Introduction

Seed germination (SG) is a complex physiological process that is affected by abiotic stresses, including temperature (T), water stress, and salinity, and by other physical and biological factors (Bradford 2002; Bakhshandeh et al. 2020; Vahabinia et al. 2019). T is a critical factor affecting both the germination percentage (GP) and the germination rate (GR, or 1/tg) (Ghaderi-Far et al. 2010; Bewley et al. 2013). During the germination phase of the plant life cycle, there exists a T below which germination is prevented, called the base temperature (Tb), an optimum temperature (To, single and/or range of Ts) at which germination is most rapid, and a T above which germination does not occur, called the ceiling temperature (Tc) (Bakhshandeh et al. 2013). These so-called cardinal Ts vary depending on species and environmental conditions under which they were produced (Hatfield and Prueger 2015). GR usually increases linearly with T between Tb and To, and declines linearly and/or curvilinearly at Ts > To (Bradford 2002; Rowse and Finch-Savage 2003; Mesgaran et al. 2017; Abdellaoui et al. 2019). Water availability is an important factor affecting SG and failure or success of plant establishment. Commonly, GP and GR increase with water availability and decrease at more negative water potential (ψ) (Mesgaran et al. 2013; Atashi et al. 2015). Salinity may limit SG through both osmotic and ion-specific effects (Zhang et al. 2010; Llanes et al. 2016). Dissolved salts decrease the ψ of saline solutions, with a 1 M solution of NaCl having a ψ of approximately − 4.4 MPa at 25 °C (Nobel 1983). In addition, accumulations of inorganic ions (e.g., sodium [Na+] and chloride [Cl−]) and compatible organic solutes (e.g., amino acids, proline, etc.) result in osmotic adjustment, or a reduction of ψ inside the seed cells, that can enable SG to occur at lower water potentials (Parihar et al. 2015; Seal et al. 2018). These factors, T, ψ and salinity, influence SG alone and/or in combination, and are relevant to crop production and to ecological distributions of species (Baskin and Baskin 2014).

The hydrotime model developed by Gummerson (1986) and Bradford (1990) has been used to quantify the influence of ψ on germination at a given T. According to the model, the time to germination is inversely proportional to the difference between the ψ of the seed environment and the physiological ψ threshold for radicle emergence (base ψ or ψb), which varies among seeds in the population. The ψb values for various germination percentiles (ψb(g)) usually exhibit a normal distribution within seed populations (Dahal and Bradford 1994; Bradford and Still 2004; Bakhshandeh and Gholamhossieni 2018), although other distributions, such as Gumbel, Weibull, Log-Normal, Logistic, Log-logistic, Inverse Normal and Gamma, have been applied to specific seed populations or mixtures of populations (Mesgaran et al. 2013; Atashi et al. 2015; Moltchanova et al. 2020).

Gummerson (1986) also proposed a hydrothermal time (HTT) model of SG stating that at Ts ≤ To (i.e., between Tb and To) and at any constant ψ, the time to germination (tg) of any germination percentile of a seed population is a function of the degree to which T and ψ exceed their respective base values, Tb and ψb, at which germination is inhibited (Gummerson 1986; Dahal and Bradford 1994; Bradford 1995). In this model, Gummerson (1986) assumed that θHT and Tb are constant and equal among seeds and only ψb(g) varies in the seed population. However, this model does not predict the inhibition of SG (both GR and GP) at Ts > To (i.e., between To and Tc). Consequently, Alvarado and Bradford (2002) suggested another model in which the relationship between GR vs. T will be negatively linear at Ts > To with a common To for all germination percentiles but differing Tc values in the population. This decline in GR and GP was due to a linear increase in ψb(g) as T increased above To, and a constant value (i.e., kT) was defined as the increase in ψb(50) per degree as T increased above To. This modeling approach successfully predicted SG in many crops, such as potato (Solanum tuberosum L.) (Alvarado and Bradford 2002), watermelon (Citrullus vulgaris) (Bakhshandeh et al. 2015), sesame (Sesamum indicum L.) (Bakhshandeh et al. 2017) and eruca (Eruca sativa) (Bakhshandeh et al. 2019).

