Forage Production and Sward Structure Dynamics of Tall Fescue (Lolium arundinaceum) Pasture Grazed to Different Sward Heights

Abstract

Sward structure and post-grazing heights (SH) significantly influence plant growth and animal intake, crucial for dairy grazing systems. However, these interactions are dynamic and vary with season, resource heterogeneity, and defoliation patterns. Seasonal effects of control (TC), medium (TM), and lax (TL) post-grazing SH of grazed Lolium arundinaceum-based pasture on forage production and utilization, herbage mass, green cover, and chemical composition were tested during autumn-winter and spring seasons and among tall (TP), medium (MP), and short (SP) patches in spring. Thirty-six lactating Holstein cows were randomized evenly to TC, TM, and TL grazing treatments to achieve 6, 9, and 12 cm of post-grazing SH during autumn-winter, and 9, 12, and 15 cm in spring. Forage production was higher on TL than TM and TC, yet utilization was similar across all treatments. The TP relative to MP on SP increased for TL compared to TC and TM. The TP-TC presented higher leaf-density and leaf-proportion, than TP-TL, without modifying leaf canopy distribution of superior-medium horizons among treatments. Grazing management modulated forage production and structural heterogeneity across SH treatments. Critically, monitoring patch-level dynamics—rather than mean height—is essential for optimizing production and harvest efficiency in temperate systems by improving grazing horizon accessibility.

1. Introduction

Cattle grazing creates heterogeneous mosaics of vegetation patches with structural characteristics varying according to the season of grazing [1], grazing system, and stocking rate [1,2,3]. Animals grazing on heterogeneous pastures are faced with dietary choices that allow selecting patches with contrasting structural and chemical attributes [4]. Grazing is a two-way dynamic process [5] driven primarily by temporal and spatial differences in the herbage mass, nutritional value, accessibility, and spatial distribution of vegetation patches [1,2,3,6]. Plant growth, and plant covariate associated with structural traits and forage chemistry are significantly altered through selective grazing [2,7]. Likewise, patch selective grazing [8] can create, augment, or even revert patterns of spatial heterogeneity [9], thereby affecting plant growth and forage production. This structural heterogeneity acts as a primary regulator of forage production; by selectively altering the sward structure, grazing animals modulate the Leaf Area Index (LAI) and the efficiency of radiation interception, creating a mosaic of patches with contrasting photosynthetic capacities. Similarly, animals grazing heterogeneous pastures can hierarchically alter feeding behaviors at specific spatio-temporal scales to increase selectivity and enhance nutrient intake rate [10,11]. In Uruguay, previous research has indicated that increasing post-grazing sward height presents an increase in tall patches that affected the selective behavior of dairy cattle [1]. Furthermore, it was observed that the structural characteristics and nutritive value of these tall patches do not represent a limitation for animal intake, suggesting they remain high-quality resources within the grazing mosaic.

Tall fescue [Lolium arundinaceum (Schreb.) Darbysh.] is a perennial grass widely used for dairy and livestock grazing on the west coast of Uruguay. Tall fescue has high potential for forage yield, but poor grazing management often limits the productivity and feed value of tall fescue in the region [12,13]. The intensity of defoliation, either too low or too high, can set either high or low post-grazing residual sward heights that affect leaf area index and radiation intercepted by pasture canopy and senescence rate, thereby affecting plant growth and forage production [14]. Nevertheless, intermediate defoliation intensities can be explored to optimize the plant-animal interface. Moderate grazing intensities—characterized by increased post-grazing heights—can maintain sufficient leaf area to favor both nutrient intake and photosynthetic capacity. In this sense, there is potential to optimize the management of post-grazing sward heights in order to increase and stabilize both tall fescue forage and milk production in Uruguay through taller post-grazing sward heights (12 to 15 cm) under moderate intensities [15,16,17]. However, in grazed tall fescue pastures in Uruguay, increasing spatial heterogeneity of sward height or herbage mass may exert more significant effects on plant growth and herbage intake than the use of post-grazing sward height or herbage mass alone [18]. Resource heterogeneity is common in most herbivore-plant–soil systems, including temperate pasture systems that are used for livestock and dairy production in South America [19]. The effects of resource heterogeneity on the ingestive behavior and performance of dairy cows have been the focus of important research in recent decades [18], yet references for practical monitoring and management of spatial heterogeneity in commercial production systems are lacking [20]. Exploring moderate grazing intensities requires an evaluation of the integrated response between animal performance and sward structure. Understanding how seasonal variables explain favorable outcomes from both the plant and animal perspectives is essential. This approach is fundamental to maximizing pasture utilization by increasing forage production and effectively capitalizing on it through enhanced animal intake. Therefore, an experiment was conducted to evaluate seasonal effects of post-grazing sward heights of tall fescue pasture on (1) forage production and utilization, (2) variation in herbage mass, canopy green cover, and chemical composition, (3) and effects on patch heterogeneity on spring grazing season. The hypothesis was that taller post-grazing sward height favoring a greater growth rate and forage yield of tall fescue pasture will result in a greater frequency of tall vegetation patches offering greater green leaf vegetation cover in spring. It was also hypothesized that a greater selectivity for taller vegetation patches will favor nutrient intake and milk production by cattle.

2. Materials and Methods

2.1. Study Site

The experiment was carried out at the Experimental Station ‘‘Dr. Mario A. Cassinoni’’ (EEMAC), Agronomy Faculty, Paysandú, Uruguay, (32° S, 58° W; 51 m a.s.l.), from April to December of 2017. In May 2016, a tall fescue (cv. INIA Fortuna) pasture was sown using a seeding rate of 10 kg ha⁻¹. At the time of sowing, the pasture was fertilized with 18 kg of N and 46 kg of P ha⁻¹. In the second year of production, 180 kg of N ha⁻¹ was applied in four equal applications from June to October. In February of 2017, the sward was mowed mechanically to 6 cm height (from ground level) and the grazing treatments were implemented in April of 2017.

