Simulation of Pastoral Management in Mongolia: An Integrated System Dynamics Model

Rangeland Ecology & Management xxx (2018) xxx–xxx

Contents lists available at ScienceDirect

Rangeland Ecology & Management journal homepage: http://www.elsevier.com/locate/rama

Original Research

Simulation of Pastoral Management in Mongolia: An Integrated System Dynamics Model Shunji Oniki a,⁎, Kazumasa Shindo b,1, Seishi Yamasaki b, Kazunobu Toriyama b a b

Social Science Division, Japan International Research Center for Agricultural Sciences, Tsukuba, 305-8686, Japan Crop Livestock and Environment Division, Japan International Research Center for Agricultural Sciences, Tsukuba, 305-8686, Japan

a r t i c l e

i n f o

Article history: Received 15 June 2016 Received in revised form 14 November 2017 Accepted 6 February 2018 Available online xxxx Key Words: biomass pastoralism policy simulation profit system dynamics

a b s t r a c t High grazing density has given rise to concerns about grassland degradation in periurban areas in Mongolia. Moreover, whether livestock can increase without harming the vegetation in these areas in Mongolia and what types of policy measures should be implemented is not documented. As such, this study develops an integrated simulation model of grassland biomass, animal growth, and livestock management for a forest-steppe area in northern Mongolia and conducts a simulation on long-term changes. The simulations show that, under current conditions, the number of animals will continue to increase, while the grassland biomass will decrease. Cooperative grassland management would lead to an increase in grassland biomass and higher incomes for herders. Furthermore, herders’ population changes would have a significant impact on animal density adjustments, while the effects of conventional economic measures, such as a tax on animals, would be limited if all other conditions remain constant. Consequently, the synergistic effects of herder population changes and cooperative management can contribute toward maintaining the herders’ income while preserving the grassland ecosystem. © 2018 The Society for Range Management. Published by Elsevier Inc. All rights reserved.

Introduction In many world areas, grassland degradation occurs due to long-term increases in the number of animals in specific pastoral areas (Oldeman et al., 1991; Suttle et al., 2005; Muya et al., 2013; Lwiwski et al., 2015). Herders believe that increasing animal numbers would raise their incomes, although grassland desertification due to excessive animal grazing can eventually result in a decline in income. Therefore, it is important to maintain a balance between the optimal grassland conservation and herders’ livelihood. There are also concerns about human-induced degradation in Mongolian grasslands (Chuluun, 2011). Particularly in the Mongolian periurban areas, the people moved with their animals from remote areas after transition to a market-oriented economy in the 1990s, which resulted in high animal densities. If the situation continues for a long-term, then grasslands could degrade, and there could be a failure to achieve sustainable livestock development. Essentially, we are interested in whether an increase in herder incomes and grassland preservation are compatible under the current conditions and, if not, what measures can be taken to correct the situation.

⁎ Correspondence: Shunji Oniki, 1-1 Ohwashi, JIRCAS, Tsukuba, 305-8686, Japan. Tel.: +81 29 838 6382; Fax: +81 29 838 6342. E-mail address: [email protected] (S. Oniki). 1 Current address: Kazumasa Shindo, NARO Institute of Livestock and Grassland Science, 768 Senbonmatsu, Nasushiobara, Tochigi, 329-2793 Japan, 0287-37-7808.

Previous studies have investigated the effects of grazing pressure and other natural factors on grassland vegetation (Zhao et al., 2004; Sasaki et al., 2005; Zhou et al., 2006; Cheng et al., 2011; Muya et al., 2013; Lwiwski et al., 2015) and the relationships among grazing pressure, animal production, and forage consumption by livestock (Smart et al., 2010). In addition, Bement (1969), Hart et al. (1988), and Manley et al. (1997) show tradeoff relationships between the productivity of grassland and the productivity per unit of animal head in their studies on experimental stations in high plains areas of the United States. On the other hand, there are growing concerns about the socioeconomic behaviors and decision making of farmers, such as profitseeking behaviors and ecological systems, in rangeland studies (Lubell et al., 2013; Marshall and Stokes, 2014; Roche et al., 2015; Wilmer and Fernandez-Gimenez, 2015). However, few previous studies fill the gap between the biophysical relationships and socioeconomic behaviors, especially in the field of pastoralism in developing countries. Grassland degradation occurs through evolutionary history of interactions between plants and animals (Milchunas et al., 1988). In addition, aboveground plant biomass and animal grass consumption are interrelated with livestock income in the long term. Essentially, plant biomass affects the number of animals and incomes, which, in turn, dictates the number of animals and amount of plant biomass. In order to analyze a complex feedback system, we use a system dynamics model, which consists of stocks and flows and incorporates feedback relationships to understand complex systems that change with time, as opposed to simple linear models (Ford, 2010).

https://doi.org/10.1016/j.rama.2018.02.003 1550-7424/© 2018 The Society for Range Management. Published by Elsevier Inc. All rights reserved.

Please cite this article as: Oniki, S., et al., Simulation of Pastoral Management in Mongolia: An Integrated System Dynamics Model, (2018), https:// doi.org/10.1016/j.rama.2018.02.003

