Introduction

Concerns over global warming make it important to consider management practices that affect carbon sequestration in soils and emissions of greenhouse gasses. In Mexico, the animal production sector is the main source of methane; in 1990 the national inventory of greenhouse gases reported emissions of 1.89 teragrams (Tg) methane, of which 1.85 Tg were produced by domestic livestock (^{González and Ruíz-Suárez, 1995}). Methane emissions in national inventories of national communications of the last decade show high growth rates: 2006 with 1.83 Tg in the Third Communication (^{INE-SEMARNAT, 2006}), 2009 with 4.89 Tg in the Fourth Communication (^{INE-SEMARNAT, 2009}), 2012 with 45.52 Tg in the Fifth Communication (^{INE-SEMARNAT, 2012}) and 2015 with 64.94 Tg in the Biennial Report (^{INECC-SEMARNAT, 2015}). Differences in emissions are very likely due to the use of different methodologies and sources of data but put into context the problem of methane emissions from the livestock sector and define it as a strategic sector to reduce carbon emissions of the country.

The need for sustaining agricultural production and the livelihoods of producers, restoration of production potential, efficient use of limited resources and reduction in production of harmful by-products, demands the development of management practices that make food production systems more efficient, not only in terms of producing more food per unit of limited resource (e.g. water) but also in terms of producing more food per unit of input and harmful by-product produced (e.g. methane). ^{O’Mara (2004)} has proposed that to reduce methane emissions, among other things, we must (1) increase the productivity per cow, which will reduce methane emissions per unit of animal product and also the total production of emissions per production system, and (2) reduce the rate of replacement of stock cows at the beginning of the breeding season, which will reduce the total production of emissions per production system.

^{De Ramus et al. (2003)} indicated that, although emissions of methane from beef cattle that result from meeting maintenance requirements cannot be modified by management practices, methane emission beyond those associated with maintenance can be reduced based on the level of productivity of the animal. These authors evaluated two grazing management systems, one using the extensive, continuous grazing scheme typical of the region, and one using an intensive grazing system with pastures that had been improved by over seeding with additional forage species. However, in areas of low precipitation, it is unwise to make large investments in seeding or infrastructure to improve pastures, thus producers should seek low-investment strategies. Since production in extensive cow-calf systems is tightly linked to the variable and unpredictable precipitation regime, it is not possible to define efficient static management practices, such as is possible in other animal production sectors.

Drought is common in arid and semi-arid regions of the world, and agricultural production systems should be adapted to local weather patterns and variability. Droughts have an enormous impact on agricultural activities, and, especially in developing countries, can be an important cause of poverty and emigration. For example, in the Borana Mesa in southern Ethiopia, the droughts of 1983 - 1985 and 1991 - 1993 caused cattle mortalities of 37 and 42%, respectively. These losses were valued at approximately 15 times the net sales of cattle in the area (^{Desta and Coppock, 2002}). During these same two droughts, ^{Oba (2001)} reported cattle mortalities of 70 - 87%, respectively, in northern Kenya. In northern Mexico and the southwestern United States, according to the projections of 19 different climate models, a climate similar to the drought period of the 1950’s will return within the next few years or decades (^{Seager et al., 2007}).

Due to this climatic variability, the establishment of fixed stocking rates for semi-arid rangelands appears unwise (^{Behnke and Scoones, 1993}; ^{Illius et al., 1998}). One common option for maintaining a constant stocking rate as climate varies is to provide supplemental feed; however, irreversible vegetation change may occur if animal numbers are held constant when natural resources are scarce (^{Van de Koppel and Rietkerk, 2000}). ^{Díaz-Solís et al. (2006)}, in a simulation modeling study, found that range condition commonly decreases as the level of energy supplementation in the form of hay and concentrates is increased because supplemented cows have a lower mortality rate, thus promoting higher stocking rates which result in even more pronounced over-utilization.

