Germination of Acacia harpophylla (Brigalow) seeds in relation to soil water potential: implications for rehabilitation of a threatened ecosystem

Experimental design

We used seeds¹ of A. harpophylla to investigate germination in relation to water potential as environmental factor. Three replicates of 50 seeds per treatment (i.e., water potential) were placed at equal distance on an absorbent substrate (Wettex®). Materials (e.g., tweezers, Wettex®, glassware) were autoclaved for 20 min at 121°C and preparation of treatments as well as any monitoring took place in a laminar flow cabinet. All treatments and replicates were then placed randomly within a germination cabinet under constant temperature (25°C) and a 12 h day and night cycle. Seeds were removed from petri dishes once a perceptible radicle emerged. The experiment ceased after five days with no further germination. No pre-treatment of seeds was required for the selected species to break dormancy (Schmidt, 2000; Turnbull & Martensz, 1982).

For controlled experiments on seed-soil relations the osmotic water potential can be used to represent soil matric potential (Carpita et al., 1979; Gray, Steckel & Hands, 1990; McWilliam & Phillips, 1971). We used solutions of polyethylene glycol (PEG 6000) to expose seeds to nine osmotic water potentials: 0, −.01, −.03, −.1, −.25, −.5, −.75, −1, and −1.5 MPa. These values capture soil water conditions ranging from water saturation (0 kPa) and field capacity (−.01 to −.03 MPa) to the permanent wilting point (−1.5 MPa). The empirical equation derived by Michel & Kaufmann (1973) and revised by Wood, Dart & So (1993) was used to set up the required water potential (ψ in kPa): (1)$\begin{array}{ccc}\psi =\left(6.3\times {10}^{-5}T-0.021\times 96\right)O^{2.2357},& & \end{array}$ where T is the temperature in K (here 298.15 K), and O denotes the osmolality in g 1000 g⁻¹ of water. Solutions measuring 15 ml of PEG 6000 were added to the seeds on the Wettex® substrate within a 90 mm petri dish and wrapped with Parafilm® to prevent evaporation. The dishes were opened each day to provide aeration and to count and remove germinated seeds.

Germinability and mean germination time

We used germinability G (%) and mean germination time $\bar{t}$ (days) as measurements to quantify germination (Ranal & Santana, 2006). While G simply represents the number of germinated seeds (g) relative to the total number of seeds per replicate (n): (2)$\begin{array}{ccc}G=\left(g∕n\right)100,& & \end{array}$$\bar{t}$ denotes the mean length of time required for maximum germination of a seed lot (Czabator, 1962): (3)${\displaystyle \begin{array}{ccc}\bar{t}=\sum _{i=1}^k{g}_i{t}_i/\sum _{i=1}^k{g}_i,& & \end{array}}$ where t_i is the time elapsed from initiation of the experiment to the ith observation day until germination ceases on the kth day.

We applied a generalised logistic regression model (GLM) on a logit scale and simple linear regression for G and $\bar{t}$, respectively, to test the significance of the relationship between germination percentage or time required for maximum germination and decreasing water potential. Assumptions were tested via diagnostic plots, which showed no violation of homogeneity of variance and normal distribution for the residuals. Separate GLM models were adopted for each species as this provided the best fit for the data. We also employed other models (linear regression and logistic regression with a probit scale) and various transformations (arcsine, square root transformations) but they either did not fit the data, or when they did, failed homogeneity of variance and normal distribution of error (p > 0.05).

We compared the results for G with those published by Reichman, Bellairs & Mulligan (2006) for salinity. Therefore, we estimated the osmotic pressure ψ_o (MPa) based on the value range of electrical conductivity used in Reichman, Bellairs & Mulligan (2006): (4)$\begin{array}{ccc}{\psi }_o=\mathrm{MRT},& & \end{array}$ where M is the molarity (mol L⁻¹) of the sodium chloride solution, T is the temperature in K (here 298.15 K), and R is the gas constant (J mol⁻¹ K⁻¹). The corresponding values of ψ_o are presented in Table 1.

