Potential to Improve Wyoming Big Sagebrush Establishment with a Root-Enhancement Seed Technology

ABSTRACT Restoration of the foundational species, big sagebrush (Artemisia tridentata Nutt.), of the sagebrush steppe biome has not kept pace with the loss of habitat, demanding new tools to improve its restoration. Seed enhancement technology (SET) is one approach that is increasingly being tested in native plant restoration as a means to overcome establishment barriers. Like many semiarid shrubs, sagebrush faces establishment barriers from inadequate moisture, competition from faster-growing grasses, and limited available nutrients. We performed a series of laboratory trials testing whether nutrient amendments could be applied to sagebrush seed using a SET to increase root length and biomass, thereby potentially increasing seedling survival. We initially tested 11 amendments applied directly to bare seeds; of these, a high-phosphorus fertilizer resulted in a 2.7x increase in root biomass and 71-mm increase in root length over the control. We then tested incorporating this fertilizer at multiple concentrations into a pellet SET and a ground dust. Although the fertilizer, particularly at higher concentrations, conferred some enhancement to seedling biomass, the pellet treatments had substantially lower emergence and survival than bare seed and dust treatments. These results indicate the potential for a “root-enhancement” SET to benefit sagebrush and other species like it; they also illustrate some of the challenges of SET development for native species. Sagebrush has small seeds that typically need light to germinate. Further work is needed to develop an appropriate technology that does not negatively impact emergence but still provides enough nutrients for enhanced root growth. Field testing is also needed to determine if increases in root growth translate into greater survival. Given the low success rate of sagebrush seeding in restoration projects, however, we suggest that it is worth considering root-enhancement SET alongside other efforts to improve sagebrush establishment success.


Introduction
Drylands around the world face challenges from degradation and poor restoration success, often due to variable precipitation patterns ( Reynolds et al. 2007 ). Seed-based restoration is the most widely used approach in these landscapes and the only approach that can be implemented at large scale, yet recruitment from seed is often low ( Kildisheva et al. 2016 ). Big sagebrush ( Artemisia tridentata Nutt.) is the foundational species of the western North American sagebrush steppe biome, providing critical habitat and forage for the imperiled greater sage-grouse, mule deer, and pronghorn, as well as songbirds like the sage thrasher and Brewer's sparrow ( Welch et al. 1981 ;Yoakum 1980 ;Knick and  Connelly 2011 ), yet it has particularly poor restoration success from seed. While the sagebrush biome historically encompassed > 60.7 million ha ( Wisdom and Rowland 2007 ), it has been reduced by about 50% and continues to shrink at a rapid pace ( Schroeder et al. 2004 ;Davies et al. 2011 ;Miller et al. 2011 ). Wildfire, invasive annual grasses, conifer encroachment, climate change, and development all threaten the sagebrush steppe ( Chambers et al. 2017 ).
One of the largest threats comes from increased wildfire frequency and intensity followed by invasive annual grass invasion, which are both exacerbated by climate change ( D'Antonio and Vitousek 1992 ; Ziska et al. 2005 ). Despite considerable effort by US federal and state natural resource agencies, restoration has not kept pace with this loss ( Lysne and Pellant 2004 ;Knutson et al. 2014 ;Brabec et al. 2015 ;Shriver et al. 2018 ). Historically, under preclimate change conditions, the time required for big sagebrush to recover after disturbance varied widely ( < 30 to > 100 yr) on the basis of subspecies, disturbance https://doi.org/10.1016/j.rama.2023.03.002 1550-7424/© 2023 The Author(s). Published by Elsevier Inc. on behalf of The Society for Range Management. This is an open access article under the CC BY license ( http://creativecommons.org/licenses/by/4.0/ ) type, and site conditions ( Baker 2006 ;Avirmed et al. 2015 ). Predicted future climate conditions may make recovery of sagebrush harder and highlights the need for innovative approaches to restoration of the sagebrush biome, including the foundational big sagebrush itself.
