Introduction

Fire and grazing are considered as major disturbances in many pastures from the historical point of view and in the modern management of our era. Humans have used fire to manipulate and manage ecosystems (^{Bowman et al., 2009}), while grazing contributes to the functioning and conservation of grasslands (^{Toombs et al., 2010}; ^{Allred et al., 2011}). In fact, fire and grazing frequently operate as an interactive disturbance with different ecological effects of each other (^{Fuhlendorf et al., 2009}), which creates heterogeneity in the structure of vegetation; important in the conservation of grasslands (^{Leis et al., 2013}; ^{McGranahan et al., 2013}). Grasslands globally are vital not only for livestock but also as suppliers of an important number of goods and ecosystem services for society: as carbon sequestration, water infiltration, oxygen production, biodiversity, among others (^{Dinerstein et al., 2007}).

Despite the importance of grasslands, one of the deterioration causes is the spread of exotic or invasive species that have displaced native species disrupting biodiversity and ecological balance (^{Enriquez and Quero, 2006}). One of these species is natal grass Melinis repens (Willd.) Zizka, which is originally from South Africa (^{Bogdan, 1977}). In Mexico, it is distributed in all states regardless of latitude, altitude, rainfall and soil (^{Davila et al., 2006}; ^{Stevens and Fehmi, 2009}; ^{Díaz et al., 2012}; ^{Flores, 2013}; ^{Melgoza et al., 2014}).

In the state of Aguascalientes it is reported to be present in virtually all municipalities (^{Díaz et al., 2012}). However, existing scientific information explaining the effect of fire both in production and quality of natal grass forage is very limited. This information is essential to determine management strategies that guide to achieve full utilization of this species. Therefore, the objective of this study was to determine the effect of fire on the production and quality of natal grass forage in the municipality of Calvillo, Aguascalientes, and in which our null hypothesis of interest was that there would be no effect of fire on productivity and quality of natal grass.

Materials and methods

The study was conducted in during 2012 and 2013 in the town called Mesa Grande, Calvillo, Aguascalientes, in an abandoned agricultural land with an area of 2 ha, and totally invaded of natal grass. Its geographical position is located between 21° 47' 31" and 102° 43' 9" north latitude, at an average altitude of 1 780 masl. The climate of the region is temperate sub-humid (^{García, 1988}; ^{Medina et al., 1998}) with an average annual temperature of 20 °C and shallow, rocky soils and poor in organic matter Regosol and Feozem type (^{INIFAP, 1998}). Rainfall is mainly present from June to September, which in 2012 and 2013 had an annual average of 640 and 768 mm respectively (Figure 1).

Figure 1. Rainfall in Mesa Grande, Calvillo, Aguascalientes, during the years 2012 and 2013. 

Within the experimental area 10 plots of 16 m² (4 x 4 m) were distributed randomly, of which five were subjected to burning (Q) and the other five were considered as control since these were not burned (SQ). To the latter (SQ), were removed the accumulated dry forage from years at ground level with the help of scissors Corona brand. Burning was carried out on April 1^st 2012 using a drip torch Sure Seal^® with 10 liters capacity, with a petrol ratio: diesel 6: 4. It was set at 7:00 h at an atmospheric temperature of 16 °C, a relative humidity of 35%, wind speed which ranged between 3 and 4 km h^-1 and an average of 479.4 g m^-2 of dry matter (combustible). To minimize the risk of fire escape to neighboring areas, mineral lines were traced around the plots and its perimeter was moistened with water using a backpack sprayer with 20 L capacity. Forage production on a dry basis was evaluated at peak of production which was at the end of October 2012 and 2013, through 5 quadrants of 0.5 m² plot (n= 50) by removing in each of them the forage at ground level, same that was dehydrated in a dryer herbarium at a temperature of 60 °C for 10 days and weighed with a digital scale OHAUS^® with 20 kg capacity.

Forage quality was evaluated by a total of 10 composite samples (n= 10), five burned and five in the control area, with a weight of 1kg/plot. Forage samplings were carried out in three phenological stages: growth (C), maturity (M) and latency (L), for a total of 30 samples (n= 30) in burned (Q) and without burning (SQ) areas.

For C stage forage was harvested in early July; m at the end of September 2012; while L in mid-February 2013. The samples were processed at the Laboratory of Animal Nutrition Center of Agricultural Sciences at the Autonomous University of Aguascalientes (UAA), with a mill Wiley^® and 1 mm mesh, and stored in plastic containers until chemical analyzes of crude protein (PC), neutral detergent fiber (FDN) and acid detergent fiber (FDA) were performed, based on the procedures to estimate nutritional content in forages (^{Tejada, 1985}). In situ ruminal digestibility of dry matter (DISMS) was performed according to the nylon bag method (^{Mehrez and Ørskov, 1977}). Using a fistulated cow rumen Holstein breed with maintenance diet and residence time of the food sample in rumen for 48 hours.

