Short-Term Forage Responses, Soil Nitrogen Availability, and Greenhouse Gas Fluxes Under Contrasting Nitrogen Sources in Cool-Season Forage Systems

Highlights

• Legume integration enhanced forage CP and labile soil N without increasing CH₄ flux.
• Inorganic fertilizer and poultry litter had limited effects on most forage and soil health responses.
• Cattle dung increased early CH₄ flux without changing soil or forage responses.

Abstract

Nitrogen (N) sources affect forage production, soil N dynamics, and greenhouse gas (GHG) emissions. This study evaluated short-term effects of contrasting N sources on forage responses, soil health indicators, and GHG emissions in a cool-season oat (Avena sativa L.) and annual ryegrass (Lolium multiflorum L.) system. Treatments were T1 (grass only), T2 (grass + N), T3 (grass + legume), T4 (grass + poultry litter), and T5 (grass + cattle dung). Crude protein was greater in T3 than in T4 (18% versus 13%; p = 0.02). In May, T3 had the greatest microbial C:N ratio (60.6) and soil inorganic N (45.5 mg N kg⁻¹; p < 0.01). Potentially mineralizable N was greater in T3 than in T1, T2 and T4 (3.96 versus ≤ 2.71 mg NH₄-N kg⁻¹ day⁻¹; p < 0.01). Mean daily CH₄ fluxes were greater (p < 0.01) under T5 (1106 g CH₄-C ha⁻¹ day⁻¹) than under T1 and T3 (≤−3.59 g CH₄-C ha⁻¹ day⁻¹). Mean daily CO₂ fluxes were lower in T3 and T4 than in T1, T2, and T5. There were no effects on N₂O emissions. Legume integration enhanced short-term N cycling and forage nutritive value, without N₂O emissions, over a 60-day experiment.

1. Introduction

Livestock operations in the southeastern United States are primarily based on perennial grasses such as bahiagrass (Paspalum notatum Flueggé), bermudagrass (Cynodon dactylon L.), and tall fescue (Festuca arundinaceae (Schreb.) Dumort.). Because forage production from perennial grasses is highly seasonal, producers commonly incorporate annual forage species to extend the grazing season [1,2]. Cool-season annual forages, such as annual ryegrass, oat, and clovers (Trifolium spp. L.) are commonly used in the region due to their productivity, seasonal distribution, and nutritive value [1,3]. The performance of these species is strongly affected by nitrogen (N) management and by forage species composition [4,5,6].

Nitrogen in forage systems can be supplied through various management strategies, including inorganic fertilizer application, organic amendments, localized deposition of cattle dung, and biological N fixation by legumes. These strategies differ in how N is introduced into the system, its spatial distribution, release patterns, and interactions with soil biological and physical processes [7,8]. As a result, they can influence forage production and nutritive value, soil carbon (C) and N dynamics, and the magnitude and pathways of greenhouse gas (GHG) emissions [9,10,11,12]. A clear understanding of how these management strategies influence forage systems is therefore essential for developing practices that optimize forage productivity and environmental sustainability while improving chemical, physical, and biological soil properties commonly used as indicators of soil health.

Inorganic N fertilization is widely used to improve forage production and nutritive value in annual cool-season forage systems [13,14], supporting animal performance and carrying capacity [15,16]. When appropriately managed, N inputs can increase crude protein and reduce fiber concentrations [17,18,19]. However, higher or poorly synchronized N inputs can also increase N losses and contribute to GHG emissions [20]. In contrast, systems that incorporate legume or organic amendments often exhibit a more closed N cycle, characterized by lower inorganic N and potentially lower emission intensity, relative to conventional fertilizer-based systems [21,22].

Legume integration is a key biological strategy for supplying N in forage ecosystems. Common cool-season forage legumes in the region include crimson clover (Trifolium incarnatum L.), arrowleaf clover (Trifolium vesiculosum Savi), and hairy vetch (Vicia villosa Roth) [23,24]. Through biological N fixation, legumes can reduce reliance on inorganic fertilizer while enhancing forage nutritive value [25,26]. In addition, legume inclusion can increase biomass and litter inputs, an important pathway for returning C and N to soils and supporting longer-term nutrient cycling [27,28].

Beyond forage responses, N source selection can influence soil properties associated with soil health. Nitrogen inputs influence soil C and N pools, soil aggregation, and microbial activity [29]. Studies have reported greater soil C accumulation, higher potentially mineralizable N, and improved aggregation in systems that receive organic amendments or legumes compared with systems that rely on inorganic N fertilizers [30,31,32]. Conversely, inorganic fertilizers can increase soil inorganic N and nitrate concentrations [33,34], which may elevate risks of runoff and deterioration of water quality. Soil biological properties, including microbial biomass and related responses, can also be sensitive to N management, with several properties responding more strongly to organic inputs than to inorganic fertilizers [35,36,37].

Many studies in cool-season grazing systems have evaluated the effects of N management on forage production, soil properties, or GHG emissions; however, these responses have often been assessed separately or over longer time scales. Consequently, limited information is available on the short-term responses of forage production, soil health indicators, and GHG emissions to contrasting N management strategies evaluated simultaneously under field conditions. Although short-term studies cannot capture long-term changes in soil health, they can identify early responses of chemical, physical, and biological soil indicators following N application.

Therefore, we hypothesized that N management strategies differing in N source, the rate of mineral N release, and spatial distribution would produce contrasting short-term responses in forage production, soil health indicators, and GHG emissions. To test this hypothesis, the objectives of this study were to (i) evaluate the short-term effects of contrasting N sources, including inorganic fertilizer, poultry litter, cattle dung, and a grass–legume mixture, on forage mass and nutritive value, soil chemical, physical, and biological indicators, and GHG fluxes in a cool-season forage system in South Carolina, and (ii) assess relationships among soil, forage, and GHG variables using correlation analyses.

