Effect of Poultry Litter Application Method and Rainfall and Delayed Wrapping on Warm-Season Grass Baleage

Abstract

Poultry litter is a widely available fertilizer in the southeast USA and subsurface application of litter can increase both forage production and nutritive value. Frequent rainfall events and high humidity often limit time available for hay curing; baled silage techniques can increase harvest time flexibility. Unfortunately, rainfall events can still occur without forecast during harvest events, resulting in delayed baling or wrapping. The objective of this study was to evaluate poultry litter amendment methods, subsurface (SUB) and surface (SURF), and the effect of no rain (NR) on bales with wrapping after 2 h compared with rained-on bales with 17 h delayed wrapping (RDW) on warm-season grass baleage fermentation and nutritive value. Data were analyzed as a randomized complete block design with two amendment treatments and two post-baling treatments. Crude protein (CP) was greater (p < 0.01) and neutral detergent fiber (NDF) was lower (p < 0.01) in both pre- and post-ensiled bales with subsurface-applied poultry litter. Rain and delayed wrapping resulted in lower pH (p = 0.03), starch (p < 0.01), and water-soluble carbohydrates (p < 0.01) in pre-ensiled bales, compared to those that did not receive rain and were wrapped within 2 h, while post-ensiled bales only differed in lower (p < 0.01) starch and slightly greater (p < 0.01) NDF in RDW. Lactic acid (p < 0.01), acetic acid (p < 0.01), and total acids (p = 0.03) were greater in SUB, while butyric acid tended to be greater (p = 0.09), and alcohols (p = 0.05) were greater in SURF. Bales from RDW and NR only differed by greater (p < 0.01) propionic acid concentrations in NR. Under the conditions of this experiment, subsurface application of poultry litter increased final nutritive value, while rainfall and delayed wrapping of 17 h had few effects on the final nutritive value of warm-season grass baleage.

1. Introduction

Frequent rain events and humidity in the southeastern USA have encouraged the use of baleage as an alternative to dry hay. Winter annual forages such as cereal rye (Secale cereale L.), annual ryegrass (Lolium multiflorium L.), and wheat (Triticum aestivum L.) are generally used for baleage because they produce large amounts of forage mass of high nutritive value in the spring [1,2]. However, longer drying times due to high summer humidity and the improved nutritive value of baleage compared to similar hay stored outdoors also provide a potential advantage for warm-season baleage over dry hay [3].

Forages in the southeastern USA are commonly fertilized with poultry litter. Generally, litter is broadcast over the soil and forage surface which can lead to nutrient losses from rainfall [4,5] and N loss through ammonia volatilization [6,7]. Injecting poultry litter below the soil surface reduces volatilization [7,8,9] and runoff [5,9,10]. Additionally, injection of poultry litter can increase forage yield [5,9,11] and forage N concentrations [11,12].

Despite the increased flexibility for forage harvest and storage with baleage, rain events may still occur during the process. Effects of rainfall events on wilting hay are well documented. Simulated rainfall on wilting bermudagrass resulted in greater dry matter (DM) losses [13] and reduced nutritive value, particularly nonstructural carbohydrates [14], although the extent is dependent on moisture of the wilting forage at the time of rainfall and level of nutritive value of the wilted forage [13]. Similarly, baled hays subjected to rainfall or outdoor storage have greater nonstructural carbohydrate losses, resulting in greater acid detergent fiber (ADF) concentrations and DM losses compared to barn stored hays [3]. However, effects on high-moisture bales and subsequent delayed wrapping are largely unknown. Responses to delayed wrapping have been documented, but are variable. Delayed wrapping (0–3 days) of alfalfa baleage resulted in greater final baleage pH, reduced volatile fatty acid (VFA) concentrations, and greater neutral detergent insoluble nitrogen (NDIN) and acid detergent insoluble nitrogen (ADIN) [15]. Additionally, greater final baleage pH and ADIN were also observed with delayed wrapping of oat silage with only a 19 h delay [16].

This project aimed to evaluate the effects of poultry litter amendment below the soil surface- (SUB) versus surface-applied poultry litter (SURF) on the nutritive value of summer forages. However, an unexpected rainfall event led to an additional comparison of poultry litter amendment method and bales wrapped within 2 h of baling (NR) and bales that were rained on and wrapped the next day (17 h delayed wrapping; RDW). After the rainfall event, our hypothesis was that SUB bales would have greater nutritive value, particularly crude protein (CP) concentrations. Furthermore, RDW bales would have lower nonstructural carbohydrates prior to ensiling and lower VFA concentration post fermentation.

