Compensatory Yield Responses of Young Native Warm-Season Grass Stands to Seasonal Changes in Harvest Frequencies

Abstract

Defoliation management can significantly affect subsequent grassland’s forage productivity and sustainability. To assess the type and persistence of native warm-season grass (NWSG) yield responses to changes in defoliation intensity, a five-year harvest trial was conducted, in a randomized complete block design, at Virginia State University’s research farm. Yield responses of newly established indiangrass (Sorghastrum nutans L.). Nash, big bluestem (Andropogon gerardii Vitman), switchgrass (Panicum virgatum L.) and eastern gamagrass (Tripsacum dactyloides L.) stands to second-year changes in harvest regimes were monitored. In 2012, seedlings of these native grasses were transplanted in clean-tilled plots, with 16 plants in each pot. The field was not irrigated, but broadleaf weeds were manually controlled, and all plots were mowed in August and mid-November of 2013. Starting June 2014, each plot had three harvest-strips assigned to single, two, or three cuts per year from early June to mid-October using a plot-harvester; forage weights were recorded. Based on the recorded fresh and oven-dry sample weights, plot forage DM yields were estimated. Cumulative forage biomass of all three-cut strips flipped to single-cut increased by ≥30% and >50% for big bluestem. The second-year single-cut yields also outperformed those cut thrice since the first year by 22–51%. The second-year biomass increases from single-cut strips that were cut thrice in the first year demonstrated that flipping-triggered compensatory yield responses overshadowed the first- year losses in plant vigor. The compensatory yield increases continued to, but not beyond, the third year and varied between species. The yield responses also showed that magnitudes of defoliation management–triggered NWSG growth responses depend more on growing conditions during the recovery period than its actual duration.

1. Introduction

In the southeast and midwestern regions of the US, native warm-season grasses (NWSGs) have gained popularity as alternative summer forage sources that can sustain vegetative growth during the hot and dry months (July–August) of the year. These NWSGs of North America are morphologically and physiologically adapted to moisture stress conditions, which render cool-season grasses unproductive during the hot summer months [1]. Besides being drought tolerant and better at nutrient acquisition, because of their deep and extensive root systems [2], most NWSGs also withstand moderate herbivory [3]. The most preferred NWSG species include indiangrass (IG, Sorghastrum nutans L. Nash), big bluestem (BB, Andropogon gerardii Vitman), little bluestem (LB, Schizachyrium scoparium (Michx.) Nash), switchgrass (SG, Panicum virgatum L.) and eastern gamagrass (EG, Tripsacum dactyloides L.). In a growing season, the NWSGs can produce more forage and for longer durations than their introduced counterparts, such as bermudagrass (Cynodon dactylon L. Pers), bahiagrass (Paspalum notatum Flueggé), and dallisgrass (Paspalum dilatatum Poir.). However, greater forage yields and quality from NWSGs may not be sustainable if subjected to inappropriate defoliation, whether through overgrazing or intensive haying. The NWSG attributes, which are largely rooted in their unique morphological features, require special management approaches to ensure sustained productivity. These attributes include the fact that NWSGs have growing points positioned high above the ground, which makes them relatively more susceptible to physical damages associated with intensive defoliation [4,5]. In managed grasslands, desirable biomass productivity and/or associated ecosystem services may not be sustainable if decisions on the timing, frequency, duration, and intensities of defoliation events overlook their potential effect on specific plant growth responses. With mixed NWSG stands, for example, such decisions and their subsequent adjustments are tailored towards management objectives, mainly resource conservation, ecosystem services, and/or biomass production [6,7,8]. For managed defoliation to fulfil the set objective(s), relevant plant responses to the treatments are usually monitored through changes in species composition–numbers, spatial distribution, growth performance, stand thickness, structural uniformity, etc. [6,7,9,10]. While species composition and structural uniformity could be the focus for ecosystem services and conservation-related defoliation management, forage production responses will include regrowth performance and stand thickness. In forage production, for example, harvesting events may be delayed or hastened and thus maximize yields or forage quality. This is partly based on known increases in forage yield following timely removal of the shading effect of senescent leaves, thus allowing more sunlight to reach the base of plants where it stimulates bud break and new vegetative growth [11,12]. Defoliation also removes the hormonal influence of reproductive tillers, which induces increased production of vegetative tillers with greater leaf biomass [13]. Appropriate defoliation management favors efficient utilization and faster shoot regrowth in recovering stands, an adaptation to grazing [14] that enables plants to repair or even compensate for physical damages and/or total loss of meristematic tissues. In such cases, increasing defoliation frequencies will most likely increase total season biomass yield. However, the degree to which plants can compensate for tissue damages or loss is usually influenced by local factors such as precipitation as well as frequency and intensity of current and past defoliation events [15,16,17,18]. There are species differences in morphological adaptations to tissue damage, which increase plants’ tolerance to defoliation and respective growth responses [19,20,21]. To be appropriate, therefore, decisions on harvest regimes must also take into consideration the most likely growth responses by the dominant species as well as prevailing weather conditions.