2. Materials and Methods

2.1. Site Description

A 2-month study was carried out at a commercial beef cattle farm in Calhoun County, South Carolina, USA, during the 2025 growing season. The study site is in a humid subtropical climate (Köppen Cfa [38]), typical of the southeastern USA, with gently rolling topography and elevations ranging from approximately 90 to 130 m above sea level. The annual rainfall at the experimental site in 2025 was 798 mm, and the average monthly temperature ranged from 5 °C in winter to 28 °C in summer. Climatic data were obtained from the National Oceanic and Atmospheric Administration (NOAA) dataset for the research site (NOAA, 2025; Available at https://www.weather.gov/wrh/climate, accessed on 10 December 2025).

The soil at the experimental site is classified as an Orangeburg series sandy loam (fine-loamy, kaolinitic, thermic Typic Paleudults) according to the USDA Web Soil Survey (Available at https://websoilsurvey.nrcs.usda.gov/app/, accessed on 10 December 2025). Initial soil chemical properties (0–15 cm), determined by the Clemson University Agricultural Service Laboratory following standard analytical procedures, included organic matter of 1.25%, soil pH (1:1 soil:water) of 6.2, cation exchange capacity of 10 meq 100 g⁻¹, sum of bases of 2.4 cmol kg⁻¹, available P of 34.5 mg kg⁻¹, available K of 161 mg kg⁻¹, Ca of 1296 mg kg⁻¹, and Mg of 158 mg kg⁻¹. Soil bulk density was 1.4 g cm⁻³. The site had been managed under a forage–livestock production system for more than a decade prior to the establishment of the study.

2.2. Experimental Design and Management

The experiment was established in an annual cool-season forage mixture consisting of oat (Avena sativa L.), annual ryegrass (Lolium multiflorum L.), and cool-season legumes. In late October 2024, this grass mixture was planted using a no-till drill (Kincaid–Great Plains Compact Drill, model 3P606NT; Great Plains Manufacturing Inc., Salina, KS, USA) with 10 cm row spacing. The seeding rate used was 65 and 12 kg pure live seed (PLS) ha⁻¹ for oat and annual ryegrass, respectively. No fertilizer was applied before planting.

Daily air temperature, rainfall, and volumetric soil water content recorded during the experimental period are presented in Figure 1A–C. These environmental conditions provide the climatic context for interpreting temporal variation in forage responses and greenhouse gas fluxes.

In February 2025, a randomized complete block design was established using five treatments [control (no N fertilizer) and four N management strategies] and three replicates (15 experimental units in total). Each experimental unit (plot) measured 3 × 3 m with 1.2 m alleys. Treatments consisted of control (T1; no N fertilizer), inorganic fertilizer at 56 kg N ha⁻¹ (T2), grass–legume mixture (T3), poultry litter application at 7.4 t ha⁻¹ (T4), and cattle dung (T5). The experimental area was initially established as a uniform grass–legume mixture. Prior to treatment imposition, legumes were selectively removed (using herbicide) from T1, T2, T4, and T5 to establish the intended treatment contrasts, whereas T3 retained the original grass–legume mixture. The sequence of experimental establishment and treatment imposition is summarized in Figure 2.

Broadleaf control was achieved using 2,4-D amine (active ingredient: 2,4-dichlorophenoxyacetic acid; Nufarm Americas Inc., Alsip, IL, USA) at 0.7 kg a.e. ha⁻¹ applied with a backpack sprayer calibrated for uniform coverage according to label recommendations. Herbicide application was performed only once in February, before treatment imposition (in April; Figure 2), to establish the intended forage composition of each treatment (T1, T2, T4 and T5). The treatment chemical composition is presented in

Table 1. Chemical composition of the nitrogen sources used in the experiment (as-sampled basis).

Treatment	N Source	Chemical Composition
		NH4-N	N	P	K
		-------- g−1 Fresh Weight (As-Sample Basis) --------
T1	no fertilizer	–	–	–	–
T2	19–19–19	–	190	82.8	157.7
	13–0–0	–	130	–	–
T3	legume	–	–	–	–
T4	poultry litter	11.1	31.5	7.41	18.1
T5	cattle dung	0.2	21.5	11.3	9.1

“–” indicates values not applicable (e.g., no fertilizer applied in T1) or not determined (e.g., N contribution from biological nitrogen fixation was not quantified in T3). Nutrient concentrations for poultry litter and cattle dung are expressed on an as-sampled (fresh weight) basis. Dry matter contents were 97.5% for poultry litter and 66.1% for cattle dung. For inorganic fertilizers, nutrient concentrations are expressed per kilogram of commercial fertilizer product. Fertilizer formulations (19–19–19 and 13–0–0) are reported as N–P₂O₅–K₂O, whereas the values presented in the table for P and K correspond to elemental P and K. NH₄–N = ammoniacal nitrogen; N = total nitrogen; P = elemental phosphorus; K = elemental potassium.

.

The inorganic N (T2) rate mimics a low-input fertilization strategy typical of regional forage production systems. The total N rate (56 kg N ha⁻¹) was split into two hand-broadcast surface applications of 28 kg N ha⁻¹ each (on 1 April and 5 May). The first application used a granular 19–19–19 (N–P₂O₅–K₂O) fertilizer to supply N, P, and K according to the site’s soil fertility recommendations, whereas the second application used a granular 13–0–0 fertilizer to supply N only (Table 1). Split N application was adopted to better synchronize N availability with forage demand while reducing potential N losses. No rainfall occurred during the first seven days following the initial fertilizer application, allowing the fertilizer to remain on the soil surface. Application timing, rates, and the experimental timeline are summarized in Figure 2.

The legume treatment (T3) included a mixture of crimson clover (Trifolium incarnatum L.), arrowleaf clover (Trifolium vesiculosum Savi), and hairy vetch (Vicia villosa Roth). All legume species were established using commercially available pre-inoculated seeds containing the appropriate Rhizobium strains. The seeding rates for each legume were 10, 5.6, and 3.9 kg PLS ha⁻¹, respectively. Poultry litter (T4) was sourced from a local commercial poultry house and applied at the recommended rate [39]. A poultry subsample was collected at the time of application for laboratory analysis of chemical composition (Table 1).

