Introduction

Dairy farming is very important for meeting the demand for high quality food for human consumption; however, it also contributes to the emission of greenhouse gases (GHGs) such as methane (CH₄), which is produced by enteric fermentation and is eliminated mostly through the burp^1,2. CH₄ is one of the main GHGs emitted by the systems of production of bovine cattle, which has been associated with the global warming and climate change, as evidenced by comparison with the pre-industrial era^1,2. The issued CH₄ also contributes to the energy leaks in the livestock production systems, as it amounts to the loss of 6 to 12% of the total gross energy (GE) consumed by the dairy producing livestock².

An important reference for estimating the CH₄ production is the methodology proposed by the Intergovernmental Panel on Climate Change (IPCC)³. This methodology is used in different countries in order to generate the inventories of GHGs, and applies the corresponding procedures depending on the information available at levels 1, 2 and 3. Level 1 (Tier 1) is applied to Mexico and other developing countries; it is based on the use of an emission factor (EF) of CH₄ yr^-1 animal^-1 that is multiplied by the national inventory within each category of livestock. A criticism made to this methodology is that it employs an EF that in many situations does not represent the reality of the particular conditions of production, particularly with regard to the characteristics of the food consumed by the dairy cattle raised in different regions of the country, which varies according to the agro-ecological, environmental, economic and health characteristics^(4.5).

CH₄ emissions have also been estimated for dairy cattle in Mexico, according to the methodology of Level 2, which requires calculating the requirement of GE and using a conversion factor of CH₄ by default, which on average is 6.5 %, amounting to the percentage of the GE of the food turned into CH₄. Based on this method, an EF of 166 and 182 kg CH₄ yr^-1 has been estimated in primiparous and multiparous cows in milking conditions of herds registered at the Holstein Association of Mexico⁶, and 115 kg CH₄ yr^-1, for the total dairy cattle of Mexico⁷. The variability in the EF obtained in these studies is probably related to factors within each production system, such as the degree of technification, genetic capacity of animals for milk production and environmental influences. Due to the diversity of factors that may influence the EF, it is recommend having an EF for each type of system, and within these, an EF for each subtype, taking into account the particular characteristics of production and, in particular, the power schemes, including the type, concentration and quality of inputs used in the formulation of rations³.

In Mexico, there are various dairy production systems, the most predominant of which is intensive production, characterized by the use of modern, efficient technologies at all stages of the production process⁶. At the other extreme is the backyard or family-run production system, which contributes 10 % of the national milk production, and which is based mainly on traditional schemes of production and is characterized by a wide range of subsystems, and varying degrees of modernization and of productive efficiency^5,8,9. In general, the backyard dairy farming systems have been little studied, and EFs have not been generated for this category of dairy production; these are relevant for estimating the GHGs corrected for each type of system, for determining the degree of impact of this kind of farming on emissions and energy losses in the form of CH₄, and for designing and implementing strategies for the research, transfer of technologies and innovations to improve the sustainability of backyard dairy farming systems. Therefore, the aim of the present study was to estimate the EF and the PCH₄, using the predictor equations of the level 2 method of the IPCC and the technical information gathered on family-run dairy farming units.

Material and methods

The study was carried out at farms in the municipalities of El Marqués, Pedro Escobedo and San Juan del Río, in the state of Querétaro. All three municipalities are nestled in the Neovolcanic Axis, which is characterized by geomorphological contrasts between hills situated at an latitude between 2 and 3 thousand meters above sea level and valleys located at 800 to 900 m asl; temperate subhumid climates are prevalent in them, with a mean annual temperature of 17.3 ºC and a mean annual precipitation of 542.9 mm. Visits were made to 20 stables of the family system belonging to the Livestock Producer Groups for Technology Validation and Transfer of the State of Querétaro, with low level of modernization and in the process of adopting new technologies. The owners were interviewed in order to obtain technical information; for this purpose, at least three visits per producer were carried out in different months of the year during the period from April to August.

The body weight of the adult animals was estimated based on the thoracic perimeter, and that of the young animals was estimated subjectively by the producers themselves; the milk production was calculated using a plastic bucket with a scale in liters. The data analysis included only 10 stables, since those participating farms that did not provide complete information or whose data were not considered to be reliable were discarded.

The structure of the herd was divided, according to the terminology used by the producers interviewed, as follows: lactating or milk-producing cows; gestating cows with a 210-270 days’ pregnancy that were not being milked; heifers weighing 200-400 kg; calves weighing 100-200 kg; stud bulls, and young bulls in fattening. Table 1 shows the number of animals per category.

