Crop and Pasture Science Crop and Pasture Science Society
Plant sciences, sustainable farming systems and food quality

Just Accepted

This article has been peer reviewed and accepted for publication. It is in production and has not been edited, so may differ from the final published form.

Sensitivity and uncertainty analysis of the HYDRUS-1D model for root water uptake in saline soils

Wenzhi Zeng , Guoqing Lei , Yuanyuan Zha , Yuanhao Fang , Jingwei Wu , Jiesheng Huang


A variance-based global sensitivity analysis (EFAST) was applied to the Feddes module of the HYDRUS-1D model, and the sensitivity indices including both main (Si) and total effects (STi) of actual root water uptake (RWUa) to seven Feddes parameters were quantified at different growth stages of sunflower (i.e. seedling, bud, flowering, and mature). The effects of soil salinity, climate conditions, and crop root growth on parameter sensitivity were explored by analyzing three precipitation frequencies (PF) and two maximum root depths (MRD), across four field locations (FL) in China’s sunflower-growing regions which had different soil salinity levels. Uncertainties for the RWUa at four stages with varying Feddes parameters in different FL, PF, and MRD were evaluated. We found that the water stress factor about ceasing root water uptake (h4), and two salt stress factors ht and Sp which indicated the salinity threshold and the slope of the curve determining the fractional root water uptake decline per unit increase in salinity below the threshold respectively were three most important Feddes parameters for RWUa estimation in HYDRUS-1D. In addition, the effects of soil salinity and precipitation frequencies (PF) were stronger than maximum root depth (MRD) on the order of the parameters’ impacts on RWUa. Our study suggested that h1, h2, h3h, and h3l might be determined by economical method (e.g. literature review) in saline soils with limited observations, but it is better to calibrate wilting point (h4) and salt stress parameters (ht and Sp) based on local measurements.

CP17020  Accepted 30 November 2017

© CSIRO 2017