Abstract
While plant material releases have typically undergone some level of testing, long-term comparisons and evaluations are relatively few. Understanding long-term environmental adaptation of species and plant releases is critical for achieving persistent site restoration and rehabilitation. Idaho NRCS established a multi-species display nursery in the semi-arid Snake River Plain at Orchard, Idaho, in 2004. The display nursery included 9 entries of bluebunch wheatgrass (Pseudoroegneria spicata (Pursh) Á. Löve [Poaceae]) and 4 entries of Snake River wheatgrass (Elymus wawawaiensis J. Carlson & Barkworth [Poaceae]), which were evaluated yearly for plant density from 2005 through 2008, and again in 2021. In this non-replicated case study, we found several bluebunch and Snake River wheatgrass accessions and releases adapted to local site conditions while not being of local origin. Our observations further support the commonly accepted belief that Anatone bluebunch wheatgrass is better adapted to lower precipitation areas than ‘Goldar’. We did not see evidence that Snake River wheatgrass was more drought tolerant than bluebunch releases at the Orchard study site, but rather, we saw significant variation among releases or accessions of each species.
The Aberdeen Plant Materials Center established a display nursery near Orchard, Idaho, in 2004 to showcase and compare adapted plant materials. Included in the display were 9 bluebunch wheatgrass (Pseudoroegneria spicata (Pursh) Á. Löve [Poaceae]) and 4 Snake River wheatgrass (Elymus wawawaiensis J. Carlson & Barkworth [Poaceae]) entries.
Promising plant materials can undergo a wide range of evaluation protocols during development depending on the class of release targeted. Historically, cultivar releases have been subjected to replicated testing at multiple sites over 2 or more generations spanning 7 to 11 y on average (USDA NRCS 2010), while pre-varietal germplasms (PVG) are evaluated to a much lesser extent (DePue and Englert 2015). For example, selected class germplasms may be evaluated for only 1 to 2 y of a single generation prior to release as illustrated by Tilley (2015) or for 3 to 6 y in some cases (LeFebvre and others 2017), whereas a source-identified release is made from a naturally growing population occupying a known or defined geographic area with minimal or no selection or testing of the parent population (USDA NRCS 2004, 2010).
Despite the time invested in the development of many plant material releases, observations lasting beyond a handful of years are few. This limited information may lead to an incomplete understanding of the scope of adaptation of plant materials being used across a broad spatial range. Conditions in the semi-arid sagebrush steppe ecosystem in the Snake River Plain of western North America, for example, are highly variable. Temperatures can range from –40 °C (–40 °F) in winter to 40 °C (104 °F) in summer (Caldwell 1985), and normal precipitation ranges from 125 to >300 mm (5–12 in) (McNab 1996). The majority of precipitation falls as snow in winter and spring, outside of the growing season (Smith and others 2012), and plant growth depends largely on soil water accumulated over the previous winter (Comstock and Ehleringer 1992). Swings in temperatures and precipitation, as well as fungal or insect outbreaks, may not be captured in short-term studies, and their impact on test materials can remain undetected prior to release and not become apparent for many years.
Practitioner understanding is increasing in terms of which species have more restricted or broader ecological adaptation; however, many prefer local ecotypes for restoration projects (Aubry and others 2005). Concern arises that the use of non-local ecotypes could result in undesired genetic swamping, outbreeding depression (McKay and others 2013), or restoration failure due to maladaptation to local site conditions (Darris 2001; McKay and others 2013). These concerns may not always be the case (Jones 2013a, 2013b; Jensen and others 2020), however. Manipulated-track releases bred or selected for improved seedling vigor, or natural-track releases originating in harsher environments, may be better adapted to the current conditions of a location relative to the local ecotype that evolved under historical conditions that have since changed. Likewise, genotypic blends and polycrosses of high genetic diversity have been suggested as a form of insurance for broad adaptation in variable or changing environments (Jones and Johnson 1998; Jones and others 2002). While genecological studies and molecular analyses can elucidate genotypic and phenotypic differences between populations, they may not be able to adequately describe long-term fitness in a changing environment. To this end, long-term observations of plant material performance remain highly valuable.
