Impacts of bracteole removal and seeding rate on seedling emergence of halophyte shrubs: implications for rangeland rehabilitation in arid environments

Additional keywords: arid and semi-arid ecosystems, direct seeding, indigenous plants, rangeland restoration, seed scarification treatments.

Most rangelands in the West Asia and North Africa (WANA) region are poorly managed, resulting in land degradation, reduced productivity, and loss of plant biodiversity (Hudson et al. 2017). Overgrazing, conversion of rangelands into rainfed cropping systems, land tenure issues, and climate change are major drivers for this degradation (Ates and Louhaichi 2012; Ouled Belgacem and Louhaichi 2013). Several government agencies have becoming increasingly aware of the magnitude of the problem and strive to provide solutions to address rangeland degradation. However, given the low and slow return on investment, governments are not able to finance large-scale projects to effectively restore and develop rangeland resources (Louhaichi et al. 2016a). Therefore, cost-effective techniques for slowing down and eventually reversing the trend of rangeland degradation are needed.

Transplanting shrub species has been extensively used for rangeland rehabilitation in the WANA region due to high transplant survival, effective soil stabilisation, and reduced soil erosion (Louhaichi and Tastad 2010; Louhaichi et al. 2016b). To improve establishment rates, shrub seedlings are usually grown in nurseries and then transplanted in the field (Franco et al. 2006). However, this process is often expensive and time consuming (Douglas et al. 2007) as it requires facilities and labour to grow seedlings for several weeks or months in nurseries. Furthermore, establishing shrubs from nursery-grown seedlings is particularly challenging in arid environments and/or remote rangeland sites due to lack of irrigation and adequate site preparation (e.g. establishment of contour furrows). Besides the irrigation cost, maintenance is needed to replace dead seedlings during the first year of establishment (Palmerlee and Young 2010).

Direct seeding might be a more effective alternative to transplantation of seedlings to improve plant species richness and the productivity of rangelands (Fernández et al. 2012). Direct seeding techniques are common practices for the establishment of most herbaceous pasture species and have been used to successfully revegetate large areas (Malcolm 1994). Although direct seeding is a cost-effective technique to implement large-scale restoration projects, its success in arid regions is limited by harsh environmental factors and poor soil properties, in addition to seed characteristics such as seed dormancy and physical structures (e.g. bracteoles) that interfere with seed germination (Ungar and Khan 2001; García-Fayos et al. 2010; Busso et al. 2012; Raizada and Juyal 2012; Prats et al. 2016).

Many halophytes, in particular Atriplex species, have chemical and morphological characteristics that regulate seed dormancy and germination (Ungar 1978; Khan and Ungar 1985). Atriplex seeds are enclosed within bracteoles that control the timing of water imbibition, a crucial step in the seed germination process (Ungar and Khan 2001). The bracts surrounding the seeds may contain chemical compounds that interfere with germination (Young et al. 1980; Jefferson and Pennacchio 2003). Removal of these bracts has been suggested as a mean of improving seed germination of Atriplex halimus and Salsola vermiculata (Osman and Ghassali 1997).

In most rehabilitation projects, managers tend to increase seeding rate assuming that seeds could have a low germination rate or could be subject to predation by wildlife (Florentine et al. 2013). Direct seeding with a higher seeding rate increases input costs and may lead to inter-plant competition for nutrients and space (Gordon et al. 1989). However, using low seeding rate greatly decreases plant population and results in low productivity (Florentine et al. 2013).

Understanding the factors that affect seedling emergence and plant establishment under rangeland conditions is critical for improving our ability to address the important rehabilitation problems such as low succession and poor spatial distribution of perennial plants (Lesica and Allendorf 1999; Ghassali et al. 2012; Kemp et al. 2013; Porqueddu et al. 2016). Therefore, the objective of this study was to investigate the combined effects of bracteole removal and seeding rate on seedling emergence of seven native and introduced halophytic shrub species that are widely used for rehabilitating degraded rangelands in the WANA region.