2

S. Oniki et al. / Rangeland Ecology & Management xxx (2018) xxx–xxx

Certain studies, such as that of Belcher and Boehm (2003), successfully developed models using an economic sector and environmental conditions. Concerning livestock production, Meadows (1970) developed a model for livestock production cycles and Davidsen and Asheim (1993) simulated long-term changes in mink farming trends. Several system dynamics studies (Moxnes, 1998, 2004; Moxnes et al., 2001; Badarch and Ochirbat, 2002; Belcher and Boehm, 2003; Bald et al., 2006; Tedeschi et al., 2011; Gies et al., 2014; Dace et al., 2015) developed models on the relationships between environmental conditions and animal stocks. Some studies incorporated economic components (Hart et al., 1988; Tanaka and Workman, 1988; Manley et al., 1997; Ritten et al., 2010; Dunn et al., 2010) and analyzed economic effects of a public land policy (Torell and Doll, 1991; Torell and Drummond, 1997) in their rangeland models of a livestock farm in developed countries. Nonetheless, few simulation models that integrate an economic model and a livestock model have been developed till date. Concerning developing countries with large areas of open-access grassland where nomadic pastoralists live, no attempt has been made to develop simulation models that integrate herders’ economic activities and grassland biomass. Consequently, the objective of this study is to simulate alternative policies on measures for preserving the grassland in a forest-steppe area in Mongolia and compare these simulations with their baseline simulation (i.e., a scenario for “business as usual”). Our aim is to seek a policy to increase herders’ incomes or prevent the incomes from dropping as much as possible. While available policy options are limited in arid areas in a developing country, we test the following scenarios: cooperative management of pasture, imposition of high tax rates on livestock, herders’ population changes, as well as a combination of these scenarios, otherwise referred to as the policy mix option. This study proposes the following hypotheses. In the case of business as usual, we speculate that the number of animals continues to increase with a decrease in the grassland biomass. This is a case of unregulated common property (Baland and Platteau, 1996). However, as Ostrom (1990) and Baland and Platteau (1996) argue, collective action or coordination among members using common natural resources can solve the problem. In Mongolia, international organizations introduced intervention projects for grassland group management, which provided groups of herders with exclusive rights to use grasslands and allowed them to control the total number of animals for the project participants. 2 We made a simulation for a scenario on group management for pasture utilization. Another policy measure that can be taken to prevent an increase in the number of animals pertains to the imposition of a high rate of tax on livestock holding. Tax on a factor that is harmful to the environment increases the marginal cost of using the factor and reduces the damage; this kind of tax is called a Pigouvian tax (Baumol and Oates, 1988). Mongolia levied tax on animals until the severe winter disaster in 2010. The tax rate levied was higher for goats than sheep and other major animals in the sheep unit because goats were considered more harmful to grassland than the other animals. However, the rate was not high enough to reduce the number of animals. Another possible policy measure that can reduce the population pertains to the restriction of inward migration and promotion of outward migration of herders from an overgrazed area. Because an increase in the human population is one of the major threats to the commons (Hardin, 1968), it is recommended to keep a check on the local population. In addition, since these policies are not mutually exclusive, we will examine the effects of mixed policies.

2 In 2013, the Peri-Urban Rangeland Project, Millennium Challenge Corporation of the USA (https://www.mcc.gov/mongolia) and Green Gold Program by the Swiss Agency for Development and Cooperation (http://www.msrm.mn/) introduced group management strategies for grasslands.

Methods Data The study was conducted in the Bornuur soum (district), Tov Province, Mongolia, located approximately 90 km north of the capital, Ulaanbaatar (Fig. 1). 3 An asphalt-paved road between Ulaanbaatar and Erdenet and Darkhan Uul (the second and third largest cities, respectively) passes through Bornuur. Due to access to large cities, many people migrated to this area. The ecological zone of Bornuur is categorized as forest-steppe, and the average annual precipitation and coefficient of variation from 1990 to 2010 in Bornuur are 247 mm and 0.32, respectively.4 Fernandez-Gimenez and Allen-Diaz (1999) characterize Mongolian “mountain-steppe” areas as equilibrium rangeland ecosystems, where animal grazing may affect grassland conditions; however, under nonequilibrium rangeland ecosystems in drier areas, grassland conditions are determined primarily by weather and not by animal grazing. 5 In the past decade, the Bornuur soum underwent some damage due to winter disasters, such as damages from heavy snow and extreme temperatures, but the magnitude of disaster was lesser than that of the steppe and desert steppe areas of Mongolia. During the period 1990–2010, the largest damage was 17.3% in 2001, followed by 13.3%, 10.7%, 12.1%, and 7.0% in 1994, 1999, 2003, and 2010, respectively. 6 Concerning all the areas of Mongolia, the largest damage that took place during 1990–2010 was 23.8% in 2010, followed by 15.7%, 11.2%, 10.4%, and 6.4% in 2001, 2002, 2010, and 1993, respectively. These statistics imply that winter disaster is relatively moderate in Bornuur. Mongolian herders often work together for their animal husbandry and other activities, and there are informal rules about pasture utilization among herders who use the same areas (Mearns, 1996); however, these collaborations or informal rules for groups do not reflect collective action, argued Olson (1965) and Ostrom (1990). Herders move freely within district boundaries as nomadic pastoralists.7 Even if herders violate informal rules, their punishment is not as strong as in a closed community of farming people. While annual variations in grass production are high in steppe or desert-steppe areas of Mongolia due to high annual fluctuations in precipitation, these variations are less in forest-steppe areas due to low fluctuations in precipitation (Fernandez-Gimenez and Allen-Diaz, 1999). The characteristics of smaller annual fluctuations in forest-steppe are suitable for investigating the relationship between the biomass of grass and animal grazing. In this study, for livestock management modeling, we use household-level data collected from 290 sampled households in 2007 and 2008 in Bornuur. Data for the aboveground plant biomass and animal weight model is obtained by grazing experiments conducted in Bornuur (Shindo et al., 2013) and a feeding experiment in Ulaanbaatar (Yamasaki et al., 2014, 2016). Model Assumptions The model consists of the following three primary modules: vegetation-animal weight, animal population growth, and livestock 3 Mongolia has 21 provinces (aimags) and the capital city, Ulaanbaatar. Each province is constituted of several soums, translated to district or county, which is divided into bags (villages). 4 The Tov Province Meteorological Station provided the data. 5 The mountain-steppe area studied by Fernandez-Gimenez and Allen-Diaz (1999), with an average annual precipitation of 230 mm and a coefficient of variation of 0.28, has similar conditions to the forest-steppe, although precipitations are higher and more stable in the Bornuur soum. 6 Data for 1990 −1995 were obtained by the National Statistical Office of Mongolia, 1996. Data for the other period were obtained by statistical data collected by the National Statistical Office of Mongolia. Mongolian Statistical Yearbook (National Statistical Office of Mongolia, 1996–2010). 7 Herders sometimes move beyond the boundary of the district (soum) when necessary. In this case, they may cause conflicts or pay additional levies for grazing.