However, for some other crops such as wild barley (Hordeum spontaneum Thell.) and little canarygrass (Phalaris minor L.) (Mesgaran et al. 2017) and Retama reatam (Abdellaoui et al. 2019), the relationship between GR and T did not transition sharply from increasing to decreasing at To. Rowse and Finch-Savage (2003) showed also that a T lower than To should be defined (Td) at which ψb(50) starts to change, i.e., [ψ − ψb(50) + kT (T − Td)], resulting in a curvilinear peak in GR around and above To (e.g., Bakhshandeh and Gholamhossieni 2019).

Seal et al. (2018) used the same methodology to quantify the SG characteristics of the halophyte Suaeda maritima under different NaCl concentrations at sub-optimal Ts. They developed a halothermal time (HaloTT) model that utilized log NaCl to replace water potential in the HTT model. That is, the germination behavior at sub-optimal Ts was governed by the salt concentration threshold distribution (NaClb(g)) relative to the NaCl of the environment. In the current work, we employ the HaloTT model and extend it to describe SG in response to T across all Ts at any constant NaCl concentration. To our knowledge, no HaloTT model to quantify SG responses at supra-optimal Ts has been proposed as yet, thus, our study is the first report in this case.

Chicory (Cichorium intybus L.; Asteraceae) is a perennial herb, native to Europe, Africa, America and many parts of Asia, including Iran (Jouad et al. 2001; Benítez et al. 2010; Guarrera and Savo 2016). The medicinal effects of chicory are well documented (Abbas et al. 2015; Sampaio et al. 2009; Yoo et al. 2011). Furthermore, this plant copes with different soil types and abiotic and biotic stresses and can be used as food and feed. Based on its broad tolerance to environmental conditions, we have investigated the SG responses of chicory to combinations of T, ψ and salinity using the models described above. The aims of this study were: (1) to study the effects of heat, water and salt stresses on SG characteristics of chicory; (2) to estimate the cardinal Ts for germination using the HTT and HaloTT models; (3) to measure the uptake of salt ions into chicory seeds in response to salinity; and (4) to calculate the amount of osmotic adjustment attributable to NaCl uptake into the seeds when imbibed in salt solutions.

Materials and methods

Seed collection and treatments

Chicory seeds were provided from Mazandaran province, Iran, in 2017. They were maintained in the dark at 5 °C for 3 months. The seed viability [assessed according to ISTA (2018)] was > 98% under optimal conditions at the beginning of the experiments. The treatments were seven constant Ts (10, 15, 20, 25, 30, 35 and 40 °C), four levels of ψ (− 0.3, − 0.6, − 0.9 and − 1.5 MPa) and four levels of salinity (0.075, 0.150, 0.225 and 0.300 M). For the control we used distilled water in all experiments. Polyethylene glycol (PEG8000, Scharlau, Spain) was applied to make solutions having different ψs (and corrected for Ts) based on Michel and Radcliffe (1995) method. An osmometer (Model 5520: Wescor Inc., USA) was used initially and also every two days during the experiments to re-check the ψs for each T. Sodium chloride (NaCl) was applied to prepare different levels of salt stress that are mentioned above. The NaCl concentration (M) at each T was converted to water potential (ψ, MPa) according to the Van’t Hoff relation (Van’t Hoff 1887), and confirmed using the osmometer as well.

Seed germination tests

Fifty seeds for each of four replicates per treatment were placed on two sheets of Whatman No. 1 filter paper within 8 cm Petri dishes with six mL of the test solutions, supported by 0.1% Thiram. An incubator with ± 0.5 °C precision was used and the Petri dishes were randomly placed within in the dark, except during recording times. The Petri dishes were put into a plastic bag to avoid water evaporation. Seeds were counted several times daily, depending on the T, ψ and salt stress in all experiments. The seeds at least 2 mm long of radicle were considered as germinated seeds. At each counting time, germinated seeds were removed to avoid an error. The experiments were ended when no additional seeds germinated for three consecutive days.

Hydrotime and halotime models

A repeated probit regression analysis as followed by Bradford (1990) were used to analyses and determine the model parameters. GR50 (h−1) was calculated by interpolation by curves fit to the time course data according to:

$${\text{GR}}_{{{5}0}} \, = \,{1}/t_{{{5}0}} ,$$
(1)

where GR50 is the GR for the 50th percentile of germination (h−1) and t50 is the time to reach 50% germination.