Cattle dung (T5) was collected from the collaborator’s farm pastures in accordance with biosecurity protocols (e.g., long-sleeve gloves and boot covers). Fresh dung from adult Angus cattle was collected immediately after defecation. The amount applied (1.5 kg on areas ranging from 0.05 to 0.09 m²) was based on the reported fecal deposition characteristics of adult cattle [40,41]. The 1.5 kg dung pat was placed in a PVC collar with an area of 0.07 m², which closely matches a deposition area of a single defecation by an adult beef cow. This PVC collar was used for greenhouse gas measurements. To minimize contamination and N losses, the dung was processed as described by Garcia et al. [42]. Another 1.5 kg of dung pat was placed within a separate 0.5 × 0.5 m mini-plot for forage and soil sampling (Figure 2). The dung treatment represented a single natural cattle defecation event, with 1.5 kg of fresh dung applied as a localized patch rather than uniformly across the plot; therefore, the treatment was expressed on a per-patch basis rather than as a field-scale application rate (t ha⁻¹).

For T1, T2, T3, and T4, treatments were uniformly established across the entire 3 × 3 m plot, and PVC collars were randomly inserted for greenhouse gas measurements. Thus, the collars represented the surrounding treatment conditions and did not receive separate applications. All treatments were imposed simultaneously during active vegetative cool-season growth, when plants have high N demand and uptake. Each N source was applied according to its typical field management or natural deposition pattern (e.g., uniform fertilizer application or localized cattle dung deposition), rather than being standardized to an equivalent N application rate, thereby representing realistic management conditions commonly found in grazing systems (Figure 2).

2.3. Response Variables

2.3.1. Forage Mass, Nutritive Value, Litter Biomass and Chemical Composition

On 5 May (34 days after treatment application) and June 2 (28 days after the first harvest), forage mass (FM) was determined by harvesting two 0.12-m² samples per plot to a 10 cm stubble height using polyvinyl chloride (PVC) quadrats. The two subsamples were processed separately, and their mean value was used as the experimental unit for statistical analyses. Sampling locations were relocated within each plot at the second harvest to avoid resampling previously harvested areas. Fresh weight was recorded in the field, oven-dried at 55 °C until a constant weight, and ground to pass a 1 mm screen using a Wiley Mill (Model T3700.002; Thomas Scientific, Swedesboro, NJ, USA). Fiber concentrations were determined using a Delta Fiber Analyzer (ANKOM Technology, Macedon, NY, USA) according to the procedures described by ANKOM Technology [43,44]. Crude protein concentration was determined by automated digestion and distillation using a Kjeltec™ 9 analyzer (FOSS Analytical, Hillerød, Denmark), based on Kjeldahl chemistry according to AOAC Official Method 2001.11.

In vitro dry matter digestibility (IVDDM) was assessed using an ANKOM Daisy II incubator (ANKOM Technology, Macedon, NY, USA). Care and handling of donor animals were approved by the Clemson University Committee on Animal Use (AUP2022-0464). A composite inoculum was prepared from rumen fluid and solids collected before the morning feeding from two rumen-fistulated lactating dairy cows fed a corn silage-, barley silage-, bermudagrass hay-, and concentrate-based diet. Samples were incubated in duplicate for 30 h following the procedures of Ferreira and Mertens [45], who describe more details about the process (e.g., fluid storage method, etc.). All forage nutritive value variables (CP, NDF, ADF, and IVDDM) were expressed on a dry matter basis.

Litter was defined as the dead aboveground plant tissue no longer attached to a stem. Litter was sampled only in June using 2 PVC quadrats (0.12 m²), oven-dried at 55 °C until a constant weight, ground to pass a 1 mm screen, and analyzed for C and N concentrations with a LECO CN 828 Analyzer (LECO Corporation, St. Joseph, MI, USA). Concentrations were used to calculate the C:N ratio.

2.3.2. Soil Health Indicators

A baseline composite sample from the experimental area was taken on 1 April 2025, prior to treatment application. Subsequent sampling occurred on 5 May and 2 June, following approximately 30-day intervals, aiming to capture short-term soil responses to N inputs. In each plot, five soil cores (200 g) were collected at 0–15 cm depth, composited, placed in plastic bags on ice, transported to the laboratory at the Clemson University Pee Dee Research and Education Center, and stored at 4 °C until analyzed within one week.

Microbial biomass C (MBC) and N (MBN) were determined using the chloroform fumigation–extraction method following Voroney et al. [46], applying conversion factors of 0.37 and 0.54 for MBC and MBN, respectively. The microbial C:N ratio was calculated by dividing MBC by MBN. Extractable ammonium (NH₄⁺; EA) and nitrate (NO₃⁻; EN) were determined colorimetrically after extraction with 1 M KCl as described by Verdouw et al. [47] and Doane and Horwath [48]. Potentially mineralizable N (PMN) was determined by anaerobic incubation at 30 °C for 7 d followed by 1 M KCl extraction. Activity of N-acetyl-β-D-glucosaminidase (NAG) was determined fluorometrically using 5 g (dry-equivalent) soil incubated in triplicate according to Ye et al. [49]. Soil organic carbon (SOC) and soil organic N (SON) were measured by dry combustion using a CN analyzer (Carlo Erba NA 1500, Carlo Erba Strumentazione, Milan, Italy). Wet soil aggregate stability (WAS) was determined by wet sieving using 20 g of sieved soil and expressed as mean weight diameter (MWD) following Six et al. [50] and Márquez et al. [51]. Basal soil respiration (BSR) was measured as CO₂ production by incubating rewetted soil samples for 24 h [52]. Detailed analytical procedures for each assay are described in the corresponding references cited above.

2.3.3. Greenhouse Gas Fluxes Measurements

Soil GHG fluxes were measured using an automated static chamber system installed on permanent PVC collars. Fifteen automated chambers (LI-8200-104, LI-COR Inc., Lincoln, NE, USA) were deployed on April 1, immediately following treatment application. Each chamber enclosed a soil surface area of 317.8 cm² with an internal volume of 4076.1 cm³. Gas concentrations were measured using an LI-7810 analyzer for CO₂ and CH₄ and an LI-7820 analyzer for N₂O, controlled by an LI-8250 multiplexer (LI-COR Inc., Lincoln, NE, USA), following the manufacturer’s specifications and the methodology described by LI-COR [53]. Soil temperature (ST) and soil water content (SWC) were simultaneously monitored using Stevens Hydra Probe sensors installed approximately 10 cm from each PVC collar at a depth of 10 cm.

