Native shrublands and their associated grasses and forbs have been disappearing from the Great Basin as a result of grazing practices, exotic weed invasions, altered fire regimes, climate change and other human impacts. Native forb seed is needed to restore these areas. The irrigation requirements for maximum seed production of four key native forb species (Eriogonum umbellatum, Lomatium dissectum, Penstemon speciosus, and Sphaeralcea grossulariifolia) were studied at the Oregon State University Malheur Experiment Station beginning in 2005. Species plots were supplied with 0, 100, or 200 mm of subsurface drip irrigation per year using a randomized complete block design with four replications. Irrigation in each plot was divided into four equal increments applied between bud and seed set with timing dependent upon the flowering and seed set phenology of each species. Seed was harvested in each year of production through 2011, and the optimal irrigation rate was determined by regression. The four native forb species differed in their responses to irrigation. Lomatium dissectum seed yields were optimized with 140 mm of irrigation. Eriogonum umbellatum seed yields were optimized with 173 to 200 mm of irrigation in dry years and progressively less to no irrigation in the wettest year. Penstemon speciosus seed yields were optimized with 107 mm of irrigation in dry years and were reduced by irrigation in wet years. Sphaeralcea grossulariifolia seed yields did not respond to irrigation. Water requirements of these species are low, and these results can be used by seed growers to produce native forb seed more economically.
INTRODUCTION
Native shrublands and grasslands have been disappearing from the western United States (D'Antonio and Vitousek 1992). The transformation of perennial grassland communities into communities dominated by invasive annual species is evident throughout much of California (D'Antonio and Vitousek 1992; D'Antonio et al. 2000; Seabloom et al. 2003; Germano et al. 2012) and similar transformations are ongoing in shrublands of the arid and semi-arid Intermountain West (D'Antonio and Vitousek 1992; Millennium Ecosystem Assessment 2005; Reynolds et al. 2007; Zald 2008). The loss of native shrublands and grasslands is a worldwide phenomenon (D'Antonio and Vitousek 1992; Seabloom et al. 2003; Millennium Ecosystem Assessment 2005).
Sagebrush-steppe and associated plant communities of the Great Basin are being replaced by invasive exotic grasses, broadleaf weeds, and the expansion of native junipers (Azuma et al. 2005; Miller et al. 2005; Wells 2006; Shock et al. 2007, 2011). Grazing practices have contributed to the spread of exotic annual grasses, which have, in turn, led to catastrophic fires (Miles and Karl 1995; Pellant 1996; Davies et al. 2009, 2010). This transformation in vegetation composition influences other species in the biome due to plant-based food webs.
Restoration of sagebrush steppe vegetation in the Intermountain West is difficult and requires seeding or transplanting native species (James et al. 2013) where native seedbanks have been lost. Commercial seed production is necessary to provide the quantity of grass and forb seed needed for restoration efforts. Most major grass species used for revegetation have long been produced under agricultural conditions, but the forb seed production industry for this region is in its infancy. Major limitations to economically viable commercial production of native forb seed are stable and consistent seed production and reliable seed marketing opportunities. The irrigation of forbs planted for seed might make seed production more consistent. In natural rangelands, wide variation in winter and spring precipitation and soil moisture result in highly unpredictable water stress at flowering, seed set, and seed development, which for other seed crops is known to reduce seed yield and quality.
Both sprinkler and furrow irrigation can provide supplemental water for seed production, but these irrigation systems also encourage weeds by wetting the soil surface. In addition, sprinkler and furrow irrigation can lead to the loss of plant stands and seed production by enhancing fungal pathogens, an extreme economic disadvantage when growing perennial forbs. The trials described here used subsurface drip irrigation. Our goal was to assure flowering and seed set without undue encouragement of weeds or opportunistic diseases by burying drip tapes at 0.3-m depth to avoid wetting the soil surface.
MATERIALS AND METHODS
Plant Establishment
Irrigation trials for Eriogonum umbellatum Torr., Lomatium dissectum (Nutt.) Mathias & Constance, Penstemon speciosus Douglas ex Lindi., and Sphaeralcea grossulariifolia (Hook. & Arn.) Rydb. (USDA NRCS 2014) were initiated in 2005 and 2006 at the Malheur Experiment Station, Oregon State University, Ontario, Oregon (Table 1). Seed from wildland collections was provided by the US Forest Service.