The study was conducted in Tel Hadya (Syria) at the headquarters of the International Centre for Agricultural Research in the Dry Areas (ICARDA) (36°01N, 36°56E) using outdoor plastic pots of 14 cm diameter and 25 cm depth. Each pot contained 3.5 kg of a 1 : 1 mixture of farm soil and sand topped with a layer of organic sheep manure. The average annual rainfall over two decades in the study site is 340 mm (ICARDA headquarters meteorological station), occurring predominantly from October to May.

Seeds of Atriplex halimus, A. canescens, A. leucoclada, A. nummularia, A. lentiformis, Salsola vermiculata and Haloxylon aphyllum were collected in the autumn of 2011 from the ICARDA pastoretum (Table 1).

Table 1.  Geographic origin of the seven evaluated halophytic rangeland shrub species. Three of the species (A. halimus, A. leucoclada, S. vermiculata) are indigenous to the study area

The pots were arranged in a randomised complete block design with five replications. The treatments consisted of removal and non-removal of bracteoles. Bracteoles were manually removed on half of the seeds (termed seeds without bracteoles) of all species by scarification using sand paper 212, P80, and aluminum oxide. The other half of each seed sample retained its bracteoles (termed seeds with bracteoles) and were considered as the control treatments.

The recommended rate for direct seeding (drill-seeding) of most Atriplex species is between 25 and 35 kg ha^–1 (seeds with bracteoles) (Watson and O’Leary 1993; Booth 2002). The average weight of 1000 seeds without bracteoles was 0.50 g for A. leucoclada, 0.64 g for A. lentiformis, 0.93 g for A. halimus, 0.72 g for A. nummularia, 1.02 g for S. vermiculata, 1.81 g for H. aphyllum and 2.82 g for A. canescens. Three seeding rates consisting of 10 (low), 30 (medium), and 60 (high) seeds per pot for each species were compared. This resulted in 21 treatment combinations for seeds without bracteoles and seeds with bracteoles. These combinations resulted in 42 three-way treatments (species × bracteole removal × seeding rate).

Seeds without bracteoles and seeds with bracteoles were sown directly in the pots at a depth of 0.5 to 1 cm and patted down firmly on 27 January 2012. Five randomly selected pots were assigned for each treatment combination, for a total of 210 pots. The pots were kept under natural environmental conditions of rainfall, temperature, and light. During the course of this experiment (27 January to 10 March 2012), the site received 121.6 mm of rainfall and the average minimum and maximum temperature varied from 1.9°C to 12.5°C. The long average of rainfall and temperature is presented in Table 2.

Table 2.  Long-term monthly rainfall (mm), average daily air temperature, average maximum and minimum temperature from 1985 to 2012 in Tel Hadya, north-west Syria

Emerged seedlings were counted at 10, 15, 20 and 33 days after seeding. Seedling emergence percentages were calculated as the ratio of the number of seedlings that emerged during the monitoring period to the number of seeds sown, and then multiplying this value by 100. The speed of emergence was calculated according to Gairola et al. (2011), using the following formula;

where, n = number of emerged seeds, t = time after seeding.

Results from the manual bracteole removal performed on A. leucoclada revealed lower seedling emergence rate compared with seed with bracteoles intact. Therefore, to provide possible explanations, machine removal of bracteoles was performed on A. leucoclada seed using a Perten Laboratory disc Mill 3310 (calibrated from level one to level six). Calibration was set at level four (4), as levels between 1 and 3 resulted in unacceptable mechanical damage to the seed. The experimental design was a randomised complete block, with bracteole removal as the blocking factor, and was conducted in March 2017 at the ICARDA facility in Terbol experimental station, Lebanon.