2.2. Treatments and Experimental Design

Three post-grazing sward heights were tested in a randomized complete block design with 4 blocks, each measuring 0.6 ha. Each block was characterized by homogeneous soil conditions and consisted of 3 paddocks, each measuring 0.2 ha. Within each block, paddocks were randomly assigned to either the control treatment (TC), medium treatment (TM), or lax treatment (TL). Heights for control (TC), medium (TM), and lax (TL) post-grazing sward height treatments were 6 cm, 9 cm, and 12 cm for the autumn-winter season, and 9 cm, 12 cm, and 15 cm for the spring season, respectively (Figure 1). The different post-grazing sward heights were achieved by maintaining cows in the paddock areas until the target sward heights were obtained. Paddocks were grazed from April to July (autumn-winter season), and from October to December (spring season) by the same dairy cows assigned to treatments at study onset. Grazing started when three fully extended new leaves per tiller were developed [21]. From August and September, all paddocks were grazed down twice (Figure 1) to a 5 cm sward height (from ground level) using growing dairy heifers [13]. This management practice was implemented for early flowering control (EFC) of elongating tall fescue tillers.

2.3. Animals and Management

Each paddock was grazed by 3 milking cows throughout the trial. A total of 36 autumn-calving Holstein cows were blocked by body weight (BW), body condition score (BCS), and parity number, and randomized to the treatments. The cows had a mean BW of 593 ± 54 kg, a BCS of 3.0 ± 0.2 on a 1–5 scale, and a mean parity of 2.6 ± 0.8. During the autumn-winter season, cows were individually fed at 18:00 h a mixed ration of corn silage (28.0 kg), commercial concentrated (5.0 kg), ground sorghum grain (4.5 kg), sunflower meal pellets (2.0 kg), and alfalfa haylage (1.5 kg), all on a fresh weight basis, and allowed to graze daily between 8:00 and 14:00 h. During the spring season, no mixed ration was offered and cows were allowed to graze daily between 8:00 and 14:00 h and 17:00 and 03:00 h. Cows were milked at 4:00 and 15:00 and were kept as a single herd while they were not grazing in the experimental paddocks.

2.4. Sward Measurements

Leaf stage of tall fescue tillers was conducted weekly throughout the study [21]. Measurements of sward height (SH), herbage mass (HM), total green cover (TGC), and fescue green cover (FGC) were recorded in each paddock both pre- and post-grazing throughout the experimental period (Figure 1). Consequently, pasture production, HM utilization and chemical composition (%NDF, %ADF, %CP) were determined for every grazing cycle in both autumn-winter and spring seasons. Additionally, during the spring season, sampling was expanded to include patch spatial heterogeneity (Short, Medium, and Tall patches) and a detailed characterization of patch canopy structure (Figure 1).

Measurements of SH (±1 cm) were conducted following a modified version of the technique described by [22] and by using a minimum of 70 readings of SH evenly spaced across four permanent transects defined in each paddock at study onset. Herbage mass (HM) was estimated both, pre- and post-grazing (Figure 1) by a rising plate meter (RPM; Ashgrove Co., Palmerston North, New Zealand). On average, 70 RPM readings were recorded per paddock during each sampling event. Calibration of the RPM was performed every 30 days during the autumn-winter period and every 15 days during the spring. This process involved recording the compressed SH and the corresponding forage mass across three contrasting SH, measured in triplicate. At each of the nine sites, compressed SH was recorded prior to harvesting the forage to ground level within a 30 × 34 cm frame. The harvested biomass was collected, content weighed, and oven-dried at 60 °C to constant weight to determine dry matter content. Finally, individual dry matter forage mass measurements were regressed against the corresponding compressed sward height values to develop forage mass prediction equations. Additionally, two 0.1 m² quadrats of ungrazed herbage mass per paddock cut to post-grazing sward heights were collected, oven-dried at 60 °C to constant weight, and ground to 1 mm for determination of forage dry matter (DM), Nitrogen (N), neutral detergent fiber (NDF) and acid detergent fiber (ADF) (Figure 1). Percent of total green cover (TGC) and fescue green cover (FGC) of the canopy, both pre-and post-grazing (Figure 1) were assessed visually in twelve 0.3 × 0.5 m quadrats of pasture distributed alongside the four permanent monitoring transects in each paddock.

Characterization of pre-grazing sward structure by patch spatial heterogeneity and patch canopy structure was assessed before the first grazing cycle in spring (Figure 1) to determine residual effects of previous autumn-winter grazing. Each 0.2-ha paddock (50 × 40 m) was sampled using a systematic grid consisting of 16 transects spaced 5 m apart. Along each transect, measurements were taken every 1 m (Figure 2), resulting in approximately 700 measurements of SH and herbage mass by the RPM per paddock, 1 day before the onset of grazing. The SH measurements were used to classify pasture quadrats into tall (TP; >30 cm), medium (MP; 15 to 30 cm), and short (SP; <15 cm) patches (Figure 1). Two representative patches per patch type were selected and used for the collection of herbage samples of 0.1 m² (30 × 34 cm) of pasture cut to ground level. Samples were stratified into a superior stratum (40% total height), middle stratum (40% remaining height), lower stratum (40% remaining height), and residual stratum (remaining height). The TP and MP were separated into 4 stratums and SP into 3 stratums of canopy and stratified samples were divided into leaf, stem, and dead material. Samples were dried at 60 °C to constant weight to determine DM content.

2.5. Analytical Methods

The DM and N of herbage samples were determined according to AOAC [23] (methods numbers 984.13). The neutral detergent fiber (NDF) and acid detergent fiber (ADF) were determined by an ANKOM200 Fiber Analyzer (ANKOM Technology Corp., Fairport, NY, USA) using procedures and detergents described by Van Soest et al. [24] with a heat stable amylase and without sodium sulphite in the NDF detergent solution.

2.6. Calculations

The accumulation rate of pasture (AR, kg DM day⁻¹) was determined by the difference between the post-grazing and pre-grazing HM of two consecutive grazing cycles. In addition, 50% of the measured AR [25] was considered during the period of occupation of paddocks.

Forage production (kg DM ha⁻¹) was determined from the AR and days of regrowth between grazing cycles. Forage utilization (kg DM ha⁻¹) was assessed from the difference between pre- and post-grazing HM in each grazing cycle.

The herbage density (mg cm⁻³) of forage and the density of forage for leaves were determined for each patch type and patch canopy stratum using forage dry weights (mg), canopy stratum depths (cm), and area of clipped quadrants (cm²). The ratio of forage between leaves and stems (leaf-stem) for each patch type and patch canopy stratum was also determined.

