Abstract

A set of twelve field lysimeters were designed and installed for deficit irrigation and crop water stress studies. Permanent rainout shelters were designed to be lightweight, portable, and inexpensive. The shelters provided continuous rain protection. However, polycarbonate roofing panels reduced solar radiation levels by 35 to 40 percent. Temperature and relative humidity differences were small among the shelters and with respect to the unsheltered area.

Cotton (Gossypium hirsutum L. cv. Stoneville 825) was grown during the summer of 1985; in the lysimeter facility. A sequential water stress study was performed to analyze the applicability of the Stress Day Index yield response model (Hiler, 1969; and Hiler and Clark, 1971) to sequential stress situations. Water stress treatments included a well watered control (CON), a stress during peak flowering (ST1), a stress during late flowering (ST2), and a sequential stress (SEQ) which encompassed peak and late flowering with one irrigation between stress periods.

Water stress during peak flowering reduced yield by 37% while water stress during late flowering reduced yield by 33%. A sequential stress reduced the total yield by 44% with respect to nonstressed conditions. Comparing sequential stress results with the results from the first stress period (ST) showed that the sequential stress reduced yield by 11% with respect to the first stress.

A stress day index (SDI) analysis was performed. Yield data from this study were used to calculate crop susceptibility (CS) factors while leaf water potential ((psi)(,L)), and available water ((theta)(,A)) data were used for stress day (SD) factor calculation. These data were normalized and placed in dimensionless terms. The modified values of (psi)(,L) and (theta)(,A) each distinctly defined the stress periods. Plots of relative yield (Y/Yp) versus SID supported the linearity of the SDI model with coefficients of determination (r('2)) that ranged from 0.953 for the (theta)(,A) based SDI model to 0.997 for the (psi)(,L) based model. The effects of the sequential stress were successfully modeled using data from the CON, ST1, and ST2 treatments. This was accomplished using either (psi)(,L) or (theta)(,A) based SD factors.