Data was subjected to analysis of variance with the lineal general model (GLM) considering a completely randomized design in factorial arrangement 2 x 2 for the production of dry matter: 2 fire treatments (Q and SQ) and 2 years (2012 and 2013). For nutritional quality was used another factorial design 2 x 3: 2 fire treatments (Q and SQ) and 3 phenological stages (C, M and L) (^{SAS, 2001}). When there were differences between treatments Tukey test was used (p≤ 0.05) for mean separation (^{Steel and Torrie, 1980}).

Results and discussion

The results of this experiment of burning rejects the null hypothesis that burning would not have effect on either productivity or quality on natal grass forage. Total production of forage on dry basis was different (p≤ 0.05) between treatments Q and SQ, and between years. There was higher production on burned areas with 168.7 and 393.7 g m^-2, while in without burning production was 53.9 and 192.7 g m^-2, in 2012 and 2013 respectively (Figure 2).

Figure 2. Forage production on dry matter basis of natal grass Melinis repens (Willd.) Zizka, in Mesa Grande, Calvillo, Aguascalientes during 2012 and 2013. 

These results agree with those reported by several authors concluding that the differences in forage production on dry basis are due to chemical effects of the combustion of plant material on the ground, provides various nutrients, primarily nitrogen, which is left available favoring vegetative reproduction, coupled with the presence of higher temperatures in burned areas during the day and vapor condensation during the night, creating near the dewpoint conditions and making available the nutrients released during combustion (^{Wright and Bailey, 1982}; ^{Frost and Robertson, 1987}; ^{Briske and Richards, 1994}).

Other authors mention that the increase in forage production is because after burning, nutrients are immediately released and available to plants from the ashes, mainly minerals (^{Sacido and Cauhépé, 1993}; ^{Bates et al., 2009}). ^{Riser and Parton (1982)}, ruled considerable increases in forage production in tallgrass prairie US after burning, through its effects on temporal pattern of nitrogen and carbon in the soil and the availability of light for photosynthesis. ^{Johnson and Mitchett (2001)} concluded that herbaceous plants that serve as food for wildlife were ten times more abundant in areas where prescribed burning was applied than in unburned areas.

The increase in plant nutrients have been reported in the first growing season after spring burnings in temperate environments (^{Bennett et al., 2002}). In general, the use of fire in grassland ecosystems favors various benefits, including increased forage production of native grasses (^{McIlvain and Armstrong, 1968}; ^{Sharrow and Wright, 1977}; ^{Scifres and Hamilton, 1993}). Fire decreased forage production in the first year of implementation in grasslands from Chihuahua, but increased in subsequent years with good grazing management (^{Sierra et al., 2008}).

Crude protein showed no significant differences (p≥ 0.05) between Q and SQ treatments, but was different (p≤ 0.05) between phenological stages (C, M, and L) (Figure 3). Crude protein values for burned areas were 13.3, 5.2 and 5.6% against 12.6, 4.6, and 3.2% for unburned areas in C, M and L stages respectively.

Figure 3. Average ratio of crude protein (CP) in forage Melinis repens (Willd.) Zizka, burned and unburned in Mesa Grande, Calvillo, Aguascalientes, 2012-2013. 

These results are consistent with those from ^{Leland et al. (1976)}, who found that crude protein of Andropogon scoparius (Michx.) a C4 grass like Melinis repens was higher in burned areas than unburned.

^{Sanderson and Wedin (1989)} stated that crude protein content in summer grass (C₄) was higher in summer and lower during latency. This coincides with ^{Chávez and González (2008)}, who found in native grasses a maximum content of crude protein in summer, which decreased in the fall, and even more in the winter, recovering in early spring in areas without burning. Meanwhile ^{Krysl et al. (1987)} found that the crude protein ranged between 15.4 and 9.6% for growth and latency stages respectively, in blue grama in New Mexico.

^{McGranhan et al. (2014)} noted significant increases of up to 4% in crude protein content in burned areas of Andropogon virginicus L., compared to unburned in Tennessee. Generally crude protein content in forages is necessary to meet the needs of rumen microorganisms to digest fiber which is around 7%, without affecting their voluntary intake, suggesting the need for protein supplementation when forage has values lower than 7%, thereby increasing voluntary forage intake (^{Pitts et al., 1992}).

In situ digestibility of dry matter (DISMS) at 48 hours was different (p≤ 0.05) between Q and SQ treatments and phenological stages C, M and L (Figure 4).

Figure 4. In situ average digestibility of dry matter (DISMS) at 48 h in Melinis repens (Willd.) Zizka, burned and unburned in Mesa Grande, Calvillo, Aguascalientes, 2012-2013. 