While the harvesting of grasses ahead of their leaf senescence often optimizes the dry matter yield and forage quality [13], severe defoliation may negatively impact stand persistency and forage quality due to differential growth responses of the plant components [22]. On average, severe defoliation may impose irreversible tissue damage and thus reduce subsequent season forage biomass. Multiple defoliations have also been found to reduce the herbage yields of warm-season grasses by over 60% [11,23]. This usually involves a reduction in plant leaf area by changing respiratory rates, growth rates, and carbon allocation patterns [22]. As a result, recovery growth in defoliated stands will mostly reflect the speed at which the plants can repair vegetative tissue damage and restore their photosynthetic capacities. In part, this depends on the proportions of photosynthetic tissues that the defoliated plants retain [22,24,25,26]. It also depends on the timing of the first defoliation [27] and the time allowed between successive defoliations for plants to restore their photosynthetic capacities [22]. Appropriate defoliation intervals should be long enough for plants to sufficiently restore carbohydrate reserves, which subsequently influences stand persistence [27]. However, in mixed stands, species differences in growth rates and duration to maturity often complicate decisions on the harvest intervals. This study, therefore, sought to generate data that may help in developing sustainable defoliation strategies for increasing the performance and biomass productivity of newly established NWSG stands managed for forage and feedstock production. Specifically, the study assessed effects of changes in harvest regimes on (i) forage yield performance of NWSG pure stands and (ii) the persistence of the defoliation-induced yield responses in subsequent consecutive years. It was hypothesized that reducing the number of harvests per year would result in faster regrowth rates enough to compensate for prior-year losses in stand vigor.

2. Materials and Methods

2.1. Location and Field Preparations

The study was conducted at Virginia State University’s research farm (Randolph Farm) located in Chesterfield county, Virginia, at 37°13′43″ N; 77°26′22″ W, about 45 m above sea level. The soils are Bourne series fine sandy loam (mixed, semiactive, thermic Typic Fragiudults) with low organic matter content. By the summer of 2013, the study area had a 20-year June, July, and August average precipitation of 92, 113, and 121 mm, with day temperatures of 30.2, 32.1, and 31.2 °C, respectively [28]. The mean monthly temperatures and precipitation amounts around the study area from April through October for the 2012 through 2018 production years are shown in Figure 1a and Figure 1b, respectively. In the mid-June 2012, 64 plots (roughly 6 m W × 7 m L) were randomly assigned to BB, GG, IG, and SG in a randomized complete block design with four replications. Two days after a heavy rainfall, seedlings of each species were transplanted using a tractor-operated seedling planter set at 30 cm × 45 cm spacing within and between rows, respectively. The planting was such that seedlings were upright, with their entire root sections completely covered with at least a 2.5 cm layer of soil.

During the second year after planting, in 2013, all plots received a common first cut at the flowering stage in early August, with a second cut in mid-November; see

Table 1. Actual harvest dates for three different cutting frequencies (cuts year⁻¹) on newly established NWSG stands recorded from 2013 to 2018.

		Harvest Dates by Harvest Regime
Year	Cuts	Three Cuts	Two Cuts	Single Cut
2013 *	1st	5 June 2013	5 June 2013	5 June 2013
	2nd	18 November 2013	18 November 2013	18 November 2013
1st (2014)	1st	24 June 2014	23 July 2014	14 September 2014
	2nd	29 July 2014	12 September 2014
	3rd	12 September 2014
2nd (2015)	1st	14 June 2015	18 June 2015	14 October 2015
	2nd	31 July 2015	14 October 2015
	3rd	14 October 2015
3rd (2016)	1st	18 June 2016	18 June 2016	24 October 2016
	2nd	05 August 2016	24October 2016
	3rd	24 October 2016
4th (2017)	1st	19 June 2017	29 June 2017	18 October 2017
	2nd	17 August 2017	18 October 2017
	3rd	17 October 2017
5th (2018) *	N/A	26 June 2018	26 June 2018	26 June 2018

* All plots experienced the same pre- and post-treatment harvests in 2013 and 2018, respectively; NA = Not applicable.

.

Starting early-June of 2014, however, three adjacent 1.5 m wide harvest strips were marked in all 64 plots and assigned to three harvest frequencies: single, two, or three cuts per year. Forage harvesting was carried out for four consecutive years followed by a single fifth-year mid-summer harvest on 26 June 2018. During the first through fourth years, the harvests for the three-cut systems were early June, late July to early August, and late September to mid-October. The first and last harvest dates were also applicable for the two-cut systems. From the second year on, however, the single- and three-cut assignments were flipped for plots 33–64, but kept the same in the other 32 plots (Figure 2). Each 8-plot row had two plots of the same NWSG species. In each 8-plot row, end-to-end aligned strips were assigned the same harvest frequency, which facilitated machine operations and limited traffic over the harvest strips to the respective harvest dates. For the forage biomass assessment, a CIBUS F Plot Forage Harvester (Wintersteiger Ag, Dimmelstrasse 9 Ried im Innkreis, Austria) was used. The harvester was equipped with a Harvest-Master weighing system (Juniper Systems, Inc., Logan UT. USA) capable of 0.01 kg accuracy. To minimize losses of growing points, the cutting height was set at about 18 cm above the soil surface. From each plot harvest, a representative forage sample was weighed before and after 3 d oven drying to constant weight at 65 °C to determine percent moisture content, as harvested, and estimate the forage DM yield ha⁻¹. In the same manner, subsequent forage biomass estimates were recorded over four more consecutive years.