Please cite this article as: Oniki, S., et al., Simulation of Pastoral Management in Mongolia: An Integrated System Dynamics Model, (2018), https:// doi.org/10.1016/j.rama.2018.02.003

S. Oniki et al. / Rangeland Ecology & Management xxx (2018) xxx–xxx

3

Bornuur

Ulaanbaatar

Mongolia

Tov Province Fig. 1. Study sites in the Tov Province, Mongolia

management. The model’s framework (Fig. 2) computes the future values of the total number of animals, grass biomass, average animal weight, and household profits.8 The model represents the entire Bornuur area. All herder households are assumed to be engaged in livestock production and having homogeneous characteristics. The basic time unit in the model is a month, and a year is considered from May to April in the following year. Grass starts growing at the beginning of May and ends in the middle of September. Furthermore, animals have offspring at the beginning of the year. The model also defines “summer pasture” and “winter pasture,” the former being used from the beginning of May until the end of October and the latter from the beginning of November until the end of April. Sheep are the representative animals in this model. The other animals (i.e., goats, cattle, and horses) are converted into sheep units (SU) 9 (see Table 1 for initial values).10

temperatures. Although drought in summer rarely causes immediate animal death in Mongolia, it is one of the major death determinants in the following winter. Winter temperatures often decrease significantly, so grass does not grow in winter. If sufficient growth does not happen from spring until fall, then animals are more likely to die during winter and early spring. If grass growth is hindered by a shortage of rainfall in the summer (from May to August), then animals become weak and cannot survive during severe winters. Although the risks of winter disasters are not significant in Bornuur, we consider these risks by using stochastic death rate values. In this case, animal death ratios are determined randomly each year. The logarithmic normal distribution is estimated using actual death ratios from 1993 to 2008 in Bornuur. Birth and death ratios are correlated because they are affected by climatic conditions in winter and spring. The correlation estimated by the actual data in 1993–2008 is

Animal Population Growth Module

Birth ratio f ¼ 0:662−0:921∙Normal death ratio=2;

The animal population growth module in this study consists of six stocks for the number of animals in the following three age-sex stages: 1-yr-old males (N1m) and females (N1f), 2-yr-old males (N2m) and females (N2f), 3-yr-old and older males (N3m) and females (N3f) (Fig. 2). The stock in each stage evolves from the previous stage by birth or aging. It implies that the stock in each stage decreases as the stock in the succeeding stage increases due to aging. For example, the stock of 1-yr-old male animals (N1m) decreases as the stock of 2-yr-old male animals (N2m) increases at the end of a year due to aging. The stock also decreases through slaughter or death. After 8 yr, all animals are slaughtered and sold; therefore, we assume that all animals live up to 8 yr. Natural disasters also affect the number of animals, the major natural disasters in Mongolia being winter-related (dzud) and drought. Most animal deaths occur in winter due to heavy snow and extremely low 8

Model details are available upon request. This paper uses the Mongolian conventional conversion factor to compute SU and obtains the following results: 0.9 SU for a goat, 6 SU for cattle, 7 SU for a horse, and 5 SU for a camel. 10 The authors can provide the program file for the model on request.

½1

where Birth_ratio_f is the birth ratio for females and males. The number of births (Birth_f) is obtained by multiplying the birth ratio (Birth_ratio_f) with the number of female animals aged 2 yr or older. The birth rate is also affected by the degradation of grassland. Although no studies have estimated the relationship between the weights of ewe and their birth rate in Mongolia, we need to incorporate it to make the model more realistic. On the basis of the data of Coop (1962), 11 we estimated the relationship between the survival rate of lambs and the weights of the ewe in such a manner that the birth rate declines to 0.404% with a 1% decrease in the weights of the ewe. In this model, the animals are slaughtered in the following manner. Male and older animals are prioritized. First, mature male animals are slaughtered. If the number of animals in this stock is not sufficient, then all the 2-yr-old males are slaughtered. Subsequently, mature females are slaughtered. If animals in this stock are unavailable, then 2yr-old females are slaughtered.

9

11 We used data on 2–7 yr old sheep presented in Table 1 in Coop’s (1962) article and estimated the relationship using the ordinary least square method.

Please cite this article as: Oniki, S., et al., Simulation of Pastoral Management in Mongolia: An Integrated System Dynamics Model, (2018), https:// doi.org/10.1016/j.rama.2018.02.003

4

S. Oniki et al. / Rangeland Ecology & Management xxx (2018) xxx–xxx

Fig. 2. Framework of the system dynamics model

Vegetation and Animal Weight Module The vegetation and animal weight module represents the relationship between the plant biomass and animal weight and the proportion percentage, while the summer pasture area is assumed to be 45% (Moyobuu and Nyamaa, 1998). The parameters for plant biomass and animal weight module are estimated using the results of a grazing experiment conducted in Bornuur (Shindo et al., 2013). In addition, the details for the feeding experiment are presented in the study by Yamasaki et al. (2016). In the vegetation submodel, the amount of biomass decreases with the rate of plant consumption per animal, which is

proportional to the current animal weight starting at 3 yr of age and the current stocking ratio, which is the number of animals per hectare in the summer pasture area. The intake ratio in summer slightly decreases as the biomass increases. This relationship was estimated by grazing experiments of our research project in Bornuur (Shindo, 2015). Because the experiments were not carried out in winter, the winter intake ratio was set at the lowest value. Degradation of the summer pasture occurs as the amount of plant consumption increases. Following the pasture degradation process proposed by Zhao et al. (2004), the plant growth rate in the subsequent