Each chamber remained closed for 5 min, consisting of 3 min of gas concentration measurements followed by 2 min of chamber ventilation before the next chamber was sampled. A complete measurement cycle across the 15 chambers required approximately 75 min and was repeated continuously throughout the 24 h period. Gas concentration data were processed using SoilFluxPro software (version 4.2, LI-COR Inc., Lincoln, NE, USA) and only measurements with R² ≥ 0.75 were retained for statistical analyses. Initial chamber conditions were estimated from the first measurements following chamber closure, and gas fluxes were subsequently calculated using the manufacturer’s non-linear regression algorithm, which accounts for the non-linear change in gas concentration during chamber deployment [53].

Greenhouse gas fluxes were monitored continuously for 60 consecutive days following treatment application. A technical malfunction caused by a short circuit in the LI-8250 multiplexer prevented reliable measurements during the final nine days of monitoring. Consequently, statistical analyses were performed using data from Days 1, 2, 3, 4, 7, 14, 21, 28, 35, 42, and 51. These evaluation intervals were defined a priori to capture both the rapid responses immediately following treatment application and the subsequent longer-term dynamics, while maintaining representative temporal resolution and reducing redundancy associated with the large number of high-frequency measurements generated by the automated chamber system.

2.4. Statistical Analysis

Plant, soil and GHG data were analyzed separately in R v4.4.0 (R Core Team, Vienna, Austria) using linear mixed-effects models. Treatment, harvest, month or day, and their interaction were considered fixed effects, whereas block was considered a random effect. LSMEANS were compared using Tukey’s test (p < 0.05). Pairwise associations among variables were further evaluated using Pearson correlation analysis.

3. Results

3.1. Forage Mass, Nutritive Value, Litter Biomass and Chemical Composition

There was no treatment × harvest interaction for FM, NDF, ADF and IVDDM concentrations; therefore, results are presented as treatment means averaged across harvest time and the average of harvest across treatments (

Table 2. Forage mass and nutritive value of cool-season forage systems under contrasting nitrogen fertilization sources in Calhoun County, South Carolina.

Treatment	FM +	NDF	ADF	IVDDM	Litter C:N Ratio
	kg DM ha−1	--------- % (DM Basis) ----------			Dimensionless
T1 *	3050	52	30	64	23.9
T2	2916	49	29	63	24.8
T3	3325	49	29	63	24.1
T4	2658	49	29	60	21.3
T5	3033	51	29	60	22.6
SE	234.93	1.23	0.94	6.53	2.56
p value	0.40	0.36	0.96	0.26	0.63
Harvest	FM +	NDF	ADF	IVDDM	Litter C:N ratio
	kg DM ha−1	--------- % (DM Basis) ------------			Dimensionless
May	1546 b	48 b	27 b	67 a	-
June	4446 a	52 a	31 a	56 b	23
SE	148.58	0.77	0.60	6.53	-
p value	<0.01	<0.01	<0.01	0.02	-

* T1, grass only; T2, grass + N; T3, grass + legume; T4, grass + poultry litter; T5, grass + cattle dung. + FM, Forage mass; NDF, Neutral detergent fiber; ADF, Acid detergent fiber; CP, IVDDM, In vitro digestibility of dry matter; SE, Standard error. “–” indicates values not applicable (single litter sampling in June). Means followed by a common letter within a column are not significantly different at the 5% level of significance, determined using the Tukey test.

). Treatment differences were not significant (p > 0.05) for FM, NDF, ADF and IVDDM concentrations, with values averaging from 3325 kg DM ha⁻¹, 52%, 30% and 64%, respectively, across treatments. Forage mass was greater (p < 0.01) in June than in May (4446 vs. 1546 kg DM ha⁻¹). No significant difference was found in NDF, ADF, and IVDDM concentrations among treatments (p > 0.05). However, in June, NDF and ADF concentrations were greater (p < 0.01) than in May (Table 2).

There was a treatment × harvest interaction (p = 0.01; SE = 1.19) for CP concentration. In May, T3 had greater CP concentration (19.2%) than T4 (10.6%), but did not differ from other treatments (Figure 3). No statistically significant difference was found for Litter C:N ratio among treatments (p > 0.05), and values ranged from 21.4 to 24.8 across treatments (Table 2).

3.2. Soil Health Indicators

There was a treatment × month interaction for microbial C:N ratio and inorganic N (Figure 4). In May, T3 had a greater (p < 0.01) Microbial C:N ratio and inorganic N than all other treatments, with values of 60.6 and 45.51 mg kg⁻¹, respectively. The remaining treatments showed similar values for microbial C:N ratio (15.2–19.4) and inorganic N concentration (16.09–20.07 mg kg⁻¹). In June, microbial C:N ratio and inorganic N did not differ among treatments.

There was no treatment × month interaction for PMN, EA, EN, SOC, SON, MBC, MBN, BSR and NAG activity. The T3 treatment had greater (p < 0.01) PMN than T1, T2, and T4, with increases ranging from 46 to 103% (

Table 3. Chemical and biological soil health indicators under contrasting nitrogen fertilization sources in annual cool-season forage systems of Calhoun County, South Carolina.

Treatment	PMN+	EA	EN
	mg NH4-N kg−1 day−1		mg NO3-N kg−1
T1 *	2.71 b	2.27 b	14.40 b
T2	2.37 b	1.87 b	14.36 b
T3	3.96 a	12.18 a	20.81 a
T4	1.95 b	1.75 b	12.40 b
T5	3.11 ab	1.79 b	13.75 b
SE	0.27	2.95	1.45
p value	<0.01	0.02	<0.01

* T1, grass only; T2, grass + N; T3, grass + legume; T4, grass + poultry litter; T5, grass + cattle dung. + PMN, potentially mineralizable nitrogen; EA, Extractable ammonium; EN, Extractable nitrate. SE, Standard error. Means followed by a common letter within a column are not significantly different at the 5% level of significance by Tukey’s test.