2.2. Data Analysis

The data were organized and subjected to analysis of variance (ANOVA) as an RCBD, with the year of assessment, defoliation management (system), species, and harvest frequency (cuts) as fixed effects. For the analysis, the PROC GLM procedure of SAS version 9.4 (SAS Institute Inc., Cary, NC, USA) was used. Because of significant two- and three-way interactions, ANOVA was done for each year separately by changing the statistical model accordingly. The ANOVA was first performed to test for interaction effects of year, species, harvest frequency (cuts), and harvest regime (systems) on cumulative forage biomass. With all interaction effects being significant, the data set was sorted and analyzed by year, species, and systems for effects of cuts. The ANOVA was then done for effects of systems within cuts, species and year (

Table 2. The ANOVA F and p values for the main and interaction effects of year, harvest regime (system), species, and harvest frequency (cuts) on single, two, and three cuts per year systems for two-year-old NWSG stands recorded from 2014 to 2018.

Source of Variation	Degrees of Freedom	F Value	p > F
Model	95	11.03	<0.001
Year	3	58.54	<0.001
Species	3	65.11	<0.001
Year × Species	9	9.81	<0.001
Frequency	2	119.17	<0.001
Year × Frequency	6	10.97	<0.001
Species × Frequency	6	20.50	<0.001
Year × Species × Frequency	18	4.34	<0.001
System	5	48.71	<0.001
Year × System	15	6.33	<0.001
System × Species	15	9.04	<0.001
Year × System × Species	45	2.55	<0.001
Error	672
Corrected Total	767

NWSG = native warm-season grass.

and

Table 3. The ANOVA F and p values for the main and interaction effects of year, harvest regime (system), species, and harvest frequency (cuts) on two- and three-cut forage systems for two-year- old NWSG stands recorded from 2014 to 2018.

Source of Variation	Two Cuts Year−1			Three Cuts Year−1
	df	F Value	p > F	df	F Value	p > F
Model	63	16.4	<0.001	95	38.2	<0.001
Year	3	100.9	<0.001	3	117.4	<0.001
System	1	0.2	0.61	1	31.4	<0.001
Species	3	39.6	<0.001	3	15.8	<0.001
Cuts	1	120.1	<0.001	2	440.8	<0.001
Year × System	3	1.3	0.28	3	5.7	0.001
Year × Species	9	7.0	<0.001	9	8.1	<0.001
Year × Cuts	3	56.5	<0.001	6	299.4	<0.001
System × Species	3	0.2	0.87	3	5.9	0.001
System × Cuts	1	3.4	0.07	2	14.5	<0.001
Year × System × Species	9	1.8	0.072	9	2.1	0.027
Year × System × Cuts	3	0.5	0.67	6	5.2	<0.001
Year × Species × Cuts	12	18.4	<0.001	24	11.8	<0.001
Year × System × Species × Cuts	12	0.9	0.49	24	2.0	0.004
Error	448			672
Corrected Total	511			767

df = degrees of freedom; NWSG = native warm-season grass.

). Then, a separate ANOVA compared the timing of specific harvest events within a year, and system for two and three cuts only. Dropping the single-cut system, which is typically for feedstock production, from the second ANOVA allowed a better comparison of the two practical forage harvest regimes. Means were compared by the Fisher’s Least Significant Difference test at α = 0.05.

3. Results and Discussion

The ANOVA results showed highly significant (p < 0.001) main effects of year, species, and harvest frequency as well as their two- and three-way interactions on the cumulative forage biomass (Table 2).

Likewise, ANOVA for the two- and three-cut systems showed highly significant main effects of year, species, harvest frequency, and harvest timing as well as their two-way interactions (p < 0.001) but at slightly lower p-values (p < 0.03) and (p < 0.004) for the three- and four-way interactions, respectively (Table 3). In both Table 2 and Table 3, the observed year differences in yield are partly attributable to variations in recorded monthly precipitation amounts and average temperatures (see Figure 1). As well, year differences in the actual harvest dates coupled with weather conditions supported the rates of recovery growth and weed competition differently. The differential species yield responses to the defoliation frequencies were partly attributable to inherent morphological differences in relative location of growing points [14,15,16]. Such differences afforded GG relatively faster recovery growths owing to its rather horizontal leaf orientation, minimizing losses to most growing points.

Because of significant factor interactions, treatment means within species and for each harvest regime are presented separately for each year of production.

3.1. Cumulative Biomass Production

Within a year, the recorded mean cumulative biomass values, by species, year, and harvest regimes, are summarized in

Table 4. Cumulative annual forage biomass of newly established NWSG † stands harvested once (X1), twice (X2), or thrice (X3) per year frequency (FRQ), over four consecutive years (Same), or with X1 and X3 frequencies flipped during the second year followed by a single fifth-year June harvest in 2018.

