This study tested the effects flash flaming or acid digestion (depending on the species) and extruded seed pellets composed of a “blank” mixture or a “topsoil” mixture (hereafter referred to as “blank pellets” and “topsoil pellets”, or “pellets” when referring to both). Both blank and topsoil pellets were tested with untreated control florets and florets treated with flash flaming or acid digestion. All SETs for each of the four study species were tested on the topsoil and mine waste substrates ( Table 1 ).

The experiment was conducted at the Controlled Environment Facility (CEF) located in Newman, Western Australia (23°21’41.8”S, 119°42’26.1”E; Figures 1A, B ). The CEF allows the trialling of different seed treatments and simulated rainfall regimes when sown in a range of reconstructed soil profiles whilst the natural conditions such as temperature, light, humidity, and wind of the location are maintained (Erickson et al., 2017, 2023). The facility contains 2 x 4 m planter beds containing different reconstructed soil types including topsoil, mine waste, and blends of these two (Erickson et al., 2017, 2023). For the present study the soil types used included a locally sourced topsoil representative of stockpiled material, and mine waste commonly extracted from the adjacent Mt Whaleback mine and nearby mine sites ( Table 2 ).

A total of six blocks (2 x 4 m planter beds) per soil type were established (12 blocks in total for the study; Figures 1C–E ). Each block contained four 0.6 x 1.2 m plots ( Figure 1C ), randomly allocated one of each of the four study species. Within each species plot, each seed treatment was replicated six times and planted within a 0.1 x 0.1 m plot planting zone (totalling 36 planting zones per plot) ( Figure 1D ). Each planting zone was separated by a buffer zone of equal size. A total of 15 filled florets (i.e., florets containing a seed) were sown per replicate for each species to account for potential losses between the germination and emergence phase. Selection of filled florets was achieved by manually separating florets from non-target material (i.e., stalks, chaff), vacuum aspiration (‘Zig Zag’ Selecta, Machinefabriek BV, Enkhuizenm the Netherlands), and X-ray analysis (Faxitron MX-20 digital X-ray cabinet, Tucson, AZ, USA). Florets incorporated within a pellet were sown on the soil surface, while florets not contained within a pellet were buried at a depth of 5–10 mm. The experiment commenced on the 16^th of February 2022 and ran for a duration of 14 weeks.

Each of the study species (Cymbopogon ambiguus, C. obtectus, Eulalia aurea, and Eriachne obtusa) are widely distributed rangeland grasses across the arid, semi-arid, sub-tropical and tropical zones of Australia ( Appendix I ; Supplementary Figure S1 ) where mean annual temperature and mean annual precipitation range between 14–29°C and 125–2000 mm, respectively ( Appendix I ; Supplementary Table S1 ). The external structures surrounding the seed (i.e., palea and lemma comprising the floret) have extensive hairs and/or awns ( Figure 2 ) (Cavanagh et al., 2019), causing the florets to become easily entangled in one another in seed processing and direct seeding equipment. These inherent handling challenges were the main reasoning for selection in this study yet each of these species also have high value in ecosystem restoration [e.g., post-mine sites, degraded rangelands (Waters et al., 1997; Erickson et al., 2017; Ling et al., 2022)], as well as cultural and forage value (Lazarides, 2002). Seed collection information is available in Appendix I ( Supplementary Table S1 ). Once received, florets were stored at 15°C and 15% relative humidity until experimental use.

Selection of flash flaming in C. ambiguus, C. obtectus, and E. aurea, and acid digestion in E. obtusa was based on previous testing which found these treatments were best suited to improving seed morphology while maintaining or improving germination in the respective species [see Berto et al. (2023)]. Pellets were tested as an alternative method of improving dispersal ability (which does not require manipulation of floret morphology), and for potential benefits to seedling recruitment via an improved ‘pelleted’ microsite and/or the delivery of beneficial compounds (in the instance of topsoil pellets). Flame + pellet and acid + pellet treatments were tested to determine whether additive benefits could be achieved through these combination treatments (e.g., germination and emergence benefits associated with flaming or acid digestion and pellet microsite, and/or enhanced survival and growth from pellet microsite and/or beneficial components of topsoil).