Land managers across the western United States are responsible for revegetating areas after disturbances like mining, wildfire, and oil and gas development. The Bureau of Land Management, the agency responsible for managing much of the sagebrush biome, spends an annual average of $21 million on postfire restoration projects alone ( Bureau of Land Management 2020 ). Sagebrush is commonly included in revegetation seed mixes, yet reseeding with sagebrush often fails to increase the cover or density of the species on restoration sites ( Knutson et al. 2014 ). Therefore, new strategies are needed if we are to keep pace with the loss of this biome and other dryland systems undergoing similar losses ( Kildisheva et al. 2016 ). Many factors can lead to failure or low seeding success in these effort s ( James et al. 2011 ). Establishment is naturally episodic ( Young and Evans 1989 ;Perryman et al. 2001 ) and contingent on temperature and precipitation following seeding ( Meyer 1992 ;Schlaepfer et al. 2014 ). If conditions are not appropriate for big sagebrush germination in the first growing season after seeding, seeds typically do not remain viable in the seed bank ( Young and Evans 1989 ). Unless additional reseeding is done, it may take decades for big sagebrush to establish naturally on the site by reseeding itself inward from the edges of disturbance boundaries ( Shaw et al. 2005 ;Schlaepfer et al. 2014 ). Big sagebrush seed dispersal distance is limited, with most seeds falling within 1 m of the mother plant's canopy ( Meyer 1992 ;Schlaepfer et al. 2014 ).
Other hurdles like limited seed resources and annual weather patterns can limit establishment for this species and other longlived shrubs in drylands around the world. Big sagebrush has small seeds, averaging about 4 100 seeds ·g −1 ( Schlaepfer et al. 2014 ), and therefore needs to attain resources from the environment that enable strong taproot growth to survive summer drought periods. Consequently, most big sagebrush seedlings die during the first year of growth ( Schlaepfer et al. 2014 ), and both seed germination and subsequent seedling survival are contingent on temperature and precipitation patterns ( Cawker 1980 ;Owens and Norton 1989 ;Schuman et al. 1998 ;Maier et al. 2001 ;Hardegree et al. 2016 ). Even short-term increases in water availability in the spring can impact the survival of seedlings ( O'Connor et al. 2020 ). An increase of just 7 d of good growing conditions in March (23 d vs. 16 d with daily water availability > −2.5 megapascals and temperatures > 0 °C) created conditions where big sagebrush could establish ( O'Connor et al. 2020 ), whereas the sites with 16 d of good growing conditions did not have any big sagebrush. These results emphasize the importance of spring water availability to seedling survival and how climate-induced drought and altered precipitation regimes are likely to negatively impact establishment.
Surviving big sagebrush seedlings have a growth strategy adapted to early soil dry-down in the late spring ( Welch and Jacobson 1988 ). They have rapid early-season growth rates, with most resources going to taproot development ( Welch and Jacobson 1988 ). Wyoming big sagebrush ( Artemisia tridentata Nutt. spp. wyomingensis Beetle & Young) puts down longer roots faster in the first 10 d of growth than other subspecies, which occur at wetter sites ( Welch and Jacobson 1988 ). In one lab experiment, 40 d after germination a Wyoming big sagebrush seedling could access water in the majority of the soil profile, with roots reaching a depth of 413 mm ( Welch and Jacobson 1988 ). However, root growth is contingent on favorable conditions and nutrient availability. In areas with low nutrient availability, fertilizer additions increased basin big sagebrush ( Artemisia tridentata Nutt. spp. tridentata ) growth rate by 119% over seedlings that were not fertilized ( Kainrath et al. 2021 ). Application of fertilizer may improve the survival of young big sagebrush seedlings by enabling early and rapid root growth, giving them access to water deeper in the soil profile.
Competition from other plants can also inhibit big sagebrush seedling growth and survival by limiting the availability of water, nutrients, and light ( Schlaepfer et al. 2014 ). Invasive annual grasses, like cheatgrass ( Bromus tectorum L.), are particularly adept at competing with native species, including sagebrush, and can reduce big sagebrush seedling biomass by > 95% ( Wijayratne 2011 ). It is estimated that cheatgrass is now present in > 24 million ha across the western United States ( Downs et al. 2016 ). Fall germination allows cheatgrass to capitalize on soil moisture and nutrients before sagebrush seedlings begin growing ( Mack and Pyke 1983 ;Pellant 1996 ). Perennial grasses, particularly when seeded on a previously disturbed site, can also outcompete sagebrush seedlings ( Davies et al. 2013 ). On mine reclamation sites in Wyoming, greater perennial grass seeding rates significantly lowered big sagebrush canopy volumes ( Vicklund et al. 2004 ) and big sagebrush seedling density ( Schuman et al. 1998 ).
