Interseeded Native Forbs Resilient Under Variable Grazing Regimen

Abstract

Reduced floral resources and habitat fragmentation have led to pollinator decline. Increased diversity of native plants in pastures could support cattle and pollinators. However, the relationship between grazing and plant diversity needs to be investigated. We explored how grazing rest periods impacted persistence and forage characteristics of Andropogon gerardii (BB)/Sorghastrum nutans (IG; BBIG) and Panicum virgatum (SG) pastures interseeded with forbs and grazed over five years. ANOVA analysis was conducted using R with significance set at p ≤ 0.05. Forb species exhibited different establishment and flowering characteristics. Coreopsis tinctoria, Rudbeckia hirta (BESU), and Coreopsis lanceolata (LCOR) established early, while Helianthus maximiliani, Heliopsis helianthoides (OSUN), and Echinacea purpurea (PURC) established the second season. Rudbeckia hirta, LCOR, OSUN, and PURC flowered most frequently, and the grazing regimen did not influence the flowering frequency of any species. Desmodium tortuosum (TTFL) was one of the most selected by cattle. Total forage mass declined in 2022, but forb mass interacted with treatment and year where mass declined each year but varied among treatments annually. Based on persistence and forage characteristics, BESU, LCOR, OSUN, PURC, and TTFL could successfully provide forage in native pastures under a variety of grazing regimens.

1. Introduction

Pastures have the potential to support a diversity of animals, including both domestic livestock and insect pollinators. Native warm-season grasses (NWSGs; Poaceae) such as big bluestem (Andropogon gerardii Vitman; BB), indiangrass (Sorghastrum nutans (L.) Nash; IG), and switchgrass (Panicum virgatum L.; SG) are productive, C₄ grasses that can provide summer pasture for cattle (Bos taurus) [1,2,3,4], and offset forage deficits associated with cool-season grasses (e.g., tall fescue, Schedonorus arundinaceus (Schreb.) Dumort., nom. cons.) [5,6]. However, like other grass species, NWSGs by themselves provide few resources for pollinators. Interseeded forbs can add both nutrient-dense forage for grazers and floral resources for insect pollinators in a pasture system. Additionally, diversification of native pastures can prevent soil erosion, reduce water runoff, and increase soil biodiversity [7,8,9], thus improving grassland persistence and long-term resource provision for cattle and pollinators. This increase in pasture longevity can improve economic outcomes for cattle producers and stabilize critical habitats for declining pollinators that support nearby agricultural sectors.

Both native and domestic pollinator populations are in decline and, as a result, could negatively affect agriculture [10]. Between 1947 and 2005, the USA experienced a 59% loss in domestic honeybees despite a >300% increase in pollinator dependency in croplands [11]. Many pollinator species are impacted by agricultural practices, like herbicide and pesticide use and mono-crop production; however, habitat fragmentation and disease are also major factors that have contributed to population decline [11,12,13,14,15]. As is the case with grass-dominated pastures of introduced species, native grass pastures could also contribute to habitat fragmentation through diversity loss and homogenous habitat structure. However, NWSGs also present an opportunity to increase plant biodiversity to support both beef cattle and pollinator species through the introduction of native forbs in a working lands conservation framework. This framework seeks to add nutritional value and habitat structure to grasslands to increase sustainability for the benefit of wild and domestic fauna [16]. Richwine et al. [17] found that forbs interseeded into NWSG pastures persisted under different rest treatments over the first three grazing seasons (2018–2020) following establishment and did not appear to negatively impact animal production. Long-term grazing and seasonal environments can affect population dynamics and forage characteristics in diverse grasslands as various species establish, populate, or diminish. Therefore, to fulfill a knowledge gap in long-term persistence and establishment of forbs under grazing, we extended the study started by Richwine et al. [17]. We measured pasture characteristics for two more grazing seasons (2021 and 2022) under the same rest treatments to compare the changes over five years after establishment using the same site as Richwine et al. [17].

Grazing management (e.g., periodic rest) can have an impact on pasture persistence and botanical composition [6,18]. Additionally, in a working lands framework, the diversity and structure of pastures largely hinge on the grazer’s interaction with the flora [16]. Hence, it is important to quantify the persistence and forage characteristics of novel forages under different grazing approaches. Therefore, the objectives of this study are to determine the impact of grazing management (e.g., periodic rest, continuous grazing, and no grazing) on plant density, forb characteristics, forage mass, and forage nutritive composition of BBIG and SG pastures interseeded with native forbs and grazed by weaned beef steers for five grazing seasons.

2. Materials and Methods

2.1. Pasture Establishment

The research was conducted at the Northeast Tennessee AgResearch and Education Center (NETREC) near Greenville, TN (36.10963°, −82.86141°). Two previously established (2008) 1.2-ha pastures with either a mixture of big bluestem and indiangrass (BBIG) or lowland switchgrass (SG) were interseeded with an 11-species blend of forbs (FORB;

Table 1. Seeding rates (kg ha⁻¹) of native forbs interseeded into established BBIG ^† and SG stands and grazed by weaned Angus cross steers, May–August 2018–2022, Northeast Tennessee AgResearch and Education Center, Greeneville, TN.