Simulated rainfall in the CEF is achieved via a fully automated, pressurised watering system which applies water at a rate of 10 mm per h⁻¹ (MP 2000 sprinkler heads; Hunter MP Rotator^®, Hunter Industries, San Marcos, CA, USA). In the first week of the experiment, a typical wet-season rainfall scenario was created by delivering discrete events of 30 mm, every second day, over four separate events (e.g., 4 events x 30 mm = 120 mm total). Follow up irrigation events of 30 mm were supplied once per week for the next three weeks of the experiment (90 mm total), then a single irrigation event of 30 mm once per month for the remaining two months of the experiment (overall total water supply of 270 mm throughout the experiment). The initial pulse irrigation event and subsequent infrequent events were selected based on previous studies which have tested a range of rainfall scenarios typical of cyclonic summer rainfall events experienced in this region (Lewandrowski et al., 2017; Stock et al., 2020; Erickson et al., 2023).

The number of emerged seedlings were counted at weeks 3 (maximum emergence) and 14 (final emergence). Upon completion of the experiment (14 weeks), above-ground biomass was harvested and dried at 70°C for 3d before being weighed. Total above-ground biomass was divided by the number of seedlings within a given replicate to provide average seedling above-ground biomass. HOBO^® data loggers (HOBO^® RX3000 Data Logger; Onset Computer Corporation, Bourne, MA, USA) were used to monitor ambient temperature (°C), relative humidity (RH; %), dew point (°C), and photosynthetically active radiation (PAR; uE) ( Appendix III , Supplementary Figure S2 ).

All data were analysed in R Studio (version 4.3.0; R Core Team, 2023). Linear mixed-effects regression models (lmer and glmer functions in lme4 package in R) was used to examine the effects of SET, soil, and the interaction SET * soil on emergence and biomass where applicable. In addition, week and the interaction SET * week were considered to assess survival of emerged seedlings between week 3 and 14. Block was included as a random effect but was removed if it did not significantly improve model fit (in which case a generalised linear model (glm function) was used in the absence of random effects). Where zero scores were common for emergence, zero-inflated beta-binomial mixed-effects hurdle models (glmmTMB package) were used instead ( Appendix IV , Supplementary Table S3 ). Biomass models were weighted using the inverse variance to account for unequal sample sizes. Q-Q Normal plots and plot of residuals were used to check model assumptions. Predictor variable significance was determined using type III ANOVA (Anova function from car package in R). Estimated marginal means and pairwise comparisons (p-values adjusted using the Benjamini-Hochberg method) were obtained using the emmeans package.

Emergence (i.e., seedling counts converted to percentage) was significantly affected by SET and the interaction between SET and soil in all species (p < 0.01; Appendix V , Supplementary Table S4 ). Soil significantly affected emergence in C. obtectus and E. aurea (p < 0.01). Week was a significant predictor of emergence counts in all species (p < 0.05) except for C. ambiguus. The interaction between SET and week was also a significant predictor of emergence in E. aurea and E. obtusa (p < 0.01).

In C. ambiguus, pairwise comparisons showed that the blank pellet treatment tended to result in lower maximum (week 3) and average (i.e., averaged across week) emergence compared to the control and flame treatments when tested on topsoil (by 24–27%, p < 0.05; Figure 3 ; Table 3 ; Appendix VI , Supplementary Table S5 ). On mine waste, the topsoil pellet treatment improved final (i.e., emerged seedlings at week 14) and average emergence (by 34–48%, p < 0.05) and the flame + topsoil pellet treatment improved average emergence (by 36%, p < 0.01) in C. ambiguus compared to the control ( Figure 3 ; Table 3 ; Supplementary Table S5 ). SET comparisons averaged across soil type were not significant for maximum and final emergence in C. ambiguus, but average emergence was lower for the blank pellet treatment compared to all treatments which included flaming (by 17–23%, p < 0.05; Supplementary Table S6 ). Emergence was higher on topsoil compared to mine waste in C. ambiguus for the control (by 21–52% for maximum, final, and average, p < 0.01) and flaming (by 22% for average only, p < 0.05) treatments only ( Supplementary Table S7 ). Significant declines between maximum and final emergence counts were only present for the control when averaged across soil type in C. ambiguus (by 24%, p < 0.05; Supplementary Table S8 ).