Innovative solutions like seed enhancement technologies (SETs) may improve the success of native revegetation efforts in drylands ( Kildisheva et al. 2016 ). SETs are commonly used in agriculture to augment seedling establishment but are only beginning to be developed for native plant restoration ( Brown et al. 2021 ;Pedrini et al. 2020 ). SETs can include but are not limited to seed priming, coating, and pelleting with compounds that can enhance outcomes like seed germination and seedling vigor ( Pedrini et al. 2020 ). SETs have the potential to improve the establishment of big sagebrush and other native species by helping overcome known establishment barriers, such as a short-term lack of water after seedling emergence ( Madsen et al. 2016a ). If successful, a SET approach could be beneficial for restoration of dryland sites with undisturbed soils where short-term drought and competition are the major establishment barriers, as well as disturbed soils where nutrient deficiencies may be compounding the moisture barrier.
In this study, we begin to explore the potential for fertilizeraddition SETs to promote the growth of a larger root system early in sagebrush seedling development, hypothesizing that this might better equip seedlings to survive short-term pulses of low water availability ( Donovan and Ehlerlinger 1991 ). We conducted a series of four laboratory experiments that progressively built off each other. Experiment 1 tested the direct application of a suite of available and expert-recommended fertilizers and soil amendments on seedling emergence, survival, and growth. On the basis of the results of the first trial, Experiment 2 focused on a high phosphorus fertilizer and further tested its effect on aboveground and belowground seedling characteristics. Experiment 3 evaluated how this fertilizer performed when incorporated into a SET, specifically extruded pellets, and Experiment 4 tested what fertilizer concentration incurred the most benefit to the roots of sagebrush seedlings. Our hypotheses for these four experiments are as follows: Experiment 1: Added fertilizers will not impact emergence but will improve sagebrush survival and increase aboveground biomass when compared with bare seed. Experiment 2: Added fertilizer will not impact emergence but will improve sagebrush survival and increase root length and biomass, as well as shoot biomass, when compared with bare seed. Experiment 3: Root-enhancement pellets will increase seedling survival and aboveground biomass when compared with bare seed but may inhibit seedling emergence. Root-enhancement dust will not inhibit emergence and will increase seedling survival and aboveground biomass when compared with bare seed. Aboveground biomass will increase as the amount of fertilizer increases.
Experiment 4: Belowground biomass and root length will increase with increasing amounts of fertilizer.

Methods
For all four experiments, we used soil collected from the reclaimed Andria Hunter mine site (42.786098 N, −107.656883 W) in the Gas Hills region of Wyoming. Soils on the reclaimed mine site are characterized as a sandy clay loam (see Table S1, available online at 10.1016/j.rama.2023.03.002 , for full soil analysis data). We sifted the soil through a 1-mm sieve to exclude rocks and litter. Wyoming big sagebrush (hereafter sagebrush) seed was purchased from Granite Seed and Erosion Control (Denver, CO).
For all experiments, we grew sagebrush seedlings at approximately 22 °C/72 °F in The Nature Conservancy Wyoming Seed Lab under Platinum P1200 LED lights (PlatinumLED, Kailua, HI). The lights were placed 61 cm (24 inches) above the top of the pots and set to Vegetation mode. The photoperiod for each experiment was 12 h light/12 h dark. All biomass samples were dried in a forcedair oven (Model 89511-410, VWR International, Germany) for 24 h at 65 °C before being weighed.
The response variables we measured in pot experiments were biomass per pot, percent total emergence, and percent survival. Biomass per pot was the total aboveground biomass of all seedlings in a pot. Percent total emergence was determined by dividing the total number of seedlings per pot that emerged throughout the trial (even if they died and were not part of the final seedling count) by the estimated number of seeds per pot. Seedling survival was calculated by dividing the final seedling count per pot by the total number of seedlings that emerged in that pot throughout the trial. The response variables we measured for PVC pipe experiments were shoot (aboveground) and root (belowground) biomass per individual (one individual per pipe), root:shoot biomass ratio, shoot length, and root length.