Common Name	Species	Abbreviation	Seeding Rate (PLS kg ha−1)	SEEDS m−2 *
Black-eyed Susan	Rudbeckia hirta L.	BESU	0.56	194
Eastern purple coneflower	Echinacea purpurea (L.) Moench	PURC	0.7	18
Illinois bundleflower §	Desmanthus illinoensis (Michx.) MacMill.	ILBF	1.26	23
Lanceleaf coreopsis	Coreopsis lanceolata L.	LCOR	1.12	54
Maximilian sunflower	Helianthus maximiliani Schrad.	MSUN	0.56	308
Oxeye sunflower	Heliopsis helianthoides (L.) Sweet	OSUN	0.28	07
Partridge pea §	Chamaecrista fasciculata (Michx.) Greene	PPEA	0.56	08
Plains coreopsis	Coreopsis tinctoria Nutt.	PLAC	0.56	392
Purple prairie clover §	Dalea purpurea (Vent.) Rydb.	PUPC	0.56	37
Dixie ticktrefoil §	Desmodium tortuosum (Sw.) DC.	TTFL	0.56	17
Upright prairie coneflower	Ratibida columnifera (Nutt.) Woot & Standl.	UPPC	0.28	46

^† Big bluestem/indiangrass mixture, BBIG; switchgrass, SG; pure live seed, PLS. * Approximate number of seeds planted m⁻². ^§ Signifies legume species.

). Species were selected based on multiple factors, including adaptation to the eastern US, plant functional group (i.e., forb or legume), life history (i.e., annual, biennial, or perennial), seed cost and availability (Ernst Conservation Seed, Meadville, PA, USA), and existing recommendations from various conservation agencies. The species selected do not represent an exhaustive list of options; rather, selection criteria prioritized species diversity, seed price, and availability to test species most likely to be adopted by producers. The mixture was interseeded on 12 June 2017 at a depth of 1 cm with a 9-row Great Plains^® no-till drill (Great Plains Manufacturing, Inc., Salina, KS, USA) with 19 cm row spacing. Prior to interseeding, pastures were harvested, with biomass removed in the fall of 2016 and burned in early April 2017 to improve the vigor of established grasses and remove thatch that could interfere with the recruitment of forb seedlings. Pastures were sprayed with PastureGard^® HL [{triclopyr: 3,5,6-trichloro-2-pyridinyloxyacetic acid, butoxyethyl ester; 45.07%}; {fluroxypyr: (4-amino-3,5-dichloro-6-fluoropyridin-2-yl)oxy]acetic acid, 1-methylheptyl ester; 15.56%}] at 2.34 L ha⁻¹ and 2,4-D Amine 4 [dimethylamine salt of 2,4-Dichlorophenoxyacetic acid; 47.3%] at 2.34 L ha⁻¹ in spring 2017 to reduce weed competition. Annual species (plains coreopsis (PLAC, Coreopsis tinctoria) and partridge pea (PPEA, Chamaecrista fasciculata)) were reseeded on 26 March 2018 due to the late initial planting date. Prior to grazing each spring 2019–2022, pastures were clipped using a rotary mower. In late March 2021, both pastures were again burned to remove thatch and to reduce weed pressure.

The soil was a Dunmore type (Fine, kaolinitic, mesic Typic Paleudults) and had a pH of 6.4 and phosphorus and potassium content of 11.9 kg ha⁻¹ and 91.5 kg ha⁻¹, respectively. Pastures were fertilized 27 March 2018, 26 March 2020, and 22 April 2021 with 67.3 kg N ha⁻¹ yr⁻¹ and 7 May 2019 with 18.2 kg N ha⁻¹ yr⁻¹ in the form of urea (CO[NH₂]₂), in late April 2022 with 18.2 kg N ha⁻¹ and 46.6 kg P ha⁻¹ in the form of di-ammonium phosphate ([NH₄]₂HPO₄) and 40.9 kg ha⁻¹ K in the form of potash (KCl). To control undesirable species, pastures were spot-sprayed with either glyphosate (N-[phosphonomethyl]glycine) or 2,4-D (2,4-dichlorophenoxyacetic acid) in September 2019, July 2020, and June 2021. Pastures were not sprayed with herbicide in their entirety following forb establishment because forb tolerance to herbicides was not well defined.

2.2. Pasture Treatments

Each 1.2-ha pasture was divided into four replicates of five rest treatments (n = 20; 0.05 ha each) randomly assigned within a completely randomized design with a 0.2-ha center alley. The treatments were as follows: early rest (EARLY), middle rest (MIDDLE), late rest (LATE), no rest (NOREST), and no grazing (NOGRAZE). Each 0.05-ha paddock was accessible to cattle during different parts of the grazing season consistent with the assigned treatment (

Table 2. Grazing treatment application schedule for BBIG ^† and SG pastures grazed 2018–2022, Northeast Tennessee AgResearch and Education Center, Greeneville, TN.