There were four replicates of 50 seeds of A. leucoclada. Half of the seeds were machine scarified (machine removed bracteoles) whereas the remainder were not (seeds with intact bracteoles). Seeds were exposed to two media (paper and river sand) for evaluation of germination and emergence (International Seed Testing Association 2017). Laboratory conditions were maintained at 20 ± 2°C, with a 16-h light at 75 347 lm m^–2 and 8-h darkness regime, for 14 days.

Seedling emergence counts were taken on each of four different dates. In order to factor in the effect of time on seedling emergence, linear regressions passing through the initial first seedling emergence day were performed, and the slopes of the regression estimated using the last observation date.

It is possible that the time it takes a seedling to emerge from the soil might surpass the period of observation, for example, 90% absolute seedling emergence percentage may not be reached within a given time period. To account for this, we rather estimated the percentage of the observed maximum emergence which is interpreted as seedling emergence potential. The slope and time to reach 90% emergence are a measure of emergence strength of the species and treatment, which explains the dynamics of emergence. These two estimated parameters were analysed using ANOVA. The pair-wise comparisons of the species under each combination of treatment and time were carried out using Bonferroni test at α = 0.01. Data of the complimentary laboratory experiment were analysed using ANOVA for the two treatments (bracteole removal and growing media). All statistical analyses were carried out using Genstat software (Genstat 2009).

Three of the species (A. halimus, A. leucoclada, S. vermiculata) are indigenous to the study area (Table 1). There were significant differences in seedling emergence among the species in response to bracteole removal (P < 0.01) and there was an interaction between bracteole removal treatment and species for seedling emergence percentage (P < 0.01). For seeds without bracteoles and seeds with bracteoles, A. halimus and S. vermiculata showed the highest seedling emergence percentages (72.6%, 69.4% and 65.2, 64.5% respectively). Bracteole removal enhanced seedling emergence of A. halimus, A. canescens, A. lentiformis, A. nummularia, and H. aphyllum. There were no significant differences in seedling emergence between seeds without bracteoles and seeds with bracteoles in S. vermiculata (Table 3). These results confirm the common belief that native perennial species (A. halimus and S. vermiculata) are well adapted to the harsh arid environments (recurrent droughts, shallow soils, crusted soil surface, high summer temperature) and perform better than most introduced species (D’Antonio and Meyerson 2002; Niane 2013). These results are similar to the findings of DePuit et al. (1980) where direct seeding of Gardner saltbush (A. gardneri) in disturbed lands experienced limited success due to the presence of seed bracteoles. Similarly, bracteoles of river saltbush (A. amnicola), and wavy-leaved saltbush (A. undulata) reduced final germination percentage and seedling emergence in Western Australia (Stevens et al. 2006).

Table 3.  Seedling emergence percentage (%) of seven halophytic species under two seed treatments (seeds without bracteoles and seeds with bracteoles), over four time-sampling periods (10, 15, 25 and 33 days)
The means are represented along with standard errors. s.e.m., Standard error of the mean. Means within the same column with the same letter(s) are not significantly different at 5% level of significance, using Bonferroni test

Overall, the presence of bracteoles reduced seedling emergence except in S. vermiculata, where emergence was similar between seeds with bracteoles (63.7%) and seeds without bracteoles (65.2%). However, seedling emergence of A. leucoclada was significantly higher for non-scarified seeds (34.6%) compared with scarified seeds (22.1%) (Table 3). This could be attributed to mechanical damage caused by the bracteole removal method. It is likely that excessive bracteole removal damage may have occurred due to low precision of manual removal of bracteoles, owed to the fact that the bracteoles of A. leucoclada seeds are much more tightly attached to the endosperm than the rest of the studied species. To test this hypothesis, a precise and adjustable seed threshing machine was used. The difference in seedling emergence for the seed with machine removed bracteoles was significantly higher (P < 0.01) for A. leucoclada, in both paper and river sand media (Fig. 1). Such differences in the response of seeds of different species to different bracteole removal techniques needs to be taken in considerations in rangeland rehabilitation using direct or indirect seeding.

