as a PDF
Transcription
as a PDF
Shiraz University Faculty of Agriculture Ph.D. Dissertation In Agricultural Extension EXPLAINING THE EQUILIBRIUM BETWEEN LIVESTOCK AND RANGELAND USING FUZZY LOGIC By HOSSEIN AZADI NASRABAD Supervised by Dr. Mansour Shahvali Dr. Nezameddin Faghih December 2005 IN THE NAME OF GOD EXPLAINING THE EQUILIBRIUM BETWEEN LIVESTOCK AND RANGELAND USING FUZZY LOGIC BY HOSSEIN AZADI NASRABAD DISSERTATION SUBMITTED TO THE SCHOOL OF GRADUATE STUDIES IN PARTIAL FULFILMENT OF THE REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY (PH.D.) IN AGRICULTURAL EXTENSION SHIRAZ UNIVERSITY SHIRAZ ISLAMIC REPUBLIC OF IRAN EVALUATED AND APPROVED BY THE THESIS COMMITTEE AS: EXCELLENT MANSOUR SHAHVALI, Ph.D., …………………………………………… ASSOCIATE PROF., DEPT. OF AGRICULTURAL EXTENSION & EDUCATION, SHIRAZ UNIVERSITY. NEZAMEDDIN FAGHIH, Ph.D., ……...……………...…………………… Supervisor (Chairman) Supervisor PROF., DEPT. OF MANAGEMENT, SHIRAZ UNINIVERSITY. JAN VAN DEN BERG, Ph.D., ..…………………………………..………… Advisor ASSOCIATE PROF., DEPT. OF COMPUTER SCIENCE, ERASMUS UNIVERSITY ROTTERDAM. HOSSEIN MARZBAN, Ph.D., ……...……………………...……………….. Advisor ASSISSTANT PROF., DEPT. OF ECONOMICS, SHIRAZ UNIVERSITY. AHMAD KHTOONABADI, Ph.D., .…………………………...…………… Advisor ASSISSTANT PROF., DEPT. OF RURAL DEVELOPMENT, ISFAHAN UNIVERSITY OF TECHNOLOGY. MOHMMAD JAVAD ZAMIRI, Ph.D., …………………………………..... PROF., DEPT. OF ANIMAL SCIENCE, SHIRAZ UNIVERSITY. DECEMBER 2005 Advisor Dedicated to: Anahita for all her sympathies Acknowledgements All praises belong to Allah, The Most Gracious and Most Merciful, for His blessings that enabled me to accomplish this dissertation. This study is a product of many hands that directly and indirectly pushed me to gait the next steps. It is my pleasure to express my gratitude to individuals for their help and supports. My special word of thanks goes to Dr. Mansour Shahvali, my first supervisor, for his continuous supports not only during the course of this study but also in all other my academic activities. I am also grateful to Dr. Nezameddin Faghih, my second supervisor for his directions to set up this research, especially during the initial stages of planning the work. It would be a great honor for me to convey my sincere gratitude to Dr. Jan van den Berg, for his great contribution to solve my educational and residential problems during my stay in the Netherlands. My sincere gratitude goes to Dr. Hossein Marzban, Dr. Ahmad Khatoonabadi and Dr. Mohammad Javad Zamiri for sharing their knowledge and experiences to conduct this study. Many thanks go to Dr. Gholam Hossein Zamani; the head of Department of Agricultural Extension and Education at Shiraz University and my other respected teachers, Prof. Dr. Ezatollah Karami, Dr. Dariush Hayati and Mohammad Bagher Lari for all their supports during my university studies. I am also thankful to Mr. Ali Kheradmand, Mr. Asghar Sahranavard, Ms. Tooran Jezghani, Khalighzadeh, Adeli, and Paydar for all their helps. Special thanks go to my dearest classmate, Dr. Kiumars Zarafshani who is the best friend for me. Also, to my other classmates, especially Dr. Nozar Monfared and Dr. Ahmad Abedi. I am grateful to the pastoralists and the experts of various administrations in Fars province, especially Talati, Riahi, Haddadi, Mansoori and Mohseni. My greatest gratitude goes to my beloved wife, Anahita Aghaei for her supports, prayers, and unlimited patience during this long-term study. My special gratitude goes to Jozef Caluwaerts and my mother-in-law for all their continual supports. Finally, to my mother, father and family for all their endless supports. Hossein Azadi Nasrabad Shiraz University, Iran December 2005 i Abstract While there is no consensus on a definition, it is widely recognized that the concept of sustainability has economic, environmental and social dimensions. We used fuzzy logic as a well-suited tool to handle the vague, uncertain, and polymorphous concept of sustainability. For recognizing the major important indicators in defining sustainability in rangeland management, several semistructured interviews with an open-ended questionnaire were held in three different areas of the Fars province in Southwest Iran. Different groups of ‘experts’ were chosen by using the ‘socio-metric’ sampling method, and were interviewed. Pastoralists’ experts recognized that sustainability in rangeland management is a function of three major components (inputs) which are the Stocking Rate in a rangeland, the amount of Plantation Density per hectare, and the Number of Pastoralists who live in a rangeland where the output of the model is the Right Rate of Stocking. Based on pastoralists’ insights we developed a model called Equilibrium Assessment by Fuzzy Logic (EAFL) which provides a mechanism for assessing sustainability in rangeland management. The EAFL model exhibits five important characteristics. First, it permits the combination of various aspects of sustainability with different units of measurement. Second, it overcomes the difficulty of assessing certain attributes or indicators of sustainability without precise quantitative criteria. Third, it supports researcher with an easy to use and interpret. Fourth, considering the sequence "crisp input – fuzzifier – inference engine – defuzzifier – crisp output", it illustrates the uncertainity that exists in such a complex vague concept as sustainable rangeland management, and fifth, it also well adjusts to usual ambigious linguistic statements of individuals. To deal with the heterogeneity of experts’ knowledge, which should be considered either as a reality or necessity, a multi-fuzzy model was developed. In order to find the final output of the multi-fuzzy model, different ‘voting’ methods were applied. The mean method simply uses the arithmetic average of the primary outputs as the final output of the multi-fuzzy model. This final output represents an estimation of the Right Rate of Stocking. By harmonizing the primary outputs such that outliers get less emphasis, an unsupervised voting method calculating a weighted estimate of the Right Rate of Stocking was introduced. This harmonizing method is expected to provide a new useful tool for policymakers in order to deal with heterogeneity in experts’ opinions: it is especially useful in cases where little field data is available and one is forced to rely on experts’ knowledge only. By constructing the three fuzzy models based on the heterogeneous knowledge and using some harmonized methods, our study tried to show the multi-dimensional vaguenesses which generally exist in rangeland management, and solve the conflict that especially exists in economical and conservational views in the Iranian rangeland management. Finally, by comparing the estimated Right Rate of Stocking, which elicited from both experts' opinions and Matlab Fuzzytoolbox Editor, with its medium range, the models verified overgrazing in the three regions of the Fars province in Southwest Iran. ii Contents Acknowledgement ……………………………………………………………………… Abstract ………………………………………………………………………………….. Chapter One – Introduction …………………………………………………………… 1.1. Prelude …………………………………………………………………………... 1.2. Population Growth ……………………………………………………………… 1.3. Current challenges in rangeland management ………………………………… 1.4. Defining the problem …………………………………………………………… 1.5. Objectives ……………………………………………………………………….. i ii 1.5.1. General goal ………………………………………………………………… 1.5.2. Specific goals ……………………………………………………………….. 1.6. The structure of dissertation ……………………………………………………. Chapter Two - Population Growth: Consequences …………………………………. 2.1. Population growth ………………………………………………………………. 2.1.1. A general view ……………………………………………………………… 2.1.2. Three possible scenarios ……………………………………………………. 2.1.2.1. Low scenario …………………………………………………………… 2.1.2.2. Medium scenario ……………………………………………………….. 2.1.2.3. High scenario …………………………………………………………… 2.2. Consequences …………………………………………………………………… 2.2.1. A historical challenge ………………………………………………………. 2.2.2. Food security ………………………………………………………………... 2.2.3. Food consumption …………………………………………………………... 2.2.4. Agricultural research ………………………………………………………... 2.2.5. Biotechnology ………………………………………………………………. 2.2.6. Arable land ………………………………………………………………….. 2.2.7. Water scarcity ………………………………………………………………. 2.2.8. Forestry and fisheries ……………………………………………………….. 2.2.9. Rangelands …………………………………..……………………………... 2.3. Conclusion ………………………………………………………………………. Chapter Three - Sustainability: Basic Challenges ………………………………...… 3.1. Importance ……………………………………………………………………… 3.2. Definitions ……………………………………………………………………… 3.3. Dimensions ……………………………………………………………………… 3.4. Modeling problems ……………………………………………………………... 3.5. Conclusion ……………………………………………………………………… Chapter Four - Rangeland Management: Basic Challenges and Principles ………. 4.1. A review of literature …………………………………………………………… 4.2. Rangeland management: Art or science? ……………………………………… 4.3. Equilibrium and disequilibrium systems in rangeland management …………. 4.4. Current challenges in rangeland management ………………………………… 4.4.1. Overgrazing ………………………………………………………………… 4.4.2. Carrying capacity …………………………………………………………… 4.5. Basic principles in rangeland management ……………………………………. 4.6. Conclusion ……………………………………………………………………… 1 1 3 4 6 9 9 9 10 11 11 11 13 14 16 18 20 21 23 26 28 30 33 37 42 43 45 48 48 51 53 55 59 61 61 63 64 66 66 68 70 77 Chapter Five - Application of Fuzzy Logic in Sustainable Rangeland Management ……………………………………………………… 5.1. Fuzzy Logic: A shifting paradigm ……………………………………………... 82 82 iii 5.2. Foundations of fuzzy logic ……………………………………………………... 5.2.1. Crisp models ………………………………………………………………... 5.2.2. Boolean vs. Fuzzy ………………………………………………………….. 5.2.3. Towards soft computing ……………………………………………………. 5.2.4. Towards fuzzy sets ………………………………………………………….. 5.2.5. Operators on fuzzy sets ……………………………………………………... 5.2.6. Linguistic variables …………………………………………………………. 5.2.7. Knowledge representation by fuzzy IF-Then rules …………………………. 5.2.8. Architecture of fuzzy systems ………………………………………………. 5.2.9. Fuzzy reasoning …………………………………………………………….. 5.3. Theoretical frameworks ………………………………………………………… 5.3.1. Architecture the EAFL model ………………………………………………. 5.3.2. Architecture of multi-fuzzy model …………………………………………. Chapter Six - Research Method ……………………………………………………….. 6.1. The population of study ………………………………………………………… 6.2. The area of study ………………………………………………………………... 6.3. Research method ………………………………………………………………... 6.3.1. Multiple-case study …………………………………………………………. 6.4. Sampling method ………………………………………………………………... 6.5. Data collection and applied techniques ………………………………………... 6.5.1. Data analysis ………………………………………………………………... Chapter Seven - Fuzzy Analysis and Discussion ……………………………………. 7.1. Development the EAFL model ………………………………………………… 7.1.1. Determining the relevant input and output variables ……………………….. 7.1.2. Defining linguistic values …………………………………………………... 7.1.3. Constructing membership function …………………………………………. 