	Year
Grazing Treatment					2022 §
	2018	2019	2020	2021	SG	BBIG
Early rest *	31 May to 20 Jun	30 May to 19 Jun	9 Jun to 26 Jun	3 Jun to 23 Jun	12 May to 2 Jun	1 Jun to 22 Jun
Middle rest	21 Jun to 11 Jul	20 Jun to 10 Jul	27 Jun to 17 Jul	24 Jun to 15 Jul	3 Jun to 23 Jun	23 Jun to 13 Jul
Late rest	12 Jul to 2 Aug	12 Jul to 31 Jul	18 Jul to 7 Aug	16 Jul to 5 Aug	24 Jun to 14 Jul	14 Jul to 3 Aug
No rest	17 May to 15 Aug	16 May to 14 Aug	28 May to 17 Aug	19 May to 20 Aug	28 Apr to 29 Jul	18 May to 18 Aug
No graze	17 May to 15 Aug	16 May to 14 Aug	28 May to 17 Aug	19 May to 20 Aug	28 Apr to 29 Jul	18 May to 18 Aug
Total grazing days	91 days	91 days	82 days	94 days	93 days	93 days

^† Big bluestem/indiangrass mixture, BBIG; switchgrass, SG. * Rest periods were structured to omit grazing for three weeks. ^§ Grazing was initiated earlier in SG pastures in 2022 due to pasture condition and available forage.

; Figure S1). Cattle grazing each pasture (BBIG or SG) had ad libitum access to all “open” paddocks each given day. The no-rest treatment was continually accessible for grazing throughout the season, while the no-graze treatment was not accessible for grazing. Treatment paddocks received the same grazing regimen each of the five grazing seasons. Grazing was precluded for each rest treatment for a 21-day period, either early, middle, or late in the grazing season, to evaluate the seasonal influence of grazing on the ability of the forbs to produce inflorescences and seeds and to loosely model a periodic rest management schedule. Grazing began, on average, on 16 May, concluded on 14 August each year, and occurred for 91, 91, 82, 94, and 93 days in 2018–2022, respectively.

2.3. Temperature and Precipitation

Weather data were collected on-site from the NETREC weather station in Greeneville, TN, and were used for observational comparison as opposed to formal analysis due to the small scale of the pastures. Mean ambient temperature was most similar to the 30-year mean in August and varied between 0.5 and 5.0 °C in May–July each year (Figure 1). Precipitation was most similar to the 30-year mean every year in May, but precipitation varied between 20 and 100 mm from June to August each year, with the greatest variability in rainfall observed in June.

2.4. Cattle Management

Weaned Angus-cross steers were stocked to include two to three tester animals per pasture (n = 2 in 2018 and n = 3, 2019–2022), and additional steers (grazers) were added as needed in a put-and-take grazing protocol to maintain a target grass canopy height of 36–41 cm for BBIG and 46–51 cm for SG [2]. Testers were weighed on two consecutive days at stocking and following removal from the pasture. Animals weighed 299 ± 36 kg at the beginning of the season across both BBIG and SG over the five years. The average daily gain ranged from 0.80 to 0.92 kg d⁻¹ for the tester animals. Steers had ad libitum access to minerals, water, and shade while grazing. All animal care and experimental procedures were approved by the University of Tennessee Institutional Animal Care and Use Committee (protocols #2258-0417 and #2258-0320).

2.5. NWSG and Forb Characteristics

Plant densities were collected prior to (17–28 May) and at the conclusion of (1–18 August) each grazing season from 2018 to 2022. Four randomly placed, 0.25 m² quadrats were sampled in each paddock. Native warm-season grasses were assessed by both plant and tiller counts, and forbs were assessed by plant counts per species.

Treatment paddocks were individually assessed for forb presence and flowering by species. Forb species were tallied as either flowering, present but not flowering, or not present based on observation of any plant of each species being present or flowering within the entire paddock. Forbs were also assessed for any grazing occurrence. Assessments were collected in conjunction with plant density measurements in spring and fall and with forage mass throughout the season.

2.6. Forage Mass

To determine forage mass, three 0.25 m² quadrats were sampled in each treatment paddock for every treatment change in 2018–2020 and at the start and conclusion of the respective treatment period in 2021 and 2022 by harvesting the available forage to a 5 cm stubble height. Forage was sorted into NWSG and FORB to determine botanical composition by mass. Samples were dried in forced-air ovens (Wisconsin Oven Corporation, East Troy, WI, USA) at 55 °C until they maintained a constant mass (approximately 72 h) to determine dry matter (DM) content. Forage mass was not analyzed in 2020 due to incomplete data.

2.7. Forage Nutritive Composition

Of the forage samples collected to determine forage mass, material above a 20 cm stubble height was retained to determine nutritive composition. Forage below this stubble height was not within the grazing horizon of the NWSG. After drying, samples that had been separated to evaluate botanical composition were recombined by paddock and sampling date to include both NWSG and FORB components. Samples were then ground using a Wiley Mill (Thomas-Wiley Laboratory Mill Model 4, Arthur H. Thomas Co., Philadelphia, PA, USA) passing through a 2 mm screen, followed by a cyclone sample mill (UDY Corporation, Fort Collins, CO, USA) ground to pass through a 1 mm screen [19]. In 2019, samples were ground to 1 mm using a Wiley Mill. Additional drying of the prepared sample in a forced-air oven at 55 °C was performed to ensure consistent moisture for scanning on a near-infrared spectrometer (NIRS) for less variability in predicted results across all samples [20]. The samples were scanned on a Foss DS2500F using ISIScan Nova v. 8.0.6.2 (Foss North America, Eden Prairie, MN, USA). Spectra were then applied to the 2021 Mixed Hay calibration licensed by the NIRS Forage and Feed Consortium (NIRSC, Berea, KY, USA). The global and neighborhood statistical tests were analyzed for accuracy across all predictions and were within the limit (H < 3.0) of fit. Units of measurement for nutritive analyses and calculated parameters are presented on a 100% DM basis.