In C. obtectus, pairwise comparisons of SETs were not significant on both soil types for maximum emergence. Final and average emergence, however, were generally higher for pellet treatments compared to the control (by 31–67%, p < 0.05), and for the flame + topsoil pellet compared to flaming treatment (by 14–25%, p < 0.05) ( Figure 3 ; Table 3 ; Supplementary Table S5 ). SET comparisons for maximum, final and average (i.e., averaged across week) emergence were not significant for C. obtectus when averaged across soil type ( Supplementary Table S6 ). Emergence in C. obtectus was higher on mine waste compared to topsoil for the flame + topsoil pellet (by 2–54%, p < 0.05; maximum, final, and average emergence), blank pellet (by 28%, p < 0.05; average emergence), and flame + blank pellet treatments (by 21%, p < 0.05; average emergence), and higher on topsoil compared to mine waste for the control (by 29%, p < 0.05; average emergence) ( Supplementary Table S7 ). Final emergence was lower than maximum emergence (i.e., survival declined) for all SETs on topsoil (by 36–69%, p < 0.05), mine waste (by 10–66%, p < 0.05) and when averaged across soil type (by 31–56%, p < 0.05) in C. obtectus ( Supplementary Table S8 ).

Pairwise comparisons for E. aurea showed that the control and flaming tended to have lower maximum and average emergence compared to pelleting treatments (except the blank pellet treatment in some instances) on both soil types and when averaged across soil type (by 6–24%, p < 0.05), though maximum and average emergence were lower for the blank pellet compared to all other pellet treatments (by 7–19%, p < 0.01), and for the topsoil pellet compared to flame + topsoil pellet treatment (by 5–9%, p < 0.05, on topsoil only) ( Figure 3 ; Table 3 ; Supplementary Tables S5, S6 ). Final emergence in E. aurea tended to be higher for the flame + topsoil pellet treatment compared to the control, flame, and blank pellet treatments depending on soil type (by 6–9%, p < 0.05), and for all pellet (except blank pellet) treatments compared to flaming on mine waste (by 5–10%, p < 0.05) ( Figure 3 ; Table 3 ; Supplementary Tables S5–6 ). Emergence (maximum, final, and average) was similar for each SET compared with itself across topsoil and mine waste for E. aurea ( Supplementary Table S7 ). Final emergence was lower than maximum emergence (i.e., survival declined) in E. aurea for pelleting treatments only on both soil types and when averaged across soil type (by 6–17%, p < 0.05; Supplementary Table S8 ).

In E. obtusa, pairwise comparisons showed that maximum and average emergence tended to be higher for the acid + blank pellet and acid + topsoil pellet treatments compared to all other treatments across both soil types and when averaged across soil type (by 6–20%, p < 0.05; Figure 3 ; Table 3 ; Supplementary Tables S5, S6 ). Final emergence was higher for the acid + topsoil pellet compared to the acid treatment and topsoil pellet treatment on mine waste and when averaged across soil type (by 3–9%, p < 0.05), and for the acid + blank pellet compared to topsoil pellet treatment on mine waste (by 6%, p < 0.05) ( Figure 3 ; Table 3 ; Supplementary Tables S5, S6 ). Emergence (maximum, final, and average) was similar for each SET compared with itself across topsoil and mine waste for E. obtusa ( Supplementary Table S7 ). Declines in survival between maximum and final emergence for E. obtusa were only observed for the acid + blank pellet and acid + topsoil pellet treatments on both soil types and when averaged across soil type (by 9–14%, p < 0.001; Supplementary Table S8 ).

