Assessment of the Response to Fertilization for the Sustainable Management of Native Grasses from Flooded Savannah Ecosystem Arauca, Colombian Orinoquia

Abstract

The native grasses of the flooded savannah ecosystem are produced under natural conditions and there is little information on the productive and nutritional response to the application of fertilizers. They are proposed as a strategy for adaptation to climate change and for the sustainable development of livestock farming. The aim of the study was to evaluate the response to low doses of fertilization of native grasses (“bank” grasses: Paspalum plicatulum, Panicum versicolor, and Paspalum sp. “Low” grasses: Leersia hexandra and Hymenachne amplexicaulis) in flooded savannah conditions. The green forage samples were taken in a 1 m² frame at 28-, 35-, and 42-day cutting intervals and biomass production was estimated with and without fertilization. After 35 days, the nutritional composition was analyzed by near-infrared reflectance spectroscopy (NIRS). The effect of fertilization and the grasses × cutting interval interaction influenced (p < 0.05) green forage (GF, t/ha) and dry matter (DM, t/ha). The effect of fertilization and the grasses × fertilization interaction on the nutritional composition only influenced the content of calcium (Ca²⁺) and magnesium (Mg²⁺) in the “low” grasses, while in the “bank” grasses, it influenced the sodium (Na) content (p < 0.05). The application of fertilizers generated significant differences in forage yield, but not in the general nutritional composition of grasses. However, some numerical variations were observed in favor of fertilized grasses. According to these results, the application of fertilizers will not be required to increase the value of the nutritional composition. Native grasses constitute an important sustainable food resource for livestock in flooded savannah ecosystems. This study constitutes the first approximation to understanding the behavior of native grasses for sustainable management in the flooded savannah ecosystem.

1. Introduction

The flooded savannahs are located on the left side of the Meta River, in the Colombian Orinoquia. This area represents an ecosystem with valuable diversity regarding native grasses and is essential for the development of cattle breeding and raising in extensive grazing [1,2]. Native grasses constitute about 90% of the food available for animals in flooded savannahs [1]. The flooded savannah ecosystem in Arauca (Eastern Colombia) is made up of different physiographic positions defined by the relief and drainage. In the flat and high topography called “banks” of the savannah, there is a diversity of adapted grasses such as black “bank” grass (Paspalum plicatulum), white “bank” grass (Panicum versicolor), guaratara (Axonopus purpussi), and native grass (Paspalum sp), among others. The low areas, known as “low”, are flooded during the rainy period and the adapted lambedora grasses (Leersia hexandra), water straw (Hymenachne amplexicaulis), and black water straw (Acroceras zizanioides), among others are present [1,2,3]. Livestock farming developed in flooded savannahs preserves the animal and plant germplasm without any anthropogenic alteration of the ecosystems and has great forage potential for feeding cattle [1].

The productivity of native grasses in the flooded savannah ecosystem occurs in natural conditions without any type of management or fertilization. The information on the productive and nutritional composition response to fertilizer application is limited [4]. According to research carried out in native savannahs, the epigeal production of native herbaceous vegetation is limited in particular by the chemical composition of the soil [5]. One of the management strategies in pastures is the use of fertilizers to increase their productivity [6], applying the appropriate amounts of nutrients according to the crop requirement and availability in the soil [7,8]. The fertilization of pastures has been extensively studied in the past and its positive effects on productive yields have been scientifically demonstrated [9]; however, studies on native grasses are scarce.

It has been reported globally that 30% to 40% of pasture yield is due to fertilizer application [10]. It is suggested that fertilization doses used should optimize the production of biomass with high nutritional value and low cost [11]. The interaction of the genetic potential of the crop, environmental factors, and management are determining factors in pasture productivity [12,13]. It has been proven that nitrogen is the nutrient that generates the greatest response in pastures in production and quality improvement, especially in terms of protein [14]. The production of dry matter and the persistence of pastures are directly related to natural fertility and nutrient supplementation to the soil. Therefore, the periodic application of fertilizers is important in order to supply the demands of the forage species and the pasture and herd management system [15]. Fertilization is an alternative to achieve an increase in forage yield in terms of quantity and quality [16], and pasture fertilization decisions must include conservation and production goals [17]. The selection of fertilizers for pastures must be made considering criteria of efficiency and economy, considering pastures as a perennial crop, which requires permanent availability of nutrients [14].

Therefore, it is essential to carry out research that allows for an evaluation of the effect of fertilization on the productive performance and nutritional components of native grasses of flooded savannahs, according to the cutting intervals. Thus, it is possible to define the optimal grazing time, in terms of production and nutritional value [18]. Likewise, other properties associated with their digestibility can be specified that promote adequate use of these forage resources in animal production systems with grazing ruminants. This information is useful to define a forage management plan according to its availability in the different physiographic positions (“bank” and “low”) and synchronize its use according to the nutritional requirements of the animals [19]. The information is also useful for making management decisions such as determining paddock rotation, adjusting animal load, and defining a nutritional supplementation program [20].