Samples of the food ingredients were obtained from all the stables, and the type and amount of the ingredients offered to the animals in the different phases of production were recorded. With this information, it was calculated the levels of inclusion of each ingredient in kilograms of fresh matter for each category of animals. Table 2 lists the food ingredients utilized and the number of samples taken. In the laboratory, the ingredients were dried in a forced air oven at a temperature of 55°C for a period of 24 to 72 h, depending on the moisture content, and ground with a Wiley type mill through a 2 mm sieve. The content of dry matter (DM) was determined in accordance with the AOAC official method 930.15¹⁰, while the GE content was determined by combustion of the samples in an adiabatic calorimetric bomb.

Results and discussion

Table 1 shows the number of animals by category and milk production of cows in milking condition. The total number of cows in milking condition and pregnant cows averaged 22.5, and the heifers amounted to 13.2 animals, amounting to 57.5 and 33.8 % of the total number of animals; the young bulls added up to merely 4 and 1.5 % of the total herd, respectively. The total number of animals was 39.1, with a minimum of 5 and a maximum of 55. Due to the fact that cows in milking condition were found in all the farms, but bulls in fattening were found only in four, since the rest of the producers sold to intermediaries; this was reflected in a lower and a higher CV in these two groups of animals. Only one farm had five cows in milking condition, while the rest of the units had 8 to 27 cows in milking condition and a total of 23 to 55 animals. The size of the herd in the present study was within the range reported in the family-run dairy farms in various states of the Republic^5,8,9. Eight farms had only Holstein animals, and Holstein and Brown Swiss American animals were found in two farms.

The average, minimum and maximum milk productions were 8.9, 3.4 and 14.4 kg d^-1, respectively. The duration of lactation was on average 253.1 d, with a minimum and a maximum of 180 and 420 d, respectively. The average, minimum, and maximum production of milk lactation^-1 kg^-1 was 2.130, 670 and 5.546 kg. Milk production in kg^-1 yr averaged 3.231, with minimum and maximum values of 1.253 and 5.263 kg. The milk production reported in other assessments of family systems was 10.7, 11 and 17.3 kg of milk cow^-1 d^-1(5,8,9, which is greater than that reported in this paper. At the same time, milk production in kg yr^-1 was lower than that reported for family systems of milk production with traditional management (3.417 kg of milk yr^-1) or improved through a program of validation and transfer of technology (4.632 kg of milk yr^-1)⁵.

The average content, minimum and maximum values, and CV of the GE are presented in Table 2. In general, the average values and CV found in this study agree with those of other reports and reflect the natural variation of the ingredients available for the animals^12,17. The minimum and maximum CV occurred in corn silage (4.9 %) and chicken dung (22.8 %). Various agro-ecological and management conditions reportedly influence the content of GE at different times of the year, depending on the weather, on whether the production is seasonal or utilizes irrigation, and on the doses of fertilization. Probably, the greatest extremes in terms of availability and quality of the ingredients occur between the dry and rainy season; in the present work it was tried to minimize this effect when carrying out the samplings between spring and summer. Other factors that influence the GE content are the variety of plants, the harvest time, the conditions of post-harvest handling and storage, and the type of processing^12,17,18.

During visits to the farms it was noted that various alfalfas occurred in adjoining plots, but the majority of the producers had no knowledge of the seed variety, the age of the crop, the number of cuts or the age of the forage at the time of the cutting; some alfalfas were offered fresh, sugar-coated or dried. Other producers who bought alfalfa in bales, usually dehydrated, were unaware of the information mentioned before and did not know for how long the bales had been stored before acquiring them. This lack of information was observed in most of the sampled ingredients.

Table 3 shows the DMC, GEC, EF and PCH₄ are shown by productive phase 3. In cows in milking condition, the DMC (11.28 and 10.42, kg d^-1), GEC (45.42 and 39.41, Mcal d^-1), EF (81 and 70, kg CH₄ yr^-1) and PCH₄ (2.95 and 2.56, Mcal d^-1) estimated with IPCC equations of the IPCC were similar to the estimates of the observations carried out at the farms. This indicates that, in cows in milking condition, the results of the IPCC equations constitute a good point of comparison or verification for the prediction of these variables with the use of procedures recommended for level 2. Concentrates were included in the food of the cows in milking condition in all stables, accounting for the largest portion of the ration, in an average proportion of 33 %; in order of frequency, they were followed by the alfalfa and corn silage. According to the majority of the interviewed producers interviewed, the family-run dairy production systems follow similar feeding schemes of cows in milking condition to that of cows in intensive systems within the study area. This is probably the reason why there were no differences in the variables evaluated at the farms and those estimated with the IPCC.