Alternative management strategies for dealing with drought might include increasing or decreasing stocking rate based on the current condition of the pasture, season of the year, and the direction and rate of change in animal body condition. However these reactive strategies have market problems because the producers need to sell when many other producers are liquidating animals and prices are low, and they then have to restock when prices are very high after the drought. Managers also might influence body condition by weaning calves early and shifting the breeding season such that the period in which nutrient requirements of animals are highest coincides with the period of highest forage quantity and quality.

The low productive efficiency of extensive cow-calf production systems is due to the fact that we use static management rules, that is, we do not adapt pasture and animal management to the variable precipitation. For example, during a drought we employ the same stocking rate that is estimated based on periods of “normal“ precipitation, which results in a larger proportion of cows that do not produce a calf, but continue producing methane (^{De Ramus et al., 2003}).

Although the arrival of a drought is unpredictable, in this study we examine whether it is possible to increase CO₂ sequestration in soils and reduce methane emissions, as well as reduce the negative effects of drought on cow-calf production systems, by using different management decision strategies in drought-prone regions. In a previous study, in which we simulated dynamics of extensive cow-calf production systems (^{Díaz-Solís et al., 2003}), we observed predictable time lags between precipitation events, production of phytomass, and changes in body condition of cattle. Specifically, loss of body condition was greater the year after a dry year, even when that dry year was followed by a year with higher than average precipitation. Kgosikoma (2006)¹ searched for correlations between time-lagged precipitation and variables affecting agricultural production to relate climatic variability and animal population dynamics in Botswana. ^{Illius et al. (1998)} conducted a similar simulation study with some different results, however, their study focused on an essentially grass-dominated system and a single species of small ruminant (sheep or goats).

This study evaluates the potential of three management strategies to increase CO₂ sequestration, reduce methane emissions, and reduce the financial and environmental negative effects of drought on cow-calf production systems in semi-arid rangelands: (1) The traditional strategy (CONTROL) uses a relatively high stocking rate that is not adjusted and open cows are not sold after breeding season; (2) the first alternative strategy (LOW SR) is similar to CONTROL but uses low stocking rate; and (3) the second alternative strategy (REPLA-PPT) uses the decision of replacement of breeding stock cows based on the amount of precipitation during the last growing season with a relatively low stocking rate. In this strategy, open cows are sold after breeding season, and replacement of breeding cows is based on the amount of precipitation during the previous growing season.

Materials and Methods

First it is described the study area selected as the test case, and the alternative management strategies, and the modeling procedures. Model SESS (Simple Ecological Sustainability Simulator) (^{Díaz-Solís et al., 2003}) was modified to represent each of three different management strategies (sets of decision rules). In the sections that follow, it is described the basic structure of the model SESS and the modifications done to SESS for this paper. Then use of the model to evaluate CO₂ sequestration by vegetation, methane emissions, and animal performance and net long-term profits under these management strategies.

Description of the study area

To evaluate the performance of the management strategies, the SESS model was parameterized to represent a region of extensive cow-calf production systems in the northeastern portion of the Mexican state of Coahuila, specifically within the municipalities of Allende, Zaragoza, Muzquiz, Monclova, Sabinas, and San Juan de Sabinas. The area has a mean annual precipitation near 400 mm, with a coefficient of variation of more than 30% (Figure 1). The carrying capacity of these lands for cattle varies significantly both seasonally and from year to year in a manner that is typical of drought-prone rangelands throughout the world. Cattle production systems in the area also are typical of semi-arid rangelands throughout the world in that producers seldom adjust stocking rates based on environmental conditions. In the study area, a governmental agency establishes permissible stocking rates (statics) based on vegetation surveys conducted almost 30 years ago (^{COTECOCA, 1979}) that does not consider precipitation variability or if the precipitation in the year of the study was higher lower or average.

Figure 1 Mean annual precipitation (mm) in the northeast of Coahuila, Mexico, during the period from 1950 to 1994. 