2. Materials and Methods

2.1. Description of Field Site and Experimental Layout

A 2.6 ha field site on the USDA-ARS Dale Bumpers Small Farms Research Center, located near Booneville, AR (35.1401° N, 93.9216° W), was selected for the experiment with harvest in July 2020. The forage of this location was primarily warm-season grass consisting of approximately 35% bermudagrass (Cynodon dactylon L.), 28% barnyard grass [Echinochloa Crus-Galli (L.) Beauv.], 22% giant foxtail (Setaria faberi Herrm.), and 15% broadleaf weeds. The soil type was a Leadvale silt loam (fine-silty, siliceous, semiactive, thermic Typic Fragiudults). Mehlich-3 extractable P, K, Ca, Mg, and SO₄-S concentrations were 22, 70, 997, 121, and 13 mg/kg, respectively; and the soil pH was 6.0 (University of Arkansas Diagnostic Laboratory, Fayetteville, AR). The site was divided into three experimental blocks from north to south with two plots per block for a total of six plots. All six plots were 0.4 ha and were separated by 6 m buffer. Plots were allocated randomly to one of two treatments (SUB or SURF) within each block.

2.2. Site Preparation and Application of Treatments

The study was a randomized complete block design with two factors, poultry litter application type and post-baling management. The poultry litter application types were poultry litter applied as SUB or SURF. The post-baling management treatments were not intentional. An unplanned rainfall event occurred on 20 July, resulting in random unwrapped bales receiving 28.6 mm of rain after baling. Bales that received rain were wrapped the following day, after a 17 h delay. Originally, the design included a moisture treatment, which required approximately half the bales on each plot to be baled first to generate the high moisture treatment. This treatment was abandoned due logistical constraints; shortly after starting baling on the second halves of each plot, the rain event occurred. The order for baling plots was random; therefore, plots with bales that were rained on were “selected” randomly. The experiment maintained a randomized design and was analyzed as such.

The site was sprayed with glyphosate on 25 March 2020 (1185 g a.e.·ha⁻¹) to kill early spring weeds. Subsequently, the site was cut and hayed on 11 June 2020 to remove any spring or early summer forage mass. On 19 June 2020, all plots were amended with poultry litter collected on site from houses for broiler (Gallus gallus domesticus) production at a rate of 6.7 Mg·ha⁻¹. The litter nutrient analysis was 4.81%, 1.99%, and 3.78% N, P₂O₅, and K₂O, respectively (University of Arkansas Diagnostic Laboratory, Fayetteville, AR, USA). To three plots, one from each block, poultry litter was applied via injection (SUB), while the remaining three plots had poultry litter applied as SURF. The equipment used was a tractor-drawn prototype described in detail by Pote et al. [9]. Briefly, the subsurface applicator was designed with four trench openers with a fluted coulter to slice the soil. Each coulter was followed by a double-disc opener, which formed the trench. Following the deposition of litter in the trench, the injected litter was covered with soil. This no-till band technique minimizes soil disturbance and has the added benefit of pulverizing the poultry litter, thus precluding the need for pre-grinding the litter [9]. The application bands were 5.1 cm wide and 7.6 cm deep, with eight bands per spreader pass, configured with 38.1 cm band spacings. To ensure a consistent application rate, all poultry litter was applied using the poultry litter injector; for SURF treatments, coulters were lifted above ground for deposition directly on the soil surface.

2.3. Harvest

Each of the six plots were mowed and roller-conditioned in swaths with a pull-type New Holland 7230 mower–conditioner (CNH America, London, UK) to a 15.2 cm stubble height at 1000 h on 20 July 2020. The maximum cutting width of the mower–conditioner was 3.05 m, and the swath width after conditioning was 1.68 m, indicating about a 45% reduction in width. After mowing, random grab samples (~2000 g per plot; wet basis) were taken from each plot by a technician walking a zigzag pattern that covered the length of the plot for determination of botanical composition. At 11:30 a.m. on 20 July, two adjacent swaths were merged, and baled at 12:30 p.m. into 1.2 m by 1.2 m round bales (884 ± 80 kg, wet basis) with a John Deere 469 Silage Special baler (John Deere, Moline, IL, USA). Bales were secured with three layers of 1.3 m CoverEdge netwrap (John Deere, Moline, IL, USA) when discharged from the baler. On 20 July, 11 bales were produced and wrapped within 2 h, and 10 bales were baled and rained on (28.6 mm) at 3:00 a.m. prior to wrapping, which was delayed until 8:00 a.m. on 21 July (17 h delayed wrapping). All forage from the plots was baled; however, bales that did not meet a threshold diameter of 1.2 m were ejected from the baler and not used in the analysis. All bales meeting the 1.2 m diameter threshold were tagged, measured, weighed, and core-sampled prior to wrapping; wrapping occurred within 2 h of baling. The number of bales per block and treatment varied: Block 1: one SURF NR, one SUB RDW, one SURF RDW, one SUB NR; Block 2: two SURF NR, two SURF RDW, one SURF RDW; Block 3: three SURF NR, two SURF NR, two SUB RDW, three SURF NR, three SUB RDW.