). Similarly, EA under T3 was greater than in other treatments (p = 0.02), exceeding them by 436 to 596%. T3 also had greater (p < 0.01) EN than T1, T2, T4 and T5, representing an increase of 44 to 68% for EN. A lack of significance was observed for SOC, SON, MBC, MBN, BSR, and NAG activity (p > 0.05), and values ranged from 17.95 to 26.75 g C kg⁻¹, 1.46 to 2.31 g N kg⁻¹, 273.27 to 339.21 mg C kg⁻¹, 14.12 to 24.71 mg N kg⁻¹, 24.91 to 40.87 mg CO₂-C kg⁻¹ day⁻¹ and 2745.93 to 3893.61 µM MUB kg⁻¹ h⁻¹, respectively (Full values per treatment are in

Table A1. Physical, chemical, and biological soil properties of a cool-season forage system under contrasting nitrogen sources.

Treatment	+WAS	SOC	SON	MBC	MBN	BSR	NAG
	mm	g C kg−1	g N kg−1	mg C kg−1	mg N kg−1	mg CO2-C kg−1 day−1	µM MUB kg−1 h−1
T1 *	0.59	23.08	1.92	339.21	23.58	30.45	3893.61
T2	0.57	21.37	1.69	344.15	24.71	25.3	3297.81
T3	0.52	26.75	2.31	334.65	14.12	40.87	4149.02
T4	0.49	17.95	1.46	273.27	18.94	24.91	2745.93
T5	0.53	23.41	1.98	356.1	21.24	26.73	3306.51
SE	0.05	2.64	0.23	32.21	4.14	4.63	607.94
p value	0.66	0.24	0.15	0.42	0.4	0.1	0.52

* T1, grass only; T2, grass + N; T3, grass + legume; T4, grass + poultry litter; T5, grass + cattle dung. +WAS, Wet aggregate stability; SOC, Soil organic carbon; SON, soil organic nitrogen; MBC, microbial biomass carbon; MBN, microbial biomass nitrogen; BSR, short-term carbon mineralization; NAG, N-acetyl-β-glucoaminidase. SE, Standard error.

).

3.3. Greenhouse Gas Fluxes

There was a treatment × day interaction (p < 0.001) for soil methane fluxes over time. Methane fluxes were greater (p < 0.001) in T5 than in T1, T3, and T4 from Day 1 to Day 4 (Figure 5). On Day 1, CH₄ flux reached 3319 g CH₄-C ha⁻¹ day⁻¹ in T5, whereas T1 had a flux of 0.93 g CH₄-C ha⁻¹ day⁻¹. The T1, T3, and T4 treatments did not differ from each other and exhibited the lowest CH₄ fluxes during this period, ranging from −0.99 to 1.36 g CH₄-C ha⁻¹ day⁻¹. From Day 7 to Day 51, CH₄ fluxes remained low across all treatments, ranging from −2.26 to −0.57 g CH₄-C ha⁻¹ day⁻¹ (Figure 5).

No treatment × day interaction was observed for CO₂ or N₂O fluxes (p > 0.05). Across the evaluation period, CO₂ and N₂O fluxes ranged from 75 to 192 kg CO₂-C ha⁻¹ day⁻¹ and from 7 to 17 g N₂O-N ha⁻¹ day⁻¹, respectively (Figure 6). Carbon dioxide fluxes differed over time (p < 0.001), reaching their highest values on Day 14, whereas fluxes on Day 14 did not differ from those on Day 28 (Figure 6A). Although N₂O fluxes reached their highest numerical values on Day 14, differences among days were not statistically significant (p = 0.087; Figure 6B).

Mean daily N₂O fluxes did not differ among treatments (p = 0.10), whereas mean daily CH₄ and CO₂ fluxes were affected by treatment (p < 0.01). Greater (p < 0.01) mean daily CH₄ fluxes were observed under T5 than under T1, T2, T3, and T4. Negative CH₄ fluxes were observed under T1 and T3, whereas T4 exhibited similarly low CH₄ fluxes (

Table 4. Average daily greenhouse gas fluxes from annual cool-season grazing systems under contrasting nitrogen sources in Calhoun County, South Carolina.

Treatment	N2O +	CH4	CO2
	g N2O-N ha−1 day−1	g CH4-C ha−1 day−1	kg CO2-C ha−1 day−1
T1 *	193	−1 c	117 a
T2	77	412 b	132 a
T3	26	−3.59 c	90 b
T4	6	0.38 c	86 b
T5	55	1106 a	175 a
SE	56.35	155.85	22.73
p value	0.10	<0.01	<0.01

* T1, grass only; T2, grass + N; T3, grass + legume; T4, grass + poultry litter; T5, grass + cattle dung. +N₂O, nitrous oxide; CH₄, methane; CO₂, carbon dioxide; SE, standard error. Means followed by a common letter within a column are not significantly different at the 5% level of significance, determined using the Tukey test.

). Lower (p < 0.01) mean daily CO₂ fluxes were observed under T3 and T4 than under T1, T2, and T5 (Table 4).

3.4. Correlation Analysis

During May, potentially mineralizable nitrogen (PMN) was positively correlated with EA, EN, IN, SOC, SON, BSR, and NAG (

Table 5. Pairwise correlation of forage, soil health and greenhouse gas responses (May 2025).

