Abstract
Basalt milkvetch and western prairie clover are native perennial legumes distributed in the northern Great Basin and Columbia River Plateau of North America, with recent germplasm releases providing seed for revegetation. Although seed germination studies have identified scarification requirements for both species, and additional cold stratification for Astragalus filipes, no research has been conducted to optimize the likelihood of seedling emergence on rangeland sites. In this study, we planted both species in 3 locations in the northern Great Basin over 3 y. Seeds were planted in both fall and spring seasons, with emergence measured each June. For both species, there was variation in seedling emergence success across locations and years, with the Powell Butte, Oregon, location providing the highest emergence and the 2011–2012 winter season leading to the highest emergence. For A. filipes, the greatest seedling emergence occurred with fall plantings with scarified seed because of a need for cold stratification and a slow germination rate. For Dalea ornata, because of its hard seedcoat yet rapid germination, the greatest seedling emergence occurred with spring plantings with scarified seed. Our data indicate a need for scarifying seed for both species, but planting A. filipes in fall and planting D. ornata in spring.
Western rangeland ecosystems are increasingly subject to large-scale disturbances from wildfires, weed invasion, soil erosion, and mismanagement (Pierson and others 2011). These rangelands can become degraded with respect to their biodiversity, ability to cycle nutrients or water, and ability to sustain wildlife or livestock to the point they may require revegetation for recovery or restoration (Sheley and others 2006). Revegetation efforts have relied mainly on plantings of grasses, but legumes and other forb plants are critical for biodiversity, pollinator support, and as nutrient sources for herbivores. Therefore, more recent revegetation efforts have focused on the collection and seed source development of forbs (Shaw and others 2005).
Two such legumes important in the Great Basin were collected and developed into germplasm releases for seed production and revegetation. Astragalus filipes Torr. ex A. Gray (basalt milkvetch [Fabaceae]) (USDA NRCS 2018) is distributed across the semi-arid western US (Bhattarai and others 2008; Bushman and others 2010), is nontoxic for livestock and wildlife, attracts a wide variety of pollinators, and has the potential for commercial seed production (Bhattarai and others 2008; Watrous and Cane 2011). As a legume, it forms nodules with many nitrogen-fixing bacteria and has high protein content and forage quality (Bhattarai and others 2008). It is also a long-lived perennial that germinates in early spring and can sprout quickly after fires (USDA NRCS 2018). The NBR-1 germplasm of A. filipes was released in 2008 from a combination of A. filipes collections spanning the Northern Basin and Range Level III ecoregion of the western US (Johnson and others 2008).
Dalea ornata (Douglas ex Hook.) Eaton & J. Wright (western prairie clover or blue mountain prairie clover [Fabaceae]) (USDA NRCS 2018) is distributed in isolated areas throughout the semi-arid regions of Oregon and Washington (Bhattarai and others 2010). It is highly palatable to livestock and wildlife with high protein and tannin contents, attracts generalist and specialist pollinators, fixes nitrogen in coordination with rhizobacteria, and has the potential for commercial seed production (Bhattarai and others 2010; Cane and others 2012). Dalea ornata also germinates in early spring but is much more difficult to find on the rangelands than A. filipes (Bhattarai and others 2010), possibly because of its high palatability. Two germplasm releases were made for D. ornata: ‘Majestic’ germplasm originated from collections made in the Deschutes River basin in north-central Oregon and ‘Spectrum’ germplasm originated from collections in southeastern Oregon (Johnson and others 2011).
One substantial challenge limiting the use of these native legumes in revegetation projects is the ability to obtain sufficient germination and seedling establishment on range sites. In controlled conditions with adequate water and sandy soil, A. filipes was found to germinate slowly and erratically despite seed scarification (Bushman and others 2015), with improvements in germination observed after scarification and cold stratification (Jones and others 2016). Under similar conditions, D. ornata germinated quickly upon seed scarification alone (Bushman and others 2015). Note that the seedling establishment of these species under range conditions, which typically lack adequate water and soil bed preparation, is unknown.
