Fig. 1

Relationship between plant traits and yield explained through quantile regression modeling: Each panel represents the estimated value of regression coefficient in black dots and the estimated 95% confidence interval in grey area as a function of quantile level for respective covariate in the model with yield as response, obtained by fitting a multiple quantile regression model (n=808). The horizontal black lines represent the zero value of the estimated coefficients. The variable condition is a dummy variable and was coded 1 for non-saline conditions and 0 for saline conditions. a*b represents the interaction between covariates a and b. a Intercept, b Plant height (cm), c Flowering time (days), d Ripening period (days), e Ear number per plant, f Grain number per ear, g Condition, h Condition*Plant height (cm), i Condition*Flowering time (days), j Condition*Ripening period (days), k Condition*Ear number per plant, l Condition*Grain number per ear