Variable	+WAS	PMN	EA	EN	IN	SOC	SON	MBC	MBN	C:N	BSR	NAG	FM	NDF	ADF	CP	IVDDM	N2O	CH4
PMN+	−0.27
EA	−0.17	0.69 **, †
EN	0.06	0.56 *	0.19
IN	−0.11	0.81 **	0.89 **	0.62 *
SOC	−0.34	0.84 **	0.56 *	0.46	0.66 **
SON	−0.32	0.86 **	0.56 *	0.48	0.67 **	0.99 **
MBC	−0.23	0.46	−0.23	0.60 *	0.09	0.48	0.46
MBN	−0.37	−0.20	−0.48	−0.37	−0.56 *	0.09	0.03	0.39
C:N	0.17	0.43	0.42	0.81 **	0.72 **	0.24	0.30	0.08	−0.76 **
BSR	−0.27	0.85 **	0.88 **	0.34	0.86 **	0.74 **	0.76 **	0.04	−0.35	0.36
NAG	−0.22	0.73 **	0.50	0.25	0.51 *	0.55 *	0.58 *	0.19	0.01	0.15	0.70 **
FM	0.06	0.27	0.16	0.04	0.14	0.12	0.13	0.12	−0.24	0.10	0.23	0.17
NDF	0.09	−0.08	0.02	−0.42	−0.18	−0.14	−0.13	−0.30	−0.01	−0.25	0.07	0.15	0.59 *
ADF	0.35	−0.11	−0.04	−0.30	−0.17	−0.26	−0.25	−0.16	0.04	−0.18	−0.09	0.18	0.39	0.70 **
CP	0.14	0.37	0.29	0.59 *	0.51	0.34	0.31	0.33	−0.28	0.58 *	0.23	0.14	0.52 *	−0.05	0.11
IVDDM	−0.08	0.11	0.21	0.38	0.34	0.29	0.31	0.01	−0.24	0.34	0.22	−0.24	−0.46	−0.66 **	−0.82 **	−0.06
N2O	−0.05	−0.13	0.10	−0.26	−0.04	−0.01	0.01	−0.26	−0.22	−0.03	0.03	−0.25	0.55 *	0.18	−0.11	0.23	0.02
CH4	−0.23	−0.14	−0.20	−0.38	−0.34	−0.31	−0.25	−0.13	−0.04	−0.16	−0.26	−0.08	0.40	0.23	0.14	−0.16	−0.42	0.51
CO2	−0.10	−0.23	−0.26	−0.33	−0.36	−0.34	−0.29	−0.13	−0.07	−0.09	−0.37	−0.24	0.39	0.18	0.08	−0.06	−0.31	0.61 *	0.95 **

+WAS, Wet aggregate stability; PMN, potentially mineralizable nitrogen; EA, Extractable ammonium; EN, Extractable nitrate; IN, Inorganic nitrogen; SOC, Soil organic carbon; SON, soil organic nitrogen; MBC, microbial biomass carbon; MBN, microbial biomass nitrogen; C:N, microbial biomass carbon-to-nitrogen ratio; BSR, short-term carbon mineralization, NAG, N-acetyl-β-glucoaminidase; FM, Forage mass; NDF, Neutral detergent fiber; ADF, Acid detergent fiber; CP, Crude protein; IVDDM, In vitro digestibility of dry matter; N₂O, nitrous oxide; CH₄, methane; CO₂, carbon dioxide. † Values are the correlation coefficient R with designated significant values; * Significant at the 5% of probability level; ** Significant at the 1% of probability level.

). Inorganic nitrogen (IN) was positively correlated with EA, EN, SOC, SON, the microbial C:N ratio, BSR, and NAG, but negatively correlated with MBN. Forage mass (FM) was positively correlated with NDF, CP, and N₂O, whereas IVDDM was negatively correlated with NDF and ADF. Methane (CH₄) was positively correlated with CO₂. During June, PMN was positively correlated with EN, IN, SOC, SON, and BSR (

Table 6. Pairwise correlation of forage, soil health and greenhouse gas responses (June 2025).

Variable	+WAS	PMN	EA	EN	IN	SOC	SON	MBC	MBN	C:N	BSR	NAG	FM	NDF	ADF	CP	IVDDM	N2O	CH4
PMN+	−0.03 †
EA	−0.18	0.39
EN	−0.25	0.69 **	0.00
IN	−0.29	0.72 **	0.81 **	0.59 *
SOC	−0.12	0.66 **	0.22	0.69 **	0.59 *
SON	−0.10	0.70 **	0.23	0.73 **	0.62 *	0.99 **
MBC	0.12	0.29	−0.47	0.61 *	−0.02	0.55 *	0.54 *
MBN	0.34	0.13	−0.46	0.49	−0.08	0.29	0.30	0.75 **
C:N	−0.41	0.28	0.69 **	−0.14	0.48	0.08	0.08	−0.50	−0.87 **
BSR	−0.32	0.55 *	0.20	0.68 **	0.56 *	0.72 **	0.75 **	0.39	0.31	0.06
NAG	−0.10	0.31	0.56 *	0.03	0.47	0.41	0.37	−0.29	−0.17	0.27	0.13
FM	0.16	0.16	−0.13	0.17	0.00	−0.24	−0.19	0.08	0.37	−0.27	0.19	−0.26
NDF	0.14	−0.27	−0.06	−0.26	−0.20	−0.46	−0.45	−0.10	−0.15	0.08	−0.18	−0.39	0.56 *
ADF	0.41	−0.26	−0.13	−0.30	−0.28	−0.37	−0.36	−0.05	0.25	−0.31	−0.15	−0.19	0.73 **	0.65 **
CP	−0.68 **	0.16	0.35	0.44	0.54 *	0.34	0.34	−0.10	−0.16	0.26	0.46	0.11	−0.27	−0.34	−0.50
IVDDM	−0.46	0.07	0.23	0.09	0.24	0.27	0.24	−0.21	−0.47	0.42	0.02	0.26	−0.77 **	−0.54 *	−0.82 **	0.69 **
N2O	0.09	0.26	0.03	−0.12	−0.04	0.12	0.13	0.06	−0.25	0.27	0.11	−0.25	−0.05	0.10	−0.13	0.03	0.14
CH4	−0.16	−0.05	−0.32	0.18	−0.16	0.25	0.21	0.68 **	0.40	−0.28	0.15	−0.32	−0.30	−0.31	−0.27	0.05	0.11	0.02
CO2	−0.10	−0.13	0.48	−0.40	0.16	−0.36	−0.35	−0.70 **	−0.50	0.33	−0.23	0.16	−0.09	0.05	−0.08	0.29	0.21	0.37	−0.53 *

+WAS, Wet aggregate stability; PMN, potentially mineralizable nitrogen; EA, Extractable ammonium; EN, Extractable nitrate; IN, Inorganic nitrogen; SOC, Soil organic carbon; SON, soil organic nitrogen; MBC, microbial biomass carbon; MBN, microbial biomass nitrogen; C:N, microbial biomass carbon-to-nitrogen ratio; BSR, short-term carbon mineralization, NAG, N-acetyl-β-glucoaminidase; FM, Forage mass; NDF, Neutral detergent fiber; ADF, Acid detergent fiber; CP, Crude protein; IVDDM, In vitro digestibility of dry matter; N₂O, nitrous oxide; CH₄, methane; CO₂, carbon dioxide. † Values are the correlation coefficient R with designated significant values; * Significant at the 5% of probability level; ** Significant at the 1% of probability level.