In this study we tested the seedling emergence of A. filipes and D. ornata as a predictor of seedling establishment in field trials that mimicked the rangeland conditions of operational revegetation plantings. Seed of both species was planted at 3 Oregon locations that varied in climate and seedbed preparation. We tested seed scarification and fall vs spring planting seasons to determine what combination of treatments provided the highest level of seedling emergence. Our results can aid rangeland managers in revegetation efforts by providing preferred planting seasons and seed treatments that can increase native legume reseeding success.
MATERIALS AND METHODS
We used 2 species in this experiment: G0 seed of NBR-1 germplasm to represent A. filipes and G0 seed of Majestic and Spectrum germplasm releases to represent D. ornata. The proportion of total viable seed was 83.3% for NBR-1, 84% for Majestic, and 97% for Spectrum. Prior to sowing, seeds were either treated with 5N sulfuric acid scarification for 5 min or were a non-scarified control. Three study locations spanned north-central and western Oregon, including major areas in the range of distribution of these 2 species (Figure 1). Study locations were the Oregon State University Malheur Experiment Station near Ontario, Oregon (43.97861N, 117.02361W; soil Nyssa silt loam, coarse-silty, mixed, mesic xeric haplodurids); the Oregon State University Powell Butte Ag Research Center (COARC, 44.25028N, 120.94778W; soil Buckbert ashy sandy loam, fine-loamy, mixed, mesic vitritorrandic haploxerolls); and an uncultivated wildland location near the Clarno bridge crossing (Oregon Highway 218) of the John Day River. The Clarno location was initially (2011–2012 season) planted adjacent to the John Day River at the USFS BLM Clarno Hardwoods Facility, but thereafter moved 2.5 km (1.55 mi) to the west and uphill (44.25583N, 120.50417W; soil Simas cobbly silty clay loam, fine, smectitic, mesic aridic palexerolls) to avoid possible spring flooding. Soil series were obtained from the Biochar Soil Data Explorer (http://www.pnwbiochar.org). A location map was generated in ArcGIS (ESRI, Redlands, California).
Locations of study in Oregon, USA.
At each location the 6 entry–treatment combinations were planted in randomized complete block designs of 2 m (6.5 ft) long rows with 240 seeds per row, and with 4 replicated blocks of rows (thus 24 rows per seeding). Seeds were planted with a Wintersteiger (Williston, Vermont) single-row seeder at a 1.3 cm (0.5 in) depth. The seedbed at the Ontario location was tilled prior to each planting, the seedbed at the Powell Butte location had been tilled only prior to the first fall planting, and the seedbed at the Clarno site did not have any tilling or other seedbed preparation. No supplemental irrigation, fertilization, or herbicide was applied to any of the locations. Plantings were conducted each October and April over 3 seasons (2011–2012, 2012–2013, and 2014–2015). For the intended 2013–2014 planting season, insufficient seed was available of D. ornata, so plantings were delayed until fall 2014. Seedling emergence data were collected each June 1, with an exception of the Clarno in 2013 when counts were made later in the season due to access issues. Because of low precipitation at Ontario in the 2012–2013 season, no seedlings emerged, so the data from that year/location combination were removed from the analysis. Monthly precipitation was obtained from the AgriMet website (https://www.usbr.gov/pn/agrimet; verified 5 February 2018) (Figure 2). For the Powell Butte and Ontario locations, the weather stations listed on AgriMet were on the respective farms; but for Clarno site the nearest weather station was at Madras, Oregon, approximately 56 km (35 mi) southwest of the Clarno location.
The sum of precipitation by month over 3 y during which A. filipes and D. ornata seedling emergence data was collected. *The precipitation for the Clarno location was extracted from a weather station approximately 56 km (35 mi) southwest.
We analyzed data using the MIXED procedure of SAS (SAS Institute 2018), with years, locations, and replications as random variables and seed treatments, planting dates, and species as fixed variables. Best linear unbiased predictions (BLUPs) were calculated for years and locations using the SOLUTION option in SAS. Least square means were calculated using SAS and differences with multiple testing were determined using Tukey-Kramer tests at the P ≤ 0.05 level.