The herbage mass of leaf (kg DM leaf ha⁻¹) in each paddock was estimated from the proportion of each patch type, the proportion of leaf in each patch type, and the herbage mass in each patch type.

2.7. Statistical Analysis

Data were analyzed as a randomized complete block design using the SAS software (version 9.4; SAS Institute Inc., Cary, NC, USA) [26]. The paddock was considered the experimental unit. Univariate analyses were performed to identify data outliers and deviations of data from a normal distribution of residuals. Proportion variables not conforming to assumptions for normality were transformed and analyzed following a log-normal distribution of data.

The HM, SH, TGC, FGC, and herbage chemical composition variables were analyzed as repeated measures using mixed models for a randomized complete block design fitted with the GLIMMIX procedure. Mixed models included the fixed effects of treatment, season, and grazing cycle (season) and their interactions, and block as a random effect. For the post-grazing TGC and FGC, the fixed effect of the grazing cycle was not considered, due to incomplete determinations related to inclement weather. The forage production and utilization were analyzed using the GLIMMIX procedure with a mixed model that included the fixed effects of treatment, season and their interaction, and block as the random effect. The herbage mass of leaf per treatment was analyzed, with a mixed model considering the fixed effect of treatment and block as random factors. The patch spatial heterogeneity and patch canopy structure were analyzed with a mixed model that included the fixed effects of treatment, patch type, stratum (patch) and their interactions, and block as the random effect. Least square means were separated using Tukey–Kramer tests with significant differences detected according to a 5%, type I, error and tendencies for differences discussed according to a 10%, type I, error.

3. Results

The study lasted 249 ± 4.6 days. In autumn-winter, the first grazing cycle started on 4 April 2017, and lasted 8, 5, and 2 days for TC, TM, and TL, respectively. The regrowth period until the second autumn-winter grazing cycle lasted for 52, 36, and 39 days for TC, TM, and TL, respectively. The third and final autumn-winter grazing cycle for TM and TL started 48 days after the completion of the second grazing cycle. In spring, the first grazing cycle started 40 days after completion of flower control in all post-grazing SH treatments. The regrowth interval until the second spring grazing cycle was 35 days and was established for the three post-grazing sward height treatments. In total, seven grazing cycles were conducted for TL and TM, and six for the TC.

The prevailing monthly weather conditions during the experiment are presented in

Table 1. Weather conditions and dekadal rainfall distribution during the trial.

	Apr	May	Jun	Jul	Aug	Sep	Oct	Nov	Dec
Mean Temperature (°C)	18	15	14	14	15	16	18	20	25
Accumulated Rainfall
1st dekad (mm)	33.5	40.5	12.8	51.6	82.9	148.0	37.7	41.2	9.9
1st dekad (mm)	37.4	129.7	0.9	2.8	94.9	2.7	79.3	4.1	58.5
1st dekad (mm)	39.9	32.8	6.7	0.5	148.6	52.4	22.5	42.9	30.2
Monthly Total (mm)	111	203	20	55	326	203	139	88	99

Apr, April; Jun, June; Jul, July; Aug, August, Sep, September; Oct, October; Nov, November; Dec, December. A dekad corresponds to a 10-day rainfall recording period.

. The overall and seasonal average temperatures recorded in the trial were similar to historical records for the north zone of Uruguay. However, precipitation showed significant deviations from the 30-year average. While June and July were drier than historical norms, the period from August to October was characterized by excessive rainfall, with August recording more than double the typical monthly accumulation. Consequently, the regrowth intervals observed were closely related to the 10-day (dekadal) rainfall distribution. The extended rest periods during spring (40 days) coincided with a period of excessive soil moisture, driven by high rainfall in October (139 mm) and the first dekad of November (41.2 mm), which likely led to temporary soil saturation. Conversely, during autumn-winter, the 200 mm recorded in May ensured adequate soil water storage, thereby allowing regrowth to proceed under low evaporative demand despite the scarce rainfall (20 mm) recorded in June.

3.1. Herbage Mass and Sward Height

Pre- and post-grazing sward characteristics are presented in

Table 2. Herbage mass and sward height pre- and post-grazing during autumn-winter and spring seasons.

Variable	Season				SEM	p-Value
	Autumn-Winter		Spring
	1	2	1	2		T	S	C(S)	T × S	T × C(S)
Pre-grazing HM (kg DM ha−1)
TC	1163	2034	2717	2559	81	0.3712	<0.0001	<0.0001	0.4635	0.1638
TM	1128	1673	2499	2534
TL	1063	1793	2495	2968
Pre-grazing SH (cm)
TC	15.7	15.8	18.0 a	11.4 b	0.37	0.1975	0.7971	<0.0001	0.1384	0.0264
TM	16.8	14.6	20.1 a	13.1 b
TL	16.3	15.2	17.1	13.9
Post-grazing HM (kg DM ha−1)
TC	443 a	1780 b	1798 a	1818 aB	54	0.0081	<0.0001	<0.0001	0.0274	<0.0001
TM	547 a	1386 b	1681 a	2061 aB
TL	848 a	1463 b	1706 a	2790 bA
Post-grazing SH (cm)
TC	6.3 B	7.5	9.0 B	6.7 B	0.33	<0.0001	0.0157	0.0001	0.8485	0.0245
TM	8.4 B	9.7	12.0 aA	8.1 bAB
TL	11.6 A	10.1	13.0 aA	9.9 bA

Mean values within a row with different lower-case letters differ within each season (p < 0.05). Mean values within a column with different capital letters differ between treatments (p < 0.05). HM, herbage mass; SH, sward heigh; TC, control post-grazing SH treatment; TM, medium post-grazing SH treatment; TL, lax post-grazing SH treatment; S, season of grazing; C, grazing cycle.

. Pre-grazing SH and HM were not different between treatments, averaging 15.7 ± 0.48 cm and 2052 ± 84 kg DM ha⁻¹, respectively. Greater pre-grazing HM was recorded for spring compared to autumn-winter (2628 vs. 1475 kg DM ha⁻¹, respectively). There was a grazing cycle by season interaction (p < 0.0001) for both SH and HM. During autumn-winter, greater pre-grazing HM was observed for the second grazing cycle compared to the first grazing cycle (1833 vs. 1118 kg DM ha⁻¹, respectively). Pre-grazing SH was reduced in the last spring grazing cycle on TC and TM and a similar tendency was observed on TL (p = 0.06).