Six core samples (2.5 cm diameter and 0.61 m deep) were obtained with a Uni-Forage Sampler (Star Quality Samplers, Irricana, AB, Canada) at evenly spaced areas from the round side of the bale, such that all of the round side of the bale was sampled. Holes from sampling were backfilled with painters’ drop cloth packed with a wooden dowel. All core samples were composited by bale, sealed in plastic freezer bags, placed on ice in insulated coolers, and then taken to the research laboratory for processing.

Each bale was wrapped with 750 mm by 1500 m Sunfilm Silage Wrap (AEP Industries Inc., Montvale, NJ, USA) (thickness = 25 μm) using a Model SBW8500 single-bale wrapper (Vermeer Corporation, Pella, IA, USA) programmed to apply eight plastic layers (32 table revolutions) to each bale. After wrapping, the bales were stored on a gravel pad until the final post-ensiled sampling date on 18 November 2020.

After a storage period of 121 days, nine evenly spaced cores (2.5 cm diameter and 0.61 m deep) were taken from the entire circumference of the rounded side of each bale using the same equipment for the pre-ensiled sampling. All samples were collected, composited by bale, placed in freezer bags, and stored on ice in insulated coolers until transport to the laboratory for processing and subsequent laboratory analysis.

2.4. Laboratory Procedures

For pre-ensiled samples, one 50 g subsample was taken from the freezer bags after thorough mixing. The subsample was dried for 48 h at 105 °C to determine the initial moisture concentration of each bale. A 25 g sample was separated and mixed with 100 mL of deionized water for determination of pH. The remaining sample (250 g) was stored in an ultralow freezer (−80 °C) pending subsequent lyophilization, and then ground through a Model 4 Wiley Mill equipped with a 1 mm screen (Thomas Scientific, West Washington Square, PA, USA). Water-soluble carbohydrates (WSC), starch, Ca, P, K, Mg, K, and S were determined in a commercial lab (Rock River Laboratory, Watertown, WI, USA) using the following methods: water-soluble carbohydrates were analyzed as described Dubois et al. [17]; starch was determined using the method of Hall [18]; for mineral concentrations, the samples were digested at 120 °C in acid to extract minerals, and then analyzed via inductively coupled plasma-optical emission spectrometry.

For post-ensiled samples, one 50 g subsample was taken from each bale and dried for 48 h at 105 °C to determine the final moisture concentration of each silage. The remainder of the sample was placed in an ultralow freezer at −80 °C. The following week, 150 g of the thoroughly mixed post-ensiled sample was shipped to the same commercial lab for determination of the pH and fermentation products using the following methods: for pH, forage samples were mixed with deionized water and read with a combination pH electrode in the commercial laboratory; volatile fatty acids were extracted from the sample in a 1:10 ratio of the sample to deionized water and centrifuged, the supernatant was combined with calcium hydroxide and copper sulfate and centrifuged again, and then the supernatant was analyzed by high-performance liquid chromatography equipped with a reverse-phase ion exclusion column and a refractive index detector (Waters Corporation, Milford, MA, USA); for ammonia-N, the same supernatant was produced as in the volatile fatty acids procedure and was analyzed using a Skalar San++ Segmented Flow SA 5000 Analyzer (Skalar, Breda, The Netherlands) on the basis of the Berthelot reaction [19]; another 100 g of the sample was lyophilized, ground as described for pre-ensiled samples, and sent to the same commercial lab for analysis following the same methods for WSC and starch described previously.

The fiber composition of the dried ground forage samples was analyzed sequentially for neutral detergent fiber (NDF) and acid detergent fiber (ADF) via the batch procedures outlined for an Ankom200 Fiber Analyzer (Ankom Technology Corporation, Macedon, NY, USA). The concentration of total N was determined via a rapid combustion procedure (Elementar VarioMax, Elementar Ltd., Langenselbold, Germany) [20], followed by conversion to crude protein (CP) with a factor of 6.25. Neutral and acid detergent insoluble CP (NDICP and ADICP, respectively) were determined through nonsequential NDF and ADF procedures that excluded sodium sulfite and heat-stable α-amylase [21,22]; residual CP was determined using the combustion procedure described previously.