SET was a significant predictor of seedling biomass only in C. obtectus (p < 0.001), while soil was a significant predictor of seedling biomass for all four study species (p < 0.05) ( Appendix V ; Supplementary Table S4 ). Pairwise comparisons of SETs on each soil type and when averaged across soil type yielded no significant results, even in C. obtectus (despite SET being a significant model predictor) ( Figure 4 ; Table 3 ; Appendix VI , Supplementary Table S10 ). When each SET was compared against itself across soil type, seedling biomass was higher on topsoil compared to mine waste for all SETs in C. ambiguus (by 0.1473–0.1875 g = 1.7–3.8 times heavier, p < 0.001), and for all SETs except the control in E. aurea (by 0.0438–0.0692 g = 1.4–5.9 times heavier, p < 0.01), and E. obtusa (by 0.0246–0.0446 g = 2.6–4.6 times heavier, p < 0.01) ( Figure 4 ; Table 3 ; Supplementary Table S11 ). Pairwise comparisons of seedling biomass for each SET across soil type were not significant in C. obtectus (despite soil being a significant model predictor) ( Figure 4 ; Table 3 ; Supplementary Table S11 ).

SETs may increase the ability of seeds to utilise often narrow and infrequent windows of opportunity for recruitment typical of dryland environments (Pedrini et al., 2020; Jarrar et al., 2023). For example, faster germination or emergence resulting from SET application may mean seedlings emerge sooner during a pulse rainfall event (e.g., cyclone) and are better able to establish roots to a depth sufficient for survival compared to untreated seeds (Pedrini et al., 2020; Eshleman and Riginos, 2023). Higher maximum germination and emergence resulting from SET application may also allow for a larger number of surviving seedlings compared to untreated seeds, even where survival rates are similar (Kildisheva et al., 2020). An early recruitment advantage (via enhanced germination and emergence) increases the ability of seedlings to utilise temporally limited resources, which may in turn confer to longer-term benefits to plant establishment (Vaughn and Young, 2015).

Flash flaming and acid digestion have previously been found to enhance maximum germination and germination rate in a range of native grasses, including the selected study species (Stevens et al., 2015; Ling et al., 2022; Berto et al., 2023). In this study, there was little evidence to suggest that flaming or acid digestion (used in isolation) benefitted emergence compared to untreated florets (i.e., maximum emergence was similar to the control, subsequently flaming and acid treatments did not result in higher survival or biomass).

Seed pellets are often associated with delayed or reduced emergence (Clenet et al., 2019; Gornish et al., 2019), which could become particularly problematic in dryland environments if rapid emergence is desirable during a narrow recruitment window. For example, Donovan et al. (2023) found that maximum emergence was three-fold higher for non-pelleted compared to pelleted Artemesia tridentata seed, and Munro et al. (2023) found that pellets reduced emergence in Acacia pulchella. By contrast, Baughman et al. (2023) found that pellets resulted in similar or increased seedling counts for native (US) rangeland grasses Elymus elymoides, Poa secunda, and Pseudoroegneria spicata. Similarly in this study, maximum and final emergence counts for pellets were often greater than the control, suggesting pellets did not restrict emergence (except in the instance of blank pellets for C. ambiguus on topsoil). The addition of topsoil (at 1:1 v/v) tended to further enhance emergence outcomes, particularly for E. aurea (whether tested on mine waste or topsoil), and C. ambiguus on mine waste. Final seedling counts were higher for pellet compared to non-pellet treatments in a number of instances due to improved survival (e.g., pellet treatments on mine waste in Cymbopogon sp.) or higher maximum emergence (e.g., acid + topsoil pellet on mine waste in E. obtusa). This highlights the value of maximising seedling emergence and/or survival via SETs to increase the odds of plant establishment.

Pellet hardness and compaction owing to the materials (e.g., clay content) and/or extrusion method are often considered as possible reasons for delayed or reduced seedling emergence from pelleting treatments (Clenet et al., 2019; Gornish et al., 2019; Erickson et al., 2021). The structure of the soil mixture used in pellets in this study may have allowed for better water penetration and/or retention compared to the substrates on which they were tested, a key consideration given that moisture availability is critical in dryland, post-mine restoration (Erickson et al., 2023). In addition, pellet production for restoration has often utilised an industrial pasta/dough machine to extrude the soil matrix containing seeds through a die, risking excessive compression and/or damage to seeds (Clenet et al., 2019; Brown et al., 2021). The pellet production method used in this study (using silicone moulds) eliminates the risk of compression and seed damage and therefore provides a valuable and efficient alternative for pellet production.

