INTRODUCTION

Smooth brome (Bromus inermis Leyss.) is an introduced grass that has invaded grasslands throughout North America (Otfinowski et al. 2007; Bahm et al. 2011a, 2011b; Bolwahn Salesman and Thomsen 2011; Slopek and Lamb 2017). Smooth brome was intentionally planted for forage production or soil stabilization but rapidly spread from agricultural fields and disturbed roadsides into grasslands throughout the Northern Great Plains (Otfinowski et al. 2007; Bahm et al. 2011b; Bolwahn Salesman and Thomsen 2011) where it threatens the structure and function of native grassland ecosystems (Murphy and Grant 2005; Vinton and Goergen 2006; Otfinowski et al. 2007; Fink and Wilson 2011; Piper et al. 2015). Smooth brome exhibits prolific early season growth and alters soil conditions thereby fostering its competitive advantage over native species, which can eventually result in the formation of smooth brome monocultures (Vinton and Goergen 2006; Otfinowski et al. 2007).

Early studies determined that smooth brome's seed and forage yields increased with nitrogen applications (Harrison and Crawford 1941), which has been confirmed by other researchers (Levang-Brilz and Biondini 2002; Vinton and Goergen 2006; DiAllesandro et al. 2013). Harrison and Crawford (1941) observed that nitrogen supplementation increased production of sterile tillers and suggested that increased nitrogen allowed smooth brome to be more competitive because it produces more and larger tillers in response. Increased smooth brome canopy (Vinton and Goergen 2006) allows individual plants to produce more carbohydrates through photosynthesis, the excess of which can be stored for the next season (Harrison and Crawford 1941). Additionally, rapid decomposition of smooth brome litter allows for further supplementation of nitrogen via nutrient cycling (Vinton and Goergen 2006), resulting in favorable conditions for smooth brome.

Land managers throughout the Northern Great Plains have attempted to adapt management strategies to target smooth brome populations and have achieved varying degrees of success (Grant et al. 2009; Bolwahn Salesman and Thomsen 2011; Kobiela et al. 2017; Slopek and Lamb 2017). Managers have employed herbicide treatments (Otfinowski et al. 2007; Bahm et al. 2011a, 2011b; Bolwahn Salesman and Thomsen 2011; Slopek and Lamb 2017), grazing (Willson and Stubbendieck 1997, 2000; Brueland et al. 2003), mowing (Otfinowski et al. 2007; Bolwahn Salesman and Thomsen 2011), and prescribed fire (Willson and Stubbendieck 1997; Bahm et al. 2011a; Bolwahn Salesman and Thomsen 2011; Kobiela et al. 2017) to reduce smooth brome's prevalence in North American grasslands.

The United States Fish and Wildlife Service (USFWS) is tasked with managing >100,000 ha of native grasslands in the Northern Great Plains region (Grant et al. 2009). The USFWS typically relies on prescribed fire and grazing to control smooth brome, in accordance with the Native Prairie Adaptive Management (NPAM) program (Gannon et al. 2013). The NPAM program, implemented on certain USFWS-owned native prairie lands in Montana, North Dakota, South Dakota, and Minnesota, utilizes a decision support tool to guide management actions to improve ecological conditions given personnel and financial constraints (Gannon et al. 2013). The NPAM program incorporates monitoring data obtained from field surveys conducted by USFWS personnel to determine an appropriate course of action for an enrolled tract of land.

In addition, USFWS personnel utilize Willson and Stubbendieck's (2000) provisional model for burning smooth brome to guide decisions regarding the timing of prescribed burning. Willson and Stubbendieck (2000) determined their provisional model is appropriate when a significant warm-season grass component is present to aid in the competitive exclusion of smooth brome. For areas with at least 20% composition of native warm-season grasses, Willson and Stubbendieck (2000) recommend prescribed burning once 50% of a smooth brome population has reached elongation and rely on the five-leaf stage as an easily identifiable cue marking elongation. The model further recommends that burning be completed prior to the maturation of inflorescence. Prescribed burning during these developmental stages is expected to have the most detrimental impacts to smooth brome, thereby depleting the plants' carbon reserves and giving a competitive advantage to the other species present (Willson and Stubbendieck 2000).