7.1.4. Determining the fuzzy rules ………………………………………………… 7.1.5. Computing degree of membership of crisp inputs ………………………….. 7.1.6. Detemining approximate reasoning ………………………………………… 7.1.7. Computing crisp output (defuzzify) ………………………………………… 7.1.8. Assessing the model performance ………………………………………….. 7.2. Development the multi-fuzzy model …………………………………………... 7.2.1. Computing the crisp primary outputs ………………………………………. 7.2.2. Implementing voting ………………………………………………………... 7.2.2.1. Method 1: Calculating the mean of outputs ……………………………... 7.2.2.2. Method 2: Minimizing the sum of squared errors ……………………….. 7.2.2.3. Method 3: Minimizing an approximation of the sum of squared errors …... 7.2.2.4. Method 4: Harmonizing the primary outputs ……………………………. 7.2.3. Comparison of Method 1 and Method 4 ……………………………………. Chapter Eight - Summary and Conclusions ………………………………………….. 8.1. Summary ………………………………………………………………………… 8.2. Conclusions ……………………………………………………………………... Bibliography …………………………………………………………………………….. APPENDIX 1 ………………………………………………………………………….... APPENDIX 2 ………………………………………………………………………….... iv 85 85 86 87 89 89 90 91 92 92 94 94 96 100 100 103 106 108 109 110 111 112 112 112 112 113 114 116 117 118 119 121 124 128 128 130 130 131 134 139 139 141 145 165 166 Tables Table 2.1. The projection of world population (Medium scenario) 1950-2050. …………. Table 2.2. Water scarcity by country groups. ……………………………………………. Table 6.1. General information of the three regions of the study. ……………………….. Table 6.2. Some personal characteristics of 9 nominated experts at the first round of study…………………………………………………………………………… Table 7.1. Linguistic values used in the EAFL model. …………………………………... Table 7.2. The complete rules base (33 = 27) used to construct the overall experts’ knowledge base. …...…………………………………………………………. Table 7.3. Assessing the performance of the EAFL model by using real data. ………….. Table 7.4. Inputs, linguistic values and fuzzy range of each experts. ……………………. Table 7.5. Characteristics of the output (RRS) for three fuzzy models. ………………….. Table 7.6. Computing the outputs of the first model with 5 cases for each region. ……… Table 7.7. Computing the outputs of the second model with 5 cases for each region. …... Table 7.8. Computing the outputs of the third model with 5 cases for each region. …….. Table 7.9. Finding the final outputs by calculating the mean of primary outputs. …….… Table 7.10. Estimating the final output RRSf by calculating the sum of weighted outputs for separated regions according to Method 4. ……………………………….. v 16 41 105 109 113 115 119 122 124 125 126 127 128 133 Figures Fig. 2.1. World population: Three possible futures. ……………………………………... Fig. 2.2. World Population and Arable Land, 1700 – 1990. ……………………………... Fig. 5.1. Diagrammatic representation of the linguistic variable stocking rate in a rangeland having linguistic values low, medium, and high defined by a corresponding membership function. ………………………………………….. Fig. 5.2. Building blocks of a Fuzzy Inference System (FIS). …………………………… Fig. 5.3. Scheme of development of the EAFL model applying approximate reasoning to assess the Right Rate of Stocking (RRSp) based on the inputs values (I1p, I2p, and I3p). …...………………………………………………………………………… Fig. 5.4. Architecture of the multi-fuzzy model to deal with different experts’ knowledge. ……………………………………………………………………... Fig. 7.1. Membership functions for a) Stocking Rate, b) Plantation Density, and c) Number of Pastoralists. ………………………………………………………... Fig. 7.2. Linguistic values and fuzzification of crisp inputs for a) Stocking Rate, b) Plantation Density, and c) Number of Pastoralists …………………………….. Fig. 7.3. Graphical illustration of the EAFL model for approximate reasoning and defuzzification. ………………………………………………………………… Fig. 7.4. Comparison of the RRSf for the harmonized method and mean method. ………. vi 13 35 91 92 95 98 114 116 118 135 Chapter One Introduction 1.1. Prelude At the beginning of this century, it is important to care the scarcity of resources and food for everyone. This shortage has made a major challenge for policy-makers. Population growth on one hand and production diversification on the other has made a dilemma in natural resources management. At the ecological level, land scarcity is causing food scarcity for the ever-increasing population. Brown et al., (2000) believe that: • Resources are becoming scarce, • Natural species and forests are destroyed which also leads to destruction of wildlife and fisheries, and • Air pollution is causing contamination faster than it can be recovered. Scientists believe that the number one cause of environmental contamination is “over consumption”. In spite of major development in 1950s and 1960s, the recent socio-cultural and environmental crises have caused a major destruction in sustaining natural resources. This has led to the new development under the label of sustainability. Sustainable Development (SD) is nowadays the goal in words at least, of most politicians and decision makers (Rigby et al., 2001). Since publication of the Brundtland report in 1987 (WCED, 1987; p. 43), the concept of SD was defined as "development that meets the needs of 1 the present without compromising the ability of future generations to meet their own needs". Demands for natural resources are recently high among rural areas. Most people in developing countries are agrarians and pastorals. In 1988 some 65 percent of the population in what the World Bank classified as low-income countries were living in rural areas. The share of the labour force engaged in agriculture in these countries was a bit higher, and agriculture accounted for about 30 percent of GDP. In industrial countries, by contrast, agriculture accounted for 6 percent of the labour force and 2 percent of GDP (Dasgupta and Maler, 1995). For most part developing countries have biomass-based subsistence economies, in that rural people live on products obtained directly from plants and animals (Dasgupta, 1996). Studies in Central and West Africa have shown how vital natural resource products are to the lives of rural inhabitants (Falconer, 1990; Falconer and Arnold, 1989). Come what may, developing countries will remain largely rural economies for sometimes. Thus, it seems obvious that any management techniques in natural resources in developing countries should take into account the enormous importance of these countries’ natural resources base. Forty years of research on maintaining equilibrium between supply and demand on the natural resources among pastorals has failed to do so. This is because the current management strategies have taken a "black and white" logic in their attitudes. In the case of range or pastures, the only way to reduce degradation is to not use them, and this is unlikely to be the right approach. However, an optimal pattern of use seems appropriate for sustaining these resources. In order to reach an optimal pattern, we need to reconsider our range management views. Thus, an informed 2 management must define values of rangeland use by pastorals as well as values of natural resources as inputs in production. Of the many factors that affect natural resource management, such as property right, population growth, and discounting and access to markets, population growth gains the most important. 1.2. Population Growth Rapid population growth directly contributes to range and pasture degradation. Also, population growth may break down social norms and resource management systems, further contributing to natural resource degradation. The relationship between population growth and natural resource degradation is positive. Pastorals depend more on basic natural resources. While, population growth rate and their consumption level affect poverty and natural resource degradation, degradation affects the capacity of the pastorals (Pender, 1999). The future scenarios on population growth and its effects can be summarized as follows: • The world’s population is growing by 200,000 people a day. It is expected to nearly double by 2050, from 5.7 billion in 1994 to about 10 billion people (Peterson, 1998). Nearly, all the growth will occur in the developing world (UNFPA, 2005)1; • Absolute number of poor will remain quite high (FAO, 2003); • Land per capita availability will go down (PAI, 2005); • Environmental pollution likely to worsen (UNEP, 2000); and 1. Between 1980 and 2030, the population of low- and middle-income countries will be more than double - to 7.0 billion, compared with 1 billion for high-income countries. In the next 35 years, 2.5 billion people will be added to the current population of 6 billion (World Bank, 1998). 3 • Global climate change will have serious effect on agriculture and rural pastorals (Kautza and Gronski, 2003). Responding to the needs of a rapidly growing population can challenge a country’s ability to manage its natural resources on a sustainable basis. People may not be able to get access to safe water because more and more households, farms and factories are using increasing amounts of water. The air may become polluted as people crowd into cities, the number of cars increases, people use more and more energy, and economies continue to industrialize. Deforestation may occur as trees are cut to provide fuel for cooking, building materials, or land for grazing and agriculture. Desertification may occur as land that has been intensively farmed becomes depleted of its nutrients or eroded when trees whose roots systems once anchored the soil are gone (World Bank, 1998). Land degradation is one of the other consequences of mis-management of land and results frequently from a mismatch between land quality and land use (Reich et al., 2001). 