Table 1 Initial values of stocks in the model. Name of stock

Description

Weight_1 Weight_2 Weight_3 N1m N2m N3m N1f N2f N3f Capital Labor Feed_stock Cost_convert Income_stock NPV_cum_income Invest_adjst_lag

Weight of a 0- to 12-mo-old animal Weight of a 13- to 24-mo-old animal Weight of a 25- to 96-mo-old animal Number of a 0- to 12-mo-old male animal Number of a 13- to 24-mo-old male animal Number of a 25- to 96-mo-old male animal Number of a 0- to 12-mo-old female animal Number of a 13- to 24-mo-old female animal Number of a 25- to 96-mo-old female animal Fixed costs of capital per household Number of labor per household Total feed supply per household (hey: kg) Total costs in the year Total gross incomes in the year Cumulative net present value of the net income Time lag of the capital investment

Initial value 0 23.9 42 66 27 35 66 35 158 2 178 2.5 0 0 0 0 0

Unit kg/head kg/head kg/head Head Head Head Head Head Head 1 000 MNT Person

Source (1) (1) (1) (2) (2) (3) (3) (3) (3) (3) (3)

Sources: (1) Peng (1985), p. 496. (2) Minjigdorj and Sambuu (1986). (3) Herders’ household surveys in Bornuur soum in 2008. MNT indicates Mongolian currency, Tögrög. 1 MNT is equivalent to 0.000418 US dollars (at the market exchange rate on 8 March, 2018).

Please cite this article as: Oniki, S., et al., Simulation of Pastoral Management in Mongolia: An Integrated System Dynamics Model, (2018), https:// doi.org/10.1016/j.rama.2018.02.003

year decreases, leading to a degradation effect. 12 Degradation occurs if the yearly cumulative ratio of plant consumption per hectare of summer pasture to the cumulative amount of plant production exceeds

The costs of livestock production are the sum of labor, capital, and feed costs, and tax on livestock. The laborers comprise hired and family workers, the initial value of the number of laborers is 2.5, and their wage is 86 000 MNT per month.17 Additionally, a family laborer’s wage is half of the hired laborer’s wage because of the lack of employment opportunities for hired labor in the study area. Capital in this model includes fixed capital goods, such as animal barns and agricultural machinery. Fuel costs for machinery are also added in this category. The initial value of the capitab7(a)]Tmc178.b-207.5(50(e)0T)18.14.875T nt7(2-14.7(i)1i)13.0(n)-8.3(f)15.5(182.3(.6(ac24.031.3(c)-8.6(u)

S. Oniki et al. / Rangeland Ecology & Management xxx (2018) xxx–xxx

500

70

Weight ( kg / sheep)

Number of animal (sheep unit)

8

400 300 200 100

60 50 40 30 20 10 0

0

5

10

15

20

25

30

35

5

40

10

300 Profit (Tg/household)

Biomass (kg/ha)

250 200 150 100 50 0 15

20

25

30

35

40

c) Weight of an animal

a) The number of animals per household

10

20

year

year

5

15

25

30

35

9000. 8000. 7000. 6000. 5000. 4000. 3000. 2000. 1000. 0. -1000

5 10 15 20 25 30 35 40 year

40

year

d) Profit

b) Biomass

Fig. 6. Scenario of animal tax

animals, but only the animal husbandry profits. We should note that the profit becomes negative in some years; therefore, if herders fail to borrow from a credit institution (bank or government) after drought, then some animals will die due to malnutrition and their number will fall below the baseline. When simulating a case without using supplemental feed, the profit reaches the zero-level 3 yr before it reaches this level in the case using supplemental feed. Although supplemental feed does not solve the problem, it helps to mitigate the impact. Cooperative Management’ s Scenario To avoid the problem of overutilization of common-pooled natural resources, economists recommend group management for a common pooled natural resource instead of open-access utilization (Ostrom, 1990; Bromley, 1992; Baland and Platteau, 1996; Meinzen-Dick et al., 2002). If all resource users agree to maximize the total profits of the region, overutilization of the resource can be prevented. On the other hand, Dunn et al. (2010) argue that stocking rates tend to exceed the recommended rates even by profit-maximization behaviors in a private rangeland. Subsequently, we simulate group management in a scenario called the “cooperative management case,” which is not necessarily limited to cooperation among members. In the model, when profit increases, the optimal number of animals increases with the addition of capital and labor, and vice-versa. This process keeps the profit at the maximum level. The results of the simulations in Figure 5 use the assumption of the cooperative management scenario. The amount of biomass is larger, and the weight of the animal is greater than the baseline simulation. The number of animals increases but at a lower rate (see Fig. 5a), the

amount of biomass also decreases at a lower rate (Fig. 5b), the decrease in the weight of an animal is slower (Fig. 5c), and the profit remains at a positive level as expected (Fig. 5d). Increasing Animal Tax Rate Scenario The economic measure taken to curb the increase in the number of animals and prevent grassland overuse serves as a tax on grassland utilization. It increases the cost of holding animals for herders, and it may reduce incentives to increase the number of animals. Since stock externality is a major overstocking determinant, imposing a tax on land utilization would internalize the externality. For pastoral production, a tax may be imposed on the amount of natural grass consumed by animals; however, it is difficult to know how much grass the animals consume. Thus, a tax on livestock can be considered as a second-best policy. In Mongolia, an animal tax existed until 2009; however, the tax rate was almost negligible in terms of effectiveness in reducing the number of animals (100 MNT per SU, approximately 0.2% of gross income). We simulate to check how an increase in the tax rate to 10% of animal value might change the number of animals, assuming the other conditions remain constant. The results of the simulation show a limited impact on the number of animals, amount of biomass, and weight of the animal (Fig. 6a–c), as simulation reduces only household profit at a faster pace (Fig 6d). This is because the simulation did not assume any extra income sources other than animal grazing; in addition, it did not consider migration to other areas without a tax increase. In fact, there is a limited opportunity for side work or work away from home in Bornuur. Herders cannot

Please cite this article as: Oniki, S., et al., Simulation of Pastoral Management in Mongolia: An Integrated System Dynamics Model, (2018), https:// doi.org/10.1016/j.rama.2018.02.003

900 800 700 600 500 400 300 200 100 0

60 50 40 30 20 10 0 5

10

15

20 25 year

30

35

40

5

10

15

20 25 year

30

35

40

35

40

c) Weight of an animal

a) The number of animals per household

300

10000.