Practical use of Willson and Stubbendieck's (2000) provisional model has been complicated in the Northern Great Plains because USFWS personnel rarely observed the five-leaf stage in situ. Instead, smooth brome populations were observed to begin elongation despite never achieving the five-leaf stage (S. Vacek, wildlife biologist, USFWS, pers. comm.). To address this issue, Preister et al. (2019) developed a method to determine elongation of smooth brome using accumulated growing degree days (AGDD) instead of the five-leaf stage to properly time prescribed burning to target smooth brome populations in the Northern Great Plains. This project observed smooth brome phenological development in a controlled setting to determine whether the relationship between developmental stage (relying on Moore et al. [1991] mean stage counts) and AGDD observed in the field was similar under greenhouse conditions. We manipulated nitrogen levels to examine whether nitrogen supplementation affects the pattern of phenological development or biomass production of smooth brome, similar to results reported in field plots (Vinton and Goergen 2006; DiAllesandro et al. 2013) and prior greenhouse studies (Levang-Brilz and Biondini 2002).

METHODS

This experiment was conducted in three trials. For each trial, 400 7.25-cm round unglazed clay pots were filled with commercial potting mix and a single smooth brome seedling, grown from an agricultural seed source. Pots were randomly assigned one of four nitrogen supplementation treatments: control (no nitrogen addition), low (28 kg/ha), medium (56 kg/ ha), or high (112 kg/ha) and supplemented with UFLEXX granular 46% stabilized nitrogen fertilizer (Koch Agronomic Services, Wichita, Kansas, USA). The nitrogen fertilizer was applied on day 1 and repeated approximately 21 d later to maintain sufficient nitrogen levels.

Greenhouse conditions were monitored and controlled by computer with the temperature held between 21 °C and 23 °C. During periods of warmer weather, the greenhouse conditions were maintained at ambient temperature minus 10 degrees. Hourly temperature readings were automatically recorded. Greenhouse lights, set on computer-controlled timers, supplemented daylight from 0600 to 2200 daily. Plants were watered on a regular basis and monitored daily for growth and phenological development, according to Moore et al. (1991), as described in Table 1. In the event that rogue seeds may have been mixed in with the smooth brome seed, suspect plants were allowed to mature and removed from the study when determined to belong to a species other than smooth brome. These individuals were excluded from all calculations.

Table 1.

Phenological development stage, index, and description, adapted from Moore et al. (1991).

At the end of each trial, all tillers in each pot were clipped at the soil level, placed in a single bag and dried in an industrial dryer for 7 d. Dried samples were stored until 2 d prior to weighing, at which time they were returned to the dryer for an additional 48 h to remove moisture that may have been added while in storage. Clipped biomass was weighed immediately following the second drying period. Analysis of variance (ANOVA) and Tukey's honestly significant difference (HSD) test (SAS Enterprise Guide 7.1, SAS Institute, Cary, North Carolina, USA) were performed on biomass samples to identify differences in response to nitrogen supplementation.

Each treatment within a trial was treated as an independent sample population and the phenological stages were totaled and the mean stage count (MSC; Moore et al. 1991) was calculated for each trial. Growing degree days began accumulating on day 1 (upon seed planting). Accumulated growing degree days were also calculated for each trial using the minimum and maximum daily temperature and a base temperature of 0 °C (selected to correspond to the methods employed in the field study as reported in Preister et al. 2019). The equation used for AGDD calculations (Akyuz and Ransom 2015) is AGDD = Σ[(Maximum temperature + Minimum temperature)/2 – Base temperature] .

Figure 1.

Mean (±SE) biomass for each nitrogen treatment: control (no nitrogen addition), low (28 kg/ha), medium (56 kg/ha), or high (112 kg/ha). Different letters are used to indicate significant differences (P < 0.05).

Linear regression was used to further examine MSC and AGDD for each treatment and trial. Percent elongation was calculated for each nitrogen treatment and trial, identifying the AGDD at which the populations reached 50% elongation or higher. This AGDD and corresponding percent elongation was used, along with nearest previous percent elongation point, to extrapolate the value of AGDD at 50% for each treatment. Average AGDD for all treatments and trials were calculated and ANOVA and Tukey's HSD test (SAS Enterprise Guide 7.1) were performed, using nitrogen treatment as the independent variable and AGDD at 50% elongation as the dependent variable. Additionally, the percentage of observations of plants at the five-leaf stage was calculated to determine how often the greenhouse populations reached the five-leaf stage to compare to the observations obtained during the field studies reported by Preister et al. (2019).