2.5. Statistical Analyses

All analyses of the initial and final bale characteristics and nutritive value were analyzed by PROC GLIMMIX in SAS Version 9.4 (SAS Institute Inc., Cary, NC, USA). The experimental design was a randomized complete block with three field blocks. Data were analyzed as a 2 × 2 factorial with two poultry litter amendment treatments (SUB and SURF) and two post-baling management treatments (NR and RDW). Amendment treatments and rain treatments were evaluated as fixed effects, whereas blocks were considered random. The LSMEANS option was used to generate the individual treatment means. The model was run with interactions; however, only a few irrelevant interactions were detected and, thus, omitted from the results and discussion. The pre- and post-ensiling changes in chemical composition were analyzed as stated for the individual components, and it was determined whether the changes differed from zero using the LSMEANS statement in SAS. Significance was considered at p ≤ 0.05 and tendencies were considered at p > 0.05 and ≤0.10.

3. Results

3.1. Pre-Ensiled Bale Characteristics and Nutritive Value

Pre-ensiled bales did not differ for any variables including initial DM, wet weight, dry weight, volume, or density (

Table 1. Initial characteristics of wrapped silage bales following application with poultry litter by surface (SURF) or subsurface (SUB) methods and wrapped in 2 h without rain (NR) and with rain and delayed wrapping after 17 h (RDW) at Booneville, AR, in 2020.

Treatment	DM 1	Wet Weight	Dry Weight	Bale Height	Volume	Density
	g/kg	kg	kg	m	m3	kg DM/m3
Application
SUB	318	871	278	1.41	1.86	150.0
SURF	325	894	289	1.42	1.88	153.3
SEM	16.8	28.5	11.6	0.012	0.032	6.44
p-value	0.71	0.55	0.30	0.64	0.64	0.60
Post-baling treatment
NR	330	862	284	1.41	1.86	152.3
RDW	313	904	283	1.41	1.88	150.9
SEM 2	16.3	27.4	11.4	0.012	0.031	6.29
p-value	0.38	0.28	0.88	0.72	0.75	0.82

¹ DM = dry matter; ² SEM = standard error of the mean.

). Nutritive value variables for pre-ensiled bales are presented in

Table 2. Pre-ensiled pH, concentrations of carbohydrates and minerals, crude protein (CP), and fiber-bound CP fractions of wrapped silage bales following application with poultry litter by surface (SURF) or subsurface (SUB) methods and wrapped in 2 h without rain (NR) or with rain and delayed wrapping after 17 h (RDW) at Booneville, AR, in 2020.

Treatment	pH	Starch	WSC 1	CP	NDF	ADF	Hemicellulose	NDICP	ADICP	Ca	P	Mg	K	S
		g/kg						g/kg CP		g/kg
Application
SUB	6.4	18.2	57.6	133	653	343	311.2	86.8	42.0	3.0	2.5	1.3	29.9	2.3
SURF	6.5	18.8	57.2	124	683	351	332.5	81.4	44.8	2.8	2.6	1.3	28.2	2.4
SEM 2	0.16	0.68	2.18	0.2	6.1	3.8	6.23	2.12	5.30	0.09	0.12	0.09	1.77	0.15
p-value	0.49	0.56	0.87	<0.01	<0.01	0.10	<0.01	0.07	0.63	0.22	0.16	0.95	0.33	0.44
Post-baling treatment
NR	6.59	21.4	64.5	127	664	341	322.9	86.5	41.2	2.9	2.7	1.3	29.9	2.3
RDW	6.25	15.6	50.3	131	672	352	320.8	81.6	45.7	2.9	2.5	1.2	28.1	2.3
SEM	0.16	0.65	2.10	2.3	5.9	3.7	6.09	2.00	5.15	0.10	0.12	0.08	1.73	0.15
p-value	0.03	<0.01	<0.01	0.08	0.27	0.02	0.72	0.09	0.43	0.62	0.02	0.55	0.30	0.76

¹ WSC = water-soluble carbohydrate; CP = crude protein; NDF = neutral detergent fiber; ADF = acid detergent fiber; NDICP = neutral detergent insoluble crude protein; ADICP = acid detergent insoluble crude protein; DM = dry matter; ² SEM = standard error of the mean.