The research field, a Nyssa silt loam (coarse-silty, mixed, mesic, Xeric Haplodurid), was leveled more than 70 years ago, and most of the topsoil was removed at that time to fill a natural ravine within the field. Analysis of a soil sample taken on 22 Nov. 2005 indicated a pH of 8.3, 1.09% organic matter, 12 mg·kg-1 P2O5, 438 mg·kg-1 K, 27 mg·kg-1 SO4-S, 4370 mg·kg-1 Ca, 456 mg·kg-1 Mg, 81 mg·kg-1 Na, 1.6 mg·kg-1 Zn, 0.6 mg·kg-1 Cu, 4 mg·kg1 Mn, 3 mg·kg-1 Fe, and 0.6 mg·kg-1 B.
The field was bedded to plant the forbs in rows 76 cm apart. Drip tapes (T-Tape TSX 515-16-340) were buried at 30-cm depth and spaced 1.52 m apart beneath alternating inter-row spaces. Emitters were spaced 41 cm apart, and the flow rate for the drip tape was 4.16 L·min-1100 m-1 at 55 kPa, resulting in a water application rate of 1.7 mm hr-1. Water was filtered through sand media filters, and application duration was controlled automatically.
Seed of E. umbellatum, L. dissectum, and P. speciosus was planted on 3 March 2005 using a custom-made small-plot grain drill with disk openers at 1.25-cm depth seeding 65–100 seeds·m-1 of row. Due to poor establishment from spring planting, seed was replanted from the same seed lots at 65 seeds m-1 using the same planter on 26 October 2005 so that natural, winter vernalization could occur. Excellent stands of all species were obtained in spring 2006. Sphaeralcea grossulariifolia was planted 11 April 2006 and was replanted 10 November 2006.
In April 2006, the areas planted to each species were divided into 12 plots, each 9 m long. Each plot contained four rows 0.76 m apart. The experimental design for each species was a randomized complete block with four replicates. The three irrigation treatments were 0 mm·year-1(nonirrigated check), 100 mm·year-1, and 200 mm·year-1. The 100-mm and 200-mm irrigation treatments received four irrigations, applied approximately two weeks apart, starting individually for each species at flowering. The dates of the start, peak, and end of flowering for each species were recorded and are reported in conjunction with the dates of the first and last irrigation in Table 2. Each irrigation delivered 25 mm ( 100-mm treatment) or 50 mm (200-mm treatment) through the drip system. The amount of water applied was measured by a water meter and recorded. In 2007, irrigation treatments were inadvertently continued after the fourth irrigation, providing additional water after seed set in proportion to the irrigation treatments.
Soil volumetric water content was measured several times each growing season by neutron probe. The neutron probe was calibrated by taking soil samples and probe readings at 0.20-, 0.50-, and 0.80-m depths during installation of the access tubes. The soil water content was determined volumetrically from soil samples taken from each depth and regressed against the neutron probe readings.
Fertilization of the irrigation trials over the six years was minimal. On 27 Oct. 2006, 56 kg·ha-1 phosphorus and 2.2 kg·ha-1 zinc were injected through the drip tape. On 11 Nov. 2006, 112 kg·ha-1 nitrogen as urea was broadcast. On 9 Apr. 2009, 56 kg·ha-1 N and 11 kg·ha-1 P were applied through the drip irrigation system. On 3 May 2011, 56 kg·ha-1 N was applied through the drip irrigation system. Natural precipitation varied considerably from 2006 through 2011 (Table 3, Figure 1).
During the first two years (2005 and 2006), weeds were controlled primarily with cultivation and hand rouging. Herbicides were screened for their effectiveness and plant tolerance in other trials (Shock et al. 2010). Even though they are not yet registered for use, we broadcasted Prowl® (pendimethalin) at 1.1 kg ai·ha-1 on the soil surface for weed control on 17 Nov. 2006, 9 Nov. 2007, 15 Apr. 2008, 18 Mar. 2009, 4 Dec. 2009, and 17 Nov. 2010. Volunteer® (clethodim) was broadcast at 0.57 L·ha-1 on 18 Mar. 2009. Hand rouging of weeds continued throughout the study.
Table 1.
Forb species planted in drip irrigation trials at the Malheur Experiment Station, Oregon State University, Ontario, OR.
Table 2.
Native forb flowering, irrigation, and seed harvest dates by species in 2006–2011, Malheur Experiment Station, Oregon State University, Ontario, OR.
Table 3.
Annual precipitation (2006–2011) and the long term average (66 years) at the Malheur Experiment Station, Oregon State University, Ontario, OR.
Seed production of P. speciosus was susceptible to losses from Lygus Hahn, (lygus bugs). Penstemon speciosus was sprayed with Aza-Direct® at 0.006.9 g ai·ha-1 on 14 and 29 May 2007 and Capture® 2EC at 112 g ai·ha-1 on 20 May 2008 for lygus control.