Bluebunch wheatgrass (Pseudoroegneria spicata (Pursh) Á. Löve [Poaceae]) (Figure 1) is the dominant bunchgrass throughout much of the Intermountain West’s semi-arid sagebrush steppe (St Clair and others 2013). It is highly palatable and is considered desired forage for a variety of livestock and ungulates (Ogle and others 2010), making it one of the most important grasses to Intermountain rangeland management (USDA FS 1937). Snake River wheatgrass (Elymus wawawaiensis J. Carlson & Barkworth [Poaceae]) is a native of the canyons of the Snake River and its tributaries in Washington, eastern Oregon, and western to northern Idaho. Similar in general appearance to bluebunch wheatgrass, Snake River wheatgrass differs morphologically in having narrower, acuminate (pointed) to aciculate (needle-like) glumes, a more imbricate spike (spikelets overlapping), and glabrate basal leaf sheaths. Snake River wheatgrass was for years misidentified as bluebunch wheatgrass (Barkworth and Carlson 1997), and though it occupies a similar niche, it has been believed to be generally more drought tolerant and to have better establishment in low precipitation areas (Monsen and others 2003; Mukherjee and others 2011). For this reason, it is often recommended as a substitute for bluebunch on drier sites (Ogle and others 2002), even though the site may be outside of the natural range of Snake River wheatgrass. However, this use of Snake River wheatgrass outside its native range may be a concern for those wishing to preserve species composition and distribution. Growing evidence also suggests that Snake River wheatgrass may be more susceptible to reduced water and is naturally limited in distribution to sites of deeper, more fertile soils (Barkworth and Carlson 1997; Jones and others 2021).
Bluebunch wheatgrass is a widely distributed and commonly seeded native perennial bunchgrass in Intermountain rangelands.
These 2 grasses combined are the most broadly seeded native species in the Great Basin in terms of acres covered and volume of seed planted (Gunnell 2021). For example, Utah Division of Wildlife Resources shipped a yearly average of approximately 18,000 kg (40,000 lb) of bluebunch wheatgrass seeds and 15,000 kg (33,000 lb) of Snake River wheatgrass seeds to restoration sites between 2006 and 2019. Over those years, bluebunch and Snake River wheatgrasses combined amounted to 9% of the total shipped seed and 18% of native species (Gunnell 2021).
To effectively match seed source to site, seed transfer zones are being developed using various analyses. Genecological studies examining variation of phenotypic traits in relation to environmental conditions have been used to develop seed transfer guidelines of many restoration species (Johnson and others 2013; St Clair and others 2013; Kramer and others 2015). St Clair and others (2013) proposed 11 seed zones for bluebunch wheatgrass based on plant size, leaf width, and phenology. More recently, Massatti and others (2018) inferred population history from genetic information and proposed 4 large genetic clusters for western North America within which seed transfer might be appropriate.
Released plant materials of both species are widely used by land management agencies, such as USDA Forest Service (FS) and USDI Bureau of Land Management (BLM), and by private landowners and conservation groups. However, long-term plant survival observations are needed to determine which plant materials can persist beyond the initial establishment years and increase and compete on the changing Intermountain landscape. The Orchard Test Site was planted in 2004 in cooperation with the Great Basin Native Plant Selection and Increase Project (GBNPSIP) to allow managers to make side-by-side comparisons of released and experimental materials. The nursery has remained, thanks to management by BLM, and now offers a rare opportunity to observe long-term trends of many materials in semi-arid sagebrush steppe.
MATERIALS AND METHODS
The Orchard Test Site is located in the Snake River Plain of Idaho approximately 37 km (22 mi) southeast of Boise at an elevation of 975 m (3200 ft) (Figure 2) and was selected as representative of conditions in the Snake River Plain ecoregion. Soils at the site are sandy loams of the Lankbush-Chardoton complex (USDA NRCS 2021a). Mean annual precipitation for the location is 290 mm (11.6 in), most of which falls as winter snow (PRISM 2021). The plant community is described by NRCS as a Wyoming big sagebrush (Artemisia tridentata ssp. wyomingensis Beetle & Young [Asteraceae])–bluebunch wheatgrass–Thurber’s needlegrass (Achnatherum thurberianum (Piper) Barkworth [Poaceae]) site (USDA NRCS 2021b). The site falls clearly within seed transfer zone 3a (St Clair and others 2013), in which bluebunch wheatgrass plants were shown to be of medium size, to have early phenology, and to feature narrow leaves. The genetic clusters delineated by Massatti and others (2018), however, did not include representatives from the lower Snake River Plain where the Orchard Test Site is located. The site would appear to fall between the Western Great Basin and the Eastern Great Basin metapopulations (Massatti and others 2018).