250

8000.

Profit (Tg/household)

Biomass (kg/ha)

9

70 Weight ( kg / sheep)

Number of animal (sheep unit)

S. Oniki et al. / Rangeland Ecology & Management xxx (2018) xxx–xxx

200 150 100

6000. 4000. 2000. 0.

50

-2000

0 5

10

15

20 25 year

30

35

5

40

10

15

20 25 year

30

d) Profit

b) Biomass

Fig. 7. Scenario of a decrease in population

abandon livestock production even if the profitability of livestock is lowered. 21 In addition, if only one district increases the tax rate, then the herders will move out from the district, thereby leading to a decrease in the herder population in that district. This model does not incorporate this secondary effect and assumes the district population to be constant.

Population Change Scenario An alternative scenario to mitigate grazing pressure is to decrease herder population in the area. Hardin (1968) argues that the primary cause of overstocking is an increase in the herders’ population density on an open-access resource. Consequently, as population density decreases, grassland areas per household increase, thereby resulting in a lower grazing rate, higher grassland biomass, animal weight gain, and household incomes.22 Reducing population is one of the most important policy measures to protect natural resources in populated areas. For example, relocation to another area can be promoted through subsidies on resettlement, which is referred to as the “environmental migration” in China (Nakao et al., 2010). Similarly, imposing a higher tax rate in an area may result in relocation to other areas. 21 However, it is possible for some herders to work for other herders, although our model does not accommodate such a case. 22 This study does not consider a disparity in the incomes levels. The results may be different if we consider a case for a small number of households who have many animals and a large number of herders who have a relatively small number of animals.

To simulate population changes, we set the following scenario: The population declines by 2% annually so that the total population decreases by half in 35 yr. As outward migration is difficult to estimate, this study runs the simulation under a given population change rate. The results show that the number of animals per household increases in the long term (Fig. 7). However, grassland biomass and the weight of an animal decreases at a lower rate than the baseline simulation. Thus, the decrease in population size has positive effects on grassland protection, while biomass and profits still decline over time. Population Change Under the Cooperative Management Scenario Previous simulations reveal that decreases in herder population and cooperative management have certain impacts on the number of animals; however, a single policy measure hardly prevents the decrease in plant biomass in the long term. To enhance policy effects, policymakers may want to implement more than one policy simultaneously; thus, we simulate a mixed-policy scenario. An example of mixed policies is the combination of population change and cooperative management. Because both result in higher amounts of biomass and higher profit levels than the baseline simulation, we expect the combination to be more effective. The results of the simulation are presented in Figure 8. The lines in the graphs indicate the case that combines the two best alternative measures. By declining population size under the cooperation scenario, the simulation demonstrates that the number of animals increases for longer amounts of time than in the baseline and the cooperation-only scenarios. Neither the biomass nor the weight of an animal has a

Please cite this article as: Oniki, S., et al., Simulation of Pastoral Management in Mongolia: An Integrated System Dynamics Model, (2018), https:// doi.org/10.1016/j.rama.2018.02.003

S. Oniki et al. / Rangeland Ecology & Management xxx (2018) xxx–xxx

700 600 Weight ( kg / sheep)

Number of animal (sheep unit)

10

500 400 300 200 100 0 5

80 70 60 50 40 30 20 10 0 5

10 15 20 25 30 35 40

10

Profit (Tg/household)

Biomass (kg/ha)

250 200 150 100 50 0

5

30

35

40

10000. 9000. 8000. 7000. 6000. 5000. 4000. 3000. 2000. 1000. 0.

10 15 20 25 30 35 40 year

b) Biomass

25

c) Weight of an animal

350 300

20

year

year

a) The number of animals per household

15

5 10 15 20 25 30 35 40 year

d) Profit

Fig. 8. Scenario of a decrease in population size under cooperative management

downward trend. The profit for the mixed policy scenario increases, owing to the decrease in population. Overall, this policy option is beneficial for sustainable livestock growth. Reduction in the number of animals is useful for maintaining the net incomes in the long run, but it is also an important proactive management tool for reducing the damage from a natural disaster. In the case of cooperative management of a pasture affected by a disaster, even with 10× the level of risks, the number of animals remains at around the same level with some fluctuation. Although the profits are reduced by the disaster years, it still remains positive in the long run. Discussion This study presents a simulation model applicable to open-access grazing areas. Previous studies incorporating economic behaviors into pasture management focus on the decision making in rangeland management (Lubell et al., 2013), the optimal level of stocking rate and grazing period (Hart et al., 1988; Manley et al., 1997; Ritten et al., 2010), economic impacts of policies (Torell and Doll, 1991; Torell and Drummond, 1997), and bush encroachment control in rangeland management (Tanaka and Workman, 1988); however, there are few studies using integrated models for long-run simulations, especially for cases of open-access grasslands, such as Mongolia. The baseline simulations in this study show that the number of animals continues to increase and the amount of grass biomass and the average weight of animals decreases for business as usual. The results show that under the situations of business as usual, the number of animals