Changes to SESS

Monthly biomass production (MNPP) depends on precipitation, RUE, RC and a temperature growth index (TGI) presented by Sikhalazo (2005):

MNPP= (PPT*RUE*RC)*TGI                                (1)

where PPT is precipitation (mm month^-1), RUE and RC are rain use efficiency and range condition respectively as described by ^{Díaz-Solís et al. (2003)}

To simulate the new alternative management strategy (REPLA-PPT), we added the following decision rule to SESS:

IF PDIF > 0 THEN REPLA ELSE 0                       (2)

where PDIF represents the difference between the total precipitation observed during the last growing season (PPT_c); (March through November) and the historical mean precipitation observed during the growing season (PPT_m):

PDIF = PPTc - PPTm                                                (3)

Thus, under this rule, breeding cows are replaced (REPLA) when growing season precipitation of current year has been higher than average. To avoid the number of cows reaching 0, for those years in which PDIF ≤ 0 and the number of cows ≤ 20% of the base-level stocking rate (BSR), a sufficient number of stock cows are replaced to achieve a stocking rate = 20% of BSR.

To estimate carbon sequestration by vegetation we used simulated aboveground and belowground biomass production. Belowground biomass production was considered as 2.8 time’s aboveground herbaceous biomass. C captured by biomass, on a basis of dry matter weight = (aboveground biomass production + belowground biomass production)*0.5 (^{Westlake, 1966}; ^{Penman et al., 2003}). Carbon sequestration by vegetation (aerial vegetation is a flux) is an indirect measure of soil organic carbon (SOC) sequestration (Paz et al., 2015) that must be used to represent carbon sequestration. Nevertheless, in this study, total vegetation (above and belowground) was used as an indirect proxy for SOC estimation. Data from Paz (2010)² show that the transition from primary to secondary condition of vegetation in northern Mexico (scrub and grassland), which implies a loss of at least one category in the range condition; generate a loss of around 9.3 Mg C ha^‑1. However, considering that the loss of soil organic carbon in the first five years account for 80% of total long-term losses (^{West et al., 2004}), in this period a loss of 1.5 Mg C ha^-1 is present, which it represents values comparable to deforestation processes.

To estimate methane production we used the total livestock biomass (kg) simulated by SESS and information from the Livestock Analysis Model (LAM 1.01) (http://www.epa.gov/rlep/library/lam/intro.html, consulted May 2009). The total livestock biomass simulated by SESS consisted of the mass of each class of animal (stock cows, calves, heifers, bulls) multiplied by the mean body weight in kg (SBW) of that class. LAM estimates methane production by beef cattle on rangelands in which both quality and availability of forage is low as 20 kg head^-1 year^-1 for heifers (200 kg) and 71 kg head^-1 year^-1for adult females (450 kg), respectively. Based on this information, we estimated:

Methane (kg head-1 year-1) = 16.18 + 0.084 * SBW     (4)

Since methane global warming potential is 21 times that of carbon dioxide, we multiplied by 21 to convert kg of methane head^-1 year^-1 to kg CO₂ eq. head^-1 year^-1 (^{Penman et al., 2003}).

A simple net present value economic sub model was added to SESS (^{Teague et al., 2008}) to generate the state variable PROFIT. The profit values shown in Table 1 are United States Dollars (US$) for the whole ranch (5000 ha) over the 45 years simulated. Revenue consists of: -weaned kilograms * $ per kg (US$2.10), -sale of cows (#) * price per cow (US$500), and -hunting income (US$12 ha^-1 year^-1). The costs of the system considered were: -replacement of stock cows (US$ 750 head^-1) and -annual cow cost (US$350 head^‑1). The costs and product prices used in this paper were taken from neighboring Texas (^{Teague et al., 2009}) because we were unable to get this information for ranches in northern México. Notwithstanding the previous statement, the financial data are representative of operations in Coahuila, Mexico.

Table 1 Simulated carbon sequestration, methane emissions and animal performance under three management strategies in a cow-calf operation in the northeast of Coahuila, Mexico during the period from 1950 to 1994 (Ranch area = 5000 ha). 