To quantify the germination time courses data in the osmotic solutions at 10, 15, 20, 25, 30, 35 and 40 °C, the hydrotime model was used. This model is expressed as (Gummerson 1986):

$$\theta_{\text{H}} \, = \,\left( {\psi \, - \,\psi_{\text{b}} \left( {\text{g}} \right)} \right)t_{\text{g}} ,$$
(2)

where θH is the hydrotime constant (MPa h or MPa d); ψ is the actual water potential of the medium (MPa); ψb(g) is the base value of ψ inhibiting radicle emergence of percentage g (MPa); and tg is the actual time to germination of percentage g. Thus, the ψb(50) is the base water potential for the median or 50th percentile SG fraction.

A similar method was used to determine the parameters of halotime model by substituting the NaCl for ψ, the NaClb(g) for ψb(g) and the θHalo for θH. For each salt treatment at 10, 15, 20, 25, 30, 35 and 40 °C, the halotime model can be written as (Seal et al. 2018):

$$\theta_{\text{Halo}} \, = \,\left( {{\text{NaCl}}_{\text{b}} \left( {\text{g}} \right) - {\text{NaCl}}} \right)t_{\text{g}} ,$$
(3)

where θHalo is the halotime constant (M h or M d); NaClb(g) is the base value of NaCl inhibiting radicle emergence of percentage g (M); NaCl is concentration of the imbibition medium (M); and tg is the actual time to germination percentage g (h or d). Thus, the NaClb(50) is the median base NaCl concentration for the 50th percentile SG fraction. We note that the original equation in Seal et al. (2018) would result in negative values for θHalo since the term used there (NaCl − NaClb(g)) [or NaClmax g as defined in Seal et al. (2018)] will be a negative number, as the magnitude of NaCl is smaller than NaClb(50) at any NaCl permitting germination. We have reversed the order of NaClb(g) and NaCl in Eq. 3 to give positive values of θHaloTT.

Hydrothermal time and halothermal time models

At Ts ≤ To, the HTT model is expressed as (Gummerson 1986):

$$\theta_{\text{HTT}} \, = \,\left( {\psi - \psi_{\text{b}} \left( {\text{g}} \right)} \right)\left( {T - T_{\text{b}} } \right)t_{\text{g}} ,$$
(4)

or

$${\text{Probit (g)}}\, = \,[\psi - (\theta_{\text{HTT}} /(T - T_{\text{b}} )t_{\text{g}} ) - \psi_{\text{b}} ( 50)]/\sigma_{{\psi {\text{b}}}} .$$
(5)

At Ts > To, the model should be modified to (Alvarado and Bradford 2002):

$$\theta_{\text{HTT}} \, = \,[\psi - (\psi_{\text{b}} ({\text{g}})\, + \,k_{T} (T - T_{\text{d}} ))](T - T_{\text{b}} )t_{\text{g}} ,$$
(6)

or

$${\text{Probit (g)}}\, = \,\{ [ (\psi - k_{\text{T}} \left( {T - T_{\text{d}} } \right) ) { }{-}\theta_{\text{HT}} /(\left( {T - T_{\text{b}} } \right)t_{\text{g}} )] - \psi_{\text{b}} \left( { 50} \right)\} /\sigma_{{\psi {\text{b}}}} ,$$
(7)

where Td represents the T at which GR begins to be affected by warm Ts, which can be lower than To (Rowse and Finch-Savage 2003).

For the HaloTT model, the model proposed by Seal et al. (2018) (with correction as above) at Ts ≤ To was used:

$$\theta_{\text{HaloTT}} \, = \,\left( {{\text{NaCl}}_{\text{b}} \left( {\text{g}} \right) - {\text{NaCl}}} \right)\left( {T - T_{\text{b}} } \right)t_{\text{g}} ,$$
(8)

or

$${\text{Probit }}\left( {\text{g}} \right)\, = \,\left[ {{\text{NaCl}}\, + \,\left( {\theta_{\text{HaloTT}} /\left( {T - T_{\text{b}} } \right)t_{\text{g}} } \right) - {\text{NaCl}}_{\text{b}} \left( { 50} \right)} \right]/\sigma_{\text{NaClb}} ,$$
(9)

where θHaloTT is the HaloTT constant (M °C h or M °C d) and σNaClb is the standard deviation of the NaClb(g) distribution.