Flowering, Harvesting, and Seed Cleaning
Floral phenology was monitored annually for each species (Table 2). Seed harvests were timed to coincide with seed maturation, which varied among species and years (Table 2). Seed from the middle 7.5 m of the two center rows in each plot was harvested and cleaned. Seed harvesting and cleaning methods differed by species (Table 4). Seed of E. umbellatum, P. speciosus, and S. grossulariifolia was harvested by combine. The seed of E. umbellatum and S. grossulariifolia required extra threshing. Seed of L. dissectum was harvested by hand and did not require additional threshing. Seed was cleaned prior to yield measurements; seed quality was not determined. Seed yield means were compared by protected analysis of variance, LSDs were calculated, and nonlinear regression of seed yield was calculated against applied water.
Figure 1.
Cumulative annual and 6-year average precipitation from January through July at the Malheur Experiment Station, Oregon State University, Ontario, OR.
Table 4.
Native forb seed harvest and cleaning by species, Malheur Experiment Station, Oregon State University, Ontario, OR.
RESULTS
Soil Volumetric Water Content
Soil volumetric water content responded to the irrigation treatments for each species and varied each year with precipitation and winter snow pack (data not shown).
Flowering and Seed Set
Eriogonum umbellatum and P. speciosus produced seed in 2006, in part from plants that had emerged in the spring of 2005. There were few S. grossulariifolia plants in 2006, but this species flowered and set seed in 2007 after the November 2006 replanting. Lomatium dissectum did not flower and produce seed in quantity until 2009.
Penstemon speciosus had poor seed set in 2007, partly due to a heavy lygus bug infestation that was not adequately controlled by the applied insecticides. In the Treasure Valley, the first hatch of lygus bugs occurs when 250 degree-days calculated at 52 °F base are accumulated. Data collected by an AgriMet weather station adjacent to the field indicated that the first lygus bug hatch would have occurred on 14 May 2006, 1 May 2007, 18 May 2008, 19 May 2009, and 29 May 2010. The average lygus bug hatch date for 1995–2010 was 18 May. Penstemon speciosus begins flowering in early to mid-May. The earlier lygus bug hatch in 2007 probably resulted in harmful levels of lygus bugs being present during a larger part of the P. speciosus flowering period than normal. Lower seed set for P. speciosus in 2007 was also related to poor vegetative growth compared to 2006 and 2008. In 2009, all plots of P. speciosus again showed poor vegetative growth and seed set. Root rot affected the wetter plots of P. speciosus in 2009, but the stand partially recovered due to natural reseeding. Reseeding of P. speciosus resulted in its long term productivity, but also increased the generations of seed increase between wild collection of seed and seed use in range restoration.
Sphaeralcea grossulariifolia exhibited prolonged flowering in 2007, its first seed production year (early May through September), possibly due to the extra irrigation water that was applied. Multiple manual harvests were necessary in 2007 because the seeds were rapidly dispersed from the capsules once they matured. Following harvest the flailing of the S. grossulariifolia vegetation was initiated in the fall of 2007 and thereafter was repeated annually to induce a more concentrated flowering period for a single mechanical harvest. Precipitation in June of 2009 and 2010 was substantially higher than average. Rust (Puccinia sherardiana Korn.) infected S. grossulariifolia in June of 2009 and 2010, causing substantial leaf loss and reduced vegetative growth.
Seed Yields
Eriogonum umbellatum
In 2006, seed yield of E. umbellatum increased with increasing water application up to 200 mm, the highest amount tested (Table 5, Figure 2). In 2007–2009, seed yield showed a quadratic response to irrigation rate. Based on the quadratic equations, seed yields were maximized by 202 mm of water applied in 2007, 181 mm in 2008 and 173 mm in 2009. In 2010, there was no significant difference in seed yield among the irrigation treatments. In 2011, seed yield was highest with no irrigation. The 2010 and 2011 seasons were unusually cool and wet (Table 3, Figure. 1). Accumulated precipitation during April through June of these years was among the highest over the years of the trial (Table 3). The relatively high seed yield of E. umbellatum in the nonirrigated treatment in 2010 and 2011 seemed to be related to the high January to March precipitation followed by high April to June precipitation. Irrigation in these two years might also have exacerbated the rust infections observed in the 100- and 200-mm treatments and contributed to the lower yields. Averaged over six years, seed yield of E. umbellatum increased with increasing water applied up to 200 mm, the highest amount tested (Figure 2). The shape of the quadratic seed yield responses observed in most years suggested that additional irrigation above 200 mm would not be beneficial. Eriogonum umbellatum responded to 173 to 200 mm of irrigation in the first year and in years with less than 300 mm of precipitation from January through June. In wetter years seed yield responses were negative.