Estimating parameters of the hydrotime model

The hydrotime model was proposed by Bradford (1990), Gummerson (1986) to “simultaneously account for both the timing and the extent of germination of a given seed population in relation to its ψ environment” (Bradford, 2002). It is defined as: (5)$\begin{array}{ccc}{\theta }_H=\left(\psi -{\psi }_b\left(P\right)\right)t_P,& & \end{array}$ where θ_H is the hydrotime constant (MPa h), t_P (h) is the time to germination of percentage P of the seed population, and ψ_b(P) (MPa) is the base or threshold water potential above which P can still complete germination. Among the seed population, ψ_b is variable and can be described by a range of frequency distributions (Mesgaran et al., 2013) of which the normal frequency distribution is the most commonly used one defined by its mean ψ_b(P₅₀) and standard deviation σ_ψb (Gummerson, 1986). Together, these parameters enable prediction of the germination time courses, i.e., both rate and extent of germination of the seed population, at any ψ at a constant temperature (Bradford, 2002): (6)$\begin{array}{ccc}{t}_P={\theta }_H∕\left(\psi -{\psi }_b\left(P_{50}\right)\right).& & \end{array}$ In this regard, θ_H “quantifies the inherent speed of germination, which can vary among species and physiological states” (Bradford, 2002), whereas ψ_b(P₅₀) indicates the average stress tolerance of a seed population, and σ_ψb describes the synchrony in germination timing among seeds in a population. We estimated these parameters based on the observed values of t_P for each ψ, i.e., we plotted the germination rates (1∕t_P in h⁻¹) as a function of ψ and fitted regressions to the observed data (Fig. 4). This resulted in straight lines with slopes of ${\theta }_H^{-1}$ and intercepts on the ψ-axis corresponding with ψ_b(P). The values of ψ_b(P) were then fitted to a normal and log-logistic frequency distribution, respectively, and their distribution parameters were estimated accordingly.

Soil and seed water conditions

Initial soil water conditions play a critical role when direct seeding is the primary approach to revegetating post-mining areas. This is even more important in a semi-arid climate, where water is limited due its scarcity or erratic occurrence (Rodriguez-Iturbe & Proporato, 2004). Seed germination is triggered by the amount of water the seed can imbibe, which is related to the water potentials of both the soil and the seed (Bewley et al., 2013; Bradford, 2002; Evans & Etherington, 1990; Williams & Shaykewich, 1971). In this regard, water can only enter the seed if the seed water potential is below the water potential of the surrounding soil material. In general, air-dry seeds have water potential values of −50 to −350 MPa (Bewley et al., 2013). The empirical results of this study indicate that seeds of A. harpophylla are very water stress tolerant, that is, their seed water potential is extraordinarily low. For example, radicles still emerged at a water potential as low as −1.5 MPa (Fig. 2A). Likewise, the mean base water potential of the seed population ranges between −1.533 and −1.451 MPa (Table 2 and Fig. 3). These are remarkable values compared with selected Mediterranean (Köchy & Tielbörger, 2007)² and agricultural plant species (Watt, Bloomberg & Finch-Savage, 2011)³ . Qualitatively, the low seed water potential corresponds well with other investigative studies on physiological water relations of A. harpophylla (Doley, 2004), which measured foliage water potentials as low as −15 MPa (Connor & Tunstall, 1968) and shoot water potentials of −7.2 to −6.8 MPa (Connor, Tunstall & Van den Driessche, 1971; Tunstall & Connor, 1981; Van den Driessche, Connor & Tunstall, 1971). Also the seed coat, which is atypically soft compared with other Acacia species, denotes an adaptive mechanism to erratic rainfall patterns of the Brigalow Belt Bioregion to rapidly overcome dormancy if soil water conditions are elevated (Johnson, 1964; Reichman, Bellairs & Mulligan, 2006; Scott, Jones & Williams, 1984).

The parameter estimates of the hydrotime model (Table 2) play a critical role for predicting the time required to germinate a fraction of the seed population under given soil water conditions (Eq. (5)). In this regard and together with the physiological parameters of the hydrotime model (i.e., hydrotime constant and distribution of the base water potential) the soil water potential governs germination and, consequently, the success of initial vegetation recruitment and early ecosystem establishment. Moreover, together with empirical data on the effect of temperature on the germination of Brigalow seeds (Reichman, Bellairs & Mulligan, 2006), the findings of this study can be utilised to parameterise hydrothermal models (Bradford, 2002; Bullied, Van Acker & Bullock, 2012; Gummerson, 1986; Köchy & Tielbörger, 2007; Watt, Bloomberg & Finch-Savage, 2011) for predictive modelling of germination in relation to the two environmental factors of temperature and water availability.

The results of this study show that germinability decreases and the time required for maximum germination increases with decreasing soil water potential (Fig. 2). Moreover, the strong negative correlation between germinability and mean germination time (Fig. 1) underpins how important a rather short germination time is to maximise germination success. Thus, topsoil restoration at post-mining land in the Brigalow Belt Bioregion should target to maximise the initial soil water potential while explicitly considering the erratic character of rainfall patterns in Central Queensland (Audet et al., 2013; Audet et al., 2012). This can be accomplished by optimising soil attributes such as depth, texture, compaction, organic amendments, etc. (Arnold, 2012; Arnold et al., in press; Zipper et al., 2013). Apart from physical soil restoration, pre-treatment of seeds, i.e., seed priming, also plays a critical role to enhance germination (Jisha, Vijayakumari & Puthur, 2013). Amongst the broad range of seed priming techniques, hydropriming and osmopriming are the most promising approaches for plant establishment in semi-arid climate to increase seedling growth (Yagmur & Kaydan, 2008), and root and shoot length (Kaur, Gupta & Kaur, 2002). Although the exact mechanisms behind pre-treatments of seeds are not fully understood, seed priming seems to activate cell signalling pathways and cellular responses to environmental stressors resulting in faster plant defence responses (Jisha, Vijayakumari & Puthur, 2013).