Post-grazing HM varied significantly according to interactions between treatments, seasons, and grazing cycles (Table 2). During autumn-winter, all treatments had similar post-grazing HM, but HM was higher in the second than in the first grazing cycle. During spring, post-grazing HM was higher for TL compared to the other treatments. The post-grazing SH was affected by a treatment, season, and grazing cycle interaction. In autumn-winter, the post-grazing SH was higher for TL vs. TC and TM in the first grazing cycle. In spring, the post-grazing SH was lower for the second grazing cycle vs. the first grazing cycle for TM and TL, and the same tendency (p = 0.08) was observed for TC. Additionally, during this season, the post-grazing SH was lower on TC than on TL in both grazing cycles. Average post-grazing SH for TC (7.4 cm), TM (9.6 cm), and TL (11.1 cm) were slightly below the targeted SH treatments. During the EFC period, pre-grazing HM and SH averaged 1503 ± 141 kg DM ha⁻¹ and 16.6 ± 1.0 cm, respectively. Post-grazing HM and SH averaged 687 ± 21 kg DM ha⁻¹ and 5.6 ± 0.3 cm, respectively.

3.2. Canopy Green Cover

Results for pre- and post-grazing TGC and FGC are presented in

Table 3. Pre- and post-grazing total green cover (TGC) and tall fescue green cover (FGC) in autumn-winter and spring for control (TC), medium (TM) and lax (TL) post-grazing sward height treatments.

Variable	Season				SEM	p-Value
	Autumn-Winter		Spring
	1	2	1	2		T	S	C(S)	T × S	T × C(S)
Pre-grazing TGC (%)
TC	58.9 b	79.7 a	78.7 a	31.2 bB	3.1	0.0118	0.0847	<0.0001	<0.0001	0.0205
TM	56.8 b	71.6 a	82.7 a	37.8 bB
TL	53.3 b	73.2 a	87.1 a	57.7 bA
Pre-grazing FGC (%)
TC	48.0	74.0	44.0	61.0	5.7	0.9113	0.6366	<0.00001	0.0169	0.4068
TM	47.1	63.0	52.4	63.1
TL	45.2	56.2	50.5	69.9
Post-grazing TGC (%)
TC	38.3 B		43.6 A		3.5	0.1002	0.0134	-	0.0098	-
TM	60.9 A		39.3 A
TL	54.3 AB		44.7 A
Post-grazing FGC (%)
TC	64.5		55.3		4.6	0.4592	0.4564	-	0.0168	-
TM	53.5		57.7
TL	52.6		63.8

Mean values within a row with different lower-case letters differ within each season (p < 0.05). Mean values within a column with different capital letters differ between treatments (p < 0.05). S, season of grazing; C, grazing cycle.

. In autumn-winter, TGC was similar among treatments, and 20.8, 14.8, and 19.8% higher in the second compared to the first grazing cycle for TC, TM, and TL, respectively. In spring, the TGC was lower for the second grazing cycle compared to the first grazing cycle in all treatments. In the second spring grazing cycle, TL had higher pre-grazing TGC than TC and TM. All treatments had 17% greater FGC in the second vs. first grazing cycle in both seasons. The post-grazing TGC varied significantly by a two-way interaction between treatments and seasons (Table 3). The post-grazing TGC was lower (p = 0.013) in spring vs. autumn-winter (42.6 vs. 51.6%). In autumn-winter, the post- grazing TGC was greater for TM vs. TC, and no difference in post-grazing TGC was detected between TC vs. TL or TM vs. TL. In spring, the post-grazing TGC was not different between treatments.

Both TC and TL had similar post-grazing TGC in both seasons, but post-grazing TGC on TM was lower in autumn-winter vs. spring. Post-grazing FGC did not differ between treatments during the entire duration of the study.

3.3. Forage Production and Utilization and Chemical Characteristics

Forage production varied (p = 0.006) among treatments, season (p < 0.0001), and their two-way interaction (p < 0.0001). Forage production was higher for TL (6042 kg DM ha⁻¹) than TM (5365 kg DM ha⁻¹) and TC (5257 kg DM ha⁻¹). Forage production was greater in spring, being 26% greater for TL compared to TC and TM. Forage utilization was similar (4196 kg DM ha⁻¹) among treatments, but greater (p = 0.0005) in spring vs. autumn-winter. Forage utilization was lower for TC and TL (892 kg DM ha⁻¹) compared to TM (1298 kg DM ha⁻¹) in autumn-winter. In contrast, the forage utilization was lower in TC (1296 kg DM ha⁻¹) than TL (1889 kg DM ha⁻¹), with no difference with TM, during EFC.

The forage NDF and ADF differed according to interactions between treatments and seasons (

Table 4. Chemical composition of forage in autumn-winter and spring for control (TC), medium (TM), and lax (TL) post-grazing sward height treatments.

Variable	Season		SEM	p-Value
	Autumn-Winter	Spring		T	S	T × S
NDF (g kg−1)
TC	518 b	617 aA	10.9	0.9935	0.0006	0.0004
TM	542	591 AB
TL	577	557 B
ADF (g kg−1)
TC	318 B	353 A	8.5	0.9059	0.1083	<0.0001
TM	341 AB	328 AB
TL	361 bA	300 aB
CP (g kg−1)
TC	122	109	6.0	0.0554	0.8423	0.2328
TM	125	129
TL	129	140

Mean values within a row with different lower-case letters differ in season (p < 0.05). Mean values within a column with different capital letters differ between treatments (p < 0.05). NDF, neutral detergent fiber; ADF, acid detergent fiber; CP, crude protein; T, treatment; S, season of grazing.

). Forage NDF was greater in spring vs. autumn-winter for TC, but forage NDF for TM and TL was not different between seasons. Differences in ADF were observed between TC and TL in both seasons. The ADF was lower for TC in autumn-winter and lower for TL in spring. The CP was not different between spring and autumn-winter seasons (p = 0.8423), but it was higher (p = 0.0459) for TL vs. TC.