Experiment 1: fertilizer screening pot trial
In Experiment 1, we tested the effects of 11 different fertilizers on sagebrush seedling emergence, survival, and growth to explore fertilizers for potential use in a root-enhancement SET. We tested a mix of fertilizers already being used in sagebrush SET trials, mine site restoration, and several additional commercially available plant growth supplements. Before seeding, we lined the bottom of 120 plastic 650 mL pots with weed cloth, filled them with approximately 470 g soil, and watered them to field capacity. Twenty-four h later, we randomly applied the 11 fertilizer treatments and a nofertilizer control to 10 pots per treatment (see Table S2 for details on fertilizer products and application rate). After an additional 24 h, we placed 12 sagebrush seeds in each pot in a grid pattern, equidistant from each other and the edges of the pot, and covered them with 3 mm of soil. We watered the pots to field capacity again immediately following seeding, after which they were watered with a 1 GPM Fogg-It spray nozzle (Fogg It Nozzle Company, Belmont, CA) to field capacity whenever the soil appeared dry at the top of the pot.
We recorded seedling emergence and survival twice a wk over the course of 57 d. At the end of the experiment, we clipped all seedlings at ground level and dried them to determine total pot aboveground biomass.
For this and all other experiments, we analyzed data using R 3.6.0 (R Core Team 2019). We used one-way analysis of variance (ANOVA) to evaluate the effects of fertilizer treatment on the total biomass of all seedlings in each pot after performing a log transformation to meet model assumptions. Percent emergence and percent survival were analyzed using generalized linear models with quasibinomial error distributions to account for overdispersion. Post-hoc pairwise comparisons between variables were performed using the Tukey method.

Experiment 2: root and grow PVC pipe experiment
On the basis of Experiment 1's results, in Experiment 2 we tested how the high-performing fertilizer, Root & Grow (subsequently RG; Bonide, Oriskany, NY), impacted both aboveground and belowground growth of sagebrush seedlings. For this experiment, we cut 5 cm x 61 cm (2 inch x 2 ft) PVC pipes in half lengthwise and used packaging tape to put them back together before filling them with soil. This enabled us to later open the pipes and extract intact seedling root systems. We mixed 50% soil with 50% sand (Quikrete Premium Play Sand, Atlanta, GA) to improve water drainage through the PVC pipes. We filled each PVC pipe with approximately 1 200 mL of the soil mixture and watered the pipes to field capacity 24 h before the start of the experiment.
For the experimental treatment, we added 10 mL of RG fertilizer solution (same concentration as Experiment 1) to each of the 13 pipes, after which we placed 6 sagebrush seeds onto the soil surface. Another 13 pipes were seeded with sagebrush without adding fertilizer for a bare seed control. If no seedlings had emerged after 13 d, we added 10 more seeds to the pipe. We tracked the date of emergence of each seedling and then retained the first seedling to survive to a true leaf stage in each pipe and pulled all other seedlings. We watered the pipes to field capacity every 2 −3 d, as needed, to keep the pipes moist. After 42 −45 d of seedling growth, we opened the pipes and removed all the soil from the seedlings with light rinsing. We measured the root and shoot length of each seedling, divided each into aboveground and belowground structures, dried, and weighed their biomass.
Before analysis, we standardized all seedling measurements by the number of days of growth. We used separate one-way analyses of variance to evaluate the effects of fertilizer treatment on shoot mass, root mass, root length, and root-to-shoot biomass ratio. Shoot mass and root length were not normally distributed and were analyzed using a generalized linear model with a quasipoisson error distribution to account for overdispersion. We used a log transformation for the root-to-shoot biomass ratio and a square root transformation for root mass to meet model assumptions and analyzed it using a linear model.

Experiment 3: root-enhancing seed enhancement technology pot trial
To evaluate the potential to incorporate root-enhancing additive into a seed enhancement technology (SET), we added RG to two seed delivery mechanisms-an extruded pellet (hereafter pellet) and dust containing the same ingredients as the pellet. Pellets have been tested in a laboratory and field setting in other SET applications ( Madsen et al. 2016b ;Brown et al. 2018 ;Davies 2018 ) but have not been evaluated in the context of fertilizer delivery. We were concerned that pellet compaction and limited light availability might negatively impact smaller-seeded species like sagebrush ( Clenet et al. 2019 ). Although dust is not a practical way to deliver a root-enhancement fertilizer at a restoration scale, it allowed us to evaluate the benefits of the root enhancement material without negatively impacting seedling germination and emergence ( Clenet et al. 2019 ).