Native and adapted grasses are drought resistant and survive in low-fertility soils [21]. It is crucial to analyze, in detail, the yields and nutritional components presented by each species and offer precise recommendations on the most suitable native forage for producers in flooded savannahs. In recent decades, it has become increasingly noticeable that native grasses are being replaced by introduced grasses with the belief that these are more productive and have a higher nutritional quality [22]. It is essential to increase the scarce information on the performance and nutritional components of native pastures and consequently facilitate the integration of these species into production systems [23,24]. Native grasses are suggested as an adaptation strategy to climate change [25,26] and are essential for the success of cattle production [27]. For example, some native grasses such as lambedora grass (L. hexandra) and water straw (H. amplexicaulis) are perennials and highly consumed by cattle in rainy and dry periods; however, information on their productivity is scarce [28].

The current study presents the hypothesis that native grasses can respond to low doses of fertilization, given the adaptive capacity regarding the environmental conditions of the flooded savannah, which could be a strategy to improve their production and nutritional composition. Therefore, the aim of the study was to evaluate the response to low doses of fertilization in terms of forage yield and nutritional composition of native grasses (Paspalum plicatulum, Panicum versicolor, Paspalum sp., Leersia hexandra, and Hymenachne amplexicaulis) from the flooded savannahs ecosystem of Arauca, Colombian Orinoquia.

2. Materials and Methods

2.1. Study Area

The study was carried out in Arauca, a department located in the flooded savannahs ecosystem of de Colombian Orinoquia (Eastern Colombia) (Figure 1). The region is characterized by a flat topography with the presence of the physiographic positions “bank” and “low” (Latitude: 7°08′17″ N; Longitude; 70°59′59″ W; and Altitude: 125 m). In the region, the soils are acidic and have a clay loam and sandy loam texture with low Ca, P, K, Cu, and CEC levels and high Fe, Mn, Zin, and B [2] levels, and are classified as Ultisols and Oxisols [29]. The region corresponds to a subhumid tropical forest zone [30].

2.2. Study Native Grasses

Five native grasses adapted or used in the “bank” (n = 3) and “low” (n = 2) physiographic positions of the Arauca flooded savannah ecosystem were evaluated (

Table 1. Experimental grasses included in the study according to physiographic position “bank and “low” in flooded savannah ecosystem of the Arauca (Colombian Orinoquia). The scientific names and characteristics of the grasses were obtained from Plants of the World Online (POWO).

Species (Scientific Name)	Common Name	Habit	Characteristic
“Low” position
Leersia hexandra Sw. (1788)	Lambedora grass	Bunch	Native
Hymenachne amplexicaulis (Rudge) Nees (1829)	Water straw	Stoloniferous	Native
“Bank” position
Paspalum plicatulum Michx. (1803)	Black “Bank” grass	Bunch	Native
Panicum versicolor (EPBincknell) Nieuwl.	White “Bank” grass	Bunch	Native
Paspalum sp. (L.) L. (1762)	Native grama	Rhizomatous	Native

) (Figure 2).

The grasses were selected as part of a project that aims to identify forage alternatives for livestock activity in the region. According to the opinion of the livestock farmers associated with the region’s Livestock Committee, these grasses are the most consumed by cattle, horses, sheep, and goats that live in the region. Native grasses grow in pastures and are resistant to periods of rainy and dry weather [1,2,4,31].

The agroclimatic variables were recorded by a portable weather station (Vantage Pro2^TM, Davis, CA, USA) located near the experimental site. The region has two climatic periods: a rainy period (May to November) and a dry period (December to April). The dry–rainy transition period corresponds to the months of April and May with a rainfall of 182 mm, a relative humidity of 91.4%, and an average ambient temperature of 25.4 °C. The rainy–dry transition period corresponds to the months of November to December with a rainfall of 176 mm, a relative humidity of 89.5%, and an ambient temperature of 26.7 °C.

2.3. Experimental Procedure

Evaluations were carried out during the rainy–dry transition period (November–December 2023) in plots established in each physiographic position (“bank and “low”). The conditions under which the plots were established can be found in Salamanca-Carreño et al. [31]. The main plot had an area of 9 m² (3 m × 3 m) and constituted the pasture, and the subplot had an area of 4.5 m² (3 m × 1.5 m) and constituted the level of fertilization (with and without fertilization). The treatments were distributed in a randomized complete block experimental design.