3.4. Patch Heterogeneity and Sward Structure

The proportion of SP, MP, and TP patches exhibited variation due to a significant interaction (p < 0.001) between post-grazing SH treatments and patch types. The proportion of SP was similar among treatments, but the proportion of TP was greater on TL compared to TC and TM (

Table 5. Characteristics of tall fescue short (SP), medium (MP), and tall (TP) patches of tall fescue pasture grazed with lactating dairy cows to control (TC), medium (TM) or lax (TL) post-grazing sward height.

Variable Patch Type	Treatment			SEM	p-Value
	TC	TM	TL		T	P	T × P
Proportion of patch type
SP	0.08 B	0.08 B	0.05 B	0.03	1.0000	<0.0001	0.001
MP	0.64 aA	0.66 aA	0.28 bB
TP	0.28 bB	0.26 bB	0.67 aA
Herbage mass (kg DM ha−1)
SP	2245 aC	2159 abC	1960 bC	8.5	0.9059	0.1083	<0.0001
MP	3668 B	3898 B	3529 B
TP	5191 A	4961 A	5381 A
Sward Height (cm)
SP	14.4 aC	12.9 bC	12.4 bC	0.66	0.1244	<0.0001	<0.0001
MP	24.7 B	25.6 B	23.6 B
TP	38.0 A	38.2 A	41.0 A
Herbage density (mg cm−3)
SP	1.45 B	1.61	1.37	0.05	<0.0001	0.0986	0.0015
MP	1.46 bB	1.83 a	1.43 b
TP	1.85 aA	1.70 a	1.32 b
Proportion of leaf
SP	0.75 A	0.70 A	0.77 A	0.01	<0.0001	<0.0001	<0.0001
MP	0.73 aA	0.60 bB	0.67 aB
TP	0.64 aB	0.55 bB	0.53 bC
Leaf density (mg cm−3)
SP	1.04	1.09 A	1.08 A	0.04	<0.0001	<0.0001	<0.0001
MP	1.04	0.99 A	0.92 A
TP	1.16 a	0.84 bB	0.68 cB
Proportion of stem
SP	0.15 A	0.17 A	0.12 A	0.01	<0.0001	<0.0001	<0.0001
MP	0.18 aA	0.26 bB	0.21 abB
TP	0.24 aB	0.27 aB	0.35 bC
Leaf-stem
SP	5.35 A	4.92 A	6.63 A	0.36	0.0002	<0.0001	0.0013
MP	4.79 aA	2.71 bB	4.14 aA
TP	2.97 aB	2.04 abB	1.75 bB
Proportion of dead materials
SP	0.05	0.10	0.09	0.004	<0.0001	0.002	0.9386
MP	0.05	0.09	0.08
TP	0.04	0.08	0.07

Mean values within a row with different lower-case letters differ between treatments (p < 0.05). Mean values within a column with different capital letters differ between patches (p < 0.05). T, treatment; P, patch type.

). The HM and SH were greater for TP and MP compared to SP and for TP compared to MP (Table 5). Average HM and SH were 2089 kg DM ha⁻¹ and 13 cm, 3632 kg DM ha⁻¹ and 25 cm, and 5104 kg DM ha⁻¹ and 39 cm, for SP, MP, and TP, respectively. Differences in HM and SH among treatments were observed only in SP due to a higher SH on TC compared to TM and TL and a higher HM on TC compared to TL. Herbage density varied by an interaction between treatments and patch types (p = 0.001), explained by the greater herbage density of MP on TM compared to MP on TC and TL, and greater herbage density of TP on TC and TM compared to TP on TL. In the control treatment, TP had greater herbage density than MP or SP (Table 5).

The proportion of leaves, density of leaves, and the leaf-stem forage ratio were affected by a treatment and patch type interaction (p < 0.001). The proportion of leaves for TP was higher in the TC treatment compared to TM and TL treatments, whereas the proportion of leaves for MP was lower in the TM treatment compared to the TC and TL treatments (Table 5). All treatments showed the same proportion of leaves in LP. The leaf proportion was reduced in TP for TC, and for TM and TL it was reduced in MP.

The density of leaves of TP patches was greater in the TC treatment compared to the TM and TL treatments (Table 5). Similarly, the density of leaves for SP, MP, and TP was similar in the TC treatment (1.08 mg cm⁻³ ± 0.07), but the leaf density of SP and MP was greater than TP in the TM and TL treatments (Table 5). The proportion of stems was affected by an interaction between treatments and patch types (Table 5) due to the lower proportion of stems for MP patches in the TC treatment and TP patches in the TC and TM treatments, respectively. The stem proportion was increased in TP for TC, and for TM and TL it was increased in MP and TP.

The leaf to stem forage ratio was highest in the SP patches (5.53 ± 0.50) and lowest in the TP patches (2.19 ± 0.18). The leaf-stem forage ratio was greater for the TC compared to TL patches, but no differences with TM were observed (Table 5). The leaf to stem forage ratio was affected by an interaction between treatments and patch types (Table 5), due to greater leaf to stem forage ratios for MP patches in the TC and TL treatments, and TP patches in the TC treatment, respectively.

The proportions of dead plant material were lower (p = 0.03) for TP vs. SP and a tendency for a similar difference (p = 0.07) was observed between TP vs. MP. The lowest value among treatments was observed on TC vs. TM and TL (0.05 vs. 0.08 ± 0.0009; p < 0.0001).

The HM of leaf per paddock was similar in all treatments (p = 0.198; 2699, 2400, 2695 kg DM leaf ha⁻¹ for TC, TM, and TL, respectively).

The interaction of grazing treatment and stratum within a patch was significant for the proportion of leaves, density of leaves, proportion of stems, and leaf-stem forage ratio (

Table 6. Patch canopy structure by stratum before grazing during the spring season of tall fescue pasture grazed with lactating dairy cows to control (TC), medium (TM), or lax (TL) post-grazing sward height.