Year	FRQ	Species and Harvest Regimes
		BB		GG		IG		SG
		Flipped	Same	Flipped	Same	Flipped	Same	Flipped	Same
		-----------------------------------------Mg DM ha−1-------------------------------------------
1st	X1	12.93 aA‡	11.71 aA	11.27 aA	10.33 aA	22.69 aA	15.7 aB	27.91 aA	18.82 aB
	X2	9.4 bA	9.39 bA	8.16 bA	8.94 aA	11.47 bA	11.16 bA	12.22 bA	13.15 bA
	X3	5.05 cA	5.31 cA	5.38 cA	5.45 cA	7.18 cA	6.73 cA	8.97 cA	7.58 cB
	p > F α#	<0.001	<0.001	0.001	0.002	<0.001	<0.001	<0.001	<0.001
2nd	X1	11.65 aA	10.85 aA	9.27 aA	8.17 aB	9.64 aA	7.92 aB	12.48 aA	10.55 aB
	X2	7.57 bA	7.17 bA	7.17 cA	6.86 bA	6.31 bB	7.19 bA	8.66 cA	6.41 bB
	X3	10.42 aA	6.84 bB	8.78 bA	6.80 bB	5.26 cA	5.48 cA	11.49 bA	5.59 cB
	p > F	<0.007	0.009	0.002	0.125	<0.001	0.002	<0.001	<0.001
3rd	X1	10.89 aA	11.24 aA	9.24 bB	11.39 bA	16.66 aA	13.18 aB	23.44 aA	26.88 aA
	X2	11.53 aA	11.77 aA	11.94 aA	13.74 aA	8828 bB	10.63 bA	15.90 bA	15.90 bA
	X3	11.32 aA	9.84 aA	12.16 aA	12.54 aA	9.25 bA	9.33 bA	11.56 cA	10.18 cB
	p > F	0.1805	0.108	0.007	0.489	<0.001	<0.001	<0.001	<0.001
4th	X1	10.73 aA	9.70 aA	11.39 bA	10.57 bA	9.29 bA	9.73 bA	28.47 aA	30.79 aA
	X2	11.39 aA	11.19 aA	13.05 aA	12.11 aA	9.91 abA	10.29 aA	18.01 bA	18.00 bA
	X3	9.64 bA	8.17 bA	11.04 bA	9.59 bB	10.98 aA	10.10 aA	13.48 cA	11.62 cB
	p > F	0.733	0.05	0.001	0.001	0.05	0.054	<0.001	<0.001
5th	X1	17.99 aA	13.34 aB	9.95 aA	8.49 aA	5.63 aA	5.95 aA	13.86 aA	9.44 aB
	X2	19.79 aA	11.20 abB	10.07 aA	8.13 aB	5.63 aA	6.36 aA	15.06 aA	7.67 abB
	X3	12.84 bA	10.32 bA	8.45 bA	7.41 aA	5.16 aA	5.36 aA	10.21 bA	5.96 bB
	p > F	<0.001	0.034	0.035	0.46	0.69	0.35	0.002	0.01

† Native warm-season grasses: BB = big bluestem (Andropogon gerardii), GG = eastern gamagrass (Tripsacum dactyloides), IG = indiangrass (Sorghastrum nutans), and SG = switchgrass (Panicum virgatum). ‡ Means in one year followed by the same lower-case letter within a column or of any species followed by same upper-case letter, within a row, are not statistically different at α = 0.05. # Probability of mean difference between harvest frequencies within a year.

. Likewise, for each harvest year, the mean biomass values are described separately for each harvest frequency. From the single-cut strips and for most species, the second-year biomass values were significantly lower than the first and from the third year on. Cumulative two-cut yields showed mostly similar trends to the single cuts, but not the three-cut strips. In strips experiencing the same harvest regime every year, the second-year three-cut yields were significantly lower than the single and double cuts for IG and SG but not for BB or GG. Except for one cut in SG and IG and three cuts in SG, the cumulative biomass values recorded in the first harvest year (2014) showed no statistical difference between the two blocks (Table 4). Apart from the unexplained block yield variation mentioned for SG and IG above, the absence of differences in the first-year yields, within treatment, was expected. This was because all had experienced a common prior defoliation management. Among the single-cut strips, the cumulative biomass values averaged from as low as 10.8 and 12.3 Mg DM ha⁻¹ for BB and GG, to as high as 19.2 and 23.4 Mg DM ha⁻¹ for IG and SG, respectively. For BB and GG, the single-cut yields were slightly over twice as high as from their respective three-cut values and even thrice as high for IG and SG (p < 0.001). Among the respective three-cut strips, pulled averages ranged about 5.3 Mg DM ha⁻¹ for BB and GG, and nearly 7 and 8 Mg DM ha⁻¹, for IG and SG, respectively. Although the two-cut yields were significantly (p < 0.001) lower and higher than their respective single- and three-cut counterparts, the values tended more to the single for BB and GG and to three cuts for IG and SG. Having the within-year cumulative biomass values in decreasing order, single cut > two cuts > three cuts, was consistent with reported yield responses to defoliation by early-succession NWSG stands [6,17,29,30]. In a more recent study with little bluestem (Schizachyrium scoparium Michx.), IG, and BB, Rushing et al. [31] found that, for a three-year stand, more frequent harvests resulted in reduced yields and less frequent harvesting is required to sustain yields. Similarly, McIntosh et al. [32] reported that cutting BB, IG, or SG stands once for biomass at end of the season produced 50% more yield compared to a two-harvest system for forage that included a first cut at the boot or seed head stage. Therefore, the yield magnitudes in this study were reasonable, given the stands were only in their third year of establishment.