). Extractable nitrate (EN) was positively correlated with IN, SOC, SON, MBC, and BSR, whereas MBC was positively correlated with MBN and CH₄ but negatively correlated with CO₂. Forage mass was positively correlated with NDF and ADF and negatively correlated with IVDDM. In addition, IVDDM was negatively correlated with NDF and ADF but positively correlated with CP. In contrast to May, CH₄ was negatively correlated with CO₂ during the June sampling period.

4. Discussion

4.1. Nitrogen Sources Effects on Forage Responses

In the Southeastern USA, cool-season annuals’ growth generally extends through May/June and is determined by changes in photoperiod and temperature [25,54]. In this study, N treatments were applied in April, likely limiting FM responses due to the short evaluation period, the advanced phenological stage of the system, and the relatively low inorganic N rate [13,39,55]. In addition, pre-existing soil N availability may also have reduced the magnitude of the forage response to inorganic fertilization. Furthermore, the first harvest evaluated forage responses approximately 34 days after the initial N application, whereas the second split application was applied on the day of the first harvest. Therefore, only the second harvest reflected the complete split-N fertilization strategy. Under these conditions, T3 (grass + legume) increased CP concentration rather than FM. Although legumes were pre-inoculated and may have supported biological N fixation (BNF), the effectiveness of commercial inoculation remains variable [56,57].

Although BNF was not directly measured in this study, the responses observed in T3 were consistent with legume inclusion; therefore, the underlying mechanisms cannot be confirmed. Legume-derived N is strongly influenced by environmental conditions, species interactions, and time, and does not consistently exceed that supplied by other N sources [58,59]. For instance, grass–legume mixtures have been shown to increase CP by approximately 2 to 5 percentage units [60,61], while effects on FM are variable and context-dependent because responses differ with species composition, environmental conditions, and management intensity [62,63]. Consequently, the responses observed in T3 should be interpreted as effects associated with legume inclusion rather than direct evidence of biological N fixation.

Consistent with this, T3 increased CP relative to poultry litter but did not differ from other N sources, indicating limited short-term benefits of legume inclusion under the conditions of this study. In contrast, T2 (inorganic N) did not increase crude protein to the same extent, suggesting that late-season plant demand and rapid uptake limited its short-term effect [64,65]. The first fertilizer application also supplied P and K according to soil fertility recommendations, which may have contributed to forage responses in T2. Therefore, responses to this treatment cannot be attributed exclusively to inorganic N. Similarly, T4 and T5 (organic amendments) did not affect any forage variable; this may be due to organic N requiring a longer period for mineralization [66,67], while part of the N from these sources may also have been lost through ammonia volatilization because of their high moisture content [42,68].

Overall, legumes were not a superior short-term N source but rather a complementary, time-dependent input, with greater effects on forage nutritive value than on biomass. The benefits of legume integration are generally more evident over longer periods as residues decompose and nutrients are gradually recycled within the system [69,70].

The correlation patterns reinforce the well-established trade-off between forage accumulation and nutritive value. As forage maturity advances, structural carbohydrates accumulate, increasing fiber concentrations while reducing digestibility through changes in cell wall composition [71,72,73]. Consequently, greater forage mass did not necessarily translate into improved forage quality, suggesting that forage maturity exerted a stronger influence on forage nutritive value than N source under the conditions of this study.

4.2. Soil Health Indicators

Nitrogen inputs differed across treatments in both quantity and form, particularly in plant- and soil available N. While T2 supplied a defined amount of readily available N, T3 relied on biological N fixation, and T4 and T5 provided N primarily in organic forms requiring mineralization. Therefore, soil responses reflect differences in N availability rather than total N applied.

The high demand for N by the plants during the evaluation period likely limited detectable changes in soil carbon and nitrogen concentrations [55]. Relative to the T1, T2, T4 and T5, only T3 increased labile nitrogen indicators (PMN, EA, EN, IN) and microbial C:N ratio, reflecting a more active nitrogen cycle [74,75]. This response is consistent with the increase in crude protein concentration observed under T3 and suggests that pre-inoculated legumes may have contributed to nitrogen inputs through biological fixation, although the effectiveness of commercial inoculation remains variable [76,77,78].

The N input from treatments T4 and T5 would be expected to affect soil biological activity, since the application of poultry litter and cattle dung in agricultural systems increases soil organic matter concentration [79,80], especially in sandy soils [81,82], such as the soil of this study. However, neither T4 nor T5 differed from the control, indicating that N from these sources did not contribute to labile soil N pools within the evaluation period. This is consistent with the slower mineralization of organic amendments [83,84] and the spatially localized and short-lived nature of dung deposition [85,86], which could limit their detection [42,87].

Similarly, no significant differences were observed in soil physical, chemical, and biological indicators, including WAS, SOC, SON, MBC, MBN, BSR, and NAG (Table A1) following the application of organic amendments T4 and T5. The short evaluation period and the high plant N demand limited detectable changes in soil organic matter pools, microbial biomass, and soil structure. Likewise, SON remained unchanged because organic N pools generally respond more slowly than labile mineral N fractions, while MBC responses often require sustained changes in carbon inputs and substrate availability beyond the 60-day evaluation period. Comparable responses have been reported in systems receiving organic inputs, where microbial biomass and soil physical indicators remained unchanged despite increased nutrient availability [88]. Changes in SOC and associated soil aggregation are primarily linked to long-term management rather than short-term inputs [83,89]. In addition, microbial and enzymatic responses depend on substrate availability and mineralization dynamics, which may not be expressed within short timeframes.