The Orchard Test Site is located 37 km southeast of Boise in the semi-arid Snake River Plain. Sources: Esri, DigitalGlobe, GeoEye, i-cubed, USDA FSA, USGS, AEX, Getmapping, Aerogrid, IGN, IGP, swisstopo, and the GIS User Community.
Tested Materials
The Orchard Test Site included 9 entries of bluebunch wheatgrass and 4 of Snake River wheatgrass from plant releases and from test materials provided by the USDA ARS Forage and Range Research Laboratory (ARS) in Logan, Utah (Table 1). Adapted varieties of bluebunch wheatgrass were Anatone Selected Class Germplasm, which is often recommended for use in locations receiving more than 250 mm (10 in) mean annual precipitation, and ‘Goldar’ and P7 Selected Class Germplasm, which are recommended in 300+ mm (12 in) precipitation sites (Ogle and others 2014). Anatone and Goldar were both originally collected in Asotin County, Washington. P-12 and P-15 are predecessor populations of P-24, which would lead to the ARS release Columbia Selected Class Germplasm after an additional breeding cycle (Jones and Mott 2016). The parent population of P-12, P-15, and Columbia was collected near Lind, Washington, in a 250 mm (10 in) precipitation area. P-7 is a multiple-origin polycross of 23 collections and 2 cultivars intended to provide high genetic diversity within a single germplasm (Jones and others 2002). Wahluke is a source-identified germplasm from BFI Native Seeds (2021) collected from a 150 mm (6 in) precipitation site in Franklin County, Washington. Jim Creek is likewise a source-identified germplasm from BFI Native Seeds (2021).
Bluebunch wheatgrass and Snake River wheatgrass plant material entries included in the 2004 Orchard Test Site planting.
Snake River wheatgrass is generally considered to be more drought tolerant than bluebunch wheatgrass in lower precipitation areas (200 to 300 mm) (Ogle and others 2014). The cultivar ‘Secar’ was originally released as bluebunch wheatgrass before further inspection revealed that it belonged to a yet undescribed species. E-26 and E-45 are test materials descended from the cultivar ‘Discovery’ by 1 and 2 cycles of selection, respectively (Jones 2021). SERDP is also a test material that has never been released.
Site Preparation and Implementation
BLM conducted a prescribed burn over the site during fall 2002. The site was later treated in May 2003 and May 2004 with 4.7 l of glyphosate and 2.35 l of 2,4-D per ha to reduce weed cover and prepare a suitable seedbed for planting. Given limited breakdown of dead grass clumps that would inhibit proper seed placement with a drill, and to ensure a clean seedbed, the planting site was cultivated with a cultipacker just prior to seeding to a depth of approximately 5 cm (2 in).
The trial was planted as a fall-dormant seeding on 16 November 2004, using a modified Tye Drill with a width of 2.0 m (80 in) configured with 8 spouts at 0.25 m (10 in) spacing. Each plot is 2.0 m (80 in) wide and 18.3 m (60 ft) in length. Seeding depths and rates followed Ogle and others (2011) and NRCS Idaho practice standards, which target a seeding depth of 13 to 19 mm (0.5–0.75 in) and a seeding rate of 270 to 320 pure live seeds (PLS) per m2 (25–30 PLS/ft2) for both species. Pure live seed values were determined by seed laboratory results. We mixed all seed with rice hulls as an inert carrier to improve seed flow following St John and others (2005). No supplemental irrigation or fertilizer was applied at any time. The test site was excluded from livestock grazing to allow visitors to view mature plants and to allow observation of plant response to environmental conditions alone.
During the first evaluation (2005), most plots contained high densities of Russian thistle (Salsola L. [Chenopodiaceae]) and moderate amounts of bur buttercup (Ranunculus testiculatus Crantz [Ranunculaceae]). Russian thistle plants were approximately 5.0 to 7.6 cm (2–3 in) tall, and the buttercup plants had already flowered. Plots seeded to grasses were consequently sprayed on 9 June 2005 with 1.2 l of 2,4-D and 0.6 l of dicamba per ha to control broadleaf weeds. At the time of the second evaluation (2006), there was a heavy infestation of tumble mustard (Sisymbrium altissimum L. [Brassicaceae]); however, no further weed control measures were taken for the duration of the trial.
