۳۱۱۱, Iran
Behrouz Mostafazadeh-Fard - Water Engineering Department, College of Agriculture, Isfahan University of Technology, Isfahan ۸۴۱۵۶-۸۳۱۱۱, Iran
Mohammad Mahdi Dorafshan - Water Engineering Department, College of Agriculture, Isfahan University of Technology, Isfahan ۸۴۱۵۶-۸۳۱۱۱, Iran

Sensitivity analysis is an essential aspect of modelling. In this study, HYDRUS-1Dsoftware was used to determine the sensitivity index of van Genuchten-Mualemparameters on cumulative infiltration under different irrigation management. Treatmentsincluding three irrigation water qualities, ECi= 0.6, 3 and 6 dS/m, two irrigationmanagements and two irrigation periods (short and long period). Two irrigationmanagements included intermittent irrigation at the depletion of 30% field capacity (M1)and daily irrigation with leaching fraction of 0.15 (M2). Infiltration tests were applied onsoil columns after 16 and 114 days. After calibrating the infiltration process byHYDRUS-1D software, the effect of a 20% change in soil hydraulic parameters (0.8 to1.2 parameter values) on cumulative infiltration was investigated. Under the intermittentirrigation that the soil was dried and wetted successively, infiltration is most sensitive atthe short irrigation period, whereas under the daily irrigation that the soil was wettedcontinuously; it is the most sensitive at the long irrigation period. In fact, with increasingsoil permeability, the sensitivity of infiltration to change parameter increases. Therefore,the highest and lowest values of the sensitivity index were determined for the daily andintermittent irrigations with poor irrigation water quality, respectively. The resultsshowed that the most sensitive parameter was Ks, followed by the θs, n, α and θr with l theleast sensitive parameters. Therefore, the number of parameters used to optimize themodel will be reduced to three parameters Ks, θs and n.

HYDRUS-1D, Inverse modeling, Sensitivity analysis