Penstemon speciosus
From 2006 to 2009, the seed yield of P. speciosus showed a quadratic response to total irrigation (Figure 3, Table 5). Seed yields were maximized by 106–108 mm of applied water in 2006, 2007, 2008, and 2009, years when there was less than 90 mm of precipitation from May through June. Seed yields did not increase with irrigation in years when rainfall from May through June was above 90 mm. In 2010 and 2011, seed yield did not differ significantly among irrigation treatments. Seed yield was low in 2007, probably due to lygus bug damage, as mentioned above. Seed yield in 2009 was low due to stand loss from root rot. The plant stand recovered somewhat in 2010 and 2011, due largely to natural reseeding, especially in the nonirrigated plots. The plant lifespan of P. speciosus was around three years in our agronomic setting, suggesting that long-term seed yields will depend upon replanting to the same or different fields.
Lomatium dissectum
Lomatium dissectum exhibited very slow vegetative growth in 2006–2008, and produced very few flowers in 2008. The species was affected by Alternaria sp. Nees. fungus, and this infection might have delayed L. dissectum plant development and reduced its seed yield. Vegetative growth and flowering for L. dissectum were greater in 2009. Seed yield exhibited a linear response to irrigation in 2009 (Table 5; Figure 4). Seed yield for the 100-mm irrigation treatment was significantly higher than for the nonirrigated check, but the 200-mm irrigation rate did not result in a significant increase above the 100-mm rate. In 2010 and 2011, seed yields of L. dissectum showed a quadratic response to irrigation rate and were maximized by 161 mm of applied water in 2010 and 127 mm in 2011. Averaged over the three years, seed yield showed a quadratic response to irrigation rate and was estimated to be maximized by 140 mm of applied water (Shock et al. 2012).
Sphaeralcea grossulariifolia
In 2007–2011, seed yield did not vary significantly among irrigation treatments for S. grossulariifolia (Table 5). This forb was well adapted to produce moderate seed yield regardless of the natural variations of rainfall or irrigation at Ontario, Oregon.
DISCUSSION
The four forb species reported in our trials differed greatly in their seed yield response to irrigation, varying from L. dissectum, an early season flowering species, which responded to irrigation in every year, to S. grossulariifolia, which failed to respond to irrigation in any year, regardless of natural precipitation. Eriogonum umbellatum and P. speciosus responded to irrigation in dry years and displayed little or no seed yield response in wet years.
Table 5.
Native forb seed yield response to irrigation rate (mm/season) 2006–2011 at the Malheur Experiment Station, Oregon State University, Ontario, OR.
This range of responses is consistent with the findings of other authors for seed yields of other plants native to arid and semiarid lands. At the Owens Lake playa in California, Sarcobatus vermiculatus (Hook.) Torr, (greasewood) seed yield was enhanced by supplemental irrigation (Breen and Richards 2008). Pol et al. (2010) demonstrated that a wide range of perennial C3 and C4 grasses responded to rainfall events with increased seed productivity in the arid grasslands of Mendoza Province, Argentina. In contrast, Fisher et al. (1988) found that irrigation of Larrea tridentata (DC.) Coville (creosote bush) was detrimental to fruit production in the northern Chihuahuan Desert rangelands of New Mexico. In west Texas, Petersen and Ueckert (2005) found that seed yields of Atriplex canescens (Pursh) Nutt, (fourwing saltbush) failed to respond to irrigation. To date, literature on irrigation impacts on seed production of forbs native to the Intermountain Region remains extremely limited (Shock et al. 2012).
Figure 2.
Average and annual Eriogonum umbellatum seed yield response to irrigation water applied in each of 6 years, Malheur Experiment Station, Oregon State University, Ontario, OR. Regression equations: 2006, Y = 154.5 + 1.244X, R2 = 0.68, P = 0.02; 2007, Y = 89.12 + 1.268X − 0.003143X2, R2 = 0.69, P = 0.02; 2008, Y = 135.9 + 1.550X − 0.004283X2, R2 = 0.73, P = 0.005; 2009, Y = 148.2 + 1.428X − 0.004124X2, R2 = 0.60, P = 0.02; 2010, Y = 283.2 + 0.4127X − 0.003299X2, R2 = 0.08, P = NS; 2011, Y = 260.7 − 0.7252X, R2 = 0.58, P = 0.004; 6-year average, Y = 194.8 + 0.2880X, R2 = 0.48, P = 0.01.