An additional key finding for emergence outcomes in this study were the additive benefits to emergence often found when flamed or acid treated florets were used in pellets (i.e., flame/acid + pellet > flame/acid or pellet). In this instance, the pellet microsite may have facilitated the expression of flaming and acid digestion emergence benefits. This was especially pronounced in E. obtusa where acid digestion and pellets when used in isolation had neutral effects on emergence, but significant benefits to emergence were seen when used in combination. There are a limited number of studies which consider the use of SETs in combination due to the increased resource investment associated with applying more than one SET, and the challenge of disentangling the effects of SET when applied in combination (Beveridge et al., 2020; Berto et al., 2021, 2024). However, it has also been recognised that applying a single SET may be redundant if it only addresses one of multiple challenges to seed-based restoration (Berto et al., 2021; Erickson et al., 2021; Berto and Brown, 2022). This concept is supported by the dependence of flaming and acid digestion on pellets to provide emergence benefits in this study.

The enhanced emergence outcomes for pellets, however, did not translate to improved seedling growth (biomass). Overall, this suggests that SET benefits were largely restricted to emergence and early survival in this study, and that these benefits weren’t so significant as to provide a longer-term or ongoing advantage. Multiple studies acknowledge that seed germination and seedling emergence present two of the most critical bottlenecks to plant establishment in seed-based restoration efforts (Merritt et al., 2007; James et al., 2011). However, once these bottlenecks are addressed (e.g., via dormancy break treatments or SETs), other challenges (e.g., soil effects) may arise which prevent successful plant establishment, thus highlighting the need to focus on long-term outcomes.

Emergence was largely unrestricted by soil type for the species selected in this study. SETs tended to improve emergence and survival on both soil types, forming reasonable justification for their adoption in restoration efforts. However, the lack of SET effects on seedling biomass within each soil type, and the significant difference in seedling biomass across each soil type, suggest that the SETs applied in this study are unable to aid in overcoming the significant challenges the mine waste substrate presented to plant establishment.

Amending mine waste substrates with topsoil (or other amendments) has been shown to improve seedling recruitment and plant establishment in mine site restoration (Kneller et al., 2018; Commander et al., 2020; Erickson et al., 2023), though the limited supply of topsoil may mean that these techniques are not always feasible over larger scales (Lamb et al., 2015). Precision delivery of topsoil to seedlings via pellets was considered as an alternative to overcoming challenges associated with seeding into mine waste substrates in this study. Pellets, particularly those containing topsoil (1:1 v/v) tended to enhance emergence (and occasionally survival) on mine waste compared to non-pelleted florets and compared to pellet treatments tested on topsoil. In a number of instances, emergence on topsoil was higher for topsoil pellets compared to the control (e.g., E. aurea, E. obstusa), suggesting that pellet microsite benefits exceeded the benefits of seeding untreated florets into pure topsoil. However, pellets (with or without topsoil) were unable to overcome bottlenecks to seedling growth (biomass) on mine waste soil, with above-ground biomass being strongly negatively impacted by mine waste compared to topsoil substrates.

Conceptually, the precision delivery of topsoil or other microbial amendments (via pellets) can provide seedlings with nutrients and microbes critical to seedling establishment in addition to providing pellet microsite benefits (Alfonzetti et al., 2023). The introduction of valuable soil biota (e.g., via topsoil contained within pellets) can provide significant benefits to seedling recruitment on mine waste soils which lack essential microbial communities (Muñoz-Rojas et al., 2018; Dadzie et al., 2022; Schultz et al., 2022). The poor growth (biomass) on the mine waste compared to topsoil substrate for seedlings from topsoil pellets in this study, however, suggests that the benefits of topsoil could not be achieved via pellets on mine waste. Given the small volume of the pellets used in this study (0.5–1 ml volume), it is possible that topsoil quantities were insufficient to provide prolonged benefits to seedling growth. While studies have shown that it is possible to establish microbial communities in post-mine environments using dilute application of inoculants [e.g., 1:100 biocrust to soil used by Schultz et al. (2022)] and over time frames [e.g., within 6 weeks of inoculation observed by Rezasoltani et al. (2023) and 12 weeks observed by Roman et al. (2018)] comparable with this study, it is possible that the unique site and mine waste substrate characteristics used in this study exacerbate the challenge of re-establishing soil microbial communities. Furthermore, the age and quality of topsoil used in the topsoil pellets may have affected the microbial activity and potential for colonisation (Lamb et al., 2015; Williams et al., 2019).