2.8. Statistical Analysis

Analysis was conducted using R software (version 4.3.2, R Foundation for Statistical Computing, Vienna, Austria) running RStudio (version 2023.12.1.402, Posit Software, Boston, MA, USA), and statistical significance was set at p ≤ 0.05. The correlation between NWSG tillers and forb density was analyzed using a Kendall correlation test due to non-normal data. Correlations between NWSG tillers and forb density were not significant for any tests run by treatment; therefore, data were pooled across years and rest treatments, with May and September analyzed separately for each NWSG pasture. Treatment differences in forb densities for 7 species (black-eyed Susan, Rudbeckia hirta, BESU; lanceleaf coreopsis, Coreopsis lanceolate, LCOR; Maximilian sunflower, Helianthus maximiliani, MSUN; oxeye sunflower, Heliopsis helianthoides, OSUN; plains coreopsis; purple coneflower, Echinacea purpurea, PURC; and dixie ticktrefoil, Desmodium tortuosum, TTFL) that were observed most often were compared using general linear models under the Poisson distribution. A chi-squared test was performed to compare the frequency of forb and flowering presence (present and flowering, present but not flowering, not present) among forb species. Data were pooled across grass types, treatments, and years. A second chi-squared test was performed to compare the frequency of forb flowering per species across grazing treatments; data were compared within grass type and pooled across years. Each forb species was ranked based on plant abundance, persistence, flowering window, and grazing observations similar to Richwine et al. [17]. Ranks were assigned ordinally and compared to other present species as opposed to a set scale. A rank of 1 was the greatest rank and corresponded with the highest abundance, most persistent, longest flowering window, and most frequently grazed compared to the other forbs. The remaining response variables (forage mass and forage nutritive composition parameters) were compared using mixed-effects ANOVAs running a Type III Wald F-test with Kenward–Roger df. A square root transformation was applied to the total forage mass from the SG pasture. Response variables were analyzed separately for BBIG and SG pastures for all models except forb forage mass, which was not impacted by grass type. The NOGRAZE treatment was excluded from the models analyzing total forage mass and nutritive component concentrations to focus on the effects of grazing and periodic rest; data were not available for these response variables in 2020. Total forage mass, FORB forage mass, and nutritive composition (crude protein [CP], neutral detergent fiber [NDF], and acid detergent fiber [ADF]) were each analyzed with grazing treatment, month, and year entered as fixed effects, and replicate was a random effect. Two- and three-way interactions of main effects were included as fixed effects when the interaction significantly contributed to the model. Response variables transformed for analysis were back-transformed for results presentation. Mean separations were compared using Tukey’s honestly significant difference test.

3. Results

3.1. NWSG and Forb Characteristics

Among all grazing treatments, a correlation between NWSG tiller density and forb density was only observed in September in SG (p < 0.01, τ = −0.33; Figure 2). As tiller density in SG increased from 200 to 600 m⁻², forb density declined from more than 20 to less than 5 plants m⁻². Densities in BBIG (spring and fall) and SG (spring) did not demonstrate strong correlations (p > 0.17). Forb species followed different temporal trends over the five grazing seasons (Figure 3). The annual PLAC, biennial BESU, and short-lived perennial LCOR all demonstrated early establishment of plant densities between 13 and 32 plants m⁻². This initial spike was followed by a reduction (p < 0.05) in population size over the latter three grazing seasons, except (p < 0.05) for a spike in BESU and LCOR in 2021. Oxeye sunflower, PURC, and TTFL produced greater (p < 0.05) population density in the second and third grazing seasons but declined in the fourth and fifth seasons. Maximilian sunflower and OSUN did not differ (p > 0.05) among treatments, while LCOR, BESU, and TTFL established best (p < 0.05) in NOGRAZE, PLAC demonstrated moderate compatibility in EARLY, NOREST, and NOGRAZE, and PURC in NOGRAZE and NOREST.

Species demonstrated varying flowering frequencies (χ² = 3953.3, df = 20, p < 0.01; Figure 4). Black-eyed Susan, PURC, OSUN, and LCOR flowered the most frequently, while MSUN and TTFL were frequently present and not flowering. Illinois bundleflower (Desmanthus illinoensis; ILBF) and purple prairie clover (Dalea purpurea; PUPC) were rarely observed. Grazing treatment did not influence the presence of flowers for any species within the BBIG (χ² = 30.714, df = 40, p = 0.85; Figure 5) or SG (χ² = 42.871, df = 40, p = 0.35) pastures.

Lanceleaf coreopsis, BESU, and PLAC were the most abundant species in SG (

Table 3. Abundance, persistence, flowering, and grazing ranks of forbs in BBIG ^† and SG pastures, grazed 2018–2022, Northeast Tennessee AgResearch and Education Center, Greeneville, TN.