Before fertilizer application, in each plot, soil samples were collected following standard procedures [32]. The chemical and physical analysis of the soil was carried out by the Soil, Water, and Foliar Laboratory of the National University of Colombia, Orinoquia headquarters [33] following the procedures established in the Colombian Technical Standards (CTS) of the Colombian Institute of Technical Standards and Certification [34]. The analysis showed a soil with a clay–loam texture with acidic pH (5.22) and content of phosphorus (P = 12.24 mg/kg), sodium (Na = < 0.13 meq/100 g), potassium (K = 0.10 meq/100 g), calcium (Ca²⁺ = 3.43 meq/100 g), magnesium (Mg²⁺ = 1.05 meq/100 g), organic carbon (7.28 g/kg), and total nitrogen (0.68 g/kg), which characterize it as a soil of medium fertility [2,14,29,32].

To begin the experiment, a leveling cut of the plots was made approximately 10 cm from the ground level using a sickle, and fertilizer was applied to each subplot. The fertilizer dosage was based on the methodology for the evaluation of forage materials recommended by Rincon Castillo et al. [35]. In each subplot (4.5 m²), a mixture composed of 180 g (400 kg/ha) of dolomitic lime (30% Ca and 10% Mg) + 180 g (400 kg/ha) of Paz de Río fertilizer (4% P, 34% Ca, 1% Mg). After 15 days of the leveling cut, a mixture of 45 g (100 kg/ha) of Urea (46% N) + 23 g (50 kg/ha) of DAP (18% N, 19.78% P) + 23 g (50 kg/ha) of KCl (50% K) was applied, values considered low for grassland fertilization.

2.4. Data Collection

To carry out forage measurements, three measurement points were randomly selected within each plot, and three cutting intervals (28, 35, and 42 days) were randomly assigned to each previously defined measurement point. Using a sickle, the available forage was cut within a 1 m² frame and the biomass production per m² was weighed using a Ranger precision balance. Then, the production of green forage (GF, t/ha) was estimated. The green forage samples were dried in a Caloric brand electric oven at 60 °C for 72 h, the dry matter (DM) value was recorded, and the dry matter yields (t/ha) were estimated.

At 35 days, nutritional composition variables were also analyzed using the procedure defined in the methodology for the selection of forage materials as recommended by Rincon Castillo et al. [35]. The dried samples (35 days) were stored in kraft paper bags and transported to the Analytical Chemistry Laboratory of the Colombian Agricultural Research Corporation (AGROSAVIA) for analysis of the nutritional composition. The nutritional composition of dried samples was analyzed using near-infrared reflectance spectroscopy (NIRS) [36]. The samples were homogenized to ensure a similar particle size; subsequently, they were placed in a 50 mm-diameter ring cup and spectra were obtained using the FOSS NIR Systems DS6500 model equipment (Foss, Hilleroed, Denmark) by scanning in the range of 400–2498 nm. The reference and the new spectra data were handled with WinISI 4.7.0.0 (Foss, Hilleroed, Denmark). The calibration equations used were constructed with spectra from the 2020 forage resources of three families (Grass forage, n = 1418; legume forage, n = 320; and other forage plants, n = 282) sampled between 2014 and 2016 from different livestock regions of Colombia [36]. The estimated variables included grasses’ nutritional composition and minerals: crude protein (CP), ash, ethereal extract (EE), neutral detergent fiber (NDF), acid detergent fiber (ADF), lignin, hemicellulose (HC), total digestible nutrients (TDN), dry matter digestibility (DMD), and minerals Ca²⁺, P, Mg²⁺, K, Na, and S.

2.5. Statistical Analysis

Data from “bank” and “low” grasses were analyzed independently. Analyses of variance were performed to identify differences between grasses. In the case of the forage yield variables (GF and DM/ton/ha), the following linear model was used:

$${\mathrm{Y}}_{\mathrm{i}\mathrm{j}\mathrm{k}}=\mathrm{\mu }+{\mathrm{E}}_{\mathrm{i}}+{\mathrm{F}}_{\mathrm{j}}+{\mathrm{C}\mathrm{D}}_{\mathrm{k}}+{(\mathrm{E}\times \mathrm{F})}_{\mathrm{i}\mathrm{j}}+{(\mathrm{E}\times \mathrm{C}\mathrm{D})}_{\mathrm{i}\mathrm{k}}+{(\mathrm{E}\times \mathrm{F}\times \mathrm{C}\mathrm{D})}_{\mathrm{i}\mathrm{j}\mathrm{k}}+{\mathsf{\epsilon }}_{\mathrm{i}\mathrm{j}\mathrm{k}}$$

where Y_ijk is the measured response, µ is the general mean, E_i is the effect of the i-th grass (“banks”: P. plicatulum, P. versicolor, and Paspalum sp.; “lows”: L. hexandra and H. amplexicaulis), F_j is the effect of the j-th fertilization (yes or no), CD_k is the effect of the k-th cutting day (28, 35, and 42 days), (E × F)_ij is the interaction of the grass by fertilization, (E × CD)_ik is the interaction of grass by cutting days, (E × F × CD)_ijk is the interaction of grass by fertilization by cutting interval, and ε_ijk is the experimental error.