1.3. Current challenges in rangeland management Literature on rangeland grazing management is being published at an accelerating rate, even though the use of rangelands for domestic stock production is increasingly questioned on conservation and sustainability grounds in the world (Walker and Hodgkinson, 2000). There are some references which indicate that the only way to reduce degradation is not to utilize natural resources, and this is unlikely to be the right approach. However, an alternative pattern of use seems to be possible where a balance is reached between grazing and conserving 4 these resources. In order to reach this pattern, we need to reconsider our range management views. An informed management must define values of pasture use by pastorals as well as values of natural resources as inputs in production. Demand for natural resources is recently high among rural areas. Most people in developing countries are agrarian and pastoralist. There are an estimated 190 million pastoralists in the world (NGO Forum for Food Sovereignty, 2002). The majority of them suffer from the effects of settlement, encroachment on their traditional rangelands, lack of infrastructure, hostile market mechanisms, and difficulties of marketing their products, forcing large numbers to abandon their rural livelihoods and seek employment in cities. Human interference keeps numbers at artificially high levels, e.g. by feeding imported fodder and household residues to animals that then have the energy to go to the last grasslands. “Without interference, livestock numbers respond to the laws of population dynamics” (NGO Forum for Food Sovereignty, 2002; p. 3). In this case, therefore, with increasing population growth (specially in developing countries like Iran), the increased number of livestock would be expected. This continuous overgrazing has changed the composition of the pasturage and is reflected in a decline in the animal quality. Thus, it seems obvious that any management techniques in natural resources in developing countries should take into account the enormous importance of these countries’ natural resources base. Forty years of research projects and activities on maintaining equilibrium between supply and demand on the natural resources among pastorals has failed to do so, because the projects have focused mostly on grazing management. Although grazing management is important, because this is where theory is put into 5 practice, however, successful grazing management will be based on the ability to accomplish three objectives: 1. To control what animal graze, 2. To control where they graze, and 3. To monitor the impact on both the environment and the animal (Walker, 1995). Governments and international donor agencies have over the years supported a variety of interventions aimed at improving the rangelands. These interventions include (i) sedentarization, i.e. the placing of pastoralists in permanent settlements and providing an alternative livelihoods such as irrigated agriculture; (ii) controlled grazing schemes, such activities established and enforced land management rules which are conducting in Iran, Kenya and Tanzania; and (iii) construction of boreholes to provide water for pastoralists and their livestock (Azadi et al., 2003; Lusigi and Acquay, 1999). 1.4. Defining the problem A claim is commonly made that the rangelands of the world are overgrazed and hence producing edible forage and animal produce at less than their potential. Globally, rangelands are also at risk from numerous pressures. The most important pressure arises from overgrazing of livestock. Livestock, therefore, have been a key factor in the SD in rangeland management. But what will their role be in the future and how should the science of rangeland management change to meet the challenges of the future? The recent literature on rangelands disequilibrium model calls into question any specific measures of carrying capacity, whether the range is stocked or unstocked, managed or mis-managed. Ideally, such 6 objections can be taken into account for any individual carrying capacity estimated by accepting that it has to be determined on a case - by - case basis in the field. Once one knows the size of the grazing and browsing animals, and once one knows the biomass production of the area, the pattern of range management, and so on, she/he can - so this argument goes - produce a site specific carrying capacity estimated for the range area under consideration. But, it cannot pack livestock into a given rangeland, without at some point deteriorating that range demonstrably. The fact get higher importance when we notice that biomass production, surely is going down on rangelands because carrying capacity has been exceeded for so long, even taking into account factors such as drought and climate changes (Hardesty et al., 1993). As a result, carrying capacity is the most important variable in range management (Walker, 1995). At a time, when the planet's limited carrying capacity seems increasingly obvious, the rationale and measures of rangelands carrying capacity are increasingly criticized. One of the elements of rangeland capacity is stocking rate. According to Roe (1997), if stocking rate is not close to the right level, then, regardless of other grazing management practices, employed objectives will not be met. Thus, a major problem facing range management is the range disequilibrium. This applies to a regular topic of books, articles and symposia, and a common justification for further research in many countries, including Iran (Conference on Sustainable Range Management, 2004; Iranian Nomadic Organization, 1992). It seems that under environmental conditions of great uncertainty, the notion of rangeland equilibrium would still be ambiguous and 7 confused. Moreover, since environmental conditions are highly uncertain for the dry rangelands of the world such as Iran, current understanding of rangeland equilibrium turns out to be all the more questionable. There is no workable, practical equation for rangeland management in general, and carrying capacity in particular, nor could there ever be. But fortunately, there is an alternative formulation for rangeland equilibrium, which is considerably more realistic if not more useful than even the conventional methods, named as fuzzy logic. Iran has approximately 90 million hectares of rangeland, 9.3 million hectares of which are considered in ‘good’ conditions while the remaining in ‘fair’ or ‘poor’ conditions. The country’s rangelands in a normal year produce around 10 million tons of dry matter (dm), of which 5.8 million tons may be available for grazing. The remaining amount is the minimum required for reproduction and soil conservation. The later amount of dm can support 38.5 million animal units (au) for duration of 8 months. At the moment there are 115.5 million of au in Iran and only 16.5 of them are fed from other sources including by agricultural products. The above figures prove that the rangelands are being utilized at three times more than their peak capacities in a non-drought year. This results in severe degradation of the rangelands and accelerates soil erosion. As the rangeland is considered by its users as "free resource" it is subject to heavy abuse, which further exacerbates the drought (FAO, 2004a; Iranian Nomadic Organization , 1992; Mesdaghi, 1995; UNCT, 2001). The purpose of this dissertation is to design a model for solving the mis-management of rangelands in Iran. Specifically, this study 8 discusses the application of fuzzy logic in rangeland management in the Fars province of Southwest Iran. 1.5. Objectives 1.5.1. General goal The above scenarios make it clear that the rangeland grazing management is a complex and confusing phenomenon. Thus, the main objective of this dissertation is “To Explain the equilibrium between livestock and rangeland by using fuzzy logic”. 1.5.2. Specific goals More specifically, we focus on the following goals: 1. Defining basic challenges and principles in rangeland management; 2. Selecting the indicators of Sustainable Rangeland Management (SRM) based on experts’ knowledge; 3. Explaining the applicability of fuzzy logic in SRM; 4. Estimating the Right Rate of Stocking (RRS) based on homogenous experts’ knowledge; 5. Estimating primary the RRS based on heterogeneous experts’ knowledge; 6. Dealing with heterogeneous experts’ knowledge to find the final the RRS; and 7. Evaluating the behaviour heterogeneous models. 9 of fuzzy homogenous and 1.6. The structure of dissertation In this dissertation, fuzzy logic is proposed as a systematic methodology for the assessment of sustainability in rangeland management. Based on this methodology, we develop several fuzzy systems, which use different experts’ knowledge to estimate carrying capacity of the Fars rangelands in Southwest Iran. A new model is developed, called Equilibrium Assessment by Fuzzy Logic (EAFL), which provides a mechanism for assessing sustainability in rangeland management in the studied areas. Furthermore, to deal with the diversity of different experts’ knowledge, we construct a multi-fuzzy model. By introducing several voting methods, we estimate the weights of each model. Finally, we evaluate the behaviour of the models. The dissertation is organized as follows: In Chapter 2, we review population growth and its consequences. Chapter 3 gives a discussion regarding sustainability and its major concerned issues. Chapter 4 explains basic challenges and principals in rangeland management, which cause (dis)equilibrium. In Chapter 5, we introduce fuzzy logic as a powerful tool to deal with these challenges. Designing theoretical frameworks in this chapter, we conduct a field study that is reported in Chapter 6. Findings and their fuzzy analyses are extensively explained in Chapter 7. Finally, Chapter 8 presents a summary and conclusions. 10 Chapter Two Population Growth: Consequences Mahatma Gandhi: "The world has enough for everybody’s need, but not for everybody’s greed" (Takle, 2001). 2.1. Population growth 2.1.1. A general view Overpopulation is the greatest problem of our age. In the next hour, 9,000 babies will be born. By this time, tomorrow, the world will have an additional 200,000 mouths to feed. Each baby will be a unique individual with the potential to live a full life and be a contributing world citizen. Unfortunately, most of these babies will be born into poverty and hunger. Many will die in infancy (Brown et al., 2000). The earth is increasing its population by 90 million people per year, and yet we still have approximately 6 billion people left to feed and to give shelter (Mitchell, 1998). Along with the increase in the population, there are also more people on Earth who are living longer lives. The global population boom has coincided with the improvement of health, and of productivity, around the world. On average, the human population today lives longer, eats better, produces more, and consumes more than at any other time period in the past. Agriculture feeds people, but will it be able to feed the expanding global population, especially with its exponential increase (Einstein, 1998). 