RESULTS

ANOVA indicated nitrogen supplementation treatments affected biomass (P < 0.0001) and number of tillers per pot (P < 0.0001). While all nitrogen supplementation treatments increased biomass over the control (Figure 1), biomass was higher in the high nitrogen treatment than the low nitrogen treatment. Similarly, all nitrogen supplementation treatments increased the number of tillers over the control (Figure 2). However, while mean tiller count was higher in the medium nitrogen treatment than the low nitrogen treatment (Figure 2) there was no difference in tiller counts between the medium and high nitrogen treatments.

MSC was regressed against AGDD for each treatment within the three trials (Table 2) and the combined data for each trial (Figure 3). Each greenhouse trial showed a linear relationship between combined treatment MSC and AGDD but not as strong as the field survey described in Preister et al. (2019). There were no differences in the extrapolated AGDD at 50% elongation (Table 2) between treatments (ANOVA, P = 0.6464). The percentage of observed five-leaf stage plants was calculated for the greenhouse population, using both total observations and only those observations after elongation was initiated (Table 3).

Figure 2.

Mean (±SE) tiller count for each nitrogen treatment: control (no nitrogen addition), low (28 kg/ha), medium (56 kg/ha), or high (112 kg/ha). Different letters are used to indicate significant differences (P < 0.05).

DISCUSSION

Observed increases in smooth brome biomass and tiller counts following nitrogen supplementation are consistent with others' research (Harrison and Crawford 1941; Frank and Hofmann 1989; Levang-Brilz and Biondini 2002; Vinton and Goergen 2006; DiAllesandro et al. 2013). Frank and Hofmann (1989) stated that air temperature primarily controlled the rate of morphological development in Northern Great Plains grasses, while nutrients impacted the quantity of production. Levang-Brilz and Biondini (2002) determined that increased nitrogen resulted in increased aboveground biomass in smooth brome. Increased biomass results in more canopy to intercept solar radiation and produce more litter that rapidly decomposes, perpetuating a nitrogen-rich environment favorable for smooth brome invasion (Vinton and Goergen 2006).

Table 2.

Linear regression R² and Accumulated Growing Degree Days (AGDD) values when the greenhouse population reached 50% elongation (for trials 1–3) and nitrogen treatment (C, control [no nitrogen addition]; L, low [28 kg/ha]; M, medium [56 kg/ha]; or H, high [112 kg/ha]).

Figure 3.

Mean stage count (circles) versus accumulated growing degree days (linear regression R²) with the percentage of daily observations in elongation phase or higher (diamonds) for Trial 1 (a), Trial 2 (b), and Trial 3 (c).

Preister et al. (2019) determined the average AGDD on field sites for the population to reach 50% elongation was 1256 (standard deviation of 155). The greenhouse population reached 50% elongation at an average of 2287 AGDD (standard deviation of 243). The greenhouse population average does not fall within the 95% confidence interval of the field study model and the difference between the two averages is hypothesized to be the result of a number of factors. The greenhouse populations were planted from seeds with all pots in each trial at approximately the same chronological age. In a field setting, perennial plants grow from existing plants that potentially have very different chronological ages or from seeds that were produced during a number of seasons. This variation in growth could be reproduced in the greenhouse if the study was extended over a number of growing seasons but because our study was completed in one season, we were unable to reproduce perennial growth variation.

Table 3.

Observed five-leaf tillers as percent of total observations and percent of observations following the initiation of elongation (for trials 1–3) and nitrogen treatment (C, control [no nitrogen addition]; L, low [28 kg/ha]; M, medium [56 kg/ha]; or H, high [112 kg/ha]). A single asterisk (*) denotes the maximum and double asterisks (**) denote the minimum.