The EF values of the cows in milking condition were between 41.6 and 65.8 % lower than those found in milk-producing cows in other studies conducted in Mexico in which the procedures recommended for level 2 were also utilized. The EF in primiparous and multiparous cows of 16 herds registered at the Holstein Association of Mexico was 166 and 182 kg CH₄ yr^-1, in lactations of 305 d, respectively⁶; in order to calculate the EF, the DE content of the diet was estimated using a default value and the content of total digestible nutrients (TDN), and subsequently converted to GE using another default value, while an EF of 115 kg CH₄ yr^-1 was estimated for the total milk-producing herd of Mexico, with a lactation period of 305 d⁷. The EF was calculated based on an estimate of the GE contents of five diets reported in studies of dairy cattle in Mexico published between 1971-2009. The lower EF estimated in the present work can have two explanations: 1) differences in the methodologies used to estimate the GEC in relation to the previous work, which emphasize that the content of DM and GE in the ingredients or food consumed by animals included in the studies was not analyzed in a laboratory, and 2) the differences in the levels of milk production, which directly affect the level of consumption of nutrients, and the GEC in the case of the estimates based on the IPCC equations. The conversion factor of CH₄ (Ym) is a value that represents a fixed percentage of the GE (6.5%), which is converted to CH₄ (Equation 1) and, therefore, the higher the GEC, the greater the production of CH₄³. In the present work, the production of milk per lactation adjusted to 305 d was 2.715 kg; for the Holstein cows, it was 9.985 kg⁶ and for the milk-producing herd of Mexico, it was 3.795 kg⁷. Thus, the EF is greater in the cows with higher milk production, intermediate in the cows with intermediate production of milk, and lower in the cows in milking condition included in this study. This same trend is observed in several reports where the EF ranged between 102 and 128 kg CH₄ year^-1 in cows in milking condition with an annual milk production of 5.365 to 8.270 kg, respectively^(19.20).

In pregnant cows, the observed DMC was 19 % lower (P<0.01), but the observed GEC, EF and the PCH₄ were 32 % higher (P<0.01) than those estimated with the equations of the IPCC (Table 3). These results indicate that, in gestating cows, the use of the equations of the IPCC overestimated the DMC and underestimated the GEC, the EF and the PCH₄ and therefore they do not constitute a good point of comparison or verification for the prediction of these variables using the level 2 method. The difference in the DMC was 2.2 kg d^-1 below the expected value, and the GEC was 8.0 Mcal kg^-1 above the expected value. In gestating cows, the most frequently utilized forages were stubble and oats hay, in a higher concentration than those used for cows in milking condition. Probably, these ingredients caused a greater filling of the rumen, which, associated with the pressure of the fetus, reduced the DMC, but at the same time contributed to a greater GEC.

The estimated and observed EF was of 47 and 62 kg CH₄ yr^-1 in pregnant cows. There are few researches on the CH₄ emissions by these animals. A report found that the EF was 55.15 kg CH₄ yr^-1 in heifers aged over one year until delivery, with a body weight of 310-520 kg¹⁹. In the present study, the weight of the pregnant cows averaged 475 kg, which is within the range mentioned above; however, the requirement of GE for non-pregnant heifers and pregnant cows was calculated separately because, at the same weight, pregnant cows have a higher requirement of GE. This is due to the use of the ENp (Equation 5), and in cows of between 210-270 d of gestation, the ENc increases in parallel with the growth of the fetus and also depends on the expected weight of the calf at birth²¹.

The DMC was similar in the heifers and the calves, but the GEC, EF and the PCH₄ estimates were lower (P<0.01) with the equations of the IPCC than those observed in farms (Table 3). The observed GEC, EF and the PCH₄ were 56 % higher in the heifers and 63 % higher in calves with respect to the corresponding estimated values. In most of the farms, corn stalks were the ingredient utilized most frequently and in the greatest proportion (on average, 50 % of the ration) and provided the largest amount of the GE in growing females. For calves up to one year of age, with body weight of 43 to 320 kg, the reported EF was 34 to 35 kg CH₄ yr^-1, and in females aged one to two years, with weights of 310 to 530 kg, the EF was 49 kg CH₄ yr^-1(19. In heifers of 499 kg of weight and kept in solitary confinement or grazing, the EF was of 77 and 67 kg CH₄ yr^-1, respectively²². The EF of heifers (estimated= 26 and observed= 46 kg CH₄ yr^-1) and calves (estimated= 19 and observed= 30 kg CH₄ yr^-1) were lower than those reported on previous occasions. This is likely due to differences in body weight and has an impact on the requirement of ENm and ENc and on the DMC^3,12.