11 Growth will not be uniform across the globe. In general, the population booms now underway in South Asia and southern Africa follow the reduction in infant mortality. According to the Population Reference Bureau, 98 percent of today's population growth is taking place in developing regions. Development and birth control, however, tend to control population growth, and population is now stabilized in the industrialized world. Asia will continue to be the largest, with 60 percent of the world's population. Asia alone will add almost 3 billion people in the 21st century. Africa will experience spectacular growth. Between now and 2100, its population will quadruple, and its percentage of the world's population will double from 13 percent to 25 percent. Only 45 years ago Africa's population was 40 percent of Europe's. Today they are equal. In 2100, Africa will be four times larger. In fact, Africa is potentially a time bomb. Its huge population growth will occur alongside the world's lowest standard of living, greatest poverty, highest illiteracy, poorest infrastructure, least industry, and shakiest history of stable governments. Latin America's percentage will remain constant at 8 percent, while North America's will drop from 5 percent today to 3 percent at the Year 2100. Europe's population will actually shrink in the next 100 years: its percentage of the world's population will fall from 13 percent now to 6 percent (van der Werff, 1998). One of the key variables determining future outcomes, the growth rate of world population, has been on the decline since the second half of the 1960s. The UN demographic assessment of 1996 has a variant projection indicating further deceleration, from 1.4 percent currently (1995-2000) to 1 percent in 2020 and to 0.4 percent by the middle of the 21st annually. However, the absolute increments in world 12 population are currently very large, about 80 million persons, over 90 percent of who are added in the developing countries. Such high annual increments may persist for another 15-20 years, but with declines in prospect for the longer-term future, falling to some 40 million (30 million in the new projections) by 2050. Demographic growth in sub-Saharan Africa will increasingly dominate the total additions to world population: it will account for one half of the world increment by 2050, compared with only one fifth currently (Alexandratos, 1998). The aboved-described projections may make individuals confused by various figures. But what will really be happened in the future? 2.1.2. Three possible scenarios Historically, as we come up to the new situations, especially by economic, scientists in different disciplines would align in different parties of prediction the world population. According to their estimations, there are three possible scenarios, which are presented by those who underestimate (low scenario), some who normal estimate (medium scenario) and the rest who overestimate (high scenario). They are shown in Fig. 2.1. Fig. 2.1. World population: Three possible futures. 13 2.1.2.1. Low scenario Many researchers expect the world's population to level off between ten and eleven billion people. They predicted that "world food production could grow significantly more slowly than the current rate, and there would still be enough food for 10 billion mouths by the time they come." They believe that the earth can provide all the food needed for the foreseeable future (Milne, 2002). So why are so many saying we must take powerful measures, like widespread abortion, to control world population? But the "present rate" was already declining, and the world now doubles about every 82 years. And more conservative scholars had pointed this out years ago. As the standard of living of a country increases, its doubling time also increases. Thus, the developed nations are close to stability now, and as less developed nations become more industrialized their population growth also slows. That is the basis on which many experts predict that the world population will stabilize at about ten to eleven billion people (Milne, 2002). The study reports that the most recent UN assessment of global population trends indicates a drastic slowdown in world population growth. The UN, for example, reset the 2010 population level of 7.2 billion people projected in 1995, in 1998 at 6.8 million, or about 400 million fewer people. This recalibration in population level is due in part to changes in the world population growth rate, which has fallen from 2.1 percent per year in the later half of the 1960's to 1.3 percent in the late 1990's. This growth rate is predicted to continue dropping over the next three decades, reaching 0.7 percent by 2030. According to the latest UN projections, the most likely scenario for population in 2050 will be around 8.9 billion, and will peak out slightly above 10 14 billion after 2200. By 2050 the global population growth rate is expected to have dropped as low as 0.3 percent. But population estimates are notoriously inexact, especially those that peer deep into the future. Even though the rate of growth is slowing, the two billion people below age 20 will be raising a lot of children over the next couple of decades. Contrary to popular belief, this group believes the world's population is not increasing exponentially. Indeed, the growth rate has fallen steadily since the late 1960’s and is now about 1.5 percent annually. The annual population increase peaked in the late 1980’s and has declined to 85 million/year. The increase will drop to 58 million/year in the second quarter of the 21st century and to 19 million/year by its end. It will be growing at an average of 1.1 percent a year up to 2030, compared to 1.7 percent annually over the past 30 years. At the same time, an ever-increasing share of the world's population is well-fed. As a result, the growth in world demand for agricultural products is expected to slow further, from an average 2.2 percent annually over the past 30 years to 1.5 percent per year until 2030. In developing countries, the slowdown will be more dramatic, from 3.7 percent for the past 30 years to an average of 2 percent until 2030. However, the developing countries with low to medium levels of consumption, accounting for about half of the population in developing countries, would see demand growth slowing only from 2.9 to 2.5 percent per year, and per caput consumption increasing (FAO, 2002). The number of hungry people in developing countries is expected to decline from 777 million today to about 440 million in 2030. This means, that the target of the World Food Summit in 1996, to reduce the number of hungry will be met by 2030 (EuropaWorld, 2003). 15 2.1.2.2. Medium scenario Interestingly, the number of births is expected to remain relatively constant at 135 million/year for the next four decades or more, whereas the number of deaths will rise steadily from the present 50 million/year (van der Werff, 1998). According to Dr. Norman Borlaugh, a Noble Peace, called the father of the "green revolution," the world will have to increase food production 50 percent in 30 years, just to feed the world at today's substandard level and double it to provide everyone with the quality and abundance of food enjoyed in America. The world scientist noted that world population stood at 1.6 billion people when he was born in 1914. In 1995 it stood at 5.7 billion. Borlaugh says, "We are adding 100 million (100,000,000) people each year, a billion per decade." But, he says, "That (doubling) will never happen. That will be impossible" (Serf Publishing Inc., 2001) (Table 2.1). Table 2.1. The projection of world population (Medium scenario) 1950-2050. Estimated population Percentage Major (millions) distribution Areas 1950 2000 2050 1950 2000 2050 Oceania 13 31 47 0.5 0.5 0.5 Northern America 172 314 438 6.8 5.2 4.7 Latin America and the Caribbean 167 519 806 6.6 8.6 8.6 Europe 548 727 603 21.8 12.0 6.5 Africa 221 794 2000 8.8 13.1 21.5 Asia 1399 3672 5428 55.5 60.6 58.2 World 2519 6057 9322 100 100 100 NOTE: Information for 2050 is from medium-fertility variant projections (UNPD, 2003). Future demand for livestock and dairy products can be met, but the consequences of increased production must be addressed. Production will shift away from extensive grazing systems towards more intensive and industrial methods. This could pose a threat to the 16 estimated 675 million rural poor whose livelihoods depend on livestock. Without special measures, the poor will find it harder to compete and may become marginalized, descending into still deeper poverty. If the policy environment is right, the future growth in demand for livestock products could provide an opportunity for poor families to generate additional income and employment." Environmental and health problems of industrial meat production (waste disposals, pollution, the spread of animal diseases, overuse of antibiotics) also need to be addressed (FAO, 2002). Climate change could increase the dependency of some developing countries on food imports. The overall effect of climate change on global food production by 2030 is likely to be small. Production will probably be boosted in developed countries. Hardest hit will be smallscale farmers in areas affected by drought, flooding, salt-water intrusion or sea surges. Some countries, mainly in Africa, are likely to become more vulnerable to food insecurity. With many marine stocks now fully exploited or overexploited, future fish supplies are likely to be constrained by resource limits. The share of capture fisheries in world production will continue to decline, and the contribution of aquaculture to world fish production will continue to grow. The capacity of the global fishing fleet should be brought to a level at which fish stocks can be harvested sustainably. Past policies have promoted the build-up of excess capacity and incited fishermen to increase the catch beyond sustainable levels. Policy makers must act to reverse this situation (FAO, 2002). 17 2.1.2.3. High scenario If world population continues to grow at the current rate, world population will be doubled from six billion to twelve billion in fifty years (Harris, 2001). Half of this growth will occur in just six countries; which are India, China, Pakistan, Nigeria, Bangladesh, and Indonesia. Each of these nations faces a steady shrinkage of grainland per person and thus risks heavy future dependence on grain imports. This raises two important questions (Larsen, 2002): Will these countries be able to afford to import large quantities of grain as land hunger increases? And will grain markets be able to meet their additional demands? In India, where one out of every four people is undernourished, 16 million people are added to the population each year. The grain area per person in India has shrunk steadily for several decades and is now below 0.10 hectares—less than half that in 1950 (EDC News, 2003). As land holdings are divided for inheritance with each succeeding generation, the 48 million farms that averaged 2.7 hectares each in 1960 were split into 105 million farms half that size in 1990, when India's grainland expansion peaked. The average Indian family, which now has three children, will be hard pressed to pass on viable parcels of land to future generations (Larsen, 2002). Pakistan, with five children per family, is growing even more rapidly. In 1988, Pakistan's National Commission on Agriculture was already linking farm fragmentation and a rising reliance on marginal lands to declining farm productivity in some areas. Since then, the country has grown from just over 100 million to almost 150 million. Its per person grain area is now less than 0.09 hectares (Larsen, 2002; Larsen, 2003). 