keeps increasing until the grasslands are degraded. If livestock farmers exceed the carrying capacity by grazing many animals, then the pasture cannot be maintained in the long run (Manley et al., 1997; Hart et al., 1988). However, as Behnke and Scoones (1993) argue, in a dryland where the intertemporal variation of vegetation is large, the concept of static carrying capacity does not hold. Herders take adaptive strategies according to changes in vegetation. The number of animals increases in a year with sufficient rain, while many animals are killed by frequent natural disasters. In terms of our target area in this study, a variation in the vegetation of grassland is less than a dryland pasture area (Fernandez-Gimenez and Allen-Diaz, 1999). The number of animals easily increases in a forest-steppe because the decreases in animals by disasters are less frequent in this region. The baseline simulation of this study is consistent with the estimation of the system dynamics model of Badarch and Ochirbat (2002), which predicts that the number of goats doubles in 20 yr owing to a 10% slaughtering rate in Mongolia. To maintain sustainable livestock development, public interventions to avoid further increases in the number of animals are necessary. Several measures to prevent the stocking rate exceeding its capacity are available; however, their effectiveness varies. Raising the tax rate on animals has a limited impact on the reduction in the rate of animal increase if the herder population in the area is constant. On the other hand, a decrease in the population leads to higher incomes per household and higher preservation of grass biomass. We may assume that an increase in the tax rate in an area would induce migration to another area with a lower tax rate. Consequently, a tax rate increase would have a positive effect on biomass if introduced in certain areas only.

Please cite this article as: Oniki, S., et al., Simulation of Pastoral Management in Mongolia: An Integrated System Dynamics Model, (2018), https:// doi.org/10.1016/j.rama.2018.02.003

S. Oniki et al. / Rangeland Ecology & Management xxx (2018) xxx–xxx

When compared with the current practice, the introduction of group pasture management will result in higher amounts of biomass and increase the weight of animals while also increasing household income. The finding that collaborative pasture management results in continuous positive profits in a long run is consistent with the study of Dunn et al. (2010), which shows that positive profits continue for 34 yr in a private rangeland while a high stocking rate remains. A more effective option is to implement a combination of policies for reduced population sizes and group pasture management. As Ostrom et al. (2007) and Roch et al. (2015) argue, a single policy may not be effective for a complex ecological and economic system. Since the problem is complex and multidimensional, land privatization is not a perfect remedy for preventing the overproduction of animals. We should note that neither complete free grazing nor controlled pasture management exist. Thus, the simulation may illustrate extreme cases. In the model, the population of herders is an exogenous parameter (i.e., the herders do not move to other areas or work even if their incomes decrease). The model simulation would be more realistic if population is considered as an endogenous variable. Implications This study shows the importance of policy mixes based on model impact simulations. No single policy can achieve both grassland conservation and the maintenance of herders’ incomes. Essentially, a combination of group pasture management and reducing herder population is effective for grassland conservation and income growth. In reality, if the number of animals increases in a region and the grassland is degraded, then herders may adapt to the changes in the environment by using more land-intensive technologies and even changing the organization of livestock production to increase the productivity of the land. Such possibilities are not included in the analysis using the system dynamics model in this study. Another limitation of this study is the assumption of homogeneity for natural conditions in the area. If the area has wide variations in natural conditions, then we would need to consider the spatial heterogeneity and evolutional history of grass communities, as argued by Milchunas et al. (1988). If animals are overcrowded, then herders’ movement will be restricted. The reduction of spatial and temporal variability in pasture utilization reduces the productivity of grazing (Fynn, 2015). We should also be aware that this study is limited to the foreststeppe areas, where grass biomass is relatively stable, and hence this simulation might not be directly applicable to the steppe or desert steppe areas. In the latter cases, intertemporal and spatial variation of the biomass is much larger; thus, the nonequilibrium rangeland ecological model may be more appropriate than the conventional rangeland model (Fernandez-Gimenez and Allen-Diaz, 1999). In summary, the simulations in this paper are mostly applicable to areas where climatic conditions are stable and few serious natural disasters occur. Acknowledgments This study is a part of the results of the research project the “Development of Sustainable Agro-Pastoral System in Northeast Asia” of the Japan International Research Center for Agricultural Sciences, Mongolian State University of Agriculture, and Inner Mongolia Agricultural University. We are grateful for the support received from professors, lecturers, researchers, and other staff of these institutes. We are also thankful to the Ministry of Agriculture and Bornuur soum for their support in data collection and analyses. Appendix A. Supplementary data Supplementary data to this article can be found online at https://doi. org/10.1016/j.rama.2018.02.003.