Analysis
Given that bluebunch wheatgrass and Snake River wheatgrass are often treated as interchangeable by restorationists, we analyzed both species together and not as separate identities. We evaluated plant density by counting the number of individual plants with at least some rooted material within a 1 m2 frame at 2 m (80 in) intervals starting at the northern end of each plot, for a total of 5 frames per plot. Canopy cover was ocularly estimated using the same 5 frames. Plant density data were collected yearly in early to midsummer from 2005 through 2008 and again in 2021. Canopy cover was evaluated on 2 April 2021.
Annual and long-term average (1981–2010) precipitation for the Orchard Test Site, Idaho (PRISM 2021).
The site was planted as a display nursery and not a replicated trial, so true blocks and replications were not possible for more rigorous statistical analysis. Because the plots were located adjacent to one another, however, some degree of site condition uniformity between plots (soil, slope, received precipitation) can be assumed. We treated the 5 frames as pseudo-replications to generate means and standard errors as an indicator of variation within the plot. Furthermore, the original demonstration design did not include wide border spaces to prevent contamination from adjacent plots. Evaluation frames were placed in the middle of each plot, approximately 0.5 m (1.6 ft) from the nearest neighboring plants from a different entry; however, migration of accessions via seed dispersal is possible. Data presented here should be viewed therefore as a case study and not as a scientific experiment. Conclusions should not be viewed as definitive.
Weather
Precipitation data were obtained from PRISM (2021). Average annual precipitation for the 2004–2020 study period was 304 mm (12.0 in), slightly above the long-term (1981–2010) average of 294 mm (11.6 in). We show annual and long-term average precipitation in Figure 3. The first 3 y of the study received near- or above-average precipitation, 276 to 332 mm (11–13 in). Five years of the study received less than 250 mm, with 2013 being the driest year at 179 mm (7 in). Precipitation exceeded 380 mm (15 in) in 4 y of the study. The wettest year was 2010, with a total of 420 mm (17 in).
We obtained daily high and low temperatures and long-term (1981–2010) monthly average minimum, mean, and maximum temperatures from PRISM (2021). We show long-term monthly averages in Figure 4. The coldest monthly mean minimum temperature, –6.1 °C (21 °F), occurs in December, and the highest monthly mean maximum temperature, 32.8 °C (91 °F) occurs in July. Across the duration of this study, yearly extreme low temperatures for the Orchard Test Site ranged from –29 °C (–20 °F) in 2017 to –12.5 °C (10 °F) in 2020. Winter temperatures below –17 °C (0 °F) occurred in more than half of the years of the study. Summer temperatures reached or exceeded 38 °C (100 °F) in every year of the study. The highest maximum temperature was 44.2 °C (112 °F) in 2013. Temperatures greater than 42 °C (108 °F) occurred in 5 y of the study.
Monthly average maximum, minimum, and mean temperatures for the Orchard Test Site, Idaho (PRISM 2021).
Average plant density of 13 bluebunch and Snake River wheatgrass accessions at Orchard, Idaho, for years 2005–2008 derived from 5 pseudo-replications within each non-replicated plot. Error bars indicate ±1 standard error.
RESULTS
2005–2008
P-12, P-24, and Wahluke bluebunch wheatgrass accessions scored the highest average plant density in the establishment year with 17.2, 13.8, and 13.6 plants/m2, respectively (Figure 5). Snake River wheatgrass accession E-45 had the highest establishment of the species with 15.6 plants/m2. The lowest plant densities in the establishment year were from E-26 Snake River wheatgrass (2.2 plants/m2), along with Goldar and P-5 bluebunch wheatgrass with 4.0 and 6.6 plants/m2, respectively. In general, plant density of most accessions decreased from 2005 through 2008, with the greatest decrease in density occurring between the first and second years. Plant densities declined among all entries from 2005 to 2006 except for Jim Creek bluebunch wheatgrass, which remained constant at 11.0 plants/m2. Many accessions dropped significantly in density in the first year. For example, P-5 bluebunch wheatgrass decreased approximately 45% between 2005 and 2006. Following the declines in density in the first year of establishment, plant densities for most accessions remained somewhat steady between 2006 and 2008. In 2008, 4 y after establishment, the highest plant densities were observed for the source-identified bluebunch wheatgrass releases Wahluke and Jim Creek (both 10.2 plants/m2).