. The poultry litter amendment method affected CP, NDF, and hemicellulose with greater (p < 0.01) CP in SUB and greater (p < 0.01) NDF and hemicellulose in SURF, while a tendency was observed for NDICP with greater (p = 0.07) NDICP in SUB. Post-baling events had several effects on pre-ensiled bales. Pre-ensiled pH was lower (p = 0.03) for RDW compared to NR. Starch and WSC were lower in RDW (p < 0.01), while ADF was greater (p = 0.02) in RDW. Phosphorus was lower (p = 0.02) in RDW compared to NR. A tendency was observed for greater (p = 0.08) CP in RDW and greater (p = 0.09) NDICP in NR.

3.2. Post-Ensiled Nutritive Value and Changes in Forage Chemical Components

Crude protein, NDF, and hemicellulose differed for poultry litter application method’; CP was greater (p < 0.01) in SUB, while NDF and hemicellulose were greater (p < 0.01) in SURF (

Table 3. Post-ensiled pH, concentrations of carbohydrates and minerals, crude protein (CP), and fiber-bound CP fractions of wrapped silage bales following application with poultry litter by surface (SURF) or subsurface (SUB) methods and wrapped in 2 h without rain (NR) or with rain and delayed wrapping after 17 h (RDW) at Booneville, AR, USA in 2020.

Treatment	pH	Starch	WSC 1	CP	NDF	ADF	Hemicellulose	NDICP	ADICP
		g/kg						g/kg CP
Application
SUB	4.4	8.6	15.3	138	619	351	266.1	55.8	35.1
SURF	4.5	9.4	16.1	129	639	354	283.9	56.2	38.5
SEM 2	0.04	0.88	3.90	2.0	3.86	4.9	3.52	3.11	2.50
p-value	0.16	0.34	0.82	<0.01	<0.01	0.43	<0.01	0.89	0.31
Post-baling treatment
NR	4.4	10.4	17.8	134	619	348	268.9	54.6	38.7
RDW	4.4	7.6	13.6	133	639	356	280.9	57.4	34.9
SEM	0.04	0.86	2.1	1.9	3.7	4.9	3.39	3.05	2.41
p-value	0.89	<0.01	0.21	0.77	<0.01	0.06	0.02	0.29	0.06

¹ WSC = water-soluble carbohydrates; CP = crude protein; NDF = neutral detergent fiber; ADF = acid detergent fiber; NDICP = neutral detergent insoluble crude protein; ADICP = acid detergent insoluble crude protein; DM = dry matter; ² SEM = standard error of the mean.

). Starch was greater (p < 0.01) in NR, while NDF and hemicellulose were greater (p < 0.02) in RDW. Tendencies were observed for ADF, with greater (p = 0.06) ADF in RDW and greater (p = 0.06) ADICP for NR.

Pre-ensiling and post-ensiling changes in forage chemical components are presented in

Table 4. Pre-ensiling and post-ensiling changes in forage chemical components for wrapped silage bales following application with poultry litter by surface (SURF) or subsurface (SUB) methods and wrapped in 2 h without rain (NR) or with rain and delayed wrapping after 17 h (RDW) at Booneville, AR, in 2020.

Treatment	pH	Starch	WSC 1	CP	NDF	ADF	Hemicellose	NDICP	ADICP
		g/kg						g/kg CP
Application
SUB	−1.98 *	−8.7 *	−40.4 *	0.3	−25.9 *	11.0 *	−36.9 *	−31.6 *	−7.1
SURF	−1.97 *	−10.2 *	−44.8 *	8.1 *	−50.4 *	3.1	−53.5 *	−25.8 *	−6.7
SEM 2	0.156	1.23	3.16	3.63	10.20	4.54	8.00	2.75	5.92
p-value	0.96	0.30	0.30	0.05	0.09	0.21	0.14	0.13	0.96
Post-baling treatment
NR	−2.12 *	−11.9 *	−48.3 *	2.8	−27.4 *	13.0 *	−40.4 *	−32.4 *	−2.3
RDW	−1.83 *	−7.0 *	−36.9 *	5.6	−48.8 *	1.1	−50.0 *	−25.0 *	−11.6
SEM	0.152	1.19	3.04	3.54	9.83	4.36	7.69	2.65	5.72
p-value	0.07	<0.01	0.02	0.46	0.133	0.07	0.38	0.06	0.2

¹ WSC = water-soluble carbohydrates; CP = crude protein; NDF = neutral detergent fiber; ADF = acid detergent fiber; NDICP = neutral detergent insoluble crude protein; ADICP = acid detergent insoluble crude protein; DM = dry matter; ² SEM = standard error of the mean. * Means differed from zero (p ≤ 0.05), indicating a statistical change in pre- and post-ensiling concentrations.

