Evaluation of Several Empirical Salinity-Yield Models for Sorghum Plant Response to Irrigation Water and Soil Salinity

Sorghum is considered as a cereal in the world but in Iran forage sorghum is preferred due to lack of forage. This research was conducted considering the importance of forage sorghum and the conditions of sorghum cultivation under salinity stress in Mazandaran province. The aims of this research were: 1) to investigate the functional response of forage sorghum (variety Speedfeed) to irrigation water salinity and soil salinity, and 2) to quantify the sorghum functional response to salinity during maturation, using several empirical models. Water salinity treatments included well water (control), ratios of 1:4, 2:4, and 3:4 sea water to well water, and pure sea water which were applied in a column study. Salinity-Yield threshold models (Mass and Hoffman, van Genuchten and Hoffman, Dirksen et al., Homaee et al., modified Weibull, bi-exponential and Gompertz( were fitted to the data obtained (relative yield of sorghum at maturation stage vs. electrical conductivity, EC) and the parameters of  models were estimated. The van Genuchten model and the Hoffman and Gompertz model were better than other models in fitting the relative yield data as a function of the EC of the saturated soil extract, due to the lower root mean square errors (RMSE) and Akaike’s information criterion (AIC) in these models (2.84 and –157.34 for the van Genuchten model and 3.11 and –152.53 for the Hoffmann and Gompertz model, respectively). The van Genuchten and Hoffmann model (–205.77, -1.09) and the corrected Weibull model (-213.21 and 0.95) were superior in fitting relative yield data as a function of EC of irrigation water. The salinity tolerance thresholds of sorghum based on fitting relative yield data as a function of EC of saturated soil extract and irrigation water were also estimated to be 3.65 and 2.33 dS m-1, respectively. Also, no significant difference was observed between the relative yield obtained from the models as a function of the EC of saturated soil extract or irrigation water. In other words, the mentioned models are capable to estimate the yield based on the salinity of the irrigation water.