Fig. 1.  Seed germination (%) of A. leucoclada under laboratory conditions. The error bars represent the standard error.

In the present study, seeds with bracteoles of S. vermiculata showed similar seedling emergence percentage to seeds without bracteoles (Table 3). This response could be attributed to environmental conditions that contribute to seedling emergence potential, with or without bracteoles (Pearson et al. 2002). Therefore, despite the presence of bracteoles, seeds with bracteoles were still able to demonstrate comparable emergence levels suggesting that bracteoles in this species do not physically hinder germination and they may not harbor chemicals that interfere with germination as occurs with some other shrubs. Ungar and Khan (2001) studied the influence of bracteoles on the germination response of Atriplex prostrata and A. griffithii. They found that attached bracteoles did not inhibit germination of A. prostrata but completely inhibited germination of A. griffthii seeds. Germination of seeds of A. griffthii was also inhibited in the presence of detached bracteoles, suggesting that the reduced germination of A. griffithii may be due to the presence of relatively high concentrations of dissolved salts in the bracteoles.

Although there were some inconsistencies, removing bracteoles from the seed improved seedling emergence for most species, possibly through breaking the seed coat, thus partially exposing the cotyledons to moisture (Table 3). Exposing the cotyledons can permit the process of hydrolysis and hormone activation, leading to an increase in nucleic acid transcription to promote protein synthesis for germination and early seedling growth (Müntz et al. 2001; Okunlola et al. 2011).

Bracteole removal substantially increased seedling emergence. Separating the seeds from the bracteoles could be beneficial for both physical and physiological reasons including (1) the inhibitory effects of the leachate from the bracteoles (many of them contain a high concentration of salt); (2) the physical barrier the bracteoles impose on the seed; (3) the presence of growth regulating substances or allelopathic compounds in the bracteoles that delay germination; (4) a reduction in oxygen diffusion due to bracteole presence; (5) the effect of the bracteoles on the quality of light falling on the seed (Koller 1957; Aizazzi and Argüello 1992; Schütz 2000; Mandak and Pysek 2001; Li et al. 2008). Similarly, Stevens et al. (2006) reported that removing bracteoles increased germination and improved seedling emergence of Atriplex amnicola and Atriplex undulata by 50% and 150%, respectively. Webb et al. (2009) also found that the removal of seed husks and bracts of sawgrass (Cladium jamaicense) enhanced the percentage and success of germination. The inhibition of germination was attributed to the presence of abscisic acid and other compounds stored in the husks and bracts. Osman and Ghassali (1997) also found that seed germination of S. vermiculata and A. halimus was significantly improved by removing fruiting bracts from the fruits (utricles).

In line with the findings of the present study, mechanical bracteole removal also increased seedling emergence of the coated seed of Tylosema esculentum (Travlos et al. 2007). Our findings add to the consensus that removal of bracteoles stimulates germination and increases seedling emergence. The bracteole removal treatments applied in this study and more specifically the mechanical treatment used for A. leucoclada seems to be very effective in protecting the embryo and hence improving seedlings emergence. Therefore, they would be useful in rangeland revegetation by direct seeding.

Since no species had reached 90% emergence during the experiment period, the time to 90% observed seedling emergence was estimated from the linear regression. This time and the speed of emergence, showed significant differences in relation to species, bracteole removal, and their interactions (P < 0.01). When compared across all shrub species, bracteole removal resulted in higher seedling emergence percentage per day. The exception was the seeds with bracteole of A. leucoclada, where the opposite was noted. There was no response to bracteole removal in S. vermiculata whereas A. halimus showed the highest seedling emergence percentages per day and together with S. vermiculata recorded the highest mean seedling emergence per day (Fig. 2).

Fig. 2.  Seedling emergence dynamics of the seven halophytic species under treatments (seeds without bracteoles and seeds with bracteoles) in terms of seedling emergence trend (% per day).