2021
Plant density data measured 16 y after planting varied from 1.2 to 7.4 plants/m2 (Figure 6), indicating some entries may be better adapted to long-term conditions at Orchard than others. Likewise, some releases showed superior long-term persistence to those that had better early establishment. Three entries, all bluebunch wheatgrass, had densities greater than 5 plants/m2 after 16 y, P-24, P-12, and Anatone with 7.4, 6.8, and 5.4 plants/m2, respectively. Goldar, which is reputed to be better adapted to higher precipitation sites than other bluebunch releases (Ogle and others 2014), had lower density (4.6 plants/m2) than P-24 and P-12 but was similar in density to Anatone. Among Snake River wheatgrass accessions, Secar, SERDP, and E-45 had mean densities of 5.0, 5.0, and 3.4, respectively. P-5, Wahluke, and E-26 had the lowest plant densities after 16 y with 1.2, 1.2, and 1.4 plants/m2, respectively.
Percent cover values after 16 y averaged 24% among entries, ranging from as much as 44% (P-24) to a low of 8% (Wahluke and P-5) (Figure 7). Snake River wheatgrass accessions E-45 and E-26 appeared to have somewhat larger plants as indicated by low plant densities but relatively high percent cover. All entries except P-5 and Wahluke had cover of greater than 20%.
Plant densities of Snake River wheatgrass and bluebunch wheatgrass accessions in 2021, 16 y after planting. Error bars indicate ±1 standard error.
Canopy cover of Snake River wheatgrass and bluebunch wheatgrass accessions in 2021, 16 y after planting. Error bars indicate ±1 standard error.
DISCUSSION
Plant densities for all accessions trended downward year by year, and we observed no significant increases. We likewise could see no obvious signs of plant recruitment within the plots. Unfortunately, because of the proximity of some of the original seed sources (for example, several of the bluebunch wheatgrass germplasms came from the same region of Washington), definitive identification using genetic markers would be difficult. We cannot state with perfect confidence that the plants examined in the final evaluation were from the original seeding. However, decreasing plant densities and no apparent recruitment or obvious variation in plant age would suggest that the likelihood of cross contamination between plots is low. Conclusions from this non-replicated evaluation should be approached with caution, and extrapolation to broader scale is tenuous.
St Clair and others (2013) found released materials of bluebunch wheatgrass to be more vigorous, more reproductively fecund, and larger in stature when compared to local ecotypes at multiple common-garden locations. Snake River wheatgrass is not indigenous to southern Idaho, so no local Snake River Plain ecotypes could be included in the Orchard display nursery. Nevertheless, we found many bluebunch and Snake River wheatgrass germplasms of non-local origin to be adapted to local conditions. Plant materials with long-term stand persistence, however, were limited to 3 entries. After 16 y, P-24 and P-12, both precursors to Columbia germplasm bluebunch wheatgrass, and Anatone had high plant density.
Anatone originated in seed transfer zone 6a (St Clair and others 2013) and is noted for large plants, early phenology, and narrow leaves. However, this study indicates that this germplasm, although from a different seed transfer zone, is well adapted to conditions at the Orchard site in the semi-arid Snake River Plain (zone 3a: medium plant size, early phenology, and narrow leaf width). P-12 and P-24 originated from Adams County, Washington, in what falls in zone 3a. Based on St Clair and others’ (2013) seed transfer zones, these would be expected to do well at Orchard, also a 3a site. However, P-12 and P-24 hail from what Massatti and others (2018) described as the Palouse/Wallowa genetic cluster, which was distinct from the Snake River Plain genotypes of the 2 Great Basin genetic clusters. This finding would indicate that while being from a distinctly different genetic cluster, at least some materials from the Palouse/Wallowa cluster can be well adapted to the Snake River Plain.