The recipe for these delivery mechanisms was modified from a base recipe used in previous research with herbicide protection pellet seed technologies ( Clenet et al. 2019 ). Root-enhancement pellets and dust consisted of 54.5% clay, 17.5% worm castings, 25.1% compost, 2.9% seed, and 0.01% fungicide by dry weight mixed with 400 mL of liquid (combination of water and RG, if applicable). The root-enhancement pellets and dust were made with four different levels of fertilizer: 3x, 5x, 8x, and 10x the application rate on the  ( Table 1 ). We mixed all dry ingredients using a commercialgrade stand dough mixer (Model M20 ETL, Eurodib, Plattsburgh, NY). We dried the dough for the dust treatment in a forced-air drying system for 24 h and then manually ground it to a fine powder that could be spread evenly within a pot. For the pellet treatment, we passed the dough through a commercial-grade pasta extruder (Model TR95, Rosito Bisani, Los Angeles, CA) with a 4.3-mm circular die to create pellets approximately 9.5 mm in length and dried them in the same manner as the dust. We prepared 72 pots as in Experiment 1 and randomly assigned 8 replicates of each of the 9 treatments. For the bare seed treatment, we applied 20 seeds to the surface of the soil and pressed them down gently to improve seed-soil contact. Using the same method, for the dust treatment we placed 20 seeds in the pots and then spread 1.78 g of root-enhancement dust over the seeds. For the pellet treatments, we placed approximately 1.78 g of rootenhancement pellets into each pot. We dissected the pellets to estimate that there were approximately 40 seeds in 1.78 g of pellets.
Previous unpublished laboratory trials with Wyoming big sagebrush had shown a significant reduction in seedling emergence from pellets. Therefore, we seeded those at a higher rate than dust and bare seed to try to get a similar number of seedlings per pot in this experiment. After treatments were applied, we watered the pots to field capacity and subsequently every 2 d or as needed using a 1 GPM Fogg-It nozzle. We monitored seedling emergence and survival twice a wk for 4 wk. At the end of the experiment, we clipped plants at the soil level and dried and weighed them to measure total aboveground biomass per pot.
We took a two-step approach to analyzing the effects of our experimental treatments on seedling response variables. First, we used a generalized linear model on all data, with bare seed and each combination of fertilizer concentration and delivery mechanism as separate treatments. This enabled us to compare all rootenhancement SET treatments to bare seed. Second, we used a separate linear model to test the effects of fertilizer concentration (four levels) and delivery (two levels) and the interaction between these two factors. In the latter analyses, we excluded the bare seed treatment since it did not fit into the factorial design. This allowed us to test more specific hypotheses about fertilizer concentration and delivery mechanism.
For biomass, one replicate pot (Pot 44, 5x pellet) had a perseedling biomass twice as high as any other pot. Because this pot was an extreme outlier, we excluded it from further analysis to be conservative (keeping this pot in the analysis would have strengthened the positive effect of increasing fertilizer concentration on biomass; therefore, we feel justified that its exclusion was unbiased and warranted as part of data quality control).
Biomass per seedling and percent emergence were normally distributed, and the data were analyzed using a linear model. Percent survival responses were analyzed using a generalized lin-ear model with a quasibinomial error distribution to account for overdispersion.

Experiment 4: root-enhancing seed enhancement technology PVC trial
On the basis of Experiment 3's results, we carried the 5x and 10x root-enhancing SETs (dust and pellet) forward into a PVC pipe trial so that we could examine their impact on belowground and aboveground seedling characteristics. Seven replicates of each SET, as well as a bare seed control, were tested in this experiment. Approximately 0.5 g of pellets, containing about 11 seeds, were added to the designated pipes since we only needed one seedling per pipe for the experiment, but notably this resulted in less than one third the fertilizer used per pot in Experiment 3. For the dust treatment, 0.5 g of dust was evenly spread over 10 bare seeds and 10 bare seeds were added to each of the bare seed control pipes. Each pipe was opened after 40 d of growth when it was anticipated that the first roots would be near the bottom of the pipe based on Experiment 2. Watering throughout the experiment followed the same protocol as Experiment 2, and at the end of this experiment the seedlings were processed in the same way as Experiment 2.