. Changes during storage on the basis of poultry application method were only observed for CP, in which the difference was greater (p = 0.05) for SURF. A similar tendency also was observed for NDF, in which the difference was greater (p = 0.09) for SURF. Following bale storage, concentration changes in starch and WSC were greater (p ≤ 0.02) for NR. Several tendencies also were observed, where pH, ADF, and NDICP differences tended (0.05 < p ≤ 0.07) to be greater in NR than RDW. Several differences from multiple variables for both poultry litter amendment method and post-baling treatment were different (p < 0.05) from zero (no change). The only variables that did not change were CP for SUB, NR, and RDW, ADF for SURF and RDW, and ADICP for all treatments. All differences trended to be reduced from pre to post ensiled, except for CP and ADF which either did not differ from zero or increased slightly.

3.3. Post-Ensiled Nutritive Value

Several fermentation parameters differed for poultry litter application method (

Table 5. Post-ensiled fermentation parameters for wrapped silage bales following application with poultry litter by surface (SURF) or subsurface (SUB) methods and wrapped within 2 h without rain (NR) or with rain and delayed wrapping after 17 h (RDW) at Booneville, AR, USA in 2020.

Treatment	DM 1	Ammonia-N	Lactic Acid	Acetic Acid	Butyric Acid	Propionic Acid	Ethanol	Total Acids	Total Alcohols
Application	g/kg	g/kg CP	g/kg
SUB	276	21.6	40.2	10.3	1.6	4.2	5.5	58.0	8.9
SURF	291	17.5	30.0	6.9	5.9	3.1	7.3	47.1	10.3
SEM 2	11.6	1.30	2.21	0.76	1.80	0.45	0.41	3.46	0.50
p-value	0.33	0.03	<0.01	<0.01	0.09	0.07	<0.01	0.03	0.05
Post-baling treatment
NR	292	19.5	34.0	7.6	2.2	4.5	6.9	50.0	10.1
RWD	275	19.7	36.3	9.7	5.2	2.7	6.0	55.2	9.2
SEM	11.2	1.25	2.13	0.74	1.78	0.44	0.39	3.33	0.48
p-value	0.28	0.92	0.40	0.06	0.21	<0.01	0.13	0.27	0.18

¹ DM = dry matter; ² SEM = standard error of the mean.

). Ammonia-N, lactic acid, acetic acid, and total acids were greater (p ≤ 0.03) for SUB than SURF. Ethanol and total alcohols were greater (p ≤ 0.05) for SURF than SUB. Tendencies were observed for butyric acid, in which SURF was greater (p = 0.09) than SUB. Furthermore, ammonia-N (% CP) (p = 0.08) and propionic acid were greater (p = 0.07) for SUB than SURF. Few variables differed for post-baling treatments with only propionic acid being greater (p < 0.01) in NR compared to RDW. A tendency was observed for acetic acid, with RDW greater (p = 0.06) than NR.

4. Discussion

4.1. Pre-Ensiled Bale Characteristics and Nutritive Value

Across all treatments, bale DM averaged 322 g/kg, lower than the recommended moisture range for baleage of 450 g/kg to 550 g/kg [23]. Appropriate baleage DM supports fermentation, and bales of greater DM concentrations (lower moisture) are generally greater in final pH as a result of reduced VFA production. Lactic acid and total acid production generally increase with bale moisture [1,24]. However, bales with an initial bale moisture in the 300 g/kg DM range may result in clostridial fermentations characterized by increased butyrate and of ammonia-N concentrations [25]. Bale density was lower than ideal bale density of 162 kg DM/m³ [26], due to the high bale moisture, as bale density is inversely related to bale moisture.

Greater CP in SUB as opposed to SURF forages has been previously documented. Kulesza et al. [12] observed greater N concentrations in orchard grass amended by SUB compared to SURF. The authors hypothesized that subsurface application increased NH₄-N capture, providing more N for forage uptake. Similar results were observed for bermudagrass forages amended with SUB compared to SURF [11]. Nitrogen application on bermudagrass also resulted in lower NDF compared to unfertilized bermudagrass [27]. Lower NDF and hemicellulose in SUB could have been a result of greater N-uptake that resulted in greater leaf growth. Mineral concentration was not improved with SUB, which was also observed previously by Nieman et al. [1].