Variable	Patch Type	Stratum	Treatment			p-Value
			TC	TM	TL	T	S(P)	T × S(P)
Leaf Proportion	SP	Superior	0.73	0.82 A	0.84 A	<0.0001	<0.0001	<0.0001
		Middle	0.81	0.75 A	0.82 AB
		Residual	0.71	0.53 B	0.65 B
	MP	Superior	0.75 A	0.79 A	0.86 A
		Middle	0.83 A	0.75 A	0.79 AB
		Low	0.77 aA	0.57 bB	0.66 abB
		Residual	0.56 aB	0.28 bC	0.44 abC
	TP	Superior	0.66 A	0.72 A	0.80 A
		Middle	0.74 A	0.70 A	0.57 B
		Low	0.70 aA	0.55 abA	0.48 bB
		Residual	0.46 aB	0.25 bB	0.29 abC
Leaf density (mg cm−3)	SP	Superior	0.99	1.43 A	1.33	<0.0001	<0.0001	<0.0001
		Middle	1.23	1.18 AB	0.98
		Residual	1.04	0.78 B	0.89
	MP	Superior	0.78 B	1.21 A	1.07 A
		Middle	1.08 AB	1.17 A	1.06 A
		Low	1.37 A	1.23 A	1.11 AB
		Residual	1.04 aAB	0.57 bB	0.60 abB
	TP	Superior	0.97	0.97 A	0.77 A
		Middle	1.15	0.98 A	0.72 A
		Low	1.50 a	1.06 abA	0.84 bA
		Residual	1.09 a	0.50 bB	0.38 bB
Stem proportion	SP	Superior	0.13	0.11 A	0.08	<0.0001	<0.0001	0.0008
		Middle	0.13	0.13 A	0.08
		Residual	0.19	0.28 B	0.20
	MP	Superior	0.11 A	0.10 A	0.06 A
		Middle	0.14 A	0.18 AB	0.12 AB
		Low	0.17 A	0.30 B	0.24 B
		Residual	0.30 aB	0.45 bC	0.41 abC
	TP	Superior	0.12 A	0.14 A	0.11 A
		Middle	0.22 aA	0.24 abAB	0.37 bB
		Low	0.24 aA	0.30 abBC	0.41 bBC
		Residual	0.38 B	0.42 C	0.51 C
Leaf-Stem	SP	Superior	6.3	8.7 A	8.6 AB	0.0002	<0.0001	0.0007
		Middle	6.4	6.9 A	10.7 A
		Residual	3.9	2.1 B	3.3 B
	MP	Superior	8.0 A	9.2 A	14.6 A
		Middle	7.2 A	4.9 AB	6.8 AB
		Low	4.9 AB	2.1 B	2.9 BC
		Residual	1.9 aB	0.6 bC	1.1 abC
	TP	Superior	6.0 A	5.5 A	8.4 A
		Middle	3.6 A	3.1 AB	1.7 B
		Low	3.0 AB	1.9 B	1.2 BC
		Residual	1.2 B	0.5 C	0.6 C

Mean values within a row with different lower-case letters differ between treatments (p < 0.05). Mean values within a column with different capital letters differ (p < 0.05). SP, tall fescue short patch; MP, tall fescue medium patch; TP, tall fescue tall patch; T, treatment; S, stratum; P, patch type.

). The leaf proportion between stratums of SP was not affected by treatments (0.74), but it was lower for the residual stratum of patches in the TM (0.53) and TL (0.65) treatments, respectively. The proportion of leaves was not different from the superior to the low canopy stratum of MP and TP in the TC treatment, but it was lower for the low canopy stratum of MP in the TM and TL treatments. The proportion of leaves for TP was lower in the residual and middle canopy stratum in the TM and TL treatments, respectively. No differences in the proportion of leaves were observed for the superior and middle canopy stratum among grazing treatments (Table 6).

No differences in leaf density among the canopy stratums were observed for SP in TC and TL, but leaf density was lower for the residual canopy stratum of SP in the TM. Leaf density was homogeneous inside the canopy in MP (1.14 mg cm⁻³ ± 0.02), similar between TM and TL grazing treatments (p < 0.05), and lower for the residual canopy stratum compared to the low, middle, and superior stratum (Table 6). In contrast, the leaf density in MP was lowest in the superior canopy stratum compared to the middle, lower, and residual canopy stratum in the TC treatment (Table 6). The leaf density in the residual canopy stratum was greater for TC (1.04 mg cm⁻³) vs. TM (0.57 mg cm⁻³) and a tendency (p = 0.05) for a difference was observed between TC vs. TL (0.60 mg cm⁻³). The leaf density did not differ from superior to low canopy stratums of TP in the TM (1 mg cm⁻³ ± 0.06) and TL (0.78 mg cm⁻³ ± 0.06) treatments and among canopy stratums of TP patches in the TC treatment (1.18 mg cm⁻³ ± 0.2). The leaf density of the low stratum of TP patches was lower in the TL treatment vs. TM and TC treatments and lower in the TM and TL treatments vs. TC treatment for the residual stratum.

The proportions of stems between stratums of SP were not affected by treatments, but it was lower for the residual stratum of patches in the TM treatments. The stem proportion was not different from the superior to low canopy stratum of MP (0.14) and TP (0.19) in the TC treatment, but it was higher both for the low canopy stratum of MP in the TM and TL treatments, respectively. The proportion of stems for TP was higher in the low and middle canopy stratum in the TM and TL treatments, respectively. No differences in the proportion of stems were observed for the superior and residual canopy stratum among grazing treatments.

No differences in leaf-stem forage ratio among canopy stratums were observed between treatments. However, the leaf-stem forage ratio across the canopy stratums was more variable for patches in the TM and TL compared to TC.

4. Discussion

Heterogeneity and sward structure are grazing determinants of plant-herbivore mediated processes such as plant growth and intake rate, and as such, they play a multifaceted role in grazing management. We measured the spatial heterogeneity of tall fescue pasture managed with different post-grazing sward heights and compared the effects on forage production and utilization. The use of taller post-grazing sward heights was expected to amplify the divergence of spatial heterogeneity among short, medium, and tall patches of tall fescue pasture, enhancing selective grazing and intake rate, without sacrificing the amount and chemical composition of utilized forage. This hypothesis was tested in a randomized complete block study where lactating dairy cows were assigned to either control, medium, or lax treatment in the autumn-winter and spring grazing seasons. Grazing to a taller sward height (lax treatment) increased the divergence in proportion, sward height, and herbage mass of tall patches compared to patches of control or medium sward height.