18 In China, the grain area per person has also shrunk dramatically to a diminutive 0.07 hectares, down from 0.17 hectares in 1950. Shifting agricultural production to higher-value crops, like fruits and vegetables, and converting farms to forest for conservation accounts for some of the grainland contraction, along with losses to nonfarm uses such as buildings and roads (Coulter, 2002). Though the shrinkage of farmland available per person in China has slowed in concert with declining family size, this country—whose population of 1.3 billion is as large as the entire world's in 1850—is still expected to add 187 million people to its ranks in the next 50 years. The robustness of China's economy enables it to turn to world markets to import grain, but this does not guarantee that those markets can support massive additional demand without hefty price increases (Larsen, 2003). The scarcity of arable cropland in sub-Saharan Africa helps to explain the region's declining production per person in recent decades. Nigeria, for example, Africa's most populous country, has seen its population quadruple since 1950 while its grainland area doubled— effectively halving the grainland per person. In northern Nigeria, pastoralists and farmers fleeing the encroaching Sahara, which annually claims 350,000 hectares of land (about half the size of the U.S. state of Delaware), have increased demands on the already scarce land elsewhere in the country, sparking ethnic tensions (Coulter, 2003; Larsen, 2002; Larsen, 2003). Most of the 3 billion people to be added to world population in the next 50 years will be born in areas where land resources are scarce. If world grainland area stays the same as in 2000, the 9 billion people projected to inhabit the planet in 2050 would each be fed from less 19 than 0.07 hectares of grainland—an area smaller than what is available per person today in land-hungry countries like Bangladesh, Pakistan, and Afghanistan (EDC News, 2003). By 2050, India and Nigeria would cultivate 0.06 hectares of grainland for each person, less than one tenth the size of a soccer field. China, Pakistan, Bangladesh, and Ethiopia would drop even lower, to 0.040.05 hectares of grainland per person. Faring worse would be Egypt and Afghanistan with 0.02 hectares, as well as Yemen, the Democratic Republic of the Congo, and Uganda, with just 0.01 hectares. These numbers are in stark contrast to those of the less densely populated grain exporters, which may have upwards of 10 times as much grainland per person. For Americans, who live in a country with 0.21 hectares of highly productive grain land per person, surviving from such a small food production base is difficult to comprehend (UNDP, 2003). With most of the planet's arable land already under the plow and with additional cropland being paved over and built on each year, there is little chance that the world grain area will rebound. At the same time, the annual rise in cropland productivity of 2 percent from 1950 to 1990 has decreased to scarcely 1 percent since 1990, and may drop further in the years ahead. This slowing of productivity gains at a time when the land available per person is still shrinking underlines the urgency of slowing world population growth (Coulter, 2003). 2.2. Consequences Population growth will create some major threats to challenge. The threats come from poverty, agriculture perhaps not sustaining its 20 productivity growth of recent decades, environmental degradation, and scarcity of water for both agriculture and human health. 2.2.1. A historical challenge Since Robert Thomas Malthus published his anonymous ‘Essay on the Principle of Population’ in 1798, people have been disputing his contention: “population grows exponentially, but food supplies grow arithmetically”. This means that the graph of population curves upward, while the graph of food supply is straight. Malthus said, shortages of food would cause chaos and famine (Office of News and Public Affairs, 1999). Malthus assumed that food supplies would always limit population growth. But in the two hundred years since he wrote, this has not been the case. The pronouncement was fearsome enough to earn economics this splendid moniker: the "dismal science." But it wasn't just economists who rebelled. Karl Marx also denounced Malthus. By one means or another farmers and agricultural scientists have always found a way to increase farm production to keep up with population growth. But we have yet to find efficient ways to get food from where it is produced to where it is needed most (Milne, 2002). Afterwards, Paul Erhlich, in his 1968 book; The Population Explosion, announced the approaching food crisis; “...Then, in 196566 came the first dramatic blow...mankind suffered a shocking defeat in...the war on hunger” (Dean, No date p.1). In 1966, while the population of the world increased by some 70 million people, there was no compensatory increase in food production. He continues by laying out likely scenarios of the world being rocked by food rebellions that will lead to nuclear war and the devastation of the planet (Milne, 2002). 21 In 1989, Erhlich wrote another book, The Population Explosion. Doom was again close: "In 1988, for the first time since World War II, the United States consumed more grain than it grew...only the presence of large carryover stocks prevented a serious food crisis. It is not clear how easy it will be to restore those stocks." Fortunately, Erhlich was wrong. In 1968, he quotes Louis H. Bean approvingly: "My examination of the trend of India's grain production over the last eighteen years leads me to the conclusion that the present (1967-68) production...is at a maximum level." But in seven years, India increased its grain production by nearly 26 percent! By 1992, it had increased it 112 percent! By 1990, world grain production, again came up by 50 percent from 1988! And it has continued to increase to the present (Williams, 2000). In the book of Genesis, Adam and Eve were given the command to multiply and fill the earth. In Genesis, Noah is given the same charge. We must consider the rest of the creation as we determine if we have yet fulfilled that command. But world population is not the problem (Milne, 2002). We share the planet with 5.7 billion people. If one could stand all the people in the world, men, women and children two feet apart, how much of the world would they take up? Famines are the exception in most countries, and even then absolute lack of food is usually not the problem. In a Scientific American article on world population one author says: "Food surpluses exist in many nations, and even when famines do occur the cause is much less the absence of food than its mal-distribution which is often accentuated by politics and civil war, as in the Sudan." This passing comment touches on the real problem. Most famines in the last twenty 22 years are a direct result of internal wars in African nations (Milne, 1995). Whether in Ethiopia, Sudan, or Somalia, the devastating famines and the hopeless faces of dying children we have all seen on TV are the result of politics. As one segment of the population wars against another, starvation is often a political weapon. And in each of the famine-torn countries of Africa one can show that it has been disrupted distribution more than low food production that has caused people to starve to death (Milne, 2002). 2.2.2. Food security1 One of the most important questions facing most of societies today is, how to produce enough food to feed the increasing human population on this planet (Kanwar, 2003), which is discussed as the main goal of development (Seers, 1982). Few who are alive today remember the "great depression" and "dust bowl" of the 1930's or the food ration of World War II in the 40's. American's, Briton's and other citizens of the highly developed countries have enjoyed an almost unprecedented abundance of the earth's blessings for over half a century. How long can it continue? World population has recently passed the sixth billion, however, the number of hungry people is still growing who are stimated approximately two billion people (Raven, 2004). FAO discussed concerning the future, a number of projection studies have addressed and largely answered in the positive the issue whether the resource base of world agriculture, including its land component, can continue 1 . The UN Food and Agriculture Organization defines food security as a "state of affairs where all people at all times have access to safe and nutritious food to maintain a healthy and active life" (Blundon, 2001). 23 to evolve in a flexible and adaptable manner as it did in the past, and also whether it can continue to exert downward pressure on the real price of food (see for example Pinstrup-Andersen, and Pandya-Lorch, 1999). The largely positive answers mean essentially that for the world as a whole there is enough, or more than enough, food production to meet the growth of effective demand, but potentially (FAO, 2004b). Since the mid-1980s, the upward trend in annual per capita food production appears to be levelling off. Per capita world food availability grew by 6 percent in the 1960s and 1970s, then by 4.6 percent in the 1980s. Per capita crop production has been in decline since 1985. Per capita grain production expanded by 40 percent between 1950 and 1984 but has declined after this period. Some economists argue that this decline is due more to economic policies and low grain prices than to natural resource limits. The issue is important, since grain provides more than half of all calories consumed by people, directly or in (PAI, 2005). These prospects, particularly the demographic ones, are somewhat different from those used some five years ago to produce FAO’s assessment of world food and agriculture prospects to 2010, with particular reference to the developing countries, in the study "World Agriculture: Towards 2010"(Alexandratos, 1998) and subsequent modifications used in the technical documentation of the World Food Summit of 1996. However, the essence of our findings as concerns key variables of food security at the level of large country groups and the world, as a whole, remains largely valid. The main findings, including selected preliminary findings from ongoing work to update 24 the study and extend the time horizon to 2015 and 2030, are summarized below: The incidence of undernutrition in the developing countries may decline in relative terms (from 21 percent to 12 percent of the population) but, given population growth, there will be only modest declines in the numbers undernourished. The current level of over 800 million persons is expected to decline to about 680 million by 2010. The end result of the detailed projections (for individual countries and crops) indicates that the growth of the average yields of the developing countries (other than China) will be slower than in the past, 1.5 percent (from 1.9 tons/ha in 88/90 to 2.6 tons/ha in 2010), compared with 2.2 percent in the preceding 20 years (average yield of wheat, rice paddy and coarse grains). Nine years into the projection period (1989-98), the average cereal yield grew as predicted at 1.5 percent, though rice yield grew by less than predicted, that of maize by more and that of wheat as predicted. Continued growth of average yields, even at the lower rates projected here compared with the past, will not come about without effort. Growth in average yields will depend crucially on policies that attach high priority to efforts at agricultural research and technology development and diffusion, as well as on a more active role of the state in the areas of infrastructure, education and the creation of conditions for markets to work (Alexandratos, 1998). The prospect that the production growth rate in the exporting countries needs to be lower than in the past does not in itself guarantee that it is a feasible proposition. In particular, environmental concerns related to intensive agriculture in the high-income countries (nitrate pollution, soil erosion, perceived risks from genetically modified organisms, 25 etc.) may contribute to slow down the rate at which progress may be made in achieving the required yield increases. In fact, the rate of growth in agricultural production is declining; world grain reserves have fallen to record lows; the demand for imported grain is increasing; and commitments of aid to agricultural development have decreased. This against a backdrop of expanding world population, intensifying demands on agricultural resources, and a growing recognition that the agri-food system is not sustainable (The Online Newsletter of the Bahá'í International Community, 1996). 