Evaluation of the SALTMED model for sorghum under saline conditions in an arid region

SALTMED model has been developed to predict yield, soil salinity and watercontent under saline conditions. A two year field experiment was carried out during2012-13 to calibrate and validate the model for sorghum. Plants were irrigated withsalinity levels of 2, 6, 10 and 14 dS m-1. Results showed that there were significantdifferences between the observed and simulated sorghum dry matter (SDM) andyield. Absolute mean differences between the observed and simulated SDM valuesfor 2, 6, 10 and 14 dS m-1 were 0.45, 1.53, 0.04 and 1.07 Mgha-1, respectively. Soilwater contents (SWC) were overestimated at different soil depths. Meandifferences between the simulated and observed SWC at 0.0-0.3, 0.3-0.6, 0.6-0.9and 0.0-0.9 m soil depths were 0.02, 0.04, 0.02 and 0.03 m3m-3, respectively. Assalinity increased the mean differences between the observed and simulated SWCwere increased. There were no significant differences between the observed andsimulated soil salinities at 0.0-0.3, 0.3-0.6, 0.6-0.9 and 0.0-0.9 m soil depths. TheWillmott index of agreement value of the observed and simulated ECe at differentsoil depth were between 0.92-0.96. It is concluded that following successfulcalibration, the SALTMED model could predict soil salinity and SWC withreasonably good accuracy at different water salinity levels. Although, SALTMEDmodel reasonably well predicted soil salinity at different soil depth, there was aweak agreement between the observed and simulated soil water content at differentsoil depths. There was a fair agreement between the observed and simulated drymatter and grain yield at different water salinity levels.