The observed correlation patterns are consistent with the conceptual framework of soil N cycling, where microbial activity regulates the transformation of organic N into plant-available mineral forms. Potentially mineralizable N represents the biologically active fraction of soil organic N, whereas extractable ammonium, nitrate, and inorganic N reflect successive products of microbial mineralization and nitrification [90,91]. Likewise, the positive association with NAG agrees with its recognized role in the depolymerization of organic N compounds and subsequent N release [92]. Together, these relationships support the close coupling between microbial activity and short-term N cycling under the conditions of this study, rather than indicating direct treatment effects on stable soil C and N pools.

Although SOC and SON were associated with indicators of N cycling, these relatively stable soil pools are not expected to change substantially over 60 days. Consequently, the observed relationships likely reflect the inherent linkage between stable and labile soil N pools rather than treatment-induced changes in soil organic matter [89].

In addition, herbicides such as 2,4-D may influence root exudation patterns and microbial community composition; these changes would ultimately be expected to affect soil microbial biomass and activity. In the present study, however, no consistent differences were observed among the non-legume treatments for microbial biomass or activity indicators (MBC, MBN, BSR, and NAG), suggesting no detectable short-term effects on the measured soil biological functions. Nevertheless, microbial community composition and root exudation were not directly evaluated and therefore remain beyond the scope of this study.

4.3. Greenhouse Gas Emissions

The high CH₄ flux in treatment T5 (cattle dung) during the first few days after application and in the mean daily values is likely due to the formation of anaerobic microsites (with high moisture concentration) favorable for methanogenesis, as well as the availability of C and N substrates [42,93]. According to Saggar et al. [94] and Chadwick et al. [95], most of the CH₄ emitted from dung occurs during the first week, which explains the initial peaks and subsequent decrease in fluxes from day 1 to 4 and from day 7 to 51; this is probably because the dung was losing moisture and labile C and N. In contrast, treatments T1, T2, T3, and T4 exhibited low CH₄ fluxes, with values remaining close to zero and occasionally negative. These results indicate that dung application, rather than the N fertilization source, was the primary factor associated with the greater CH₄ fluxes observed in T5. However, unlike inorganic fertilizer and poultry litter, cattle dung represents an internal nutrient recycling process in grazing systems through localized nutrient return [42].

Furthermore, although the N fertilization sources affected CH₄ fluxes from day 1 to 4, this was not the case for CO₂, which varied over time, indicating that environmental conditions and short-term microbial activity primarily influenced soil respiration rather than the N source [96]. In contrast, the N source has affected the mean daily CO₂ values, which reached their highest peaks under T2 and T5, suggesting rapid microbial respiration driven by readily available, more labile carbon substrates. However, soil CO₂ fluxes in vegetated systems largely represent biogenic respiration and may reflect differences in root activity and microbial decomposition rather than environmental impacts per se. Therefore, these responses should not be interpreted as a direct measure of overall environmental performance. In field studies, cattle dung has been shown to substantially increase CO₂ emissions relative to urea, with daily fluxes 42% higher under solid dung applications [97]. Based on field measurements in temperate wheat–maize–soybean rotations, legume systems typically emit 2.6 to 3.0 kg CO₂–C ha⁻¹ day⁻¹, compared with 3.4 kg CO₂–C ha⁻¹ day⁻¹ under mineral N fertilization [98], a pattern consistent with the lower CO₂–C emissions observed under the legume treatment in this study.

In contrast, the lack of significance in N sources and over time in N₂O fluxes likely indicates that rapid plant uptake and N microbial immobilization limited nitrate accumulation and the denitrification process [99,100]. A study in winter annual pastures reported temporary N₂O peaks within 2 to 5 days of application, with daily emissions generally remaining below 1 kg N₂O–N ha⁻¹ over multi-week periods [101].

The contrasting correlation patterns between CH₄ and CO₂ across sampling dates suggest that the processes regulating these gases changed during the experimental period. Early in the experiment, both gases likely reflected the intense microbial activity associated with fresh dung deposition, whereas later their responses became less synchronized as dung patches decomposed. Similar temporal patterns have been reported for cattle dung, where CH₄ emissions decline progressively following deposition and soils beneath decomposed dung pats may eventually shift to net CH₄ oxidation. These findings reinforce that CH₄ emissions from localized dung patches are primarily governed by short-term biophysical conditions during dung decomposition.

The CH₄ fluxes reported for T5 represent emissions measured directly over individual fresh dung patches enclosed by the automated LI-COR chambers. Therefore, although expressed per unit area following standard chamber methodology, these values should not be interpreted as whole-pasture methane emissions because fresh dung occupies only a small fraction of the grazed area.

While this study provides important information on short-term soil–plant interactions under different nitrogen fertilization sources in cool-season forage systems, it is important to acknowledge some limitations that may have influenced the observed responses of the studied indicators. The short time frame of this study may not allow for changes in some slowly responding soil properties. The lack of interpretation of nitrogen and carbon dynamics is restricted by the inability to directly measure biological nitrogen fixation or the treatment-specific soil organic matter fractions. Thus, future research should build on this information by evaluating several seasons over a more extended time period, obtaining direct measurements of biological nitrogen fixation, assessing the composition of soil organic matter fractions, and conducting a longer-term comparative evaluation of greenhouse gas fluxes.

Finally, although none of the treatments increased forage mass in the short term, T3 increased forage crude protein and several indicators of labile soil N without increasing CH₄ or N₂O fluxes during the evaluation period. These findings highlight the potential contribution of legume inclusion to short-term N cycling, while longer-term studies are required to evaluate whole-system environmental performance.

5. Conclusions

Alternative N sources may help diversify N management strategies in cool-season forage systems. Under the conditions of this 60-day field study, the grass–legume system improved forage crude protein concentration and enhanced selected indicators of soil N availability without increasing mean daily CH₄ or N₂O fluxes. Inorganic N and poultry litter produced limited short-term effects on forage nutritive value and soil biological indicators. Cattle dung was associated with greater CH₄ fluxes immediately after application but showed no consistent differences from the other N sources during the remainder of the evaluation period. These findings demonstrate that different N sources produce distinct short-term responses in forage nutritive value, soil N dynamics, and greenhouse gas fluxes. However, this study evaluated only the initial 60-day response, longer-term and multi-season; thus, future long-term studies are needed to determine the persistence of these effects and their implications for soil C and N dynamics, nutrient cycling, and overall system sustainability.