2.2.3. Food consumption By 2050, some 4.2 billion people may not have their daily basic needs met (Council for Biotechnology Information, 2004). All the primary sources of food crops, livestock and seafood depend on resources that are renewable but finite. In each case, limits now coming into view undermine the prospects for meeting future food demand (PAI, 2005). Concurrent with a decreasing population growth rate, individual food consumption rates (measured as Kcal/person/day) will continue to raise in developing countries. Citing the latest FAO assessment of under nourishment, the study reports that the percent of the world's undernourished has been dropping since the late 1960s. Projections of food consumption will continue to rise in developing countries over the next 30 years, moving from an average of 2626 kcal in the 1990s to nearly 3000 kcal in 2015. The average daily consumption rate in developing countries is expected to exceed 3000 kcal by 2030. By 2015, the report estimates, 6 percent of the world population (412 million people) will still live in countries with very low food consumption levels (under 2200 kcal) (FAO, 2000a). 26 The per-person food availability of the developing countries as a whole will continue to increase from 2580 Kcal/day (in 1994/96) to about 2750 Kcal/day by 2010. However, there will be only very modest gains in the currently very low average food availability of sub-Saharan Africa, while South Asia may still be in a middling position by 2010. The other developing regions, already starting from better levels now, are expected to be close, or above, 3000 Kcal/day (Alexandratos, 1999). One does not need sophisticated analytics to prove this point: any country starting with per-person food supplies of 2000 Kcal/day (and some countries start with less) and a population growth rate of 2.5 percent-3.0 percent would need a growth rate of aggregate food demand of about 5 percent in 15 years if, by 2010, it were to have 2700 Kcal/day, a level usually associated with significantly reduced undernutrition (provided inequality of distribution is not too high) (Alexandratos, 1998). Obviously, this kind of growth rates of aggregate demand for food can only occur in countries with "Asiantiger" rates of economic growth sustained over decades. Few of today’s poorest countries with very low food consumption levels face such prospects. As noted, the recent crisis that hit several economies of East and South-East Asia will also take its toll. The rapid pace of progress of the recent past, particularly in diet diversification towards livestock products, is being interrupted and some countries (e.g. Indonesia) are suffering outright reversals (FAO, 1999). At the end of the 1980s, about 251 million people lived in 14 countries where average per capita food supply was less than 2,100 calories per day. Another 1.4 billion people lived in 31 countries in which daily per capita availability was between 2,100 and 2,400 calories. In most 27 of these food-scarce countries, per capita production is declining. The averages also hide wide disparities within countries. However, the World Bank and the FAO are optimistic that global food supplies will be adequate and prices low at least through 2010, but analysts with both organizations assume that growth in yields and production will continue along current trend lines. A recent independent analysis that also considered potential environmental and socio-economic constraints predicted that without increased investment in agricultural research and technology, the next two decades will see food production shortages and accelerated environmental degradation (PAI, 2005). We simply cannot say for certain whether current or future populations can be fed sustainably. No single ingredient can provide such a guarantee. We know, however, one thing is the most important: how we can increase the yield of agricultural production by research? 2.2.4. Agricultural research One way for the over-population of today and tomorrow to live in harmony in regards to nourishment provided by the environment is to be able to intensify agricultural yields. With a projected population of 10 billion people, an increase of global average grain yields from 2 to 5 tons of grain per hectare would ensure a per capita diet of 6,000 calories and would save a land area twice the size of Alaska (Waggoner, 1994). Most of the world’s increased output is no longer a result of expansion of area used in agriculture, but resulting from the intensification of production on existing agricultural areas (Einstein, 1998). The "Green Revolution" of Norman Borlaugh's day only served to delay the growing crisis (Serf Publishing Inc., 2001). 28 In fact, some believe that agricultural researches can counteract the population time bomb. Perhaps an extension agent will bring a more productive variety of rice or demonstrate a more efficient method of irrigation. Mark Rosegrant, Senior Research Fellow at the Washington-based International Food Policy Research Institute and co-author of a study of agricultural production in India believes that investment in agricultural research and extension is and will continue to be the driving force in India's increased ability to feed their growing population. He also discussed, increased family incomes by growth in agricultural production, together with improved education, results in a reduction in population growth. This double benefit of agricultural research will alleviate misery today and gradually ease the world’s population woes (Rao, 1999). FAO (2000a) reports that there is a need for continued support of agricultural research and policies in developing countries. The report states that by 2030 three-quarters of the projected world crop production will occur in developing countries compared to just over half of world production in the early 1960s. Most of these future increases in crop productivity will come from a further intensification of crop production. The bulk of the increases in production will come from increasing plant yield and through more intensive land use (e.g., multicropping or high cropping intensities). These projections and complex challenges facing the world’s future food supply are prompting international food and agricultural experts and policymakers - including the U.N. Food and Agriculture Organization and the World Health Organization - to call plant biotechnology a critical tool to help feed a growing population in the 21st century (Council for Biotechnology Information, 2002). 29 The Green Revolution’s success in fending off starvation even as Asia and Latin America’s population doubled, from less than 2 billion to nearly 4 billion people, was a remarkable feat. Millions of human beings would not be able to survive today without the key innovations that launched the revolution. Foremost among these were advanced techniques of cross breeding that allowed development of rice, wheat and corn strains with increasingly higher yields per hectare. With sufficient access to irrigation water, fertilizers and pest controls, farmers could gain higher yields and, often, multiple crops in the same year, all with less labour. But there have been tradeoffs. Some Green Revolution technologies accelerate soil erosion, often beyond the thresholds of how much soil loss the land can tolerate without losing productivity. Fertilizers and pesticides have polluted groundwater supplies, while crop pests have developed resistance to common pesticides. Irrigated land is being abandoned as soils become waterlogged or contaminated by salt. Small-scale farmers, many of them women, have been pushed from ownership to tenancy, or off the land altogether, because the expenses and economies of scale required by the Green Revolution favour large farms and affluent farmers (PAI, 2005). 2.2.5. Biotechnology According to the Consultative Group on International Agricultural Research (CGIAR), world crop productivity could increase by as much as 25 percent through the use of biotechnology to grow plants that resist pests and diseases, tolerate harsh growing conditions and delay ripening to reduce spoilage. Biotechnology also offers the possibility for scientists to design “farming systems that are 30 responsive to local needs and reflect sustainability requirements,” said Calestous Juma, director of the Science, Technology and Innovation Program at the Center for International Development and senior research associate at the Belfer Center for Science and International Affairs, both at Harvard University (Council for Biotechnology Information, 2004). Scientists are developing crops that can resist against diseases, pests, viruses, bacteria and fungi, all of which reduce global production by more than 35 percent at a cost estimated at more than $200 billion a year (Council for Biotechnology Information, 2004). Nowadays, scientists are busy with developing such crops, which can tolerate extreme conditions, i.e. drought, flood and harsh soil. For instance, researchers are working on a rice that can survive long periods under water (Shah and Strong, 1999) as well as rice and corn that can tolerate aluminium in soil (Council for Biotechnology Information, 2004). A tomato plant has been developed to grow in salty water that is 50 times higher in salt content than conventional plants can tolerate and nearly half as salty as seawater (Owens, 2001). About a third of the world’s irrigated land has become useless to farmers because of high levels of accumulated salt (Council for Biotechnology Information, 2004). Biotech crops could significantly