In turn, forage production was greater on the tall post-grazing sward height treatment and in spring, but the forage utilization was unaffected by the post-grazing sward height treatments as it was originally speculated.

4.1. Sward Structure and Forage Production

There were marked treatment differences between tested post-grazing SH but this did not affect the pre-grazing SH and herbage mass across grazing treatments. The pre-grazing SH was reduced by 51, 40, and 30% for pastures that received the TC, TM, and TL, indicating strictly different defoliation intensities despite marked treatment differences in post-grazing SH throughout the trial. Interestingly, the differences in post-grazing SH were not related to post-grazing HM, likely because treatments altered the composition and structure of pasture in lower stratums of the canopy. The defoliation intensities imposed in this experiment did not affect pre-grazing HM and SH but did affect forage production resulting in a similar level of forage utilization as it was originally predicted.

Previous studies testing the effects of rotational stocking on tall fescue pasture [1] also reported variations in post-grazing SH not related to differences in post-grazing herbage mass. In contrast, other tall fescue grazing studies observed marked increases in post-grazing herbage mass associated with increases in SH [27,28]. These contrasting results could be attributed to changes in sward structure induced by prescribed manipulations in intensity and frequency of defoliation, suggesting plastic adaptations in sward tiller and sward structure in response to grazing treatments [29,30]. However, it was apparent that variations in tall fescue SH treatments associated with changes in herbage mass were more evident as tall fescue plants grown faster in spring.

Although there were no differences in pre-grazing HM and SH between treatments (Table 2), there were differences in TGC among post-grazing SH treatments (Table 3). Grazing at lax defoliation intensities (Lax treatment) resulted in greater TGC at the end of the study, resulting in 23% greater TGC compared to the medium (TM) and control (TC) sward heigh grazing treatments. As plants actively grow in spring, the ratio of live to dead plant material may also have been altered in this season, but data has been unclear. Although greater defoliation has decreased the proportion of stem and dead material of cool season grasses in spring [31,32], the defoliation intensities achieved with the present post-grazing SH treatments have not been sufficiently contrasting to have a clear effect on both total and tall fescue green cover. Thus, in spring, the increasing TGC associated with increasing HM may have determined a greater production and accumulation of green leaf HM in the TL treatment than in the TC and TM treatments. In spring, the greater forage production recorded on TL (3651 kg DM ha⁻¹) than on TC and TM (2695 ± 175 kg DM ha⁻¹) may have explained the differences in forage production in favor of the TL treatment. Previous research suggested greater tall fescue forage production associated with the use of taller post-grazing sward heights compared to more intense defoliation [33]. Similarly, Brink et al. [34], observed lower productivity and persistence of cool season grasses as post-grazing SH decreased from 10 to 5 cm. The use of taller post-grazing sward heights may enhance tillers’ leaf dynamics and growth rate through enhanced leaf area and improved net photosynthesis rate [35].

The results of this study also suggest that taller post-grazing SH in the TL treatment may had a larger LAI compared to the TC and TM, which may have also allowed for greater interception of photosynthetically active radiation and growth rate in spring. Agnusdei and Assuero [36] also observed increasing LAI of continuously grazed tall fescue pasture when SH was increased from 6 to 13 and 19 cm in spring. Likewise, Saldanha et al. [37] reported increasing green leaf herbage mass (kg MS ha⁻¹) of Lolium perenne with increasing post-grazing SH from 3 to 12 cm. Grazing to a taller sward height may also increase the rate of leaf senescence, resulting in the accumulation of dead material in lower canopy stratums, which may negatively reduce the chemical composition and nutritional value of forage [35]. Nevertheless, greater senescence of taller residual plant material should be considered a programmed improve of soil cover and cost to promote rapid development of new leaves [38], thereby reducing the time to maximum average pasture growth rate or optimal interval for grazing [14].

In spring, grazed forage in the TL treatment had greater CP and lower NDF and ADF than forage in the TC (Table 4), indicating better forage feed value [24]. This may partially be due to the consistent use of management decision criteria considering a three-leaf phenological stage for the initiation of grazing [21]. In addition, the prescribed early flower control implemented early in the spring [13], may have also prevented early loss of forage quality associated with early flowering and stem elongation phases in spring [32,39]. According to Donaghy et al. [40] and Insua et al. [41], the optimal defoliation interval for tall fescue was identified between a two- and four-leaf stage, which avoids abrupt loss of forage nutritional quality while supporting satisfactory regrowth. In our study, the frequency of defoliation was maintained on three leaves throughout the experiment. Additionally, herbage mass samples for wet chemical analysis of forage chemistry were collected according to treatment post-grazing SH levels, excluding lower canopy stratums with a lower proportion of leaves, a greater presence of stem and cumulative senescent material which could have increased with light grazing in the TL treatment.

The tested SH treatments did not alter FGC in autumn-winter or spring. These results align with previous research in similar environments [2,3,27], which indicates that tall fescue botanical composition and ground cover remain relatively stable under conservative to moderate grazing intensities, irrespective of the stocking method used.

4.2. Patch Heterogeneity and Canopy Structure

The differential SH treatments created different degrees of patchiness in the sward structure. The changes in sward height and herbage mass between short, medium, and tall patches were unrelated because changes in treatment grazing intensities created a strictly different canopy structure. For example, grazing tall fescue to a lower SH increased the leaf density and proportion of leaves in tall patches and in the lower and residual stratums of the canopy of the same tall patches (Table 6).