To our knowledge, no HaloTT model to quantify SG responses at Ts > To has been proposed as yet. Thus, we developed the model below as the first report:

$$\theta_{\text{HaloTT}} \, = \,\left[ {{\text{NaCl}}_{\text{b}} \left( { 50} \right) - \left( {{\text{NaCl}}\, + \,k_{T} \left( {T - T_{\text{d}} } \right)} \right)} \right]\left( {T - T_{\text{b}} } \right)t_{\text{g}} ,$$
(10)

or

$${\text{Probit }}\left( {\text{g}} \right)\, = \,\left[ {\left( {{\text{NaCl}}\, + \,k_{T} \left( {T - T_{\text{d}} } \right)} \right)\, + \,\left( {\theta_{\text{Halo}} /\left( {T - T_{\text{b}} } \right)t_{\text{g}} } \right) - {\text{NaCl}}_{\text{b}} \left( { 50} \right)} \right]/\sigma_{\text{NaClb}} .$$
(11)

Thus, SG response can be quantified at all Ts and ψs by the HTT models (Eqs. 4–7) and at all Ts and NaCls using the HaloTT models (Eqs. 8–11).

Seed moisture contents

Three replicates of 1 g chicory seeds were sown on two sheets of Whatman No. 1 filter paper in 8 cm Petri dishes containing 6 mL of water, 0.075, 0.150, 0.225 and 0.300 M NaCl at 10, 15, 20, 25, 30, 35 and 40 °C. Immediately prior to the initiation of germination (estimated from the cumulative GP vs. time curve, Fig. 2), seeds were rinsed several times with distilled water and blotted with filter paper to remove any solution from the testa. The seeds weighted, then dried in an oven at 103 °C for 17 h (ISTA 2018) and seed moisture content (SMC) was expressed on a dry weight basis.

Seed Na+ and Cl− concentrations and estimated osmolality

A flame photometric method as described by Estefan et al. (2013) was used to measure Na+ concentration in the chicory seeds obtained from the experiment described above (seed moisture contents section). Dried chicory seeds (103 °C for 17 h) were powdered by a mill and 0.5 g for each replicate was oxidized at 550 °C for 5 h and 2 N hydrochloric acid was used to digest the dry residue. Na+ was analyzed using a flame photometric instrument (model PFP7, Jenway, UK) and standard curves were constructed using 1000 ppm Na flame photometry standard solution (Product No: 025021, Jenway, UK) for quantification. To determine chloride (Cl−) concentration in the seeds, 0.2 g of dry powdered seeds for each replicate was placed in 20 mL of distilled water at > 90 °C for 1 h, then the amount of Cl− was measured by AgNO3 (Mohrs’s titration) method (Estefan et al. 2013). The summed concentrations of Na+ and Cl− (mmol kg DW−1) were divided by the seed water content (kg H2O kg DW−1) at the time of sampling to estimate the total osmolality (mmol/kg H2O) contributed by these ions, making the simplifying assumption that all of the water in the seed was symplastic. We then used the Van’t Hoff relation to convert the osmolality of salt ions in the seeds to ψ (MPa).

Statistical analyses

All statistical analyses were performed using the Statistical Analysis System ver. 9.4 (SAS Institute 2015), Excel and Sigma Plot ver. 11 software (Systat Software Inc., San Jose CA, USA, www.systatsoftware.com). A two-way ANOVA with the least significant difference test (LSD) at 0.05 probability level was used to analyze the effects of T, ψ, salinity and their interactions on SG, seed Na+ and Cl− concentrations and SMC.

Results

Effects of water potential, temperature and salinity on GP and GR

Water potential, NaCl and their interactions significantly affected GP and GR50 (p > 0.001) (Table 1). Supra-optimal T reduced GP (in water) while ψ and/or NaCl reduced GP at both lower and higher T. Both GP and GR50 decreased with decreasing ψ (Fig. 1) and with increasing NaCl in the medium at each tested T (Fig. 2). The relationship between GR50 and T was linear at T < 30 °C (sub-optimal Ts) under both conditions and shared a common Tb (Fig. 3). In general, GP and GR50 were more sensitive to changes in ψ and/or NaCl than to changes in T (Figs. 1, 2).