During the second year, flipping the harvest frequencies for the single- and three-cut strips resulted in significant (p < 0.05) changes in yield performance (Table 4). Three-cut strips flipped to one-cut in the second year exhibited significant (p < 0.05) yield increases that reached more than 30%. In fact, for BB, cumulative forage biomass was more than doubled from 5.05 Mg DM ha⁻¹ to 11.65 Mg DM ha⁻¹. One-cut strips that were flipped and cut thrice in the second year saw significant (p < 0.05) reduction in biomass yield, reaching more than 60% in SG and IG. The dramatic increase in yield achieved by flipping the three cuts in year 1 to a single cut in year 2 can be attributed to a possible defoliation-induced stimulation of dormant tiller buds and larger crowns formed during the first year. This could have resulted in greater tiller density, and a more dense and robust growth in year 2, which with extended growth before the single cut, led to the greater yields. Having larger crowns also enabled the stands to close canopies early in spring, which choked off more weeds and thus availed a favorable share of resources for regrowth of the perennial grasses. Skipping the early- and mid-season harvests allowed the stands enough recovery time to compensate for their lower prior-year energy reserves and thus catch up with those on their second single-cut year. As for the reduction in biomass yield upon flipping the one to the three cut, the observation may have been due to reduced performance prior to the second and third cuts. After the first cut, grass regrowth takes place during mid and late summer, periods usually associated with high temperatures and moisture stress, which might have affected plant regrowth. Similar severe reductions in tiller numbers in spring are reported from BB stands harvested three times in the prior year and not for those harvested twice or less [12,33].

The flipping effects are compared within years among paired strips (columns) in Table 4. The potential increase in growing points (buds) due to multiple cuts in the prior year, coupled with increased energy reserves due to the flip to a single cut, may explain the better performance of the flipped three-cut strips against their non-flipped three-cut counterparts, and its resultant biomass yields, which are comparable to those continuing the single cut. Except for BB, all single-cut yields from strips harvested thrice in the first year outweighed their counterparts harvested the same for a second year. Except for IG, the effects were similar among the three-cut systems with strips harvested only once in the first year, outyielding those in a second year on the same frequency by 34% (BB) and 22% (GG), to as high as 51% for SG. In fact, for these three native grasses, the second-year yield from current three-cut strips harvested only once in the first year outnumbered their respective two cuts by up to 27%. The reverse was true for the strips that did not change harvest frequencies.

During the third harvest year, yields from the single-cut strips were similar in magnitude to those of the second year for BB and GG but not IG or SG, whose values were larger by about 6 and 13 Mg DM ha⁻¹, respectively. Although all two-cut yields, which ranged from 8.8 Mg DM ha⁻¹ (IG) to 15.8 Mg DM ha⁻¹ (SG), were notably greater than their respective second-year performances, the increases were uniquely larger in SG plots (~8 Mg DM ha⁻¹) compared to about 4 Mg DM ha⁻¹ for other grasses. Forage biomass from the three-cut strips was statistically similar to the respective two-cut strips, except for SG, whose values trailed the two-cut’s performance at <4 Mg DM ha⁻¹. During the third year of harvest, there were no more yield differences due to flipping of harvest frequencies, except for SG, whose flipped three-cut strips produced 1.4 Mg DM ha⁻¹ more forage biomass. The fact that both the single- and three-cut third-year yields tended to match their respective second-year performances suggests that the positive flipping effects on the NWSG vegetative growth performance may not be lost in a single harvest year. Additionally, the increased biomass yield for three-cut systems, which maintained the same harvest frequency, can be attributed to continued maturity of the grasses since establishment.

The fourth year biomass yields, except for GG and SG, did not exhibit significant differences between harvest regimes (Table 4). In all grass species under the single- and two-cut systems, the yield values for the non-flipped and flipped cut frequencies were comparable. Among the three-cut strips, however, the flipped ones still outyielded their counterparts by about 1.5 Mg DM ha⁻¹, but only in GG and SG plots. The two-cut yield increases, through the fourth year, also outweighed their respective three-cut values for all species except IG. The absence of a flipping advantage in the fourth-year yields seems to accentuate the importance of strategic defoliation management decisions in ensuring desirable subsequent growth responses. Furthermore, the data indicate that, while defoliation-induced NWSG yield performance responses may be observable beyond the treatment year, they may be lost over two consecutive harvest years, depending on prevailing growing conditions and subsequent defoliation management. It is reported that at the initiation of harvest frequency treatments, plots harvested three times initially produced yields similar to those harvested once for all species [34]. However, two years after, plots harvested three times per growing season produced significantly lower yields than those harvested once. Multiple defoliations are also known to increase forage biomass during the first year but lead to reduced biomass after three years [33,35].