reduce malnutrition, which still affects more than 800 million people worldwide, and would be especially valuable for poor farmers working marginal lands in sub-Saharan Africa (UNDP, 2001). While the Green Revolution kept mass starvation at bay and saw global cereal production double as a result of improved crop varieties, fertilizers, pesticides and irrigation, its benefits bypassed such regions as sub-Saharan Africa. The new hybrids needed irrigation and 31 chemical inputs that farmers there couldn’t afford (Council for Biotechnology Information, 2002). In contrast, the benefits of biotechnology are passed on through a seed or plant cutting, so that farmers anywhere around the world can easily adopt the technology. That’s why biotechnology is particularly attractive to scientists and rural development experts in poor countries where most of the people farm for a living (Owens, 2001). Biotech crops are “tailor-made for Africa’s farmers, because the new technology is packaged in the seed, which all farmers know how to handle," said Florence Wambugu, a Kenyan plant scientist who helped develop a virus-resistant sweet potato (Council for Biotechnology Information, 2004). Agreeing with Wambugu, the International Society of African Scientists issued a statement in October 2001 calling plant biotechnology a major opportunity to enhance the production of food crops (ISAS, 2001). Despite, technology has been a viable part of higher productivity in agriculture and innovations such as tractors, seeds, chemicals, irrigation measures, fertilizers, pesticides, and genetic engineering have played a major part in raising yields, however, is technology the key to ensuring sustainable agriculture for a growing population? (Einstein, 1998). Many countries have tripled or even quadrupled the amount of grain they produce. Unfortunately, yields have been decreasing while population continues to increase. Grain yields per hectare have been slowing since 1990, rising only 3 percent from 1990-1996 or 0.5 percent per year. This does not keep pace with population growth which is at 1.6 percent per year (Brown, 1997). 32 Along with population growth, there is a growing demand for a more calorie-filled diet, especially with the unprecedented rise in affluence in Asia. Meat is becoming the food of choice rather than low calorie wheat or vegetables. Since it takes more grain and water to produce animal protein than vegetable protein, added pressure is placed on the environment. From 1990 to 1995, China’s grain consumption increased by 40 million tons; 33 million tons were consumed as animal feeds (Brown, 1997). As economies grow, especially in developing countries, consumption rates of resources rise in parallel. For these countries, a choice needs to be made between slowing population growth or sacrificing any hope of dietary improvement in order to lessen the pressure that agriculture creates on environments (UNFPA, 2003). 2.2.6. Arable land Another important question is: how much crop land would be needed to feed the growing population and what is the potential to further expand land area for food grain production? Increases in food production will have to come from existing agricultural land (UN, 2000). Arable land could in theory be increased by 40 percent, or 2 billion hectares, but most of the uncultivated land is marginal, with poor soils and either not enough rainfall or too much. Bringing it into production would require costly irrigation and watermanagement systems and large-scale measures to enrich the soil. Much of this land is now under forest, and clearing it would have unforeseeable consequences for erosion, degradation and local climate change, among others (UNFPA, 2003). 33 As population grows, subsistence farmers without access to new land are forced to intensify production to feed their families. This can be positive, if farmers can make the yield improvements last. Often, however, the reality is quite different. As the National Research Council notes, "fallow periods are often shortened to the point where the land becomes so badly degraded that it is virtually useless for any agricultural activity" (National Research Council, 1993). The practice of leaving some land fallow was once universal in agriculture, but in those areas where fallowing remains it is disappearing. When farmers lack access to information and technology—especially fertilizers—the abandonment of fallowing is unsustainable and a contributor to declining yields (PAI, 2005). Over the past three centuries, world population has increased eightfold while the amount of arable land has increased only fivefold. More intensive use of arable land has allowed food production to keep pace with population growth despite the slower expansion of arable land. There is limited potential, however, to expand arable land much further. Continued population growth could result in unsustainable demands on the earth's agricultural land and water resources in the coming decades (Fig. 2.2). 34 Fig. 2.2. World Population and Arable Land, 1700 – 1990 (Richards, 1990). Of the total of 13 billion hectares of land area on Earth, cropland accounts for 11 percent, rangeland 27 per cent, forested land 32 percent, and urban lands 9 per cent. Most of the remaining 21 percent is unsuitable for crops, pasture, and/or forests because the soil is too infertile or shallow to support plant growth, or the climate and region are too cold, dry, steep, stony, or wet (Richards, 1990). In 1960, when the world population numbered only 3 billion, approximately 0.5 hectare of cropland per capita was available, the minimum area considered essential for the production of a diverse, healthy, nutritious diet of plant and animal products like that enjoyed widely in the United States and Europe. But as the human population continues to increase and expand its economic activity and related artifacts, including transport systems and urban structures, vital cropland is being covered and lost from production (Pimentel and Wilson, 2005). 35 The decline of per-capita cropland is aggravated by the degradation of soils. Throughout the world, current erosion rates are higher than ever. According to a study for the International Food Policy Research Institute, each year an estimated 10 million hectares of cropland worldwide are abandoned due to soil erosion and diminished production caused by erosion. Another 10 million hectares are critically damaged each year by salinization, in large part as a result of irrigation and/or improper drainage methods. This loss amounts to more than 1.3 percent of total cropland annually. Most of the additional cropland needed to replace yearly losses comes from the world's forest areas. The urgent need to increase crop production accounts for more than 60 percent of the massive deforestation now occurring worldwide (Pimentel et al., 1996). Land expansion will continue to be a significant factor in the growth of agriculture in those developing regions where the potential for expansion exists (many countries in sub-Saharan Africa and South America) and the prevailing farming systems and more general demographic and socio-economic conditions favour land expansion (FAO, 1999). It is estimated that the developing countries outside China have some 2.5 billion ha of land of varying qualities, which has potential for growing rainfed crops at yields above an "acceptable" minimum level. Of this land, some 720 million ha (plus another 36 million ha of desert land reclaimed through irrigation) are already in cultivation in the developing countries outside China (arable land and land in permanent crops). Most of the remaining 1.8 billion ha is in Latin America and sub-Saharan Africa (FAO, 2003). At the other extreme, there is virtually no spare land available for agricultural expansion in South Asia and the Near East/North Africa region. Even 36 within the relatively land-abundant regions, there is great diversity among countries and sub-regions as concerns land availability per person, both quantity and quality. For example, in sub-Saharan Africa land is scarce in East Africa and relatively abundant in the Central Africa. Land expansion may add some 90 million ha to the above estimates of cultivated land of the developing countries (other than China). Such expansion will account for about 20 percent of the increase in their aggregate crop production (Alexandratos, 1998). The intensification of agriculture, especially under irrigated conditions, has brought new environmental problems including soil erosion, land degradation and decreased water quality. Intensive agricultural production systems were introduced in 1960’s with the advances in improved crop varieties, mechanization, and increased availability of pesticide and fertilizers. More recent experiences, in the developed countries, especially Europe and USA, have shown that modern and intensive agricultural production systems have increased land degradation and water contamination (Kanwar, 2003). 2.2.7. Water scarcity Another parallel question facing the society is, how much water would be needed to produce enough food to feed the increasing population in the world? Answer to this question is also not easy. Increased population rates have added more than 4.4 billion people on earth between 1900 and 2000, and average food production has kept pace with the increases in population. Also, between 1900 and 2000, irrigated area has increased from about 50 million hectares to 250 million hectares. Agricultural water use continues to make 85 percent of all consumptive use on a global basis (Kanwar, 2003). 37 As a country struggles to feed its people, on one hand, vital resources such as water for irrigation are dwindling (UNFPA, 2003). Scientists are finding ways to replenish groundwater aquifers and use irrigation in less wasteful ways that will not reduce food production (Rao, 1999). On the other hand, conflicts within countries are also of mounting concern to national governments (UNFPA, 2003). Among all contaminations, water pollution is a major threat to maintaining ample fresh water resources. Although considerable water pollution has been documented in developed nations like the United States, the problem is of greatest concern in countries where water regulations are not rigorously enforced or do not exist. This is common in most developing countries, which (according to the World Health Organization) discharge 95 percent of untreated urban sewage directly into surface waters. For instance, of India's 3,119 towns and cities, only 209 have even partial sewage treatment facilities, and a mere eight possess full facilities. Downstream, the polluted water is used for drinking, bathing, and washing (Pimentel and Wilson, 2005). Also, of all the environmental trends, water shortages may be the biggest. The report predicts, "By 2050, fully two-thirds of the population could be living in regions with chronic, widespread shortages of water." Agriculture must have adequate supplies of water if we are to meet future food needs, and those needs are huge (Truelsen, 2003). "Water scarcity is now the single greatest threat to human health, the environment, and the global food supply," said David Seckler, director general of the Water Institute and an author of the study with Randolph Barker and Upali Amarasinghe. "It also threatens global peace as countries in Asia and the Middle East seek to cope with shortages."