Bracteole removal treatments enhanced significantly (P < 0.01) the dynamics of seedling emergence in terms of time to 90% of maximum seedling emergence (seedling emergence potential) compared with seeds with bracteoles, except for Salsola vermiculata. A. lentiformis showed the greatest seedling emergence potential and was the fastest species whereas A. canescens was the slowest (Fig. 3). Overall, all except one species have shown better seedling emergence in response to bracteole removal. Having bracteoles, native halophyte species seem to be more adapted to the arid harsh environmental conditions of the WANA region. Of particular interest is the native S. vermiculata that showed great potential with high seedling emergence even without bracteole removal. This shrub is known for its high palatability and ability to regenerate (Louhaichi et al. 2014) and should be recommended for large-scale restoration in the WANA region. The only problem is that it has a very short seed longevity and therefore should be collected and seeded in the same year (Niane et al. 2013).

Fig. 3.  Seedling emergence dynamics of seven halophytic species under treatments (seeds without bracteoles and seeds with bracteoles) in terms of time to 90% of maximum observed seedling emergence.

The predicted seedling emergence rate on the last observation date was significantly higher for seeds without bracteoles compared with seeds with bracteoles for all the species except for A. leucoclada and S. vermiculata. The highest and lowest predicted rate of emergence were recorded for A. halimus and A. leucoclada respectively (Fig. 4).

Fig. 4.  Predicted germination on the last observation day of seedling emergence for seeds without bracteole and seeds with bracteole of seven shrubs under natural conditions.

Seedling emergence includes two biological steps: seed germination, initiation followed by radicle and shoot elongation. Each step requires different environmental conditions (Roman et al. 2000). The relationship between environmental factors such as climatic conditions, soil, seed, and seedling emergence response, varies with species and the physiological status of the seed (Finch-Savage et al. 2001). However, the interactions among these factors are complex (Vleeshouwers and Kropff 2000). The low percentage and speed of seedling emergence of arid and semi-arid species have been attributed to seed sensitivity to extreme temperatures and moisture availability (Roundy and Biedenbennder 1996; Chesson et al. 2004). Findings from this study suggest that the interaction of species and the presence of bracteoles may significantly contribute to the variability in seedling emergence responses. The initiation of seedling emergence is often more sensitive to water availability (Alvarado and Bradford 2002). This might explain the positive effect of treatment on seedling emergence as the bracteole removal may increase the access to water by the seeds. However, the results did not show any significant benefits of higher seeding rates on seedling emergence, confirming the reports of Linhart (1976) and Florentine et al. (2013). Bracteole removal, in contrast, enhanced both the percentage and speed of seedling emergence of most of the studied species. This is a strong indication of successful establishment of shrub species (Raveneau 2012).

In arid and semi-arid environments, seedling emergence of preferred species may be limited by several factors that regulate seed dormancy, germination and emergence. The success and efficiency of shrub seedling survival and growth under arid conditions depends on removing barriers that inhibit germination and stifle seedling emergence. This study found that bracteole removal from seeds of some shrub species increased seedling emergence percentage and speed. Therefore, seed bracteole removal of shrub species should be considered for large-scale rehabilitation of degraded arid rangelands under the ongoing climate change patterns. This is a crucial step, especially when direct seeding is the only viable option for reversing the trend of rangeland degradation.

Another key finding of this study is the fact that even with the presence of bracteoles, the native halophytic species S. vermiculata and to a lesser degree A. halimus have greater tolerance of extreme temperatures and moisture availability and exhibit high seedling emergence than the introduced species studied. Therefore, these native and well adapted species should be recommended for rehabilitating degraded WANA rangelands.

Nevertheless, to gain a greater understanding of the role of bracteoles in regulating emergence of these rangeland species, further physiological studies are needed to identify factors that minimise seed dormancy. Furthermore, seeding rate should be investigated thoroughly at a larger scale, in the field rather than in pots, and for a longer period to assess factors influencing seedling emergence and establishment.

The authors declare no conflicts of interest.