Predicting the Rooting Depth, Dynamic Root Distribution and the Yield of Sunflower under Different Soil Salinity and Nitrogen Applications

•High salinity induced rapid fine root elongation after sunflower bud initiation.•The effect of nitrogen fertilizer on fine root growth depended on soil salinity levels.•Belowground root development was nonlinearly related with aboveground canopy growth.•A model was proposed to describe the flexible fine root distribution under salt stress. Soil salinity is a major abiotic stress affecting crop production in arid and semi-arid regions, and nitrogen (N) is the main yield-limiting nutrient in most agroecosystems. To find out the effects of different soil electrical conductivity (EC) and N application rates on sunflower (Helianthus annuus L.) rooting depth, dynamic root distribution and its yield, a three-year field experiment was conducted from 2015 to 2017 in Hetao Irrigation District, China. The experiment consisted of three initial salinity (IS) levels expressed as the EC of a saturated soil extract (ECe) (S0: ECe < 3 dS m-1; S1: 3< ECe 6 dS m-1) and five N rates (45, 90, 135 180 and 270 kg N ha-1), referred as N0, N1, N2, N3 and N4 rates, respectively. The dynamics of sunflower root growth was in-situ monitored by minirhizotrons and revealed flexibility in its distribution among different treatments. Salt stress obviously decreased root penetration in the seedling stage, while it accelerated root elongation below 28.6 cm depth after bud initiation and prolonged its active development to later stages. In particular, the highest fine root length density (FRLD) of S2N1 and S2N3 were rapidly increased to 5.27 and 6.44 cm cm-3 before anthesis, respectively, which were more than three folds of those at S0 and S1 levels. Meanwhile, the N3 rate in this study obviously aggravated salt stress and accelerated sunflower root senescence at S2 level, but it could promote root vertical development after anthesis. In addition, our study proposed a dynamic root distribution model in saline fields (DRD-SF) to predict the dynamics of FRLD distribution based on IS levels, effective nitrogen application (ENA) and the observed leaf area index (LAI) at different growth time. The data in 2015 and 2017 were used to validate our model developed from 2016 data, demonstrating acceptable simulation of the flexible distribution of sunflower fine roots under the interactions of different IS levels and N rates. Moreover, this study also associated sunflower canopy development with yield production, by proposing a yield response function in the form of bivariate quad