We used the same two-step approach to analyze the effect of treatment, as well as delivery and concentration on seed morphometric measurements. No transformations of the data were necessary to meet model assumptions and linear models were used for the analysis.

Experiment 2: root and grow PVC pipe experiment
Adding RG fertilizer to the soil before seedling emergence had a strong positive effect on seedling biomass. Average root mass was about 2.7x greater for seedlings treated with fertilizer than untreated seeds ( Fig. 2 ; F = 23.99, N = 24, P < 0.01). There was also strong evidence that RG had a positive effect on root length compared with untreated seeds (see Fig. 2 ; F = 10.25, N = 24, P < 0.01). The average root length of bare seed was 583.91 mm, but when RG was added, the average root length increased to 655.29 mm. Shoot mass also varied strongly by treatment, with seedlings grown with RG having three times greater mass than those grown without (F = 33.63, N = 24, P < 0.01). The mean shoot mass for bare seeds was 0.03 g and for fertilizer was 0.10 g. There was no evidence that fertilizer treatment influenced the root-to-shoot biomass ratio (F = 0.01, N = 24, P = 0.92).

Experiment 3: root-enhancing seed enhancement technology pot trial
When all delivery (pellet and dust) and RG concentration (3x, 5x, 8x, or 10x) combinations and bare seed were considered as in-dividual treatments, treatment had a strong effect on percent total emergence ( Fig. 3 A; F = 8.40, N = 72, P < 0.01), most notably because bare seeds emerged to a greater extent than any other treatments except 5x dust. When we separately tested the effects of delivery type and concentration (excluding bare seed), the interaction between delivery type and concentration was not significant. We found that dust treatments had 57.6% higher emergence than pellet treatments (F = 17.51, N = 64, P < 0.01). Concentration also had a moderate impact on percent emergence (F = 3.85, N = 64, P = 0.01), with 5x treatments showing higher emergence than any other concentration-apparently driven primarily by high emergence in the 5x dust treatment.
Treatment also strongly impacted percent survival of seedlings when all treatment combinations were considered (see Fig. 3 B; χ 2 = 146.19, N = 72, P < 0.01), with the bare seed treatment having the highest percent survival, 97.2% (see Fig. 3 B). When the effects of delivery and concentration were analyzed separately, the interaction between delivery and concentration was not significant ( χ 2 = 6.85, N = 64, P = 0.37). There was strong evidence that delivery impacted percent survival of seedlings ( χ 2 = 38.32, N = 64, P < 0.01). The dust treatment had 89.5% survival while the pellet treatment had 77.1% survival. Concentration of fertilizer had a moderate effect on survival with a clear trend toward lower survival at highest concentration ( χ 2 = 18.87, N = 64, P = 0.03).
When all treatments were considered, there was only weak evidence that treatment affected biomass per seedling ( Fig. 3 c; F = 2.07, N = 72, P = 0.05). Biomass per seedling was lowest in the 3x pellet treatment, highest in the 10x dust treatment, and intermediate in all other treatments. When analyzed without bare seed, delivery method had an effect on the aboveground biomass of seedlings, with dust treatments showing higher biomass than pellets (F = 7.94, N = 64, P < 0.01). There was no evidence that concentration and the interaction between concentration and delivery impacted seedling biomass. However, overall there was a trend, albeit with weak evidence, for higher fertilizer concentration to be associated with higher biomass (F = 2.29, N = 64, P = 0.09).

Table 2
Analysis of variance results assessing different root-enhancement SETs for Wyoming big sagebrush in Experiment 4. Analysis 1 looked at the effect of treatment (bare seed and each combination of delivery and concentration), while analysis 2 split treatment into concentration and delivery. Asterisks represent significant differences among treatments ( P ≤ 0.05).  3. Experiment 3 results. A, Percent total emergence ± 1 standard of error (SE) across all treatments, B, percent survival ± 1 SE across all treatments, and C, aboveground biomass (g) per seedling ± 1 SE across all treatments. Circle represents bare seed, triangle represents dust, and square represents pellet. Fertilizer treatments sharing the same letter are not significantly different at the P = 0.05 level.