Rain and delayed wrapping affected several nutritive variables. For RDW bales, initial pre-ensiled samples were taken 17 h after baling, and the lower initial pH of RDW indicates that fermentation started occurring internally prior to sampling. This is supported by reductions in WSC and starch, which could have been fermented, contributing to the lower pH, although some losses of starch and WSC could have occurred because of exposure to oxygen and continuation of the respiratory process [28]. Acid detergent fiber was greater in RDW, which may have been a dilution effect caused by losses of nonstructural carbohydrates, resulting in an increased proportion of cell-wall components. The respiratory process can also cause heating to temperatures that reduce protein availability [29], which have been observed to increase ADF, NDF, and ADICP in silages with spontaneous heating [15]. Linear trends of increased ADF, NDF, and ADICP have been observed in pre-ensiled bales with delayed wrapping from 1 to 3 days in annual ryegrass baleage, while nonstructural carbohydrate and total digestible nutrients decreased linearly [30]. Phosphorous was lower in RDW bales. Mineral leaching can occur in wilting forage when rainfall events occur during wilting [28]; minerals can make up 14% of the total dry matter lost as a result of a rain event on wilting alfalfa [31]. Although the rain event did not occur until after baling, and although the net wrap likely prevented most leaching, a small amount of mineral leaching could have occurred prior to wrapping, but this is unlikely.

4.2. Post-Ensiled Nutritive Value and Changes in Forage Chemical Components

For amendment type, greater CP was observed in SUB bales, and greater NDF and hemicellulose was observed in SURF bales, likely for the same reasons as discussed for pre-ensiled bales. For post-baling management type, starch was greater in NR compared to RDW, but WSC did not differ between treatments on a post-ensiled basis. Neutral detergent fiber and hemicellulose were greater in RDW, and trends were observed for greater ADF in RDW and greater ADICP in NR. Nia and Wittenberg [16] also observed few post-ensiled differences among bales with delayed wrapping times of 2, 10, and 19 h after baling, noting lower WSC at 19 h compared to wrapping delays of 2 or 10 h, and greater acid detergent insoluble N (ADIN) in bales with 19 h delayed wrapping compared to 2 or 10 h delayed wrapping. The increased ADIN was attributed to the greater bale temperature in bales with 19 h delayed wrapping, as greater oxygen exposure allowed more oxidation to occur, increasing bale temperatures [16]. Coblentz et al. [15] also observed linear increases in ADICP and NDICP for alfalfa bales with wrapping delays ranging from 0 to 3 days. However, Crook et al. [30] did not observe an increase in ADICP in ryegrass baleage with delayed wrapping ranging from 0 to 3 days. Despite a 17 h delayed wrapping, ADICP was not increased in RDW in this study.

For differences between pre- and post-ensiled conditions, only differences in CP were observed for SUB and SURF, with increased CP in SURF. Although differences between SURF and SUB in NDF were not significant, CP may have increased in SURF because of numerically greater losses of NDF and hemicellulose. For post-baling treatment, losses of starch and WSC were greater for NR than RDW as pre-ensiled values were lower for RDW, because of oxidation or fermentation losses that occurred prior to sampling. Changes in pH between NR and RDW were not significant, but there was a tendency and a change of 0.29 pH units. Previous research showed increased pH in cases of delayed wrapping at 19 h compared to 2 and 10 h delays for barley (Hordeum vulgare L.) baleage [16], as well as a linear response of increasing pH with delayed wrapping over 0–3 days for alfalfa baleage [15]. Crook et al. [30] observed a trend for a linear increase in pH from annual ryegrass with delayed wrapping from 0 to 3 days. The increase in pH related to delayed wrapping may be explained by the oxidation of nonstructural carbohydrates by aerobic microbes as a result of the longer exposure to oxygen, which limits the availability of the sugars for fermentation and the production of organic acids that reduce silage pH [32].

However, two larger trends were observed across treatments, a reduction in pH despite low WSC and reductions in NDF, hemicellulose, and NDICP from pre-ensiling to post-ensiling conditions. Bales in all treatments ensiled well, reducing pH to an average of 4.4. Adequate reductions in pH can be challenging when ensiling warm-season grasses [33] because of low WSC concentrations [34], although pH values similar to those observed here were observed in laboratory silo studies (4.08–4.40 for control treatments) [34,35], and slightly greater pH (5.37 in uninoculated treatments) was reported in bermudagrass baleage with moisture of 487 g/kg [36]. Water-soluble carbohydrates are the main substrate for microbes during fermentation process [28]. Although Adesogan et al. [34] and Dean et al. [35] achieved similar pH with bermudagrass in laboratory silos with lower levels of pre-ensiled WSC (15.0 g/kg WSC and 6.9 g/kg for Adesogan et al. [34] and Dean et al. [35], respectively), the reduction in NDF and hemicellulose in the current study indicates hydrolysis of hemicellulose, resulting in sugar units that support fermentation [37]. Generally, successful fermentations result in increased NDF, ADF, and ADL [38] as the proportion of structural fiber components increase with fermentation losses of nonstructural carbohydrates. However, pre- to post- ensiling reductions in NDICP were previously observed in baleage, particularly in wetter silages [1,24,39] indicating a release of CP from structural fiber components during fermentation.