All of the seed yield responses reported in our trials were for forbs grown without competition with weeds or associated species. In a natural setting, these forbs would compete for resources with other sagebrush steppe plants in their immediate vicinity. As a result, the supplemental water required for high seed yield for these species may be substantially different in natural settings. In addition, plants were grown on soil with modest nutrient content, with minimal fertilization in 2006, 2009, and 2011. These supplements should have contributed to plant health. Native forbs were susceptible to disease and insect pests. Seed yields can be reduced by pests and diseases and labeling of fungicides, insecticides, and herbicides for these seed crops would be beneficial for commercial seed production.
Figure 3.
Annual and 6-year average Penstemon speciosus seed yield response to irrigation water, Malheur Experiment Station, Oregon State University, Ontario, OR. Regression equations: 2006, Y = 183.1 + 3.812X − 0.01767X2, R2 = 0.66, P = 0.01; 2007, Y = 2.797 + 0.1444X − 0.0006839X2, R2 = 0.48, P = 0.05; 2008, Y = 105.3 + 5.092X − 0.02035X2, R2 = 0.56, P = 0.04; 2009, Y = 7.628 + 0.1962X0 − 0.0009186X2, R2 = 0.54, P = 0.03; 2010, Y = 164.9 − 1.199X + 0.003826X2, R2 = 0.35, P = 0.13 (NS); 2011, Y = 403.8 − 0.1261X, R2 = 0.01, P = NS; 6-year average, Y = 161.5 + 0.1218X; R2 = 0.46, P = 0.06.
In the present trials, forb plant stands from spring plantings were generally poor while those from fall plantings were considerably better, an observation which may be useful both for the establishment of commercial seed production fields and rangeland restoration projects. Seed production for range restoration should include documentation of the original seed source and tests of seed viability. Restoration efforts may have greater chances of success if the percent of viable seed is known to help guide planting densities.
CONCLUSIONS
Subsurface drip irrigation systems were tested for native seed production because they have two potential strategic advantages; (1) efficient, precise, and uniform water application, and (2) the buried drip tape provides water to the plants at depth, precluding stimulation of weed seed germination on the soil surface and keeping water away from native plant tissues, which typically are not adapted to a wet environment.
Due to the semi-arid environment at the seed production location, supplemental irrigation may often be required for successful flowering and seed set of some forbs because soil water reserves may be exhausted before seed formation. The total irrigation requirements for these arid-land species were low and varied by species (Table 6). Lomatium dissectum required approximately 140 mm of irrigation for optimal seed yield. Penstemon speciosus and E. umbellatum responded quadratically to irrigation in dry years with the optimum of 173 to 200 mm and 106 to 108 mm, respectively, but irrigation in wet years provided no advantage and could be detrimental. Seed production of S. grossulariifolia did not respond to irrigation in our trials, suggesting that natural rainfall was sufficient to maximize seed production in the absence of weed competition with this species. Such variation in seed yield response among species may be expected because native forbs of the Great Basin exhibit a wide array of plant traits (e.g., growth habit, root structure, and phenology).
Figure 4.
Annual and 3-year average Lomatium dissectum seed yield response to irrigation water, Malheur Experiment Station, Oregon State University, Ontario, OR. Regression equations: 2009, Y = 80.32 + 1.552X, R2 = 0.30, P = 0.07; 2010, Y = 297.7 + 3.441X − 0.01066X2, R2 = 0.51, P = 0.04; 2011, Y = 635.6 + 14.31X − 0.05623X2, R2 = 0.86, P = 0.0001; 3-year average, Y = 330.0 + 6.905X − 0.02466X2, R2 = 0.72, P = 0.003.
Seed of native forbs is required to restore diverse native communities in the Great Basin. However, growers are reluctant to attempt production of new species without knowledge of the necessary cultural practices and potential seed yields. Delayed seed production, differences in plant sizes and growth requirements, pollinator needs, and technology and equipment required to produce each species are among the multiple challenges encountered when bringing these species into agronomic production. Elucidating irrigation requirements to maximize seed production of individual native forbs provides growers with key data required to select and economically produce seed. Increased use of native forbs will necessitate additional research on basic biology and agronomic requirements of individual species.
Table 6.
Amount of irrigation water for maximum native forb seed yield, years to seed set, and life span under the agronomic conditions at Ontario. A summary of multiyear research findings, Malheur Experiment Station, Oregon State University, Ontario, OR.
ACKNOWLEDGMENTS
The authors are thankful for support of several grants from the BLM and US Forest Service through the Great Basin Native Plant Project and support of Oregon State University, the Malheur County extension service district, and Hatch funds.
