
Parameter Estimation of a Root Water Uptake Model under Salinity Stress
Author(s) -
Fujimaki Haruyuki,
Ando Yoshitake,
Cui Yibin,
Inoue Mitsuhiro
Publication year - 2008
Publication title -
vadose zone journal
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 1.036
H-Index - 81
ISSN - 1539-1663
DOI - 10.2136/vzj2007.0025
Subject(s) - transpiration , salinity , reflectometry , soil science , irrigation , environmental science , soil water , arid , stress (linguistics) , evaporation , hydrology (agriculture) , dns root zone , greenhouse , mathematics , chemistry , time domain , agronomy , geology , geotechnical engineering , meteorology , physics , philosophy , linguistics , oceanography , computer science , biology , paleontology , biochemistry , photosynthesis , computer vision
Accurate prediction of root water uptake under salinity stress contributes to efficient water management in arid and semiarid lands. We present a cost‐effective and reliable method to determine parameter values in a widely used macroscopic root water uptake model. We conducted column experiments using soybean [ Glycine max (L.) Merr.] in a greenhouse. Six columns with one plant each were used: three were under salinity stress, the others provided potential transpiration. Three time‐domain reflectometry probes were inserted into each of the three columns to observe water content and electrical conductivity. In the daytime, the soil surface was covered to prevent evaporation. Weight of the columns was manually measured to obtain daily transpiration. After the stress period, root density distributions were obtained by dismantling the columns. Two parameter values were inversely determined by minimizing the sum of square difference between observed and calculated daily transpiration rates. Water uptake at each depth and time was calculated by substituting linearly interpolated osmotic potential into a stress response function. Optimized daily transpiration agreed well with the observations. In Addition, deviation in the three optimized response functions was small at low‐to‐moderate stress, indicating the reliability of the method.