4.3. Fermentation Parameters

The poultry litter application method affected several fermentation parameters. Ammonia-N was lower, but within the range of previous bermudagrass silage research of similar pre-ensiled CP levels [34]. Greater ammonia-N values were observed in SUB, a result of the greater CP concentrations in pre-ensiled samples, although the trend for greater ammonia-N also suggests that SUB bales underwent greater proteolysis. Nieman et al. [1] did not observe greater CP or ammonia-N levels in SUB versus SURF bales in rye/ryegrass/tall fescue baleage, although the lack of difference in that study may have been due to differences in botanical composition.

Concentrations of VFAs were within the range of previous bermudagrass silage from laboratory silos, as other authors observed lactic acid concentrations of 50.0 g/kg [35], 20.5 g/kg [40], and 10.4 g/kg [36]. Of the common VFAs, lactic acid most quickly reduces silage pH which has implications for stability and long-term storage [32]. Nonstructural carbohydrates serve as the main substrate for lactic acid-producing bacteria [28]. Although WSC and starch did not differ between SUB and SURF, the lower lactic acid, acetic acid, and total acids in SURF, and the trend for greater levels of butyric acid, ethanol, and total alcohols in SURF indicate that SURF may have been influenced by clostridial bacteria or other undesirable bacteria. Clostridia ferment carbohydrates and organic acids (particularly lactic acid) and produce butyric acid [32]; alcohol production can result from clostridia, but can also be produced by other undesirable bacteria [41]. Most pre-ensiled variables were similar between SUB and SURF, but SURF bales may have had greater exposure to clostridial or other undesirable bacteria from contamination from poultry litter. Coblentz et al. [42] observed clostridia populations 10-fold greater in alfalfa baleage fertilized with dairy slurry (4 weeks, 3 weeks, and 2 weeks prior to harvest) compared to control plots, although the authors did not observe differences in butyric acid concentrations among the control and slurry treatments. In the current study, poultry litter was applied 31 days prior to baleage making. Clostridia prefer high-moisture silages [32], which could have contributed to the trend for increased butyric acid in the current study compared to Coblentz et al. [42] in which silage DM ranged from 581 to 551 g/kg. Clostridia were not enumerated in the current study; however, bacterial contamination may be a concern as poultry litter fertilization has been associated with outbreaks of Clostridium botulinum [43]. Nevertheless, a laboratory silo experiment in which tall fescue was amended with poultry litter in late May and harvested for 2 years (3–4 harvests per season) did not result in greater butyric acid production or indications of clostridial fermentation, nor was Clostridium botulinum detected in the silage [44].

Few fermentation parameters were affected by post-harvest management. Propionic acid was greater in NR, and acetic acid tended to be greater RWD. Coblentz et al. [15] observed reduced total acids, and VFA with increased delay in wrapping alfalfa baleage from 0 to 3 days after baling. Crook et al. [30] observed decreased lactic acid and increased acetic acid, propionic acid, and butyric acid in annual ryegrass baleage with delayed wrapping from 0 to 3 days after baling. However, with a shorter delay (19 h), Nia and Wittenberg [16], observed reduced lactic acid and increased propionic acid, but observed no differences in acetic and butyric acid.

5. Conclusions

Subsurface application of poultry litter had several positive effects on final bale nutritive composition including increased crude protein, lactic acid, acetic acid, and total acids compared to baleage from plots with surface-applied litter. Increased butyric acid, ethanol, and total alcohol in bales with surface-applied poultry litter indicate lower overall nutritive value and possibly increased activity of clostridial bacteria or other undesirable bacteria due to contamination from surface-applied poultry litter. Although several pre-ensiled nutritive variables differed, including reduced water-soluble carbohydrates and starch in rainfall plus delayed wrapped bales, few differences were observed in final bale nutritive composition. Rainfall and delayed wrapping of 17 h caused reduced starch and increased fiber components, but lactic acid and total acid concentration did not differ in the final bale nutritive composition. Overall, amendment of subsurface applied poultry litter increased final bale nutritive composition, while rainfall plus delayed wrapping had negligible effects on final bale nutritive composition of warm-season grass baleage.