Overall DISMS decreased (p≤ 0.05) as natal grass maturity advances both in Q and SQ treatments, although this reduction is more marked in SQ treatment with a total of 61.6, 46; and 48.3% and 50.3, 42.5, 44% respectively in the three phenological stages C, M, L. ^{Krysl et al. (1987)} in blue grama Bouteloua gracilis (Willd. Ex Kunth) Lag. ex Griffiths., in New Mexico reached similar results by pointing out that DISMS declined as forage maturity advances 61.8% during growth to 47.9% in latency, this concomitant increase in fiber fractions in forage (fiber acid and neutral detergent) in unburned areas.

^{Smith et al. (1960)}, for Andropogon scoparius (Michx.) found DISMS values ranging between 40.3 to 65.7% in unburned and burned plots respectively. In humid savannas from Cameroon, ^{Klop et al. (2007)} concluded that grass DISMS of unburned and burned areas ranged from 50.2 to 57.1% respectively. ^{Núñez (1971)} worked in unburned areas of native grasslands in Chihuahua finding that the percentages of DISMS fluctuated between 31.6 and 40.2% at the beginning of the growth stage and 43.3 and 49.5% at the end of it. ^{Villanueva et al. (1989)} reports that this fluctuation pattern in DISMS and other nutritional components of forage is common in native grasses from arid and semiarid areas of Mexico contrasting areas subject to burning and those not subject to it.

Neutral detergent fiber (FDN) represents most of the material contained in the cell walls of the available forage in the diet and is composed of cellulose, hemicellulose, and lignin (^{Holechek et al., 2011}). Our results show generally a significant increase (p≤ 0.05) in FDN content as natal grass maturity advances, but not between Q and SQ treatments (p≥ 0.05) (Figure 5). Data related with FDN were different (p≤ 0.05) among the three phenological stages under study (C, M, L) with a total of 67.1, 76.5, 78%; and 66.4, 76.3, and 79.6% in treatments with Q and SQ respectively.

Figure 5. Average ratio of neutral detergent fiber (FDN) in forage Melinis repens (Willd.) Zizka, burned and unburned in Mesa Grande, Calvillo, Aguascalientes, 2012- 2013. 

This agrees with ^{Ulyatt (1980)} and ^{Van Soest (1982)}, who widely determined that the contents of FDN in forage increases as maturity advances. ^{Krysl et al. (1987)} concluded that FDN in grassland forage from New Mexico fluctuated slightly from 77.4% to 75.8% between growing seasons and latency respectively in areas without burning. In central grasslands from Chihuahua, ^{Chávez et al. (1984)}, found FDN values ranging between 72.1 and 76.5% in native grasses during growing season, while ^{Ortega et al. (1984)}, obtained FDN content in fodder shrubs consumed by goats in central Chihuahua ranging between 33.9 and 38.3% in unburned areas.

FDA fraction contains the most resistant substances to ruminal degradation contained in the cell wall (cellulose, lignin and silica) (^{Holechek et al., 2011}). Our results show generally a significant increase (p≤ 0.05) in FDA content as natal grass maturity advances, but not between treatments Q and SQ (p≥ 0.05) (Figure 6).

Figure 6. Average ratio of acid detergent fiber (FDA) in forage Melinis repens (Willd.) Zizka, burned and unburned in Mesa Grande, Calvillo, Aguascalientes, 2012- 2013. 

Data related to FDA were different between growth stages (C, M and L) with a total of 35.8, 50.6, 57.4%; and 33.2, 54.7; and 56.3% in Q and SQ respectively. FDA increased as natal grass maturity advances. In central grasslands of Chihuahua, ^{Chávez et al. (1984)}, found FDA values ranging between 45.6 and 50.7% in native grasses during growing season, while ^{Ortega et al. (1984)}, obtained FDA contents in fodder shrubs consumed by goats in central Chihuahua ranging between 28.9 and 31.6% in unburned areas.

Is possible to achieve increases in forage quality using prescribed burning by the effect of improving nutrient availability in some soil elements (^{Wright and Bailey, 1982}; ^{Knapp and Seastedt, 1986}; ^{Ojima et al., 1990}; ^{García, 1992}); this is due to mineralization and fixing of the same (^{Hobbs et al., 1991}), in the first five centimeters of soil (^{Haubensak et al., 2009}). ^{Short et al. (1974)} concluded that the FDN in native forages is generally always greater than the FDA.

Conclusion

Overall can be concluded that burning Natal grass forage had a positive effect on improving its production and quality, therefore fire as management strategy could be used to exploit this species in areas where its presence is constantly growing and considered a threat to be an invasive species that rapidly and practically establishes in any soil. This kind of work opens the option of using prescribed burning as management strategy of grassland for optimal and comprehensive management of rangelands with presence of Melinis repens and increase its productivity and quality, besides to achieve increases in voluntary consumption from animals used under extensive conditions this type of forage.