	Species	Abundance (Total Plants) *	2022 Plants m−2 §	Abundance Rank	2022 Plants Rank (Persistence)	Flowering Window Rank ‡	Grazing Rank ⁑	Rank Average	Overall Rank ⁂
SG	BESU	1550	0.8	2	2	1	6	2.8	1
	LCOR	2639	0	1	5	3	3	3.0	2
	OSUN	305	0.8	5	2	4	2	3.3	3
	PURC	364	0.2	4	4	2	4	3.5	4
	TTFL	245	0.9	6	1	8	1	4.0	5
	MSUN	112	0.4	7	3	5	5	5.0	6
	PLAC	741	0.2	3	4	6	7	5.0	7
	PPEA	32	0	8	5	7	7	6.8	8
	UPPC	20	0	9	5	9	7	7.5	9
	ILBF	5	0	10	5	10	7	8.0	10
	PUPC	2	0	11	5	10	7	8.3	11
BBIG	TTFL	762	3.1	3	1	6	1	2.8	1
	PURC	243	0.9	6	2	1	3	3.0	2
	LCOR	937	0.4	1	4	4	4	3.3	3
	BESU	828	0.8	2	3	2	6	3.3	4
	OSUN	311	0.8	5	3	3	2	3.3	5
	MSUN	55	0.2	8	5	5	5	5.8	6
	PLAC	341	0.1	4	6	7	7	6.0	7
	UPPC	63	0	7	7	8	7	7.3	8
	ILBF	12	0	9	7	10	7	8.3	9
	PPEA	6	0	10	7	9	7	8.3	10
	PUPC	0	0	11	7	10	7	8.8	11

^† Big bluestem/indiangrass mixture, BBIG; switchgrass, SG; black-eyed Susan, BESU; lanceleaf coreopsis, LCOR; oxeye sunflower, OSUN; purple coneflower, PURC; dixie ticktrefoil, TTFL; Maximilian sunflower, MSUN; plains coreopsis, PLAC; partridge pea, PPEA; upright prairie coneflower, UPPC; Illinois bundleflower, ILBF; purple prairie clover, PUPC. * Total abundance is the sum of all plants counted May and September 2018–2022. ^§ Mean plant density m⁻² May and September 2022 used to rank persistence. ^‡ Flowering window rank (1–10; 1 = longest flowering window, 10 = shortest flowering window or not observed flowering) based on length of flowering observed; ranks with the same number bloomed for similar durations but were not necessarily blooming at the same time of the season. ^⁑ Grazing rank (1–7; 1 = most grazed, 7 = least grazed or not grazed) based on plant grazing occurrence. ^⁂ Overall rank (1–11; 1 = most recommended forb, 11 = least recommended forb) based on average rank; abundance separated ties in average rank [17].

), while LCOR, BESU, and TTFL were the most abundant in BBIG. Black-eyed Susan, OSUN, and TTFL were among the most persistent species in both NWSGs. Lanceleaf coreopsis, BESU, OSUN, and PURC flowered the longest, and TTFL, OSUN, and PURC were grazed the most frequently. Of the 11 species, BESU, LCOR, OSUN, PURC, and TTFL ranked the highest according to abundance, persistence, flowering window, and grazing selectivity in both BBIG and SG.

3.2. Forage Mass

Total forage mass had a two-way interaction between grazing treatment and year in BBIG (p < 0.01;

Table 4. Model results for ANOVAs of forage mass and forage nutritive composition of BBIG ^† and SG pastures interseeded with FORB under different grazing management treatments, 2018–2022, Northeast Tennessee AgResearch and Education Center, Greeneville, TN.

	Predictor	F	df	p-Value
Total forage mass
SG	Grazing treatment	9.34	3	<0.01
	Month	4.89	3	<0.01
	Year	35.53	3	<0.01
BBIG	Grazing treatment	8.48	3	<0.01
	Month	22.71	3	0.03
	Year	3.06	3	<0.01
	Grazing treatment × Month *	5.74	8	<0.01
	Grazing treatment × Year	2.76	9	<0.01
Forb forage mass
SG &	Grazing treatment	14.94	4	<0.01
BBIG	Month	0.81	3	0.49
	Year	19.84	3	<0.01
	Grazing treatment × Month	4.19	12	<0.01
	Grazing treatment × Year	6.91	12	<0.01
CP
SG	Grazing treatment	2.68	3	0.05
	Month	141.11	3	<0.01
	Year	31.04	3	<0.01
	Month × Year	13.97	9	<0.01
BBIG	Grazing treatment	4.22	3	<0.01
	Month	56.71	3	<0.01
	Year	13.77	3	<0.01
	Month × Year	11.13	9	<0.01
NDF
SG	Grazing treatment	5.15	3	<0.01
	Month	130.41	3	<0.01
	Year	40.73	3	<0.01
	Month × Year	12.11	9	<0.01
BBIG	Grazing treatment	7.66	3	<0.01
	Month	39.71	3	<0.01
	Year	3.60	3	0.01
	Month × Year	11.76	9	<0.01
ADF
SG	Grazing treatment	2.10	3	<0.01
	Month	103.03	3	<0.01
	Year	5.14	3	<0.01
	Month × year	18.63	9	<0.01
BBIG	Grazing treatment	1.81	3	0.15
	Month	29.40	3	<0.01
	Year	10.19	3	<0.01
	Month × Year	6.93	9	<0.01

^† Big bluestem/indiangrass mixture, BBIG; switchgrass, SG; native forb mixture, FORB; crude protein, CP; neutral detergent fiber, NDF; acid detergent fiber, ADF. * Interactions of higher order that did not significantly impact the model were dropped from further analysis, and only significant interactions were included.

and

Table 5. Mean total forage mass of BBIG ^† and SG pastures interseeded with FORB grazed 2018–2022 *, Northeast Tennessee AgResearch and Education Center, Greeneville, TN.