^† SR = stocking rate.

Results and Discussion

REPLA-PPT was the best strategy in terms of C sequestered, methane emissions, range condition, animal performance and economic profit. The CONTROL was the worst strategy. The total carbon sequestered by vegetation (on 5000 ha during a 45-year period) under the adjustable stocking rate strategy (REPLA-PPT) was over 0.55 million tons, and only 0.39 million tons under the constant stocking rate strategy (CONTROL) (Table 1). The difference of sequestration between strategies increased over time (Figure 2), due to an improvement in range condition (Table 1) and higher biomass production, which resulted from a lower overall stocking rate and, hence, a lower level of utilization of the available biomass with adjustable stocking.

Figure 2 Cumulative carbon sequestered by vegetation under three management strategies in a cow-calf operation in the northeast of Coahuila, Mexico during the period from 1950 to 1994 (Ranch area = 5000 ha). 

According to Table 1, the range condition is regular for CONTROL and it is excellent for REPLA-PPT, involving a change between a “primary” and well preserved rangeland and a “secondary” and no preserved, which would imply losses of around 1.5 Mg C ha^-1 organic carbon in the soil (30 cm), according to what is previously discussed.

Since methane emissions are intrinsically correlated with animal numbers, emissions always were lower with REPLA-PPT and LOW SR strategies than with CONTROL (Figure 3). Emissions were lowest with REPLA-PPT during the drought from 1951 to 1959, and the two or three years after a dry year (64-65, 67‑71, 80-81, 85-86) when stocking rate was lowered by not replacing culled, non-pregnant cows. Mean annual precipitation from 1950 to 1956 was 282 mm, and annual precipitation during 1956 was 147 mm, which is only 35% of the long-term historical mean annual amount (Figure 1). The main seasonal differences in emissions with CONTROL were due to large seasonal changes in the number of animals on the ranch, which resulted from high mortality (Table 1) and subsequent high annual restocking to maintain the base-level stocking rate (250 AUY).

Figure 3 Monthly dynamics of methane emission by animals under three management strategies in a cow-calf operation in the northeast of Coahuila, Mexico during the period from 1950 to 1994 (Ranch area = 5000 ha). 

The relation between methane emissions by animals and CO₂ sequestration by vegetation in extensive, adjustable stocking, cow-calf production systems results in a simulated net decline in these substances which it is assumed would result in a decline in human contributions to global warming. Quite likely, the current contribution of cattle production systems as high sources of greenhouse gases is due to feedlot operations and intensive dairy cattle operations in which high animal densities are maintained.

Calf production efficiency, measured as kg ha^-1year^-1 of calf production per Mg CO₂ eq of methane emission, was more than three times higher with REPLA-PPT as compared to CONTROL (102.47 versus 29.30, respectively). Average body condition score was 4.78, 3.79 and 2.85 for REPLA-PPT, LOW SR and CONTROL respectively. Results for profit had the same trend and only REPLA-PPT showed a positive economic balance with the CONTROL showing the largest economic loss (Table 1).

The advantages of adjustable stocking based on replacement are due to its flexibility to adapt ranch management to current conditions, specifically, those resulting from the variable precipitation regime. The REPLA-PPT strategy reduces stocking rate during unfavorable precipitation conditions by selling more livestock and by not replacing livestock that have died. When favorable precipitation conditions return, this strategy replaces only enough stock cows to reach the low base-level stocking rate (125 AUY).

High and constant stocking rates have both short- and long-term negative effects on coupled social-ecological systems. In the short term, high stocking levels increases competition for forage resources. This results in less selective grazing and lower total consumption per cow. In the long term, the resulting overgrazing lowers range condition and, hence, forage productivity, which, in turn, reduces vegetation cover and increases soil erosion. In the long-term this reduces the ecosystem services that rangelands provide and ability of the population to feed itself.

Conclusions