In the case of the nutritional composition variables, which were measured at 35 days, the linear model used was:

$${\mathrm{Y}}_{\mathrm{i}\mathrm{j}\mathrm{k}}=\mathrm{\mu }+{\mathrm{E}}_{\mathrm{i}}+{\mathrm{F}}_{\mathrm{j}}+{(\mathrm{E}\times \mathrm{F})}_{\mathrm{i}\mathrm{j}}+{\mathsf{\epsilon }}_{\mathrm{i}\mathrm{j}}$$

where Y_ij is the measured response, µ is the general mean, E_i is the effect of the i-th grass (“banks”: P. plicatulum, P. versicolor, and Paspalum sp.; “lows”: L. hexandra and H. amplexicaulis), F_j is the effect of the j-th fertilization (yes or no), (E × F)_ij is the interaction of grass by fertilization, and ε_ij is the experimental error.

The minimum significant difference test (p < 0.05) was used to differentiate means. The analyses were carried out with the Infostat software version 2022 [37].

3. Results

3.1. Effect of Fertilization on Forage Yield of “Bank” Grass

Figure 3 shows the relationship between “bank” grasses and the cutting interval. It was found that GF (t/ha) and DM (t/ha) were significantly influenced (p < 0.05) by the effect of fertilization and by the grass × cutting interval interaction. Fertilization positively influenced GF (3.71 vs. 2.73 t/ha) and DM (1.21 vs. 0.89 t/ha) yield. In terms of GF and DM, the grass P. plicatulum was superior to the other grasses (p < 0.05). For P. plicatulum, the average production of GF from 28 to 42 days increased from 3.08 to 6.28 t/ha, while the other grasses presented values between 1.68 and 2.4 ton/ha (Figure 3a). In terms of DM, the same trend was observed, with an average range between 0.79 and 2.08 t/ha vs. 0.57 and 0.91 t/ha for the grass P. plicatulum (Figure 3b).

3.2. Effect of Fertilization on Forage Yield of “Low” Grass

Figure 4 shows the performance in terms of GF and DM between “low” grasses and the cutting interval. Significant effects were found due to fertilization and cutting interval (p < 0.05).

The average production of GF (2.33 vs. 1.69 t/ha) and DM 0.82 vs. 0.69 t/ha) was higher in the fertilized grasses, with a cutting interval of 35 days (2.50 and 1.02 t/ha, respectively). The GF was similar between H. amplexicaulis and L. hexandra, while the DM was higher in L. hexandra. The effect of fertilization showed an increase in GF until 35 days, remaining constant until 42 days (Figure 4a), while in DM, high production was observed up to 35 days with a decreasing trend until 42 days (Figure 4b).

3.3. Fertilization Effect on the Nutritional Composition of “Low” Grass

When analyzing the effect of fertilization on the nutritional composition of native grasses managed with and without fertilization in the “low” physiographic position of the flooded savannahs, the analysis of variance showed that the level of fertilization significantly affected the contents of DM, CP, EE, ADF, lignin, HC, TDN, DMD, Ca, P, Mg, K, and Na (

Table 2. Average and standard deviation of native grasses nutritional composition variables with and without fertilization of the “low” physiographic position obtained at 35 days of interval cutting.