The greenhouse environment allowed us to control a number of factors that would not be as uniform in the natural prairie setting, such as length of daylight, temperature, and moisture. Greenhouse plants did not experience the more extreme temperature dips that may occur naturally or the variation in sunlight caused by a cloudy day. The greenhouse lights automatically turn on to supplement sunlight on cloudy days and create a uniform span of “daylight” hours, even when naturally occurring daylight hours would be increasing or decreasing, depending on season. Additionally, the populations planted in the greenhouse were seeded during the first three weeks of July, after field populations of smooth brome had already begun to flower and produce seed. Starting the seeds after the field populations had completed their initial growth could contribute to the variation in the average AGDD at 50% elongation.

Willson and Stubbendieck (2000) recommended using the five-leaf stage to determine elongation and population readiness for prescribed burning. Willson and Stubbendieck's (2000) provisional model recommends burning only when there is at least a 20% native plant component in order to allow those native plants to utilize the newly available resources following the burning of smooth brome, resulting in increased competition for smooth brome. Without enough competition from other species, smooth brome is expected to grow back, perhaps more vigorous than before, on the disturbed site.

Preister et al. (2019) field surveys observed minimal plants at five-leaf stage. Following the initial field survey season, we decided to monitor phenological development in the greenhouse to determine if the lack of five-leaf stage was an anomaly. The greenhouse population showed a higher incidence of five-leaf stage observations than the field survey (average observation of 2%, and less than 1%, respectively) but the total observation of five-leaf stage was quite low in both cases. Statistical significance was not calculated for the greenhouse treatments because all treatments had so few five-leaf stage tillers observed throughout the experiment. Smooth brome cultivars exhibited a variety of characteristics and a great deal of phenotypic plasticity (Otfinowski et al. 2006). Although unlikely, we cannot discount that lack of five-leaf stage observations could be attributed to the cultivars that have been used throughout this region and in this study.

Phenological variability in smooth brome was observed in field (Preister et al. 2019) and greenhouse populations. The pattern of linear correlation between AGDD and phenological development, as measured by MSC, exists in both instances and can be used to predict the approximate number of AGDD required for a population to reach elongation. Our greenhouse study confirmed smooth brome populations may not reach the five-leaf stage before initiating elongation. The contributing factors to this developmental progression are yet to be determined, but the results of our study showed that nitrogen supplementation treatments did not significantly affect the number of AGDD required for a smooth brome population to reach the elongation stage. As such, we recommend the use of elongation, as determined by the formation of palpable nodes, be used as a cue for smooth brome population control by prescribed burning. Palpable node identification is a simple and more consistent signal to the onset of elongation than the five-leaf stage.

CONCLUSIONS

Bromus inermis (smooth brome) is a perennial, invasive grass that responds to available nitrogen by increasing overall biomass and tiller production, thereby outcompeting many native species typically encountered in the grasslands of the Northern Great Plains. Managers have employed many strategies in an attempt to reduce smooth brome populations, including burning, grazing, mowing, and applying herbicides. The USFWS has relied on prescribed burning and grazing to manage grassland tracts and is especially concerned with targeting smooth brome populations to improve grassland plant community composition (Gannon et al. 2013). In order to reduce smooth brome populations, the USFWS employs prescribed burning of grassland tracts with significant native plant species components. The USFWS has timed prescribed burning to observations of smooth brome populations reaching the five-leaf developmental stage, in recognition of the point in time when smooth brome would be most likely to be harmed by prescribed burning (at the onset of elongation). However, identification of the five-leaf stage appears an unreliable indicator of smooth brome elongation. Our greenhouse experiment corroborates results of Priester et al. (2019), which concluded that many populations throughout the Northern Great Plains were observed to begin elongation without ever achieving the five-leaf stage.

A recent field study (Preister et al. 2019) suggested that relying on accumulated growing degree days (AGDD) provides a more reliable indicator of smooth brome's phenological development in the Northern Great Plains. In the current study, three trials were conducted in a greenhouse to evaluate whether the relationship between AGDD and smooth brome's phenological development holds across various nitrogen conditions intended to replicate field conditions encountered throughout the Northern Great Plains. It was determined that while nitrogen supplementation did increase biomass and tiller production, nitrogen supplementation did not affect the phenological progression of development. Thus, relying on AGDD to trigger management activities provides a reliable indication of smooth brome's phenological development and a cost-effective alternative to the reliance on labor-intensive field surveys.