Table 1 Analysis of variance (mean squares) and means comparison for germination percentage (GP) and median germination rate (GR50) of chicory as influenced by temperature (T), water potential (ψ) and salinity (NaCl)
Full size table
Fig. 1
figure 1

Chicory seed germination at different temperatures under various water potentials

Full size image
Fig. 2
figure 2

Chicory seed germination at different temperatures under various concentrations of NaCl

Full size image
Fig. 3
figure 3

Hydrotime constant (θH, a), base water potential (ψb(50), closed circles and solid line) and base NaClb(50) after converting to MPa by van’t Hoff equation (open circles and dashed line; b), halotime constant (θHalo, c) and base NaCl concentration (NaClb(50), d) as a function of temperature (T) for chicory seeds. In b and d, the predicted lines are for the 50th percentile intercepts ψb(50) = 0 MPa and NaClb(50) = 0 MPa at Tc (43 °C), respectively. Inset a1: the inverse of θH vs. T, which will be linear below To (and extrapolates to Tb at the intercept on the x axis) and then is constant at T > To, as this maximized the fitting of the model. Inset c1: the inverse of θHalo vs. T, which continues increasing linearly up to 40 °C

Full size image

Hydrotime and halotime models

The hydrotime model successfully described germination data across all ψs at each T, with R2 > 0.88 (Table 2; Fig. 1). The θH values declined with increasing T from 62.9 MPa h at 10 °C to 9.0 MPa h at 30 °C and was fixed to a constant value of 9.0 MPa h at Ts > 30 °C (i.e., Ts > To) because this optimized the model (Table 2; Fig. 3a). The inverse of θH values (1/θH) was linear below To (and extrapolated to Tb = 4.8 °C, on the x axis) (Fig. 3a). The ψb(50) remained approximately constant at sub-optimal Ts (i.e., ≤ 30 °C), ranging from − 0.813 to − 0.839 MPa, then increased linearly to − 0.573 and − 0.210 MPa at 35 and 40 °C, respectively, and intersected 0 MPa at 43 °C (i.e., Tc(50)) (Table 2; Fig. 3b) The ψb(50) values increased with increasing T above To with a kT (slope) value of 0.066 MPa °C−1 (Table 2; Fig. 3b).

Table 2 The hydrotime (Eq. 2) and halotime (Eq. 3) models parameters for describing chicory seed germination under different Ts and ψs and Ts and NaCl concentrations, respectively. R2 is the coefficient of determination
Full size table

We used the same method to estimate the parameters of the halotime model at each T, with R2 ranging from 0.82 to 0.97 (Table 2; Fig. 2). The θHalo values non-linearly decreased with increasing T from 14.1 M h at 10 °C to 2.1 M h at 40 °C (Table 2; Fig. 3c). The inverse of θHalo values (1/θHalo) were linear with T up to 40 °C and intercepted the x axis at Tb = 4.8 °C (Fig. 3c). The NaClb(50) for chicory seeds remained approximately constant at sub-optimal Ts (i.e., < 30 °C), ranging from 0.294 to 0.326 M, then decreased linearly (kT = 0.024 M °C−1) with increasing T above To until it intersected 0 M at 43 °C (i.e., Tc(50)) (Table 2; Fig. 3d). That is, the NaCl required to prevent germination decreased by 0.024 M per degree above To (Table 2; Fig. 3d).

The median GR50 (1/t50) across Ts at different ψs increased linearly at Ts ≤ Td, the T at which ψb(50) begins to change (which is equal to To in this case), and then decreased non-linearly at Ts above To until Tc (Fig. 4a). The maximum GR50 (0.0973 h−1) was observed at To when ψ = 0 MPa and then decreased linearly as ψ increased, reaching zero at − 0.83 MPa with a rate of decrease of 0.111 h−1 per MPa with decreasing ψ (Fig. 4b). The common intercept at Tb = 4.8 °C (Table 3; Fig. 4a), indicates that base temperature was not affected by ψ. In contrast, the linear decrease in Tc(50) values predicted by the HTT model as T increases above To (i.e., 30 °C) is illustrated in Fig. 4c.

Fig. 4
figure 4

The effect of temperature (T) and water potential (MPa) (a) and T and NaCl concentration (M) (d) on the median germination rates (= 1/t50) of chicory seeds. The symbols represent the actual data and the lines drawn through these points are based upon the parameters of the hydrothermal time and halothermal time models (Table 3). The effects of ψ and of NaCl on GR50 at To (b, e, respectively) intersect the calculated ψb(50) and NaClb(50) values at GR50 = 0 or Tc = Td = To and the linear relationships were highly significant (P < 0.0001). The Tc(50) values predicted by the hydrothermal and halothermal models at each ψ or NaCl also are shown (c, f, respectively), and therefore fall exactly on the modeled lines