Despite plant available water being the primary environmental factor in the Brigalow Belt Bioregion (Arnold et al., in press), due to hot summers and highly saline mine spoils secondary stressors such as temperature and salinity may also play a critical role for the germination success of A. harpophylla, which was tested by Reichman, Bellairs & Mulligan (2006). Their findings indicate no significant trend in germination among the temperature range of 15–38°C.⁴ Likewise, seeds of A. harpophylla showed remarkable tolerance to salinity up to an electrical conductivity of 30 dS m⁻¹, however, with significant reduction in germination at salinity greater than 20 dS m⁻¹. The authors speculated “it seems unlikely that the reduced germination at 25 and 30 dS/m were due to osmotic effects” (Reichman, Bellairs & Mulligan, 2006). While the present study confirms this conclusion for high values of salinity (Table 1) corresponding to osmotic pressure values greater than −.4 MPa (Fig. 4), no significant differences could be found between treatments of sodium chloride and PEG (Materials and Methods) for small values of osmotic pressure corresponding to high salinity (Fig. 4). That said, we conclude that under low to moderate levels of salinity the osmotic pressure plays the primary role for germination of A. harpophylla seeds rather than the toxic nature of the predominant salt (here sodium chloride), whereas salinity becomes the primary environmental stressor under high salt concentrations. More generally, seeds of A. harpophylla seem to germinate quite uniformly over a broad range of environmental conditions related to temperature, salinity, and water availability.⁵

Alternative rehabilitation approaches

Despite the opportunistic germination capability of A. harpophylla seeds, the question remains whether direct seeding denotes the optimal approach to rehabilitate native Brigalow ecosystems given the hydropedological,⁶ climatic, and plant physiological attributes and conditions of the Brigalow Belt Bioregion.

Hydropedological processes are fundamental for the proliferation of Brigalow plant communities (Arnold et al., in press). The predominant clay soils (Isbell, 2002) comprise fine-textured non-cracking Grey and Black Dermosols (Lixisols (Rees et al., 2010; IUSS Working Group WRB, 2006)), and uniform dark cracking Grey and Black Vertosols (Vertisols (Rees et al., 2010)). The latter form “gilgais” (Cowie, Thornton & Radford, 2007; Radford et al., 2007; Thornton et al., 2007), which denote ephemeral water storages if filled during intensive storm events. The climate is characterised by erratically distributed rainfall patterns with short and intensive storm events occurring during summer (Cowie, Thornton & Radford, 2007), associated with the risk of water logging or soil erosion; whereas the very arid conditions during the winter season generally involve periods of water deficit (Audet et al., 2013; Audet et al., 2012; Dalton, 1993). In combination with these erratic climatic conditions the soil properties facilitate rather low soil water potentials on the long-term (Tunstall & Connor, 1981), and as a consequence, these landscapes can be colonised by a few plant species only, which is reflected in relatively low species richness and total vegetative biomass (Isbell, 2002; Johnson, 1980). As elaborated in the previous section, among the species in the Brigalow Belt Bioregion, A. harpophylla developed very well adapted mechanisms to proliferate under these harsh conditions. The plants flower only sporadically (Benson et al., 2006) and thus seeds and seedlings are only produced in large numbers during years of extraordinary rainfall (Butler, 2007; Johnson, 1997). However, A. harpophylla is able to reproduce asexually, i.e., through root suckering or sprouting even if aboveground parts of the plant are damaged dramatically, as long as belowground biomass and hydropedology stay intact (Arnold et al., in press and references therein). The extent of this vegetative reproduction is most pronounced if trees are young and severely damaged, and under dry conditions (Johnson, 1964). In Colorado, USA circumstances are similar with regard to the re-establishment of vegetatively regenerating Aspen (Populus tremuloides) on surface-mined land (Musselman et al., 2012), which initially failed due to severe damage to the root system and thus the limited access to water and nutrients (Shepperd & Mata, 2005). However, rehabilitation was more successful when more comprehensive soil-root compartments were transplanted from local sources in combination with weed control and light irrigation with non-saline water (Musselman et al., 2012). In this regard, it seems to be crucial to keep the delicate balance between soil attributes (depth, compaction, texture) and root extension in balance. This innovative rehabilitation approach of transplanting the whole soil-root compartment may also be of interest for restoration ecologists engaged with re-establishment of Brigalow ecosystems in Central Queensland to achieve safe, stable, and non-polluting ecosystems (Commonwealth of Australia, 2006). However, the advantage of direct seeding over asexual propagation such as transplantation is the gain of genetic diversity – critical if ecosystems are forced to adapt to the projected changing climate of Central Queensland (Low, 2011).