3.2. Sustainability of Yield Responses

In another assessment of how prolonged recovery growth might affect forage productivity, mean comparisons were performed on yields from a single late-June harvest in the fifth year (2018). All four species showed no significant (p > 0.05) differences in forage biomass between those that received a single or two cuts the previous year, and a majority in both outperformed those that were cut thrice the previous year (Table 4). The observed drastic reduction in forage biomass for the three-cut systems probably resulted from longer-term costs of short-term allocation of biomass to new rhizomes, similar to earlier reports on SG [36]. For IG and non-flipped GG strips, the non-significant numerical yield magnitudes observed were attributable to variations in species sward structure and growth stages at harvest. As well, the biomass values also showed no treatment effects (p > 0.05) in the IG plots or between non-flipped GG strips across harvest regimes. Except for BB, the late-June biomass values seemed notably below their respective fourth-year records—not surprising for a mid-season harvest. However, within treatment, yield differences between strips on flipped and same harvest frequencies were only significant in BB and SG stands. Although, magnitude wise, the single late-June harvest, recorded about 10 days past the harvest date for first cuts in the other years, could not represent the year’s total, it still suited the stand recovery assessment.

The observed similarities in the fifth-year mid-season yields for the single- and two-cut systems suggest that the ≥10 d delay in the first harvest event allowed enough time for compensatory growth to overcome their differences in past defoliation on early-summer stand vigor. This often results from combined effects of increased tiller density in the more frequently harvested stands and reduced vegetative growth on their less frequently harvested counterparts, which experienced sooner preferential allocation of resources to root growth and self-shedding [17,37,38,39]. The data also show that for the two-cut systems, the >10 d harvesting delay was still within the active vegetative growth but long enough for the single cuts to transition into the seed-setting and energy reserve phase. Furthermore, delaying the first harvest during a time when growing conditions (soil moisture and nutrients) were highly favorable also made it possible for compensatory growth to mask residual treatment effects on biomass yields. However, the sustained inferiority of the three-cut systems to their respective single- and two-cut systems in all except the IG plots suggests that the severity of preceding defoliation damage may compromise the reliability of compensatory growth to make up for the lost early-season vigor within a year of imposing harvest treatments. The demonstrated species differences in compensatory growth potentials underscores the importance of implementing species-specific defoliation management plans.

3.3. Differential Species Responses

In designing the sustainable defoliation management of NWSG stands, one must also consider the species’ ability to withstand intensive defoliation. In this study, the results (Table 4) did confirm differences in species growth responses to defoliation. Usually, differential species growth responses to defoliation are due to unique inherent morphology, strategic repairs of tissue damage, and restoration of lost physiological functions. For the tall-growing native grasses, and depending on the severity of sustained injuries, this may involve preferential allocation of stored reserves to replace lost leaf area, increased tillering, and/or photosynthesis rates on the residual or regrowth tissues [17]. The observed inconsistencies in GG yield patterns suggest that the interaction of environmental and species-specific factors may be dominant in influencing yield. The GG forage biomass values suggested a preferential allocation of energy reserves towards restoring lost leaf area and higher rates of photosynthesis on the residual leaf bases. The prostrate stem orientation in GG helps to keep a larger proportion of growing points below the set cutting height, making the growing points inherently more likely to escape defoliation damage and the grass less dependent on new tillers for recovery. Similar insensitivity of GG to frequent defoliation has previously been reported [40]. In this study, while the likelihood of IG growing points precluding defoliation damage is relatively lower, its uniquely loose crowns logically allowed the replacement tillers better access to solar radiation and, therefore, supported relatively better and faster stand recovery rates than in the BB and SG stands. The yield patterns in the study also show that the stands’ biomass production after two successive post-flip harvest years was more reflective of within-year management than the residual effects of their second-year harvesting.

The observed species differences in growth responses to common harvest regimes stresses the importance of paying attention to stand recovery in subsequent defoliation management for sustaining the desired species composition. However, while uniformity in post-defoliation stand recovery is desirable in forage production, the reverse is often the case for restoring ecosystem services, where heterogeneity in vegetation structures is preferred. In addition, the observed species differences in potential compensatory growth calls for a recognition of and a need for a differential approach for both harvest frequency and for the time period prior to resting a given grass species.

3.4. Forage Yield Dynamics in Three-Cut Systems

In order to assess the likely effects of second-year defoliation management on the NWSG subsequent forage productivity, mean comparisons within and between harvest-regimes were performed. Due to significant year–species–harvest regime interactions, the results for the first- through fourth-year harvests are summarized by species and harvest regimes in

Table 5. Forage biomass at the first, second, and third harvest from newly established NWSG† stands harvested once or thrice during the first year (2014) and thereafter flipped to thrice year⁻¹ (X1,3) or continued the same (X3,3) for three more years.