Full size image
Table 3 Estimated parameter values using the hydrothermal time (HTT; Eqs. 4, 7) and halothermal time (HaloTT; Eqs. 8, 11) models for describing seed germination of chicory at seven constant Ts (10, 15, 20, 25, 30, 35 and 40 °C) at each of four different ψs (0, − 0.3, − 0.6 and − 0.9 MPa) and/or five different NaCls (0, 0.075, 0.150, 0.225 and 0.300 M)
Full size table

The relationship of GR50 with T at different NaCl concentrations showed a similar pattern as was observed for ψ (Fig. 4d). GR50 reached its maximum (0.0979 h−1) at To when NaCl = 0 M and then decreased linearly as NaCl concentration increased, reaching zero at 0.304 M (− 1.37 MPa, after converting to ψ) with a rate of decrease of 0.32 h−1 per M of increase in NaCl (Fig. 4e). The linear decrease in Tc(50) values predicted by the HaloTT model relative to the control (NaCl = 0 M) as NaCl increased between 43 and 30 °C is also shown (Fig. 4f).

When the initial salinity levels tested were converted into ψ values (using the Van’t Hoff equation), the calculated ψb(50) values were much lower (more negative) in comparison with PEG data (Table 2; Fig. 3b). For example, at 40 °C, the estimated ψb(50) was − 0.21 MPa in PEG but was − 0.45 in NaCl. This could be due to Na+ and Cl− ions entering into seed cells, lowering their osmotic potential and increasing embryonic turgor, enabling the seeds to germinate at lower ψs. The relative osmotic adjustment of the chicory seeds imbibed in NaCl was calculated as [(1 − (estimated ψb(50) using hydrotime model/estimated ψb(50) using the halotime model after converting NaCl to ψ)) × 100]. This Osmotic Adjustment Value was relatively constant (32 to 40%) at T < To, then increased above To as the value of ψb(50) became less negative (Tables 2, 3).

The σψb and σNaClb values varied little across Ts (Table 2), indicating that to the variation of ψ and NaCl thresholds among individual seeds did not change across the range of temperatures tested.

Hydrothermal time model, halothermal time model and cardinal temperatures

To quantify the response of chicory SG to across all Ts under different ψ and NaCl, HTT (Eqs. 4, 7) and HaloTT (Eqs. 8, 11) models were applied to the respective data. The HTT model described the SG data of chicory well, with R2 of 0.86 (Table 3; Fig. 5a). The estimated parameters were 220 MPa °C h for θHTT, − 0.83 MPa for ψb(50), 0.211 MPa for σψb and 0.066 MPa °C−1 for kT (Table 3).

Fig. 5
figure 5

Linear regression of probit (g) plotted against base water potentials (ψb(g), MPa) and adjusted base water potentials (ψb(g) + kT(T − Td)) (a) and linear regression of probit (g) plotted against base NaCl concentrations (NaClb(g), M) and adjusted base NaCl concentrations (−NaClb(g) + kT(T − Td) (b). Symbols are the observed values of probit germination fraction at each T versus the calculated base values for different germination fractions (g). The highly significant linear relationships with probit (g) indicate that the base thresholds are normally distributed in the seed population and that the hydrothermal and halothermal models account for 83–86% of the variance in the germination behavior across all conditions

Full size image

The HaloTT model fit the data for chicory SG in NaCl with a high R2 value (0.83) (Table 3; Fig. 5b) and an estimated θHaloTT value of 74.9 M °C h (Table 3). The osmotic adjustment value was 36% when the ψb(50) value estimated by the HTT (− 0.83 MPa) was compared to the ψb(50) value estimated by the HaloTT (− 1.30 MPa) (Table 3).

To determine the cardinal Ts for chicory, GR50 vs. T was plotted in both conditions. The relationships between GR50 and Ts ≤ To were linear at all ψs and all NaCls and were limited to a single Tb of 4.8 °C (Fig. 4a, d). GR50 reached its maximum at To (30 °C) in both conditions (water and/or salt stresses) and also was not influenced by ψ and NaCl. Similarly, Tc(50) was 43 °C based on both the HTT and HaloTT models (Fig. 3b, d).

Seed moisture content and osmotic adjustment to salt

SMC was measured just prior to radicle emergence at different Ts and NaCls (Fig. 6a). At 10 °C, SMC was unchanged under salt stress relative to control (NaCl = 0 M). However, SMC decreased with increasing T from 15 to 30 °C under salt and control treatments, but with less decrease for salt treatments. SMC tended to increase with NaCl at each T, but decreased as T increased to To then increased again. The reason for the decrease in SMC with increasing T is unclear, but may be due to more rapid GR50 around To and less time to achieve maximum moisture content prior to radicle emergence. At Ts ≥ 35 °C, the SMC increased and tended to increase with increasing NaCl concentrations relative to control (Fig. 6a).