	Grazing Treatment	Year
		2018	2019	2021	2022
		–––––––––––––––––––––kg ha−1––––––––––––––––––––––
BBIG	Early rest	2011 bcd⁑	2673 bcd	7917 a	1910 bcd
	Middle rest	1368 cd	2586 bcd	3502 b	832 cd
	Late rest	1372 cd	2189 bcd	3567 bc	454 d
	No rest	1398 bcd	2422 bcd	1957 bcd	999 cd
SG	Early rest	4031 a§	4442 a	5773 a	1527 a
	Middle rest	2271 b	2582 b	3618 b	565 b
	Late rest	2323 b	2637 b	3683 b	540 b
	No rest	2017 b	2310 b	3295 b	420 b

^† Big bluestem/indiangrass mixture, BBIG; switchgrass, SG; native forb mixture, FORB. * Forage mass was not analyzed for 2020 due to incomplete data and NOGRAZE was excluded from analysis due to confounding factors. ^⁑ Mean total forage mass with different letters differed (p < 0.05) among treatments and years in BBIG based on Tukey’s HSD test. ^§ Mean total forage mass with different letters differed (p < 0.05) across treatments within years in SG based on Tukey’s HSD test.

). Total forage mass in BBIG was greatest (p < 0.05) in EARLY in 2021, while most other treatments produced similarly each year. The EARLY treatment produced the greatest (p < 0.05) total forage mass each year in SG. Forb forage mass was influenced by a two-way interaction between grazing treatment and year (p < 0.01;

Table 6. Mean forb forage mass pooled across BBIG ^† and SG pastures interseeded with FORB and grazed 2018–2022 *, Northeast Tennessee AgResearch and Education Center, Greeneville, TN.

Grazing Treatment	Year
	2018	2019	2021	2022
	–––––––––––––––––––––kg ha−1––––––––––––––––––––––
Early rest	2006 b⁑	825 cd	227 cd	0 cd
Middle rest	1064 bcd	870 bcd	839 bcd	167 cd
Late rest	1058 bc	520 cd	1563 bc	361 cd
No rest	1078 bc	1056 bc	459 cd	132 d
No graze	3154 a	951 c	1052 bc	798 cd

^† Big bluestem/indiangrass mixture, BBIG; switchgrass, SG; native forb mixture, FORB. * Forage mass was not analyzed for 2020 due to incomplete data. ^⁑ Mean forb forage mass with different letters differed (p < 0.05) among grazing treatments and years based on Tukey’s HSD test.

) and decreased (p < 0.05) from 2018 to 2022 for the EARLY, NOREST, and NOGRAZE treatments. The interactions between grazing treatment and month (p < 0.01) additionally affected forb forage mass, where all grazed treatments decreased in mass over the summer while the NOGRAZE treatment increased in mass from May through August.

3.3. Forage Nutritive Composition

Crude protein, NDF, and ADF were each reliant on two-way interactions between month and year in both BBIG (p < 0.01; Figure 6a) and SG (p < 0.01). Crude protein tended to decrease as each season progressed, while both NDF and ADF increased. Crude protein and NDF were also dependent on grazing treatment (p < 0.05; Figure 6b).

4. Discussion

Switchgrass and BBIG pastures were successfully interseeded with a diverse blend of forbs that established, persisted, and produced blooms based on their life history, grazing selectivity, and strategies for competition. Tiller density of NWSG had limited impact on forb density, and grazing regimen did not influence the blooming frequency, even though species produced varying blooming patterns. However, rest based on blooming patterns and stand quality (e.g., tiller density, forage height, and maturity) could produce desirable long-term forage mass even after five grazing seasons.

4.1. NWSG and Forb Characteristics

Plant density and relationships among species are complex in mixed grasslands. Although native grasses are bunchgrasses that provide an opportunity for the establishment of a variety of other species between plants, successfully interseeding desirable companion species like red clover (Trifolium pratense L.) among established NWSG has shown to come with challenges [1]. Due to rhizomatic spread, plant density proved to be difficult to measure in the mature (planted in 2008), robust NWSG stands in this study. However, tiller density is a reliable tool to measure vigor and persistence [21]. Established stands of NWSG grazed to maintain vigorous canopies (e.g., >36 cm) increase in tiller density as the plants increase in circumference [3]. This spread may influence the success of interseeding other species, which led us to hypothesize a negative association between NWSG tiller density and interseeded plant density. There was not a correlation between NWSG tiller density and FORB density, except for SG (September), which indicated greater forb density where NWSG tiller density was lower. Thus, it appears NWSG tiller density influenced forb establishment and persistence later in the season in SG, but not in the spring or the fall for BBIG. We hypothesize that the forb diversity in SG may have been enhanced through greater representation from late-season forbs such as OSUN or MSUN that took advantage of gaps between SG plants. Additionally, due to its erect growth habitat and large crowns, patches between SG plants in grazed pastures may provide unobstructed areas for forb establishment later in the season.