Year	Cut	Species and Harvest Regimes
		BB		GG		IG		SG
		Once	Thrice	Once	Thrice	Once	Thrice	Once	Thrice
		----------------------------------------------Mg DM ha−1----------------------------------------------
1st	1st	NA	1.25 b	NA	0.84 c	NA	1.16 c	NA	1.21 b
	2nd	NA	1.16 b	NA	1.20 b	NA	1.789 b	NA	1.27 b
	3rd	12.93	2.91 a	11.27	3.40 a	2.27	3.87 a	2.79	5.09 a
		X1,3	X3,3	X1,3	X3,3	X1,3	X3,3	X1,3	X3,3
2nd	1st	6.99 aA	3.67 aB	5.11 aA	3.62 aB	1.34 bA	1.43 bA	6.50 aA	1.82 bB
	2nd	2.77 bA	2.63 bA	2.64 bA	2.44 bA	2.63 aA	2.93 aA	3.08 bA	2.49 aB
	3rd	0.66 cA	0.55 cB	1.02 cA	0.74 cB	1.28 bA	1.13 bB	1.90 cA	1.28 cB
3rd	1st	6.93 aA	5.31 aB	7.15 aA	6.81 aA	3.08 bA	3.10 bA	5.15 aA	3.69 bB
	2nd	4.26 bA	4.37 bA	4.17 bA	4.76 bA	5.03 aA	5.04 aA	4.72 aA	4.98 aA
	3rd	0.12 cB	0.15 cA	0.84 cA	0.97 cA	1.15 cA	1.18 cA	1.62 cA	1.50 cA
4th	1st	6.64 aA	5.51 aA	7.23 aA	5.95 aB	6.23 aA	5.88 aA	9.05 aA	7.60 aB
	2nd	2.84 bA	2.49 bA	2.97 bA	3.09 bA	4.12 bA	3.37 bB	3.14 bA	2.85 bA
	3rd	0.16 Ac	0.17 cA	0.84 cA	0.79 cA	0.66 cA	0.75 cA	1.21 cA	1.16 cA
	p > F α#	<0.001	<0.001	<0.001	<0.001	<0.001	<0.001	<0.001	<0.001

† Native warm season grasses: BB = big bluestem (Andropogon gerardii), GG = eastern gamagrass (Tripsacum dactyloides), IG = indiangrass (Sorghastrum nutans), and SG = switchgrass (Panicum virgatum). Means in one year followed by the same lower-case letter within a column or same upper-case letter within a row, for the same species, are not statistically different at α = 0.05. # Probability of mean difference between harvest frequencies, within a year.

. However, the biomass in the first cut row in the first harvest year for plots that are to be flipped in the second year represents the respective total annual yield under a one cut/year regime. For each species, the first-year forage biomass at the third harvest was significantly greater than any of the earlier two, which, for BB and SG, were also statistically similar. For the third harvest, SG produced the greatest biomass yield, about 5.1 Mg DM ha⁻¹, followed by IG (3.9), GG (3.4), and BB (2.9). In the second year during which the harvest frequencies were flipped, forage biomass values were the least, at the third harvest, ranged from as low as 0.6 Mg DM ha⁻¹ (BB) to nearly 2 Mg DM ha⁻¹ (SG). With an exception of the first and second harvests for IG, the second-year yields were significantly greater from strips harvested only once in the first year than their counterparts harvested thrice.

While the second-year forage biomass from all BB and GG as well as the flipped SG strips was significantly greater at the first than the second harvest, the reverse was true for all IG and the non-flipped SG strips. In similar ranking patterns, the third-year forage biomass at the first and second harvests was also substantially greater than the corresponding second year values. By the fourth year, however, yield differences due to the flipping of harvest frequencies had mostly disappeared, except for GG and SG at the first harvest and IG at the second. The yield values within species and harvest regimes were also consistently ranked in the order first > second > third harvest. This is consistent with the findings of Rogers et al. [41], who reported that some warm season grasses, including SG, achieved more than one-third of the total biomass during the first cut. These within-year yield patterns were also consistent with growing conditions (soil moisture, microbial activities, nutrient releases, and temperatures) being better for the NWSGs early on than later in the season, when combinations of drought, temperature, and nutrient depletion are relatively harsh. Additionally, early-season stand biomass includes relatively bigger proportions of annual weeds, which occupy voids between the bunch grasses; this effectively makes up for the loss in vigor for the NWSGs. In late-season, proportions of the annual weeds are mostly negligible, and stands have more voids that further contribute to evaporative moisture losses and thus exacerbate moisture stress conditions, with impact on plant growth.

3.5. Forage Yield Dynamics in Two-Cut Systems

To assess the sustainability potentials of the two-cut forage systems, species biomass weight means for the first and second harvests were compared within year and between years (

Table 6. Sequential forage biomass production of two-cut systems in newly established NWSG† stands recorded for four consecutive years starting June 2014.