Experiment 4: root-enhancing seed enhancement technology PVC trial
When all delivery (pellet and dust) and RG concentration (5x or 10x) combinations were treated as individual treatments, there was no strong evidence that treatment impacted any of the seedling morphometric measurements ( Table 2 ). There was weak evidence that treatment affected root length, with the greatest difference between the 5x dust treatment and bare seed, average root length of 617 cm versus 602 mm, respectively.
When the effect of concentration and delivery were analyzed separately from bare seed, the interaction term was not significant for any response variable. The impact of concentration and delivery differed depending on the morphometric measurement. Delivery method had a moderate impact on root mass ( Table 2 ) while there was no evidence that concentration had an effect. Dust treatments had 54.9% greater root mass on average than pellet treatments. For root length, both delivery and concentration had a moderate effect (see Table 2 ). Roots were 14.3% longer in the dust treatment on average when compared with the pellet treatment. The lower fertilizer concentration, 5x, produced 11% longer roots than the higher concentration. This seemed to be driven by the longer length of roots in the 5x dust treatment. Shoot mass was weakly impacted by delivery but not by concentration (see Table 2 ). The dust treatment yielded 48.3% higher shoot mass than the pellet treatment. There was no evidence that delivery or concentration influenced the root-to-shoot ratio of the seedlings.

Discussion
This series of studies demonstrated that a fertilizer amendment can increase aboveground and belowground biomass of Wyoming big sagebrush seedlings under controlled laboratory conditions. However, incorporating fertilizer into an SET for sagebrush that can be used at scale may be difficult given the amount of fertilizer needed and the species-specific germination and emergence requirements of a small-seeded species like sagebrush.
In our preliminary screening of several fertilizers (Experiment 1), we saw that the type of fertilizer impacted the emergence, survival, and aboveground biomass of sagebrush seedlings. The fertilizers that negatively impacted seedlings (Biosol and seaweed) had high rates of nitrogen or potassium, respectively, compared with other nutrients (see Table S2). The availability of phosphorus early in plant development has been shown to be critical for many species ( Grant et al. 2001 ), and the poorly performing fertilizers in this experiment have low phosphorus concentrations. The mine-site derived soils used in these laboratory experiments had low levels of phosphorus (see Table S1), and when combined with fertilizers also lacking this crucial nutrient, reduced biomass is not surprising. Nutrient additions are beneficial to native plants in depleted soils like in cheatgrass-invaded rangelands. For example, the addition of phosphorus increased the growth of basin big sagebrush ( Artemisia tridentata spp. tridentata ) transplants by 119% when added to low phosphorus soils in Utah ( Kainrath et al. 2021 ). Root & Grow was the one fertilizer we tested that had a higher proportion of phosphorus than either nitrogen or potassium, which may be why it outperformed the other amendments.
Further testing (Experiment 2) showed that the addition of liquid RG solution resulted in greater aboveground biomass and also increased root length and biomass, which is the result we are trying to achieve with a root-enhancement SET. A challenge emerged, though, when we tried to incorporate RG into a SET in Experiment 3 rather than adding it directly to sagebrush seeds. Applying the fertilizer directly is not feasible in a real-world restoration. The equipment used to spread seed during large-scale restoration projects is not set up to deliver liquid fertilizer applications, and widespread fertilizer application is not recommended because it enhances weed growth ( Kay and Evans 1965 ;Wilson et al. 1966 ). A pellet SET provides a convenient delivery mechanism and has been tested with grasses without negatively impacting emergence ( Madsen et al. 2016b ;Brown et al. 2018 ;Davies 2018 ); however, Experiment 3 showed there was a significant cost to sagebrush emergence from pellets. Many populations of sagebrush produce seeds that need light to germinate ( Meyer et al. 1990 ); being encased in a pellet prevents light from reaching the seeds and is likely one of the causes for reduced seedling emergence that we saw in all pellet treatments. Ideally, pellets would disintegrate, leaving the seed on the surface. However, we often observed pellets remaining intact well after sowing or pellets dissolving but staying mounded, leaving seeds buried inside. Sagebrush seeds are small and have limited capacity to emerge from compacted or crusted soils ( Clenet et al. 2019 ). If they are compressed within a pellet, breaking out of the undissolved or mounded pellet presents a challenge. The dust treatment used in Experiments 3 and 4 was an attempt to provide RG fertilizer to the sagebrush seeds while reducing pressure on the seed and allowing the seed access to light to stimulate germination. Although there would be challenges to the real-world application of this treatment, it did have greater percent emergence over the pellet treatment, but for reasons we don't understand it was still markedly lower than the bare seed treatment. It is possible that a lower-integrity pellet or the freezethaw cycles of the field would help seeds to emerge from pellets.