Fig. 6
figure 6

Seed moisture content (a), concentrations of Na+ (b), Cl− (c) and calculated osmotic potential due to NaCl (d) into the chicory seeds. Lower case letters show significant differences among NaCl treatments within a temperature and upper case letters show significant differences among temperatures, using the least significant difference test (LSD) at 0.05 probability level

Full size image

At all Ts, Na+ concentration in the seed tissues increased significantly with increasing NaCl (p < 0.05) when compared to the control, attaining 132 mmol kg−1 DW at 0.300 M NaCl (Fig. 6b). Cl− content of chicory seeds also increased remarkably as salinity levels increased (p < 0.05) at all Ts (Fig. 6c). Seed Cl− content reached its highest levels at supra-optimal Ts (> 30 °C) (Fig. 6c).

As the median base water potentials preventing germination were − 0.5 MPa lower in NaCl in comparison with PEG (Table 2; Fig. 3b), we estimated the osmotic adjustment in the seeds that could be attributed to additional Na+ and Cl− ion uptake. Using the ion contents and water contents on a dry weight basis, we calculated the total NaCl concentrations in the seeds for each treatment and converted these into osmotic potential values. These values were approximately − 0.5 MPa in water and became more negative as NaCl concentration increased, reaching approximately − 1.3 MPa across Ts at 0.300 M (Fig. 6d). Thus, sufficient Na+ and Cl− ions were taken up by the seeds to account for the observed ability of seeds to germinate to lower osmotic potential in salt as compared to PEG, which does not enter the cells. Osmotic adjustment due to NaCl uptake is also consistent with the increase in SMC with increasing NaCl concentration (Fig. 6a), while SMC decreases with increasing PEG concentration (Bradford 1986). There is uncertainty as to the locations of the ions and the water inside of the seed tissues (i.e., apoplastic or symplastic), and only symplastic ions would presumably contribute to increasing turgor and therefore to germination capacity. Nonetheless, the data indicate that salt ion uptake into the seeds was at least sufficient to account for the differences between ψb(50) and NaClb(50) (converted to osmotic potential) (Table 3; Fig. 3b).

Discussion

In general, the temperature responses of seeds can be defined by their cardinal Ts (i.e., Tb, To and Tc) (Bewley et al. 2013). The relationship between GR50 and T identified values of 4.8, 30 and 43 °C for Tb, To and Tc, respectively (Table 3). Similar Tb values have been reported for witloof chicory (Cichorium intybus; 5.3 °C) (Bierhuizen and Wagenvoort 1974), chicory (Cichorium intybus cv. Grasslands Puna; 3.7 °C) (Moot et al. 2000) and rapeseed (0–5 °C, depending on cultivars) (Farzaneh et al. 2014). The maximum GR50 was observed at 30 °C (To) in the control treatment (0 MPa and 0 M), similar to previous reports for chicory (29.9 °C, Zarghani et al. 2014; 25 to 30 °C, Corbineau and Come 1989), although Balandary et al. (2011) reported a value of 25.5 °C after fitting an empirical beta model. Our findings showed also that Tb and To were relatively unaffected by PEG and NaCl stresses, which corroborate results for To of Silybum marianum and Calendula officinalis (Parmoon et al. 2018) and flax-leaf alyssum (Alyssum linifolium Steph. ex. Willd.) (Mobli et al. 2018). The median Tc value estimated in this study (43 °C based on both hydrotime and halotime models) was approximately equal to the values reported for chicory by Zarghani et al. (2014) (46.3 °C) and Vahabinia et al. (2019) (40.2 °C) based on different models. At T > To, the value of Tc varies among seeds in the population (Alvarado and Bradford 2002), and also depends upon the stress level (Atashi et al. 2014; Bakhshandeh et al. 2017; Parmoon et al. 2018; Abdellaoui et al. 2019). As stress intensity increases, Tc approaches nearer to To, as illustrated here for chicory (Fig. 4), because of the threshold sensitivities to ψ or NaCl increase as the T increases above To (Fig. 3c, d).