Specie	F	DM (%)	CP (%)	Ash (%)	EE (%)	NDF (%)	ADF (%)	Lignin (%)	HC (%)
H. amplexicaulis	WF	35.82 ± 1.48	14.14 ± 0.92	14.91 ± 0.54	1.83 ± 0.07	60.73 ± 1.01	34.74 ± 0.77	8.33 ± 0.26	25.98 ± 0.69
	WTF	45.92 ± 1.48	8.93 ± 0.92	15.35 ± 0.54	1.52 ± 0.07	65.67 ± 1.01	37.08 ± 0.77	9.53 ± 0.26	28.59 ± 0.69
L. hexandra	WF	38.99 ± 1.48	13.71 ± 0.92	16.63 ± 0.54	1.68 ± 0.07	63.32 ± 1.01	36.5 ± 0.77	9.39 ± 0.26	26.81 ± 0.69
	WTF	45.71 ± 1.48	10.19 ± 0.92	16.65 ± 0.54	1.6 ± 0.07	65.14 ± 1.01	37.3 ± 0.77	9.71 ± 0.26	27.83 ± 0.69
Specie (p-value)		NS	NS	0.0248	NS	NS	NS	0.0438	NS
Fertilization (p-value)		0.0005	0.0015	NS	0.0199	0.01	NS	0.019	0.0303
Interaction (p-value)		NS	NS	NS	NS	NS	NS	NS	NS
Specie	F	TDN (%)	DMD (%)	Ca2+ (%)	P (%)	Mg2+ (%)	K (%)	Na (%)	S (%)
H. amplexicaulis	WF	54.64 ± 0.84	59.86 ± 0.91	0.34 ± 0.02 a	0.25 ± 0.01	0.20 ± 0.01 a	1.99 ± 0.08	0.04 ± 0.0017	59.86 ± 0.91
	WTF	50.11 ± 0.84	54.96 ± 0.91	0.25 ± 0.02 b	0.19 ± 0.01	0.15 ± 0.01 c	1.62 ± 0.08	0.03 ± 0.0017	54.96 ± 0.91
L. hexandra	WF	53.81 ± 0.84	58.95 ± 0.91	0.27 ± 0.02 b	0.27 ± 0.01	0.18 ± 0.01 b	1.88 ± 0.08	0.03 ± 0.0017	58.95 ± 0.91
	WTF	50.98 ± 0.84	55.9 ± 0.91	0.27 ± 0.02 b	0.23 ± 0.01	0.16 ± 0.01 c	1.66 ± 0.08	0.03 ± 0.0017	55.9 ± 0.91
Specie (p-value)		NS	NS	NS	NS	NS	NS	NS	NS
Fertilization (p-value)		0.0023	0.0023	0.0293	0.0039	0.0002	0.006	NS	0.0023
Interaction (p-value)		NS	NS	0.0211	NS	0.0158	NS	NS	NS

F: Fertilization, DM: dry matter, CP: crude protein, NDF: neutral detergent fiber, ADF: acid detergent fiber, HC: hemicellulose, EE: ether extract, TDN: total digestible nutrients, DMD: dry matter digestibility. WTF = Without fertilization; WF = with fertilization. Different letters in the same column differ statistically (p < 0.05). NS: not significant. Statistical differences are shown for the interaction (the same as species × fertilization; p < 0.05), which allowed us to see if the species respond differentially to the application of fertilization.

). The grasses evaluated presented significant differences (p < 0.05) for the contents of ash and lignin (Table 2). The species × fertilization interaction was not significant (p > 0,05) for the variables DM, CP, Ash, EE, NDF, ADF, lignin, HC, TDN, DMD, P, K, Na, and S; however, it was significantly different (p < 0.05) for the Ca²⁺ and Mg²⁺ content (Table 2).

In “low” grasses at 35 days of the cutting interval, differences were found in the contents of ash (16.63%) and lignin (9.39%) (p < 0.05), with higher values for the grass L. hexandra (Table 2). Fertilization significantly increased (p < 0.05) the levels of CP (13.93 vs. 9.56%), EE (1.75 vs. 1.56%), TDN (54.22 vs. 50.54%), DMD (59.41 vs. 55.43%), Ca²⁺ (0.31 vs. 0.26%), P (0.26 vs. 0.21%), Mg²⁺ (0.19 vs. 0.15%), K (1.94 vs. 1.64%), and S (59.41 vs. 55.43%) but reduced NDF (62.02 vs. 65.4%), lignin (8.86 vs. 9.62%), and hemicellulose (26.40 vs. 28.21%). The fertilization effect decreased DM levels (45.82 vs. 37.41%) (Table 2).

The grass × fertilization interaction was only significant (p < 0.05) in the Ca²⁺ and Mg²⁺ components. Fertilized H. amplexicaulis presented the highest value of Ca²⁺ (0.34%) and Mg²⁺ (0.20%) and L. hexandra presented the lowest value of (0.27% and 0.18%) (Table 2).

3.4. Fertilization Effect on the Nutritional Composition of “Bank” Grass

The nutritional composition of “bank” grasses at 35 days of the cutting interval showed differences (p < 0.05) for the CP, ash, EE, NDF, lignin, hemicellulose, TDN, DMD, Ca²⁺, P, Mg²⁺, Na, and S components. The grass P. versicolor showed the highest concentrations of ash (11.20%), EE (1.71%), TDN (50.63%), and DMD (55.52%). The grasses P. versicolor and Paspalum sp. presented the highest levels of CP (8.4 to 9.8%), P (0.22 to 0.23%), and Na (0.03 to 0.04%). The grasses P. versicolor and P. plicatulum presented the highest concentrations of Mg²⁺ (0.25%) and S (0.17 to 0.19%); while in Paspalum sp. and P. plicatulum, the highest levels of NDF (71.8 to 73.6%) and lignin (9.16) were observed. Finally, P. plicatulum presented the highest concentrations of hemicellulose (34.9%) and Ca²⁺ (0.54%) (

Table 3. Average and standard deviation of the variables of nutritional composition of native grasses with and without fertilization in the physiographic position “bank” obtained at 35 days of interval cutting.