RESULTS AND DISCUSSION
Best linear unbiased predictions of the 3 locations showed that the most seedlings emerged generally at the Powell Butte location, with more than 6 seedlings emerging in a row of 240 seeds as a prediction. Both the Clarno and Ontario locations showed negative BLUP values not different than zero (Figure 3A), highlighting wide variation for seedling emergence among replications at these locations, with most plots showing no emergence regardless of seed treatment, planting date, or species. In addition, the lack of seedbed preparation at the Clarno location prior to planting resulted in uneven planting on hard surfaces similar to what might be expected in no-till rangeland plantings. Because of the low precipitation and planting challenges, seedling emergence values were mostly at or near zero at Clarno.
Best Linear Predictions (BLUPs) of seedling emergence of A. filipes and D. ornata at 3 locations (A) and over 3 years (B). Columns with different letters differ according to least significant difference (LSD) multiple mean separations at P ≤ 0.05.
The experiment spanned 3 fall–spring seasons. The largest number of seedlings emerged after the 2011–2012 fall–spring season, with a BLUP value greater than zero (Figure 3B). The emergence BLUPs for each year were similar to the spring precipitation levels of each year. The 2012 spring season received the most precipitation of the 3 y (Figure 2). In 2013, all 3 sites received less than 30 mm (1.2 in) of precipitation from March–May. In the 2015 spring season, the 3 sites received between 40 mm (1.6 in) and 60 mm (2.4 in) of precipitation from March–May (Figure 2). The correlation between spring precipitation and the total number of seedlings that emerged each year was high but not significant due to the small number of data points; r = 0.93 (P = 0.24).
For each year and at each location, the 2 species were planted in both October (fall seeding) and April (spring seeding), with emergence data collected each June 1. Seeds were also treated with and without scarification. Of the main effects (species, planting date, and seed treatment), statistical significance was detected only for the seed treatment main effect (P = 0.005), indicating that scarification was beneficial for germination and seedling emergence of both species under all treatments and conditions. For D. ornata in particular, the Majestic and Spectrum sources were not different and negligible emergence was detected without scarification (Figure 4). The species × planting date interaction (P ≤ 0.001) showed that while D. ornata scarified seeds germinated and emerged more when planted in spring, A. filipes showed a trend for more seedling emergence when planted in fall (Figure 4). Almost no seedling emergence occurred in fall-planted D. ornata seeds regardless of seed treatment whereas approximately 4 seedlings emerged on average in the spring-planted A. filipes plots (Figure 4).
Seedling emergence means and differences given planting date and scarification treatments for A. filipes and D. ornata.
The three-way species × planting date × seed treatment interaction (P = 0.007) was primarily a result of D. ornata, which as mentioned above showed no emergence from fall plantings regardless of other treatments and only negligible emergence upon spring plantings without seed scarification (Figure 4). Thus D. ornata seeds had the most seedling emergence upon seed scarification and spring plantings. These seedling emergence data are consistent with previous greenhouse studies, which showed > 90% germination of scarified D. ornata seed with sufficient moisture (Bushman and others 2015). In this study, we found that seed scarification in combination with early spring planting was critical for D. ornata seedling emergence, regardless of location or year. For A. filipes, previous studies found a need for both seed scarification and cold temperature stratification for germination, with a slow but constant germination rate (Jones and others 2016). Data in this study confirm those earlier controlled-environment studies, wherein a combination of fall-planted, scarified seed most likely allowed for cold stratification over winter and produced the highest trend of seedling emergence in years and locations with sufficient spring moisture. Additionally, the trend but lack of significance of treatments or conditions on A. filipes seedling emergence further confirms the observation of slower germination and lower emergence in this species.
CONCLUSIONS
Although seedling emergence varied significantly across locations and years, with dry spring seasons in particular minimizing emergence success, patterns emerged for each species. Dalea ornata seedlings emerged at the highest rate with seed scarification and spring planting. Astragalus filipes seedlings emerged at the highest rate with seed scarification and fall planting. This information can assist seed growers and rangeland revegetation managers in planting and seed preparation decisions.
ACKNOWLEDGMENTS
We acknowledge Chris Jensen, Marlo Fisher, and Ramsey Buffam for assistance in planting and data collection. We also acknowledge the Central Oregon Agriculture Research Center.
Footnotes
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.