The SH treatments created a mosaic of tall, medium, and short patches, each with different structural characteristics (Table 5). This effect was previously described by Faber-Díaz [1] and Cid et al. [2] for tall fescue pasture receiving rotational stocking. Although plant patches primarily result from soil heterogeneity and other environmental influences, livestock grazing is one of the most significant biotic factors [8] that can either increase or revert patterns of plant community heterogeneity [42]. As suggested by Li and Reynolds [9], spatial heterogeneity can be characterized by the number of contrasting patch types and the proportion of each patch type. In this study, quadrats of grazed pasture were classified into short, medium, and tall patches. Therefore, patch type was controlled by experimental design. However, the proportion of patches was altered by treatments with proportions of tall patches increasing when tall fescue was grazed to the taller sward height, TL treatment. Menegazzi et al. [11] reported increasing spatial heterogeneity of sward height over time when tall fescue pasture was grazed to taller sward height. Furthermore, grazing to a taller sward height also altered the distribution of short, medium, and tall patches compared to pasture grazed to a short or medium sward height. Pasture heterogeneity is strictly a scale-dependent phenom [18]. Therefore, grazing treatments in this study may have amplified or reduced other scale-dependent effects such as fractal aggregation, fragmentation, or dispersion [42,43,44] that were not detected by the present experimental design and scale and methods of measurements. The heterogeneity of pasture has a major effect on feeding site selection, feeding time, and intake rate, allowing selectivity for diets that are better than the average diet for animals that were grazing randomly [18]. A recent simulation model [19] explored the effects of spatial heterogeneity on intake rate, individual production and production per area, and herbage allowance. All other factors equal, intake rate increased with increasing coefficient of variation (10 vs. 50% CV) for increasing sward heights. Authors predicted lower selectivity for sward height in pastures with a lower degree of spatial heterogeneity. Conversely, greater selection for taller pasture was predicted with increasing heterogeneity in sward height. Despite taller sward heights’ maximized intake rate, effects of bivariate relationships between taller pasture and mature pasture of declining forage feed value were not tested, which may warrant careful interpretation of simulation results. Faber-Díaz [1] compared a 6 vs. 12 cm post-grazing SH in tall fescue pasture and observed a greater selection for tall over short vegetation patches in spring. Similar findings were reported by Menegazzi et al. [11] and Giles et al. [45] in the setting of this experiment. Hypothetically, different horizontal and vertical sward structures generated by SH treatments may have triggered a different array and organization of foraging behaviors (bite mass, biting rate, grazing time, rumination time) that were organized hierarchically across spatio-temporal scales (feeding station, patch, feeding site), largely influencing diets, dry matter intake, and milk production [11]. In a companion analysis of the same data, we detected a greater exploration of feeding stations associated with a shorter residence time per feeding station (i.e., fewer bites and depletion per feeding station). This resulted in a higher intake rate of dry matter, leading to a higher nutrient intake and milk production on TL compared to TM and TC [11]. Grazing management has a multifaceted role on pasture and milk production, modulating key interrelated processes defined mostly at the animal–plant interface [18], such as the above discussed feedback between spatial heterogeneity and forage selection and nutrient intake [5].

Previous autumn-winter tall fescue management affected the spring forage production of tall fescue. As argued by Gastal and Lemaire [29], defoliation management has a major impact on sward structure, affecting herbage growth rate and utilization. Short patches typically have a higher relative growth rate than taller patches, but this may not be enough to increase the average growth rate of pasture. According to Cid et al. [2], the compensatory growth rate of short patches should be doubled in order to achieve the same growth rate of tall patches. In turn, the latter are more vulnerable to defoliation [46]. In this study, the higher proportion of TP observed in the TL treatment in spring could explain the higher forage production of TL in spring. This response may be due to the effects of different defoliation intensities on leaf elongation and leaf tissue turnover rates and associated effects on plant tissue chemical composition [7].

Tall, medium, and short patches had different structural characteristics (Table 5), but most components evaluated were not affected by SH treatments. Cid et al. [2] observed no differences in SH and total and live herbage mass in short and tall patches at a stocking density between 3.6 and 6.6 animals ha⁻¹. Faber-Díaz [1] reported that SH treatments did not alter the species cover, among dead plant tissue residue, and lamina-sheath ratios of short and tall patches in spring. The proportion of leaves in TP was different between SH treatments, but the proportion of leaves always exceeded 50% in all patch types. Szymczak et al. [28] reported a linear relationship between tall fescue herbage mass and leaf mass with increasing SH from 14 to 26 cm. The proportion of lamina was 57 to 66% of the total herbage mass, similar to the observation for short, medium, and tall patches of the present study, which would imply the offering of a higher leaf mass in patches with greater herbage mass. Therefore, in addition to the proportion and availability of leaf, the accessibility of leaf plant material may also have effects on the ingestive behavior of animals. The TP presented greater variances in the canopy distribution of leaves (Table 6), which could have influenced the differences in the ingestive behavior and dry matter intake reported by Menegazzi et al. [11]. However, the treatments presented a similar leaf distribution in the upper canopy strata of TP (superior and medium canopy stratum). In general terms, leaves were the predominant component in the upper stratum of patches (>55%) and presented a leaf density of 0.93 ± 0.16 mg cm⁻³ and the stems were concentrated in the lower canopy stratums. The management of defoliation intensity also influences the composition and structure of potential grazable horizons, exerting influences on ingestive behavior, diet selection, and herbage dry matter intake [6,47,48]. The high proportion of TP combined with the high proportion of leaves in the upper canopy stratum of TP may have provided greater accessibility to leaf tissue, which could have positively influenced forage dry matter intake and milk production of cows in the TL treatment [11]. This emphasizes the importance of grazing management to create and maintain a pasture structure, both horizontally and vertically that optimizes animal forage intake without sacrificing pasture production and utilization.

5. Conclusions

The evaluated post-grazing sward height treatments differentially affected pasture production and sward structure in autumn-winter and spring. Despite SH treatments, the pre-grazing SH and HM were unaffected but forage production and patch heterogeneity were modified by treatments. The lax grazing intensities (TL) increased forage production and the proportion of tall patches offering a greater amount of green leaf tissue in spring. The high proportion of TP combined with the high proportion of green leaf tissue in the superior stratum of a tall patch canopy may provide greater accessibility to nutritious leaf tissue, thereby enhancing forage intake and milk production in spring tall fescue pasture grazed to a taller sward height. Likewise, grazing to a taller sward height was associated with a modified mosaic of short, medium, and tall patches that may have favored greater horizontal patch selection for taller sward. However, these findings are based on a single year of evaluation and spring-only patch characterization. Future multi-year research across all seasons is needed to account for climate variability and long-term persistence. Despite these limits, these results emphasize the need to consider the dynamic effects of grazing on sward structure, both vertically and horizontally. Monitoring patch-level heterogeneity—rather than mean height alone—is essential to optimize animal production without sacrificing herbage yield and utilization.