Specie	F	DM (%)	CP (%)	Ash (%)	EE (%)	NDF (%)	ADF (%)	Lignin (%)	HC (%)
P. plicatulum	WF	29.62 ± 6	7.06 ± 0.64	9.5 ± 0.29	0.91 ± 0.08	73.18 ± 1.22	38.43 ± 0.77	9.13 ± 0.16	34.75 ± 0.63
	WTF	29.62 ± 6	6.47 ± 0.64	9.29 ± 0.29	0.87 ± 0.08	73.96 ± 1.22	38.89 ± 0.77	9.18 ± 0.16	35.08 ± 0.63
P. versicolor	WF	29.58 ± 6	9.99 ± 0.64	11.08 ± 0.29	1.78 ± 0.08	66.79 ± 1.22	37.8 ± 0.77	8.51 ± 0.16	28.99 ± 0.63
	WTF	29.6 ± 6	9.59 ± 0.64	11.3 ± 0.29	1.65 ± 0.08	66.93 ± 1.22	37.2 ± 0.77	8.64 ± 0.16	29.73 ± 0.63
Paspalum sp.	WF	31.95 ± 3	8.48 ± 0.91	9.68 ± 0.42	1.36 ± 0.11	71.77 ± 1.72	39.72 ± 1.09	9.11 ± 0.22	32.05 ± 0.9
	WTF	29.23 ± 3	8.28 ± 0.91	11.08 ± 0.42	1.38 ± 0.11	71.87 ± 1.72	39.81 ± 1.09	9.22 ± 0.22	32.06 ± 0.9
Specie (p-value)		NS	0.0004	<0.0001	<0.0001	<0.0001	NS	0.0018	0.0001
Fertilization (p-value)		NS	NS	NS	NS	NS	NS	NS	NS
Interaction (p-value)		NS	NS	NS	NS	NS	NS	NS	NS
Specie	F	TDN (%)	DMD (%)	Ca2+ (%)	P (%)	Mg2+ (%)	K (%)	Na (%)	S (%)
P. plicatulum	WF	48.33 ± 0.67	53.04 ± 0.73	0.51 ± 0.03	0.19 ± 0.01	0.25 ± 0.01	1.61 ± 0.07	0.03 ± 0.001 bc	0.18 ± 0.01
	WTF	47.77 ± 0.67	52.43 ± 0.73	0.57 ± 0.03	0.19 ± 0.01	0.25 ± 0.01	1.58 ± 0.07	0.03 ± 0.001 c	0.16 ± 0.01
P. versicolor	WF	50.68 ± 0.67	55.59 ± 0.73	0.43 ± 0.03	0.24 ± 0.01	0.25 ± 0.01	1.85 ± 0.07	0.04 ± 0.001 a	0.19 ± 0.01
	WTF	50.57 ± 0.67	55.46 ± 0.73	0.38 ± 0.03	0.22 ± 0.01	0.24 ± 0.01	1.62 ± 0.07	0.03 ± 0.001 c	0.19 ± 0.01
Paspalum sp.	WF	49.01 ± 0.95	53.77 ± 1.03	0.40 ± 0.04	0.22 ± 0.01	0.22 ± 0.01	1.68 ± 0.1	0.04 ± 0.002 ab	0.16 ± 0.02
	WTF	48.83 ± 0.95	53.58 ± 1.03	0.38 ± 0.04	0.22 ± 0.01	0.20 ± 0.01	1.68 ± 0.1	0.04 ± 0.002 ab	0.14 ± 0.02
Specie (p-value)		0.0029	0.0029	<0.0001	0.0011	0.0012	NS	0.0077	0.046
Fertilization (p-value)		NS	NS	NS	NS	NS	NS	0.022	NS
Interaction (p-value)		NS	NS	NS	NS	NS	NS	0.035	NS

F: Fertilization, DM: dry matter, CP: crude protein, NDF: neutral detergent fiber, ADF: acid detergent fiber, HC: hemicellulose, EE: ether extract, TDN: total digestible nutrients, DMD: dry matter digestibility. WTF = Without fertilization; WF = with fertilization. Different letters in the same column differ statistically (p < 0.05). NS: not significant. Statistical differences are shown for the interaction (the same as species × fertilization; p < 0.05), which allowed us to see if the species respond differentially to the application of fertilization.

).

The effect of fertilization and the grass × fertilization interaction on the nutritional composition of bank grasses only influenced the Na content (p < 0.05). The grass Paspalum sp. (with and without fertilization) and fertilized P. versicolor presented the highest Na concentrations (0.04%) (Table 3).

In general, although the production of green and dry forage showed differences in cutting intervals with the application of fertilizers (p < 0.05), the nutritional composition did not present differences (p > 0.05) with the application of fertilizers, although some tendency was evident in favor of fertilized grasses.