SET effects tended to be consistent on both soil types (i.e., flaming and acid digestion had similar emergence and pellets enhanced emergence compared to the control on both soil types). In the instance of C. ambiguus, however, interactions between SET and soil were observed. In this case, emergence counts tended to be lower for the control and flamed florets on mine waste compared to topsoil, while pellets had similar emergence counts across the two soil types. This indicates that the control and flamed florets may have been more vulnerable to the detrimental environment of the mine waste substrate, and/or that the mine waste substrate was more inhospitable than the topsoil substrate, creating a recruitment barrier such as a physical crust that prevented seedling emergence. Again, this illustrates the promising microsite benefits provided by pellets to mitigate an edaphic challenge which would otherwise contribute to an emergence bottleneck.

Although flaming and acid digestion used in isolation did not always enhance emergence, survival or biomass when compared to the control, these treatments were able to successfully improve seed morphology while maintaining seedling recruitment potential. Flaming and acid digestion pose a potential risk to seeds, if applied excessively, as these SETs can result in the removal or partial disintegration of floret structures surrounding the seed (Pedrini et al., 2019; Barberis et al., 2023; Berto et al., 2023). With this partial or complete removal, excessive treatment application can chemically alter and/or physically damage seeds which may impede seedling emergence and growth (Kelly and Van Staden, 1985; Loch et al., 1996; Pedrini et al., 2019; Barberis et al., 2023). Furthermore, the structures surrounding the seed may be important in providing protection from predation, physical damage during mechanical seeding and processing, and environmental stressors (e.g., suboptimal temperatures) (Loch, 1993; Cavanagh et al., 2019; Berto et al., 2023). The findings of this study highlight that, when applied effectively, flash flaming and acid digestion can improve seed morphology without making seeds more vulnerable to potential stressors in the environment (e.g., extreme heat, degraded soil) and/or resulting in compromised growth. A subsequent benefit of this finding may be that flamed or acid treated florets can be mechanically seeded with greater efficacy and precision, increasing the likelihood of seeds being positioned at a soil depth or location most conducive to seedling growth and survival (Masarei et al., 2021; Gibson-Roy et al., 2021b; Ling et al., 2022).

Pellets also provided a valuable dispersal unit which effectively concealed the challenging seed morphology of the study species. The increased size and shape of these pellets may also prove to be beneficial for native grass seed which, like other species with small seed mass and high wind resistance (created by structures surrounding the seed), are often challenging to disperse effectively and are easily lost from targeted sites (Gornish et al., 2019; Hoose et al., 2019; Jarrar et al., 2023). In addition, the compatibility of pellets with surface-sowing may allow for successful aerial (e.g., drone) seeding, a particularly important consideration in mine-restoration where site attributes (e.g., terrain, soil rock content, slope) are often unsuited to conventional seeding equipment (Masarei et al., 2019, 2021). The large-scale uptake of and ability to handle pellets may still be limited in some instances due the significant weight and volume increases of pellets compared to non-pelleted seed (Erickson et al., 2021). However, it may be possible to somewhat mitigate these logistical constraints if seeding rates can be reduced as a result of pelleting techniques sufficiently enhancing plant establishment outcomes and reducing seed losses for small seeded or highly wind-resistant species.