11

References Badarch, D., Ochirbat, B., 2002. On sustainbale development of cashmere production and goat population in Mongolia. In: Oka, H. (Ed.), CNEAS Monograph Serie. Center for Northeast Asian Studies, Tohoku University, Sendai, Japan. Baland, J.M., Platteau, J.P., 1996. Halting degradation of natural resources: is there a role for rural communities? Oxford University Press, Oxford, England, p. 423 Bald, J., Borja, A., Muxika, I., 2006. A system dynamics model for the management of the gooseneck barnacle (Pollicipes pollicipes) in the marine reserve of Gaztelugatxe (northern Spain). Ecological Modelling 194 (1–3), 306–315. Bardhan, P., Udry, C., 1999. Development microeconomics. Oxford University Press, Oxford, England. Baumol, W.J., Oates, W.E., 1988. The Theory of Environmental Policy. Cambridge University Press, Cambridge. Behnke Jr, R.H., Scoones, I., 1993. Rethinking range ecology: implication for rangeland management in Africa. In: Behnke, R.H., Scoones, I., Kerven, C. (Eds.), Range ecology at disequilibrium: new models of natural variability and pastoral adaptation in African savannas. Belhaven Press, London, England, pp. 1–30. Belcher, K.W., Boehm, M., 2003. Evaluating agroecosystem sustainability using an integrated model. In: Rapport, D.J., Lasley, W.L., Rolston, D.E., Nielsen, N.O., Qualset, C.O., Damania, A.B. (Eds.), Managing for healthy ecosystems. Lewis Publishers, Boca Raton, FL, USA, pp. 1209–1226. Bement, R.E., 1969. A stocking-rate guide for beef production on blue-grama range. Journal of Range Management 22, 83–86. Bromley, D.W. (Ed.), 1992. Making the commons work: theory, practice, and policy. Institute for Contemporary Studies, San Francisco, CA, USA. Cheng, Y., Tsubo, M., Ito, T.Y., Nishihara, E., Shinoda, M., 2011. Impact of rainfall variability and grazing pressure on plant diversity in Mongolian grasslands. Journal of Arid Environments 75 (5), 471–476. Chuluun, T., 2011. From vulnerability to sustainability: environment and human development. United Nations Development Programme, Ulaanbaatar, Mongolia. Coop, I.E., 1962. Liveweight-productivity relationships in sheep: liveweight and reproduction. New Zealand Journal of Agriculture Research 5 (3–4), 249–264. Dace, E., Muizniece, I., Blumberga, A., Kaczala, F., 2015. Searching for solutions to mitigate greenhouse gas emissions by agricultural policy decisions. Application of system dynamics modeling for the case of Latvia. Science Total Environment 527, 80–90. Davidsen, P.I., Asheim, L.J., 1993. A system dynamics approach to the structure and economy of fur farming and trading. System Dynamics Review 9 (3), 265–285. Dunn, B.H., Smart, A.J., Gates, R.N., Johnson, P.S., Beutler, M.K., Diersen, M.A., Janssen, L.L., 2010. Long-term production and profitability from grazing cattle in the northern mixed grass prairie. Rangeland Ecology & Management 63 (2), 233–242. Fernandez-Gimenez, M.E., Allen-Diaz, B., 1999. Testing a non-equilibrium model of rangeland vegetation dynamics in Mongolia. Journal of Applied Ecology 36 (6), 871–885. Ford, A., 2010. Modeling the Environment. 2nd ed. Island Press, Washington D.C., USA, pp. 189–208. Fynn, R.W.S., 2015. Functional resource heterogeneity increases livestock and rangeland productivity. Rangeland Ecology & Management 65 (4), 319–329. Gies, L., Agusdinata, D.B., Merwade, V., 2014. Drought adaptation policy development and assessment in East Africa using hydrologic and system dynamics modeling. Natural Hazards 74 (2), 789–813. Hardin, G., 1968. The tragedy of the commons. Science 162 (3859), 1243–1248. Hart, R.H., Samuel, M.J., Test, P.S., Smith, M.A., 1988. Cattle, vegetation, and economic responses to grazing systems and grazing pressure. Journal of Range Management 41 (4), 282–286. Lubell, M.N., Cutts, B.B., Roche, L.M., Hamilton, M., Derner, J.D., Kachergis, E., Tate, K.W., 2013. Conservation program participation and adaptive rangeland decision-making. Rangeland Ecology & Management 66 (6), 609–620. Lwiwski, T.C., Koper, N., Henderson, D.C., 2015. Stocking rates and vegetation structure, heterogeneity, and community in a northern mixed-grass prairie. Rangeland Ecology & Management 68 (4), 322–331. Manley, W.A., Hart, R.H., Samuel, M.J., Smith, M.A., Waggoner, J.W., 1997. Vegetation, cattle, and economic responses to grazing strategies and pressures. Journal of Range Management 50 (6), 638–646. Marshall, N.A., Stokes, C.J., 2014. Influencing adaptation processes on the Australian rangelands for social and ecological resilience. Ecology and Society 19 (2), 14. Meadows, D.L., 1970. Dynamics of commodity production cycles. Wright-Allen Press, Cambridge, MA, USA. Mearns, R., 1996. Community, collective action and common grazing: the case of postsocialist Mongolia. Journal of Development Studies 32 (3), 297–339. Meinzen-Dick, R., Knox, A., Place, F., Swallow, B., 2002. Innovation in natural resource management: the role of property rights and collective action in developing countries. John Hopkins University Press, Baltimore, MD, USA. Milchunas, D.G., Sala, O.E., Lauenroth, W.K., 1988. A generalized-model of the effects of grazing by large herbivores on grassland community structure. The American Naturalist 132 (1), 87–106. Minjigdorj, B., Sambuu, G., 1986. Mongol honinii genefond, tuuniig ashiglah ni (Genes of the Mongolian sheep and its utilization). National Publishing House, Ulaanbaatar. Moxnes, E., 1998. Not only the tragedy of the commons: misperceptions of bioeconomics. Management Science 44 (9), 1234–1248. Moxnes, E., 2004. Misperceptions of basic dynamics: the case of renewable resource management. System Dynamics Review 20 (2), 139–162. Moxnes, E., Danell, O., Gaare, E., Kumpula, J., 2001. Optimal strategies for the use of reindeer rangelands. Ecology Modelling 145 (2–3), 225–241. Moyobuu, L., Nyamaa, N., 1998. Mongol Ulsin Belcheeriin Daats Hureltseenii Talaar Uildsen Tootsoo (The basic estimation of the pastoral carrying capacity in Mongolian State). mimeo, Mongolian State University of Agriculture, Ulaanbaatar, Mongolia.