4. Discussion

4.1. Production of Green Forage and Dry Matter in “Bank” Grasses

In this study, the production of green forage (GF) and dry matter (DM) of grasses established on “bank” and “low” physiographic positions in flooded savannahs responded significantly to fertilization (p < 0.05) in the evaluated periods (28, 35, and 42 days). These results coincide with what was found in previous research that has shown that fertilization has a significant impact on the production of GF and DM in grasses in flood savannahs [38,39]. The production of GF and DM of P. plicatulum and P. versicolor is highlighted, while Paspalum sp. showed the lowest performance for these variables. This may be due to their lower ability to absorb and utilize nutrients, as well as a lower tolerance to flooding stress [40,41].

In general, P. plicatulum showed a tendency to increase the yield of GF and DM up to 42 days, while the other grasses showed a tendency to decrease after 35 days. Other studies have found similar results; for example, in Colombia, it was found that the application of NPK fertilizer to Urochloa brizantha pastures in flood savannahs significantly increased the production of green and dry forage [42]. Similarly, in a study in Brazil, it was observed that fertilization with nitrogen and phosphorus increased the biomass production of Urochloa mutica in várzea areas [43]. Differences in the response of grasses to fertilization may be related to edaphic and physiological factors. For example, P. plicatulum and P. versicolor may have a greater capacity to absorb nutrients from the soil than Paspalum sp, possibly due to their root growth. Additionally, these grasses may have greater efficiency in the use of nutrients for biomass production. On the other hand, the lower performance of Paspalum sp. in terms of GF and DM could be associated with their lower tolerance to flooding stress and their lower efficiency in nutrient utilization [40,44]. Regarding the tendency of P. plicatulum to increase the yield of GF and DM up to 42 days, the other grasses showed a tendency to decrease after 35 days. This could be explained by the growth cycle of each species. P. plicatulum is a species that grows more slowly than the others, so its peak production is reached after a longer duration [45,46].

A benefit of fertilization is achieved when forage species specific to each productive system are used [47], which was observed in the present work, where grasses that had adapted to the conditions of the flooded savannah region showed their productive performance with the fertilizer application. Therefore, native grasses are important in the productive systems of the flooded savannah ecosystem since they guarantee sustainable production with minimal application of fertilizers. The positive response of P. plicatulum and P. versicolor to fertilization can be attributed to their greater capacity to absorb nutrients, especially nitrogen (N), phosphorus (P), and potassium (K), results that have been found in the “bank” grass Urochloa mutica [48,49]. This suggests that native grasses possess genetic adaptations that enable them to efficiently utilize available resources, thereby minimizing the need for excessive fertilizer inputs. By strategically employing native grasses and tailoring fertilization practices to their specific requirements, producers can establish sustainable forage systems that are both productive and environmentally responsible. This approach not only reduces the reliance on synthetic inputs but also promotes biodiversity and ecosystem resilience in the flooded savannah region. The application of adequate amounts of nutrients is a key aspect in increasing forage production [7], and the response of a crop varies under different agroclimatic conditions [47]. Therefore, pasture fertilization decisions must include both production and conservation goals [17].

It is important to mention that the type of fertilizer used is a factor of great influence on forage yield, and the effectiveness depends on correct application [6,50]. Some authors have mentioned that the low response to fertilization may be due to a lack of adaptation to the climatic conditions of the experimental site [51]. In the present study, all grasses responded positively to fertilization in productive terms, indicating their adaptation to the conditions of the flooded savanna. Likewise, it is stated that the response of grasses to fertilization can be optimized if they are planted in a mixture with legumes [51].

4.2. Production of Green Forage and Dry Matter in “Low” Grasses

The grasses of the “low” physiographic position (H. amplexicaulis and L. hexandra) also responded significantly to fertilization (p < 0.05), in this case, with greater production at the 35-day cutting interval. These results are possibly reflected by the fact that these grasses depend exclusively on the amount of surface water available, and the cuts were made in the rainy–dry transition period, which is characterized by decreased rainfall. The increase in DM with plant age can be explained by the fact that the plant increases the photosynthetic process and, with it, the synthesis of structural carbohydrates, generating a greater accumulation of dry matter [52]. The positive response to fertilization in these grasses, with greater production after a 35-day cutting interval, coincides with what was observed in previous studies. In a trial with Setaria adendrotricha, the application of nitrogen and phosphorus significantly increased dry biomass production [53].