Year	Cut	Species and Harvest Regimes
		BB	GG	IG	SG
		----------------------Mg DM ha−1-----------------------
1st	1st	5.62 aC‡	4.57 aB	6.29 aA	7.36 aB
	2nd	3.77 bA	4.37 bA	4.87 bB	5.86 bC
2nd	1st	4.45 aD	4.95 aB	2.27 bD	2.29 bC
	2nd	2.71 aB	1.91 bB	4.91 aB	4.12 aD
3rd	1st	7.74 aB	9.01 aA	3.14 bC	7.66 bB
	2nd	4.02 bA	4.67 bA	7.49 aA	8.24 aA
4th	1st	8.87 aA	7.83 aA	5.71 aB	11.52 aA
	2nd	2.33 bC	4.28 bA	4.58 bB	6.48 bB
	p > F α#	<0.001	<0.001	<0.001	<0.001

† Native warm season grasses: BB = big bluestem (Andropogon gerardii), GG = eastern gamagrass (Tripsacum dactyloides), IG = indiangrass (Sorghastrum nutans), and SG = switchgrass (Panicum virgatum). ‡ In a column, means followed by the same lower-case letter within a year or same upper-case letter for the same cut, are not statistically different at α = 0.05. # Probability of mean difference between harvest frequencies, within a year.

). For the next two harvest years, only BB and GG registered greater (p < 0.001) forage biomass from the first cuts. Biomass was greater only for the first cut during the last year of study in IG and SG. By the fourth year, percentage forage biomass deficits between the first and the second cuts ranged 20% (IG) to 73% (BB). Averaged across harvest regimes, all fourth-year first cuts surpassed their respective first year records by 3 to 5 Mg DM ha⁻¹. While, for GG, these fourth-year yields more or less equaled their third-year performances, they were greater by nearly 2 Mg DM ha⁻¹ for BB, and twice as much for IG and SG. In general, for all grasses, total biomass increased progressively, with biomass in the third and fourth years being of greater magnitudes than the first two years. This progression is consistent with stand establishment and crown expansion since establishment.

Unlike the three-cut systems, the two-cut systems were relatively free of annual grasses or short-growing broad leaf weeds. Although a few plots had Canadian Horseweeds (Conyza canadensis (L.)) or Dog Fennel (Eupatorium capillifolium (Lam.)), their proportions in the biomass were insignificant, implying that the timing of the mid-season harvest kept them under control. Therefore, the observed annual increases in the forage biomass production may be attributed to crown expansion and tiller density. The yield increases were also attributable to their extensive root systems, which exploited larger soil volumes in absence of weed competition and, therefore, allowed a better stand and potential for more yields in later years (Table 6). Similar observations on SG, BB, and IG have been reported, where by the third year, yields increased by 113, 67, and 89%, respectively [34]. In addition, delaying the first cut for up to 2–3 weeks in two- compared to three-harvest systems increased yields by more than 20%, an observation similar to that found in other NWSG studies [32]. As for the three-cut systems, having greater biomass (up to 50% or more of the total) from the two-cut strips at the first rather than the second harvest was consistent with growing conditions being more conducive for faster growth early on than later in the growing season. It has previously been observed that warm season grasses like SG and Bermuda grass can produce more than one-third of the total biomass during the first harvest in late spring [41]. This is probably due to better soil moisture conditions during this time, when evaporative demand is low. Additionally, vegetative regrowth in the second half, though slow, produces substantial biomass. There may also be preferential allocation of assimilates for seed setting, initiation of dormant tillers, and belowground energy reserves for growth during the following spring growth season.

4. Conclusions

The study demonstrated the ability of strategic defoliation management of NWSG stands to induce desirable biomass yield responses for forage and feedstock production. For GG, IG, and SG, the prolonged recovery with changing three cuts into a single season-end cut optimized the prior-year increases in tiller numbers, which compensated for prior shortages in energy reserves. The reverse was true for the single-cut strips flipped into three cuts, with over 60% yield reduction for IG and SG. All four grasses in the study exhibited third-year yield increases from the single-cut strips that were on a three-cut regime during the first year, although the magnitudes varied among species.

The fourth-year yield differences between corresponding strips on flipped and same harvest regimes were only significant for GG and SG, whose three-cut strips maintained a slight flipping superiority. Furthermore, the overwhelming leveling of performance during the fourth year indicates that, while defoliation-induced NWSG yield responses may last beyond the treatment year, they may disappear after two consecutive harvest seasons of an altered harvest regime. Additionally, the trends and magnitude of NWSG yield responses to defoliation reflect the within-year harvest management approach and the prevailing growing conditions. Compared to the three-cut systems, harvesting twice per year suppressed the proportion of annual weeds in the mid-season regrowth, which favored better stands and potential for subsequent yield increases.

The general continued inferiority of the three-cut strips to their respective single- and two-cut systems in the fourth and fifth year biomass yields indicated that defoliating NWSGs excessively might compromise the reliability of compensatory growth or increases in tiller numbers and crown diameter to make up for losses in stand vigor. However, reverting to a one-cut- from a three-cut-per-year harvest management of NWSG will restore the stand productivity and increase forage yields. As well, most yield increases will be to levels comparable to those of one cut and two cuts per year within 1–2 years. Therefore, a three-cut system, the management method that ensures better quality forage, may be feasible provided a 1–2 year rest/rotation with one-cut per year is included to restore stand productivity. It is also worth noting that the effectiveness of strategic defoliation management in inducing the desired NWSG growth responses is more dependent on the timing of the harvest events with respect to growing conditions than the actual duration of recovery intervals.