The diversity of forbs in the mixture likely compensated for the varying establishment success among these species. This difference was demonstrated by the early success and persistence of the biennial BESU and short-lived perennial LCOR, followed by increased densities of MSUN, OSUN, PURC, and TTFL by the second year. The more rapidly establishing BESU and LCOR may have also occupied space between the NWSG plants, reducing weeds and other competition sources as the perennial species developed over the later grazing seasons as the dominance of BESU and LCOR diminished. Regardless, biennials (BESU), short-lived perennials (LCOR) and moderate to tall-growing perennials (MSUN, OSUN, PURC, and TTFL) established best. Conversely, smaller annuals (PLAC and PPEA) and perennials (ILBF, PUPC, and upright prairie coneflower, Ratibida columnifera, UPPC) did not establish well. Competition from the NWSG and weeds may have reduced the germination of smaller-growing forbs like PUPC. Poor establishment and persistence of PUPC, ILBF, and UPPC have been previously demonstrated [22], even without grass competition or grazing pressure, suggesting that other factors may affect the establishment or persistence of these forbs. Ecotypes and variants of forb species have also shown varying success in establishment and persistence, particularly through competitive susceptibilities and seed dispersal [23]. The multi-year rest in the NOGRAZE treatment also produced some of the highest densities of BESU, LCOR, PURC, and TTFL. However, each of these four species followed similar trends over the years across grazing treatments (except for BESU in 2021). This similarity suggests that although rest could improve the establishment of more frequently grazed species, a combination of time after establishment and grazing treatment effected forb populations jointly in a complex relationship.

Flowering frequency and duration have also been shown to vary. Long-flowering forbs typically flower earlier in the season and belong to annual or biennial classifications [22]. However, in the current study, the perennials OSUN and PURC both displayed long flowering windows along with BESU and LCOR. All four of these species bloom early- to mid-summer compared to the late-season bloomer MSUN and provide critical floral resources for pollinators coming out of hibernation [24]. Although Harris et al. [24] described a high floral demand in spring, late summer and early fall typically provided the greatest floral deficit to pollinators and is a period when late-season bloomers like MSUN could bridge the gap between floral supply and demand. Despite variations in flowering windows and the total flowering observations made for each species, grazing treatment did not affect the flowering presence of each species. Ravetto Enri et al. [25] showed that excluding cattle during peak blooming periods resulted in three times greater flower cover than areas where cattle were not excluded. Their work also found that cattle grazing did not influence flower cover between continuously and rotationally grazed pastures. Flower cover increased over the three grazing seasons and heterogeneity was greater in pastures grazed by cattle compared to sheep. Although our study did not quantify flower density, it similarly concluded that varying rest regimens and continuous grazing, without overgrazing, should produce viable blooms. Including a variety of species with variable blooming windows provides season-long blooms and floral resources for pollinators without any detriment due to grazing.

Black-eyed Susan, LCOR, OSUN, PLAC, PURC, and TTFL were the most frequently observed species, but BESU, MSUN, OSUN, TTFL, and PURC were the most persistent by the fifth grazing season (2022). The great reduction (BBIG) or elimination (SG) of LCOR demonstrated its characteristic behavior as a short-lived perennial but also likely reflected moderate grazing selectivity by cattle for this species. Although LCOR is a prolific reseeder, competition and/or grazing selectivity likely kept the forb from reestablishing in subsequent years in all treatments. These temporal trends indicated challenges with persistence in both grazed and ungrazed paddocks. A lack of grazing produced challenges with competition from tall-growing grasses, while grazing resulted in defoliation and trampling. Both LCOR and BESU exhibited a population spike in 2021, likely in response to the prescribed burn in March and the stratification benefits that this disturbance provided. Although these species are prolific reseeders, stratification, such as provided by burning, can lead to changing population density over the years [26,27] and often spur greater germination the following season for annuals and perennials [28]. The increase in the perennial MSUN was likely due to its need for crown establishment during the earlier seasons, and once well established, its tall growth enabled it to compete with the NWSG. The development of basal rosettes and crowns allow perennial forbs to develop a robust root system and increase competition for space through horizontal growth before dedicating nutrients to vertical growth and reproduction in subsequent growing seasons [29]. High grazing selectivity of slower-developing plants diminishes such species’ persistence, as demonstrated by PURC and PLAC, both of which were moderately preferred by grazing cattle. Dixie ticktrefoil, however, demonstrated the highest persistence rank while also being the most preferred species. Its low-growing stature and quick regrowth likely allowed TTFL plants to regenerate and persist throughout the five grazing seasons. The inclusion of grazing has been shown to improve interseeded forb establishment [30]; however, a balance of some rest with grazing, combined with other disturbances like fire [28], could improve establishment and persistence in the grazing seasons immediately after interseeding. The change of grazing management year-to-year after interseeding has not been assessed but may have implications for improved persistence. Grazing strategies influence forb establishment, persistence, and flowering patterns [18] and should be adapted to not only manage the existing stand but also allow new plants to establish. In this study, we seeded at a rate between 7 and 392 seeds m⁻² depending on the species, much of which likely added to the seed bank since observed forb densities were below this range in some areas. These dormant seeds, along with others naturally seeded from blooms, may contribute to future plant establishment. Incorporating spring burns to provide stratification and grazing during peak germination intervals to reduce competition should contribute to improved seedling recruitment and vigor during mid-season rests. In addition, allowing pastures to rest during peak blooming periods [25] of target species should also allow for natural seed production.