Please cite this article as: Oniki, S., et al., Simulation of Pastoral Management in Mongolia: An Integrated System Dynamics Model, (2018), https:// doi.org/10.1016/j.rama.2018.02.003

12

S. Oniki et al. / Rangeland Ecology & Management xxx (2018) xxx–xxx

Muya, S.M., Kanmeya, A.M., Muigai, A.W.T., Kariuki, A., Ngene, S.M., 2013. Using range condition assessment to optimize wildlife stocking in Tindress wildlife sanctuary, Nakuru District, Kenya. Rangeland Ecology & Management 66 (4), 410–418. Nakao, M., Konagaya, Y., Shinjilt (Eds.), 2010. Ecological migration: environmental policy in China. Peter Lang, Bern, Switzerland. National Statistical Office of Mongolia, 1996–2010. Mongolian Statistical Yearbook, 1996– 2010. National Statistical Office of Mongolia, Ulaanbaatar, Mongolia. National Statistical Office of Mongolia, 1996. Agriculture in Mongolia 1971−1995: a statistical profile. National Statistical Office of Mongolia, Ulaanbaatar, Mongolia. Oldeman, L.R., Hakkeling, R.T.A., Sombroeck, W.G., 1991. World map of the status of human-induced soil degradation: an explanatory note. International Soil Reference and Information Center (ISRIC), Washington, D. C, USA. Olson, M., 1965. The logic of collective action, public goods and the theory of groups. Harvard University Press, Cambridge, MA, USA, p. 186. Ostrom, E., 1990. Governing the commons: the evolution of institutions for collective action. Cambridge University Press, Cambridge, England, p. 280. Ostrom, E., Janssen, M.A., Anderies, J.M., 2007. Going beyond panaceas. Proceedings of the National Academy of Sciences of the United States of America 101 (39), 15176–15178. Peng, Q.Q., 1985. Xumuye Changyong Shuju Shouce (Livestock Industry Common Data Handbook), Neimenggu Renmin Chubanshe. Inner Mongolia People's Publishing House, Hohhot, Inner Mongolia. Ritten, J.P., Bastian, C.T., Frasier, W.M., 2010. Economically optimal stocking rates: a bioeconomic grazing model. Rangeland Ecology & Management 63 (4), 407–414. Roche, L.M., Schohr, T.K., Derner, J.D., Lubell, M.N., Cutts, B.B., Kachergis, E., Eviner, V.T., Tate, K.W., 2015. Sustaining working rangelands: insights from rancher decision making. Rangeland Ecology & Management 68, 383–389. Ruth, M., Hannon, B., 1997. Modeling dynamic economic systems. Springer, New York, NY, USA. Sasaki, T., Okayasu, T., Takeuchi, K., Jamsran, U., Jadambaa, S., 2005. Patterns of floristic composition under different grazing intensities in Bulgan, South Gobi, Mongolia. Grassland Science 51 (3), 235–242. Shindo, K., 2015. Grazing experiment to determine animal intake and affect on biomass and vegetation of forest steppe zone in Mongolia: the results and future target of the reseearch. Proceedings of Workshop on the GrassRISK Project 2015: Development of Resilient Agro-Pastoral Systems against the Risks of Extreme Weather Events in Arid Grasslands in Northeast Asia. Japan International Research Center for Agricultural Sciences, Tsukuba, Japan, pp. 49–55.

Shindo, K., Yamasaki, S., Shimoda, K., Toriyama, K., Hirano, A., Baasanjalbuu, B., 2013. Effects of grazing intensity and livestock species on vegetation change in grasslands in the forest-steppe zone of Mongolia. Development of a Sustainable Agro-Pastoral System in the Dry Areas of Northeast Asia, JIRCAS Working Report, No. 78. Japan International Research Center for Agricultural Sciences, Tsukuba, Japan, pp. 87–100. Smart, A.J., Derner, J.D., Hendrickson, J.R., Gillen, R.L., Dunn, B.H., Mousel, E.M., Johnson, P.S., Gates, R.N., Sedivec, K.K., Harmoney, K.R., Volesky, J.D., Olson, K.C., 2010. Effects of grazing pressure on efficiency of grazing on North American Great plains rangelands. Rangeland Ecology & Management 63 (4), 397–406. Suttle, J.M., Leynolds, S.G., Battello, C. (Eds.), 2005. Grasslands of the world. Food and Agriculture Organization of the United Nations, Rome, Italy. Tanaka, J.A., Workman, J.P., 1988. Economic optimum big sagebrush control for increasing crested wheatgrass production. Journal of Range Management 41 (2), 172–178. Tedeschi, L.O., Nicholson, C.F., Rich, E., 2011. Using system dynamics modelling approach to develop management tools for animal production with emphasis on small ruminants. Small Ruminant Research 98 (1–3), 102–110. Torell, L.A., Doll, J.P., 1991. Public land policy and the value of grazing permits. Western Journal of Agriculture Economy 16 (1), 174–184. Torell, L.A., Drummond, T.W., 1997. The economic impacts of increased grazing fees on Gila national forest grazing permittees. Journal of Range Management 50 (1), 94–105. Wilmer, H., Fernandez-Gimenez, M.E., 2015. Rethinking rancher decision-making: a grounded theory of ranching approaches to drought and succession management. Rangeland Journal 37 (5), 517–528. Yamasaki, S., Baasanjalbuu, B., Zolzaya, S., Kamiya, Y., Nonaka, K., 2014. Difference of a process of fermentation by the different period and mixing ratios of brewer’s grain and wheat bran or whet residue. Global climate change and its impact on food and energy security. Proceedings of the Eleventh International Dryland Development Conference, pp. 642–648. Yamasaki, S., Bayaraa, B., Sugirjargal, B., Gantumr, C., Kamiya, Y., Sed-ochir, Z., Nonaka, K., 2016. Potential use of brewers’ grain as an animal feed in and around population centers in Mongolia. Journal of Arid Land Studies 26 (2), 59–63. Zhao, H.L., Li, S.G., Zhang, T.H., Ohkuro, T., Zhou, R.L., 2004. Sheep gain and species diversity: in sandy grassland, Inner Mongolia. Journal of Range Management 57 (2), 187–190. Zhou, H.K., Tang, Y.H., Zhao, X.Q., Zhou, L., 2006. Long-term grazing alters species composition and biomass of a shrub meadow on the Qinghai-Tibet Plateau. Pakistan Journal of Botany 38 (4), 1055–1069.

Please cite this article as: Oniki, S., et al., Simulation of Pastoral Management in Mongolia: An Integrated System Dynamics Model, (2018), https:// doi.org/10.1016/j.rama.2018.02.003