In general, the results shown in the present study are consistent with those reported in other grasses in relation to the response to fertilization in GF and DM. In Costa Rica, using star grass (Cynodon nlemfuensis Vanderyst.), researchers recorded that GF and DM were similar in cuts of 35 to 45 days, and production doubled after 55 days of cutting [54]. In a study with Brachiaria brizantha, the cutting age and time of year were evaluated, showing higher DM/ton/ha yields at 56 days [55]. In the African Star grass, they found an increase in biomass as the ages of regrowth increased, mainly for ages greater than 35 days [56]. Work carried out in Mexico on Megathyrsus maximus showed that DM production was favored by fertilization and recommended that applying high doses of N favors greater productive performance of the grass [57,58]. In Bolivia, they evaluated the effect of cutting and different fertilization levels on DM yield in the native grass Nasella sp, finding differences (p < 0.05) in the different fertilization levels [59].

4.3. Nutritional Compositions of “Bank” and “Low” Grasses

Although the nutritional composition results are very similar, they tend to be superior in treatments with fertilization. The few differences are possibly due to the low levels of fertilizer applied and the low response of native grasses to fertilization. The low response may be due to the fact that the plant obtains a good part of the minerals from the unavailable fraction of the soil [8]. The grass × fertilization interaction influenced the content of Mg²⁺ and Ca²⁺ (p < 0.05) for “low” grasses and the content of Na (p < 0.05) for “bank” grasses. The highest values were presented in H. amplexicaulis. On the other hand, in bank grasses, the interaction only influenced Na concentrations (p < 0.05) with greater increases in Paspalum sp. and P. versicolor grasses.

Studies carried out under humid tropical conditions in the grasses Paspalum notatum, Brachiaria humidicola, Brachiaria brizantha, and Brachiaria hybrid reported concentrations of Ca²⁺ (0.35%), Mg²⁺ (0.27%), and Na (0.12%) [60], with results similar to those of the current study. In another study, also in a tropical environment, they evaluated the effect of fertilization on nutritional quality in C. nlemfuensis grasses, finding a significant response regarding the concentrations of Mg²⁺ (0.11%) and Ca²⁺ (0.38%) [61]. Variations in mineral content in forages may correspond to mineral concentrations in the soil [62]. Concentrations can also vary between climatic periods of the year; for example, lower concentrations of Ca²⁺, Mg²⁺, and Na have been reported in the rainy season [63].

In evaluations carried out on the nutritional composition of L. hexandra and H. amplexicaulis, protein levels between 6.6% and 17.2% and ash levels between 12.3% and 16.1% were observed, which are considered acceptable values that are difficult to match by some forage species introduced under these natural conditions [1]. These results, although lower than those of our study, demonstrate the importance of native grasses in flooded savannah conditions.

Although the concentrations of CP, ADF, and NDF improve with the application of fertilizers [57,64,65], in our results, the grass × fertilization interaction did not show differences (p > 0.05) in grasses of the two physiographic positions. It is worth mentioning that the nutritional contents of pastures differ depending on the origin and the forage species, so constant evaluations of the nutritional quality of the different forage species under different production scenarios are necessary [24]. On the other hand, the use of fertilizer depends on the type of soil, the source, and the dose applied [61].

Native grasses, tolerant to drought and heat, are suggested as an adaptation strategy to climate change scenarios [25,26], and due to their forage potential, they should be integrated into production systems [21,66] as the main source of food for ruminants [27,67]. Native grasses develop under natural conditions depending only on rainfall and under extreme temperature variations [59]. Likewise, modern agriculture depends heavily on fertilizers, which is why it is considered an inevitable threat to agriculture. Overuse hardens the soil, reduces its fertility, pollutes the air, water, and soil, and depletes soil nutrients and minerals, which creates risks for the environment [68], hence the importance of using native grasses in the ecosystem where they grow, which is, in our case, the flooded savannah.

5. Conclusions

The application of fertilizers generated significant differences in forage yield but not in the general nutritional component of the grasses, although some numerical variations were observed in favor of the fertilized grasses. The variations in forage yield and nutritional composition (with and without fertilization) may be due to the fact that native grasses of flooded savannah have lower nutrient use efficiency and lower tolerance to flooding stress. In the native grasses in the “bank” and “low” physiographic positions, a positive response to fertilization was observed in terms of the production of green and dry forage according to the cutting interval, but they did not respond to fertilization regarding their nutritional composition. In low grasses, the interaction with fertilization had a significant effect on the concentrations of calcium (Ca²⁺) and magnesium (Mg²⁺); while in low grasses, the interaction influenced only sodium (Na) concentrations.

According to the results of this study, native grasses constitute an important sustainable food resource for livestock in flooded savannahs, and the application of fertilizers will not be required to increase the value of the nutritional composition. While the results of this study suggest that native grasses in flooded savannahs can be a sustainable food resource without excessive fertilizer inputs, further research is needed to fully understand the long-term implications of fertilization and the ecosystem’s overall self-sustaining capacity. This study is considered the first approach to understanding the behavior of native grasses, a practice that can lead to the sustainable management of native grasses under different ecosystems.