In bulls, the estimated DMC was 53 % higher (9.32 vs 9.75 kg d^-1; P<0.02) than that observed (Table 3). The estimated and observed values of GEC, EF, and PCH₄ were similar, despite the large difference in the DMC. For bulls aged over two years, a DMC of 8.6-9.2 kg d-¹, a GEC of 37.95 Mcal d^-1 and an EF of 62.18 kg CH₄ yr^-1(19 were estimated; these values are consistent with those observed in the present study. In four of five farms whose herds included bulls, these were fed food concentrates and, in general, the composition of the rations for bulls was more similar to that for cows in milking condition. Probably, the inclusion of concentrates and more digestible ingredients caused the observed DMC to be lower than expected, and the GEC to be similar to the expected value.

In young bulls in fattening, the DMC (6.95 and 7.40, kg d^-1) was similar, but the GEC (15.55 and 22.88, Mcal d^-1), EF (27.74 and 40.82, kg CH₄ yr^-1) and the PCH₄ (1.01 and 1.49, Mcal) estimates were 69 % lower (P<0.05) than those observed in the farms (Table 3). In young bulls aged one to two years, a DMC, 8.3 kg d^-1, GEC 36.0 Mcal d^-1 and an EF of 60.1 kg CH4 yr^-1(19 were estimated. These values are greater than those found in the present study, probably due to the greater weight of the growing bulls (on average, 540 kg) and quality of the diet in the above-mentioned study. In the present work, the average weight of the growing bulls was 290 kg, and in three of four farms, corn stalks were the predominant ingredient in the food.

Table 4 shows the weighted averages of the EF, HCP₄ and ML farm-1. The EF (49.06 and 54.09, kg CH₄ yr^-1) was similar, but the PCH₄ (1.11 and 1.97, Mcal d^-1; P<0.01) and ML (0.13 and 0.27, Mcal kg milk^-1; P<0.03) estimates were 44 and 52 % lower, respectively, than those observed in the farms. The lack of difference in the weighted EF between the estimated and the observed values coincide with the results observed in cows in milking condition; the estimates appear to be appropriate when applied to the herd as a whole, probably due to the specific weight of lactating cows, when the IPCC methodology is used. The PCH₄ and ML are not included in the estimates of the IPCC; however, they make it possible to know the impact of nutritional practices on the inefficiency of energy use and can be used to estimate economic inefficiencies associated with these losses, with additional information on input costs. The strength of these variables is that their estimation takes into account both the productive and the non-productive animals in the herd, in addition to the cows in milking condition. The results suggest that, using the technical information of the farms, it was possible to demonstrate a greater PCH₄ and ML farm^-1, and, consequently, a lower efficiency, with respect to the IPCC methodology, probably due to the greater PCH₄ observed in pregnant cows, heifers, calves and young bulls (Table 3).

The findings suggest that the estimates of DMC, GEC, EF AND PCH₄ with the level 2 methodology recommended by the IPCC were similar only to those observed in cows in milking condition, but there were contrasting results in the rest of the animals of the herd. This is probably due to the fact that most of the case studies where they have been derived from the equations have been conducted in cows in milking condition, but there is little work done on animals at other stages of production. The largest PCH₄ and ML observed in the farms may be due to the fact that predictor equations have been derived from studies on animals with higher levels of production and in experimental conditions with adequate nutritional, environmental and sanitary control; while the visited farms used the ingredients available according to the time of year, they do not apply a proper balancing of rations, their facilities are poor, and they do not implement appropriate preventive actions^5,8,9.

Conclusions and implications

The estimates of DMC, GEC, EF and PCH₄ using the procedures proposed by the IPCC level 2 method agreed with the observed values only in cows in milking condition; however, the results were inconsistent when applied to the rest of the animals in the herd. In general, the use of technical information of the farms made it possible to estimate the EF and demonstrate a greater PCH₄ and ML farm^-1, and, consequently, a lower efficiency, with respect to the IPCC methodology. There are several nutritional factors that influence the degradability of food, the patterns of enteric fermentation and ruminal production of CH₄ -such as the amount and type of starches, the type and solubility of proteins, the type and concentration of the fractions of fiber, among others-, that should be considered in future work to improve the precision of the estimates of CH₄. Other limitations of the study were that the producers did not have systematic records of information; part of the technical data were estimated on the basis of the subjective perception of producers, and there was no means of verification.