We saw a distinct trend in Experiment 3-greater amounts of RG led to greater aboveground biomass. We did not see a similar trend in Experiment 4. We hypothesize that this is because of the differences in leaching between the pots and pipes. For the pipes, soil was mixed with sand to improve drainage within the long column of the pipes. This could have changed not only the drainage of water but also the leaching of fertilizer through the pipe, so seedlings had less time to use the additional nutrients before they were washed out of the pipe. An alternative explanation is that the overall amount of RG per experimental unit was about one-third less in Experiment 4 than Experiment 3. The amount of RG per seedling, however, was higher in Experiment 4. Without further testing, we do not know whether the total amount of fertilizer or fertilizer per seedling is more influential. Albeit not statistically significant, in Experiment 4 RG still provided a boost to the seedlings, particularly in the dust treatment. We hypothesize that if we could incorporate greater amounts of RG into a single rootenhancement SET unit, like a seed coating or equivalent amount of dust, we would continue to see enhanced aboveground and belowground biomass. This may come at the cost of seedling survival immediately post emergence though, as we saw in Experiment 3.
The root-to-shoot mass was not negatively impacted by any fertilizer treatment in these experiments, which was a critical piece of information for exploring the viability of this root-enhancement SET. Ideally, root-enhancement SETs would produce seedlings with a greater root-to-shoot mass ratio than untreated seed. In a literature review of how root traits are impacted by drought, root-toshoot ratio was the trait that most frequently increased in response to drought, particularly in woody plants ( Garbowski et al. 2020 ). In arid environments, access to water and therefore a large root system plays an outsized role in seedling survival. While this SET did not increase the root-to-shoot mass ratio of seedlings, it did not decrease it, which would, in field conditions, likely lead to large aboveground plants with an inadequate root system.

Implications
This series of experiments shows that there is potential for a root-enhancement SET to benefit establishment of sagebrush and other dryland shrubs for which restoration from seed is a challenge. The addition of crucial nutrients, like phosphorus, significantly increased root length and biomass. Increasing root growth might help seedlings survive short ecological droughts and prolonged summer droughts when most new sagebrush seedlings die ( Schlaepfer et al. 2014 ) by providing access to water lower in the soil profile. The key to creating an effective root-enhancement SET for sagebrush and other small-seeded species will be developing a mode of delivering enough nutrients without negatively impacting the germination and emergence of seeds. Negative impacts to emergence have been seen across other seed pelleting studies ( Gornish et al. 2019 ). Seed coatings, which encompass a single seed in less material than a pellet, may be a better fit for the germination and emergence criteria of sagebrush and other similarly smallseeded or photo-dormant species. We suggest that a film coating or rapidly dissolving encrusting be tested for future sagebrush root-enhancement SETs ( Pedrini et al. 2020 ). While we saw some promising results in a controlled laboratory environment, testing under field conditions will be a crucial step toward determining if root-enhancement SETs can improve sagebrush establishment. SETs are still an emerging tool for dryland restoration ( Madsen et al. 2016a ;Brown et al. 2018 ;Clenet et al. 2019 ), although they are commonly used in agriculture. This research has shown the impor-tance of tailoring SETs to a specific species by taking into account factors like germination requirements and seed size, and it underscores the notion that species-specific requirements will need to be considered in the development of native plant SETs ( Baughman et al. 2021 ;Brown et al. 2021 ).
As threats to the sagebrush steppe and drylands globally increase with climate change, it is imperative that we improve our techniques for restoring native plants. Even incremental improvements in establishment success are likely to be cost-effective given the inherent challenges of dryland restoration and the current low success rates for many species, including sagebrush. SETs and the concept of a root-enhancement technology are one potential pathway to achieve this.

Declaration of Competing Interest
The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.