The SETs selected in this study, in particular pellets, show great promise for overcoming seed dispersal and seedling recruitment challenges on both topsoil and mine waste substrates. The enhanced benefits of pellets containing topsoil and/or using flaming and acid digestion in combination with pellets provides a proof of concept for precision delivery of topsoil, and additive benefits of combination SETs. However, there are still a number of considerations for the feasibility, improvement, and scalability of these treatments.

Higher emergence following SET application was highlighted as a way to capitalise on limited resources (i.e., pulse rainfall events), thus improving growth and survival. However, the counter scenario is that increased emergence associated with SET treatments could mean seedlings emerge at times not conducive to survival. In these instances, delayed germination and emergence can be favourable (Lewandrowski et al., 2018; Richardson et al., 2019). Because environmental variables (e.g., rainfall events) are often challenging to predict in restoration contexts, several studies have suggested the value of using a mixture of SET treated and untreated seed when implementing seed-based restoration as a means of bet-hedging (Davies et al., 2018; Brown et al., 2021).

A further challenge which persists is the inability of the selected SETs to provide long-term benefits. Pellet benefits observed for maximum and final emergence did not confer an advantage for seedling growth (biomass). Most significantly, precision delivery of topsoil via pellets was unable to overcome the limitations to seedling growth associated with the mine waste substrate. Increasing the volume of topsoil and/or the inclusion of other beneficial additives may allow prolonged benefits of precision topsoil delivery. A potential challenge is that pellets may become too thick for seedlings to readily emerge, and optimal seed positioning within pellets may become essential for seedling emergence (Baughman et al., 2021; Brown et al., 2023; Lieurance et al., 2024). Field-deployed extruded pellets provide an alternative which allows larger volumes of pellet material (e.g., topsoil or other amendments) to be delivered with minimal increases in pellet thickness (Stock et al., 2020). These large ‘pat’ shaped pellets may also further enhance the pellet microsite benefits observed in this study by increasing the size of the microsite and therefore warrant further investigation in overcoming longer-term limitations to plant establishment (e.g., seedling growth) on mine waste substrates.

An additional consideration for future assessment of long-term SET effects (particularly on mine waste substrates) is monitoring of root development. The present study only considered above-ground biomass to assess seedling growth. Multiple studies have highlighted the importance of root development for successful plant establishment (Larson et al., 2020), particularly for native grasses in dryland environments (Fort et al., 2013; Ferguson et al., 2015; Leger et al., 2019). Below-ground biomass may therefore be a more informative outcome variable. In particular, isolating the stage of seedling development at which soil effects overwhelm SET effects would be valuable in developing suitable treatments to address this barrier to plant establishment.

It should also be acknowledged that the rainfall regime utilised in this study represents a high/above average rainfall scenario for the targeted environment (Lewandrowski et al., 2017; Erickson et al., 2023). Given that water availability is the most critical driver of seedling recruitment in dryland, post-mine environments (Erickson et al., 2023), the absence of soil effects on emergence in this study may be explained by the high watering regime used. Future efforts would benefit from testing under a lower rainfall regime to understand the broader range of effects of these SETs and soils on seedling recruitment.

Finally, the scalability of the SETs tested in this study is a key consideration for their uptake and implementation. Flash flaming is already being adopted at scale in mine-site rehabilitation in the Pilbara (Erickson et al., 2017, 2019), and extruded pelleting techniques have been widely implemented in agriculture making them readily scalable (Brown et al., 2021; Svejcar et al., 2021; Jarrar et al., 2023). However, acid digestion techniques are currently laborious, costly (e.g., compared to flaming), and present a risk to the technician (Berto et al., 2023). Engineering solutions will be necessary to enable to scaled application of this technique. Combination SET treatments have not yet been adopted at scale as they are poorly studied and the increased cost of application further reduces their appeal (Berto et al., 2021; Erickson et al., 2021; Berto et al., 2024). However, combination SET treatments may become unavoidable in situations where multiple barriers to seed-based restoration exist. In these contexts, applying a particular SET in isolation may achieve little to no effect whereas applying multiple SETs could have additive benefits, as supported by the data in this study. Therefore, the increased investment in applying combination SETs may result in greater returns and restoration success compared to applying a single SET and will continue to be an important consideration in future SET development and uptake.