View More View Less
  • 1 Kermanshah Branch, Islamic Azad University, Kermanshah, Iran
  • | 2 Faculty of Agriculture, University of Tehran, Karaj, Iran
  • | 3 Agricultural Research, Education and Extension Organization (AREEO), Kermanshah, Iran
  • | 4 Agricultural Jihad Organization, Qazvin, Iran
Restricted access

In the present study, efficiency of the artificial neural network (ANN) method to identify the best drought tolerance indices was investigated. For this purpose, 25 durum genotypes were evaluated under rainfed and supplemental irrigation environments during two consecutive cropping seasons (2011–2013). The results of combined analysis of variance (ANOVA) revealed that year, environment, genotype and their interaction effects were significant for grain yield. Mean grain yield of the genotypes ranged from 184.93 g plot–1 under rainfed environment to 659.32 g plot–1 under irrigated environment. Based on the ANN results, yield stability index (YSI), harmonic mean (HM) and stress susceptible index (SSI) were identified as the best indices to predict drought-tolerant genotypes. However, mean productivity (MP) followed by geometric mean productivity (GMP) and HM were found to be accurate indices for screening drought tolerant genotypes. In general, our results indicated that genotypes G9, G12, G21, G23 and G24 were identified as more desirable genotypes for cultivation in drought-prone environments. Importantly, these results could provide an evidence that ANN method can play an important role in the selection of drought tolerant genotypes and also could be useful in other biological contexts.

Supplementary Materials

    • Supplementary Material