The HTT model (Gummerson 1986), generalized as population-based threshold models, describes well the observed GRs and GPs response to increasing stress, aging, hormones and other factors affecting seed germination (Bello and Bradford 2016). GR50 slows as the stress intensity increases due to reduced ψ or increased NaCl concentrations, and GP also declines as the stress exceeds the threshold that prevents germination for a given fraction of the seed population. The HTT and HaloTT models applied here described these patterns well for chicory seeds (Figs. 1, 2, 5). The time constants in these models (e.g., θH or θHalo) provided useful overall values for germination kinetics for estimating the cardinal Ts under both water and salt stress (Fig. 3). Following the work of Seal et al. (2018) on the halophyte Suaeda maritima, we used the NaCl concentrations to fit halotime and HaloTT models to our germination data for chicory seeds. In the suboptimal T range, NaClb(50) was not affected by NaCl concentration (Fig. 3d), which is in contrast to the results reported by Seal et al. (2018), who found that NaClb(50) values significantly decreased with increasing T (between 5 and 25 °C). In contrast, threshold values in the halophyte Chloris virgata and the glycophyte Digitaria sanguinalis tended to increase with T before declining (Zhang et al. 2012). Thus, these responses may vary among species and the environments to which they are adapted (Huarte 2006). In addition, we extended the HaloTT model for quantifying SG responses at Ts > To. The NaClb(50) values for chicory decreased linearly above To, consistent with seed germination responses to ψ at high temperature (Alvarado and Bradford 2002). Thus, germination patterns across the entire range of temperatures from Tb to Tc could be described by the HaloTT model on the basis of NaCl concentration (Figs. 4a, d; 5b).

However, SG characteristics of chicory were more influenced by ψ in comparison with NaCl at equivalent osmotic potentials (Table 2). This agrees with the findings for other species, such as soybean (Khajeh-Hosseini et al. 2003), alfalfa (Tilaki et al. 2009), barley (Zhang et al. 2010), sunflower (Luan et al. 2014) and rapeseed (Bakhshandeh and Jamali 2020). This is likely due to the uptake of salt ions by the chicory seed, enabling osmotic adjustment to maintain a ψ gradient allowing water uptake during imbibition and germination (Heshmat et al. 2011; Seal et al. 2018). SMC, Na+ and Cl− concentrations increased in chicory seeds with increasing salinity (Fig. 6), consistent with results with barley (Zhang et al. 2010), several halophytic species (Khan et al. 1985; Seal et al. 2018; Song et al. 2005) and rapeseed (Bakhshandeh and Jamali 2020). In our case, the measured uptake of Na+ and Cl− ions during imbibition in salt solutions was sufficient to account for the lower apparent ψb(50) values when NaCl was converted to osmotic potentials. As the halotime model described germination behavior well up to the highest NaCl tested (0.300 M), chicory SG is controlled primarily by the osmotic rather than toxic effects of salt, as also reported for barley (Zhang et al. 2010), Acacia harpophylla (Arnold et al. 2014) and Atriplex halimus (Shaygan et al. 2017).

In conclusion, the models used in this work accurately and successfully describe chicory SG across all Ts at a range of ψs and NaCls and provide reliable tools for assessing germination under these conditions. We developed a HaloTT model for quantifying the SG responses (both GR and GP) in the supra-optimal T range, which has not been reported previously. This model was also able to predict germination responses of salt tolerant (rapeseed), moderately salt-sensitive (cucumber), and salt-sensitive (green bean) species (Bakhshandeh et al. unpublished data). Thus, the hydrotime, halotime, HTT and HaloTT models enable prediction of SG behavior across entire the range of temperature, water and salt conditions.

Of course, conditions can vary in the field and performance may not exactly match model predictions. However, we note that Liu et al. (2020) used the HTT model to characterize germination behavior of 13 native desert annual species. They found that relative germination sensitivities to T and ψ determined from laboratory tests were highly correlated with average seedling emergence data in desert conditions over 25 years of field observations in Arizona. Both HTT and HaloTT models use parameters based on mechanistic assumptions about the underlying physiology of SG to describe patterns of germination timing. Therefore, we believe that while field conditions can vary and influence germination at a particular time, the HaloTT model developed here can quantify and predict the relative germination performance that can be expected in field conditions.

Author contribution statement

EB and KB designed the experiments. EB and FV performed the experiments. EB and KB conducted the modeling and interpreted the data. EB, KB, HP, FV and RA co-wrote all drafts of the paper and also approved the final draft for submission.