An experiment comprising two irrigation regimes (110 and 160 mm cumulative evaporation from Class A pan) and four salinity levels of irrigation water (5, 10, 20 and 30 dS/m) was conducted on bio-saline production of Panicum antidotale Retz. Results showed that the artificial neural network model could predict dry matter yield variability by 99% using soil chemical properties. The sodium content of 30-60 cm depth had the most important factor affecting dry matter production of blue panic grass. And acidity, concentration of sodium and potassium in the depth of 90-60 cm, pH in the depth of 30-60 cm and organic matter in the depth of 0-30 cm were other important factors affecting the production of dry matter. Establishment of bio-saline system in this condition had a significant effect on soil chemical properties during short time and should be paid attention for productivity, profitability and sustainability of commercial in arid zones.

Keywords