Our results did not support the notion that Snake River wheatgrass is more drought tolerant or establishes better than bluebunch wheatgrass in semi-arid rangeland. Establishment and persistence of Secar, the only formally released Snake River wheatgrass tested, was similar to that of bluebunch wheatgrass releases. Secar has been described as one of the most drought-tolerant native perennial grasses available and is able to survive in sites receiving as little as 200 mm (7.9 in) of precipitation annually (Ogle and others 2002). Our results indicate that suitable bluebunch wheatgrass alternatives can perform equally well under the semi-arid conditions of southwestern Idaho, specifically on sites receiving greater than 290 mm (11.4 in) of annual precipitation.
NRCS guidelines indicate that suitable plant densities of rangeland seedings in sites with loamy soils receiving <300 mm (<12 in) mean annual precipitation should range between 5.4 and 16.1 plants/m2 (0.5 and 1.5 plants/ft2) (Ogle and others 2011). Based on these parameters, numerous materials tested showed adequate densities for the first 4 y (SERDP, P-12, P-7, Wahluke, Jim Creek, P-24, and Anatone); however, only P-24, Anatone, and P-12 would be considered successful after 16 y. BLM Twin Falls District standards indicate a successful mature stand with >3 deep-rooted perennial grasses/m2 (Jirik 2022). Based on these criteria, all entries save E-26, P-5, and Wahluke would meet the standard of success at 16-y post seeding.
Canopy cover is not commonly used to assess rangeland health as it is seen by many researchers as too variable in plant size due to yearly precipitation; however, it can be useful in comparing performance among ecotypes. Differences in plant size among releases and accessions make it possible to achieve desired cover objectives with lower plant densities, that is, larger, more robust plants may create adequate cover with fewer numbers. This finding would indicate that density alone might not be sufficient as an indicator of adaptation. Some agency-specific goals have been suggested, but these can vary widely by management objectives and local environmental conditions. For sage-grouse (Centrocercus urophasianus Bonaparte [Phasianidae]) nesting habitat in sagebrush steppe, for example, BLM recommends ≥ 10% cover from native bunchgrasses (USDI BLM 2019). Other restoration projects, such as that of the Soda fire in southwestern Idaho, have targeted that foliar cover of large-statured perennial grasses (excluding Sandberg bluegrass) should meet an average of greater than or equal to 20% (Jirik 2022). By both standards, 11 of the 13 evaluated entries produced adequate cover 16 y after planting, including some entries that had relatively low plant density (for example, E-26).
CONCLUSION
Data from the Orchard Test Site would seem to indicate that plant materials of bluebunch wheatgrass and Snake River wheatgrass vary as much or more than the 2 species vary themselves. These results align with those of target-neighbor trials conducted by Jones and others (2021) wherein bluebunch and Snake River wheatgrass seedlings were grown with an annual grass neighbor. We saw a wide range of plant density and canopy cover among entries at the Orchard Test Site, both long- and short-term, indicating a range of phenotypes for both species (Figure 8). We found several bluebunch and Snake River wheatgrass accessions and releases adapted to the long-term site conditions despite originating from distinctly different recommended seed transfer zones than that of the planting site. Our results further suggest that Anatone bluebunch wheatgrass may be better adapted than Goldar to lower precipitation areas. We did not see evidence that Snake River wheatgrass was more drought tolerant than bluebunch releases at the Orchard Test Site, but rather we saw significant variation among releases or accessions. Greater differences in adaptation between the 2 species may be more apparent at lower precipitation sites. More long-term studies examining adaptation of specific germplasm sources to specific conditions of the various seed zones or ecoregions are needed to elucidate adaptation (St Clair and Johnson 2004). Ultimately, understanding the adaptive limitations of specific plant releases may be more informative than generalized seed transfer zones for making seeding recommendations and decisions.
A healthy stand of bluebunch wheatgrass and Snake River wheatgrass at Orchard, Idaho, 16 y after planting.
ACKNOWLEDGMENTS
We thank the BLM for providing the Orchard Test Site and fencing. We also thank the Great Basin Native Plant Selection and Increase Project (GBNPSIP) for partially funding this project. Finally, the authors thank Tom Jones for his thoughtful review and suggestions.
Footnotes
NOMENCLATURE
Photos by Derek Tilley, NRCS
This article was prepared by a U.S. government employee as part of the employee’s official duties and is in the public domain in the United States.
Jump to section
Related Articles
Cited By...
- No citing articles found.