Effect of temperature and water potential on Carthamus tinctorius L. seed germination: Quantification of the cardinal temperatures and modeling using hydrothermal time

Abstract

Seed germination, a physiological process, experiences large deviations based on temperature and water potential. Each factor can, separately or jointly, affect the germination percentage and germination rate. The response of plant growth rate (including germination rate) to both temperature and water potential can be described as a non-linear function of the thermal time, hydrotime, and the hydrothermal time models. We compared 3 non-linear regression models (Dent-like, segmented and beta) and thermal time, hydrotime and hydrothermal time models to describe the germination rate temperature and water potential relationships of safflower (Carthamus tinctorius L.). An investigation on an experiment for the joint effect of nine constant temperature regimes (5, 10, 15, 20, 25, 30, 35, 40 and 45 °C) and nine water potential values (0, −0.2, −0.4, −0.6, −0.8, −1, −1.2, −1.4 and −1.6 MPa) on safflower seed germination (SSG) was conducted. The outcome revealed that the segmented function (RMSE = 0.28, R2 = 0.82 and r = 0.90) was suitable for use in describing SSG response to temperature and water potential. Using this function, the base, optimum and ceiling emergence temperature and the physiological hour were estimated to be 3.9, 39.3 and 45 °C, and 5.7 h respectively. The hydrotime constant for SSG (θH), the middle value for the base water potential (Ψb(50)), the standard deviation in the water potential (σyb) and the thermal time constant for SSG (θT) spanned the values of 2.54–137.1 MPa h, −1.68–0.06 MPa, 0.32–0.72 MPa and 308.1–3963.1 °C h, respectively. The hydrothermal time parameters were 379.7 MPa °C h, −1.28 MPa and 3.10 °C for θHT (hydrothermal time constant), and Ψb(50) (median base water potential), Tb (base temperature) in temperatures and water potentials, respectively. All model parameters may be readily used in safflower simulation models.