4.2. Forage Mass

Total forage mass was similar in 2018 and 2019 among grazing regimens in BBIG and peaked in 2021 following the prescribed burn but then declined in 2022. The susceptibility of BBIG and SG to overgrazing, grazing repeatedly below the 36 cm grazing horizon, is high, particularly when management is driven by calendar dates and not grass conditions [31]. Although we used a put-and-take method to adjust stocking density, having paddocks open for extended periods throughout the summer, combined with the rigid three-week rest schedules, likely led to a weakening of the BBIG stands. The introduction of cattle to SG two weeks earlier than BBIG in 2022 also led to a numerical reduction in forage mass throughout that grazing season in all grazed paddocks compared to the previous season.

Despite a decline in forb mass from 2018 to 2019 in EARLY and NOGRAZE, treatments produced similar forb mass each year. The density of large-stature species that made up much of the forb forage mass, like MSUN and OSUN, did not vary among grazing treatments. The use of fire in March 2021 likely contributed to the increase in forage production numerically in LATE; however, most treatments were not significantly more productive in 2021 than in 2019, and all declined in 2022 numerically, nonetheless. Although fire disturbance is used to manage native prairies [28,32], fire characteristics, grazing pressure, and annual climate conditions introduce complex interactions that variably determine stand diversity and biomass production [28]. The ultimate decline in mass production in 2022 was likely due to a combination of inflexible (experimentally constrained) grazing, reduced forb density, and climate conditions like higher ambient temperature and reduced rainfall in June and August compared to the 30-year average.

4.3. Forage Nutritive Composition

The inverse trends between CP and fiber (NDF and ADF) were expected as the grasses became more mature throughout the season [33]. Crude protein for both pastures ranged from 80 to 220 g kg⁻¹ DM, NDF from 340 to 700 g kg⁻¹, and ADF from 250 to 450 g kg⁻¹ DM, similar to some previous findings of NDF and ADF in NWSG pastures [3]. Both NDF and ADF were also generally lower than some reports [1,2] of NWSG and NWSG/red clover pastures. Additionally, CP was higher than most NWSG-dominated pastures [1,2,3]. This difference in both CP and fiber values can be attributed to both the forage maturity at harvest and stand composition. Forbs provide more nutrient-dense forage than some grass species, 105 to 189 g kg⁻¹ DM CP, 221 to 410 g kg⁻¹ DM NDF, and 217 to 368 g kg⁻¹ DM ADF [22], and can also greatly influence the forage nutritive composition of pastures. On average, most grazing regimens produced forage that met a growing 227 kg steer’s required daily intake of 128 g kg⁻¹ CP [34], particularly early in the season and in BBIG. Although forb forage mass was not impacted by pasture type, the greater abundance of BESU in SG and TTFL in BBIG could have led to slight differences in CP between the pastures.

4.4. Sustainability and Environmental Implications

This study demonstrates the influence of structured periodic rest on NWSG pastures and FORB establishment and persistence. Although consistencies of treatment schedules among BBIG and SG were important to demonstrate the influence of rest, applying grazing management strategies should be adjusted to best protect pasture persistence [31], soil health [9,35,36], and forage quality. Overgrazing, or even continuous understocking and, thus, high selectivity, reduces plant leaf area, root volumes, [36] and stand productivity, leading to increasing weed encroachment and, ultimately, plant death [37,38,39]. Concurrent work to our study validated these phenomena when we examined continuous grazing of NWSG with and without interseeded forbs and concluded that forb establishment was successful, but the continuous low stocking rate and greater grazing selectivity led to a weakened stand after three grazing seasons [40]. Allowing pastures to rest when below the grazing horizon (<36 cm), grazing before NWSG exceed their grazing horizon (~65 to 75 cm, dependent on NWSG species), and applying an adequate stocking rate for uniform grazing allow for ample regrowth and maintain immature stands longer into the season, thus improving forage quality [33]. In addition to improving NWSG persistence and soil health, forbs that are allowed to rest during their respective flowering window could improve blooming quantities and pollinator resources [25]. Disturbance is important for forb establishment and forbs provide high-quality forage. However, a complex balance between burning, grazing, and rest allows forbs to establish, feed grazers, and produce floral resources in a diverse system.

4.5. Limitations

This study highlights the establishment and persistence of native forbs interseeded into BBIG and SG pastures and grazed for five summers under structured rest periods. Although the diversity of forbs established and blooming presence was not different across the grazing regimens, structured rest periods, as implemented, likely reduced the longevity of the NWSG stands. Resting pastures based on stand condition, sward height, or climactic factors like drought could increase stand vigor and persistence compared to rest periods predetermined on a calendar. Ravetto et al. [25] demonstrated that rest or exclusion periods centered on key blooming windows produced greater flower cover compared to grazing without rest periods. Further work should examine the effects of rest structured around key blooming windows and stand characteristics to quantify strategic elements for determining rest in diverse native pastures.

5. Conclusions

We demonstrated that periodic rest did not affect the establishment and persistence of forbs in native grass pastures. Across treatments, BESU, LCOR, OSUN, PURC, and TTFL were some of the most frequently observed, while BESU, MSUN, OSUN, TTFL, and PURC were among the most persistent after five grazing seasons. Additionally, cattle preferentially grazed LCOR, OSUN, TTFL, and PURC over other species. Black-eyed Susan, LCOR, OSUN, and PURC had the longest flowering windows, though grazing management did not influence the blooming frequency of any species. Based on these qualities, BESU, LCOR, OSUN, PURC, and TTFL could be interseeded into either SG or BBIG pastures and will successfully establish similarly across a variety of grazing management regimens.