…"Water scarcity is already a major 38 destabilizing force within countries because different sectors of the economy are vying for scarce water resources," said Seckler (Wilson, 1999). Within the next 25 years, there is great potential for more water conflict not just within countries but also between them. Historically, Egypt has threatened to go to war to protect its water supplies if necessary. And just last week, President Gaddafi of Libya warned that, “the next Middle East war would be over dwindling water supplies” (Mesbahi, 2004). The study divides the countries into four categories (Table 2.2). The first category includes those countries that are most water scarce and in 2025 will not have enough water to maintain 1990 levels of per capita food production from irrigated agriculture and meet industry, household, and environmental needs. The countries, defined as facing "absolute water scarcity," include 17 countries in the Middle East, South Africa, and the dryer regions of western and southern India and northern China, which account for more than 1 billion people today and are projected to account for as many as 1.8 billion in 2025. The study notes that while India and China will not have major water problems on average, there will be massive regional variations in water availability. The second category includes countries that have sufficient potential water resources to meet projected 2025 requirements, but will have to more than double their efforts to extract water to do so. Twenty-four countries, mainly in Sub-Saharan Africa, are defined as extremely water scarce and include some 348 million people today and are projected to include some 894 million in 2025. Because it will be extremely difficult for these countries to find the financial resources to build enough water development projects, such as dams and irrigation 39 systems, they are classified as having "economic water scarcity" (Wilson, 1999). The remaining countries of the world are in categories three and four and include North America and Europe. For these countries, there will be substantially less pressure on water supplies with moderate needs to increase water development efforts. The single greatest impact of water scarcity will be on the food supplies of the poor. To meet the world food supplies in 2025, the study provides two scenarios a "business as usual scenario" where no increases in irrigation efficiency are foreseen, and a scenario where irrigation efficiency is dramatically increased. Under the business-asusual scenario, 60 percent more water will be required for irrigation to meet the world food supplies in 2025. Even if irrigation efficiency is greatly increased, between 13 and 17 percent more water will be needed and still 2.7 billion people will remain short of water. The study uses the United Nations "medium" projection for population growth (Wilson, 1999). Many countries, including China, India, Iran, Pakistan, Mexico and nearly all of the countries of the Middle East and North Africa, have literally been having a free ride over the past two or three decades by rapidly depleting their groundwater resources, said Seckler. This could have catastrophic results in terms of limiting their ability to produce enough food to feed their populations (Table 2.2). 40 Table 2.2.Water scarcity by country groups (Wilson, 1999). Category 1 Afghanistan Egypt Iran Iraq Israel Jordan Kuwait Libya Oman Pakistan Saudi Arabia Singapore South Africa Syria Tunisia United Arab Emirates Yemen (China)* (India)* Category 2 Angola Benin Botswana Burkina Faso Burundi Cameroon Chad Congo Cote d'Ivoire Ethiopia Gabon Ghana Guinea-Bissau Haiti Lesotho Liberia Mozambique Niger Nigeria Paraguay Somalia Sudan Uganda Zaire Category 3 Albania Algeria Australia Belize Bolivia Brazil Cambodia Central African Republic Chile Colombia El Salvador Gambia Guatemala Guinea Honduras Indonesia Kenya Lebanon Madagascar Malaysia Mali Mauritania Morocco Myanmar Namibia Nepal New Zealand Nicaragua Peru Senegal Tanzania Turkey Venezuela Zambia Zimbabwe Category 4 Argentina Austria Bangladesh Belgium Bulgaria Canada (China)* Costa Rica Cuba Denmark Dominican Republic Ecuador Finland France Germany Greece Guyana Hungary (India)* Italy Jamaica Japan Mexico Netherlands North Korea Norway Panama Philippines Poland Portugal Romania South Korea Spain Sri Lanka Surinam Sweden Switzerland Thailand UK Uruguay USA Vietnam Definitions: Category 1: These countries face "absolute water scarcity." They will not be able to meet water needs in the year 2025. Category 2: These countries face "economic water scarcity." They must more than double their efforts to extract water to meet 2025 water needs, but they will not have the financial resources available to develop these water supplies. Category 3: These countries have to increase water development between 25 and 100 percent to meet 2025 needs, but have more financial resources to do so. Category 4: These countries will have to increase water development modestly overall on average, by only five percent to keep up with 2025 demands. *These countries have severe regional water scarcity. A portion of their populations (381 million people in China in 1990 and 280 million people in India in 1990) are in Category 1. The rest of their populations are in Category 4. 41 2.2.8. Forestry and fisheries Forest management goals will increasingly shift from wood production to safeguarding the environmental functions of forests. The role of industrial forest plantations to provide timber is expected to increase strongly, with its share reaching one-third of total supply by 2015. Use of fuel wood is expected to continue to grow over the next two decades before stabilizing or even declining marginally. More than 60 percent of the wood harvested globally in 1995 was used as fuel (FAO, 2004b). Average world consumption of fish per person could grow from 16 kg a year in 1997 to 19-20 kg by 2030, raising total food use of fish to 150-160 million tonnes. The yearly sustainable yield of marine capture fisheries is estimated at no more than 100 million tonnes. "The bulk of the increase in supply therefore will have to come from aquaculture" (FAO, 2000b). In some developing countries, fish provide a critical source of highquality protein and needed oils. Fisheries currently provide less than one percent of the calories the world consumes, however, and this percentage is almost certain to decline. The annual fish catch peaked at 100 million metric tons in 1989 and has been stable at slightly lower levels since. The Food and Agriculture Organization of the United Nations has concluded that annual marine yields are adversely affected at extraction levels exceeding 80 million tons, and the organization predicts no growth in catches from lakes, streams and inland seas. Aquaculture supplies about 12 million tons of fish, but its growth is constrained by competition for fresh water and the challenge of keeping fish free from disease (PAI, 2005). 42 2.2.9. Rangelands With limited potential to expand food production either from fisheries or range-fed livestock, cultivated land will be pressed to supply ever increasing proportions of the world’s food. But arable land, too, faces severe constraints. We appear to be entering a period in which all food-producing systems must function well almost all the time. Calculations of the sunlight, water and plant nutrients available for growing crops suggest that farmers could feed—at least in theory— many more people than the current world population. In the real world, however, weather and pests may not cooperate, and farmers and other food suppliers perform at human rather than theoretical levels of efficiency. Storage and transportation of food on massive scales contribute to spoilage (PAI, 2005). For many food deficit low-income countries, feeding a growing population means coaxing more food out of the same amount of land. Canadian geographer Vaclav Smil estimated that the minimum amount of land needed to supply a vegetarian diet for one person without any use of artificial chemical inputs is 0.07 hectare, or slightly less than a quarter of an acre. Based on this, Population Action International estimated that currently some 420 million people live in land-scarce developing countries. If fertility and population growth in developing countries continue to fall, there could be 560 million by 2025. If not, there could be 1.04 billion such people (UNFPA, 2003). According to IFPRI, a "demand-driven livestock revolution is under way in the developing world with profound implications for global agriculture, health, livelihoods and the environment". IFPRI projects that meat demand in the developing world will double between 1995 and 2020 to 190 million metric tons. Demand for meat in the 43 developing world is expected to grow much faster than for cereals— by close to 3 per cent per year for meat compared with 1.8 per cent for cereals. In per capita terms, demand for meat will increase 40 per cent between 1995 and 2020 (Pinstrup-Andersen, and Pandya-Lorch, 1999). What this means is that demand for cereals to feed livestock will double in developing countries over the next generation. By 2020, feed grain demand is projected to reach just less than 450 million metric tons. Given this trend, well under way in much of Asia, demand for maize (corn) will increase much faster than any other cereal, growing by 2.35 per cent per year over the next 20 years. Nearly two thirds of this increased demand will go towards feeding livestock (UNFPA, 2003). In China, rising incomes and changing diets have resulted in a tremendous demand for meat, particularly poultry and pigs. Over the next two decades total demand for meat will double, increasing pressure on grain producers. It takes 4-5 kilograms of feed to produce 1 kilogram of meat (Pinstrup-Andersen and Cohen, 1998). As urbanization develops and incomes grow, the world food economy is increasingly fuelled by a demand for livestock products. The last 20 years have witnessed spectacular growth in meat demand in developing countries - expanding at an annual rate of 5.5 percent although many countries with the greatest need for higher protein consumption did not participate in this process. The poultry sector has seen dramatic gains, with the share in meat output more than doubling to 28 percent over the last three decades. As the developing world's demand for meat begins to level off and with consumption slackening in industrial countries, FAO projects a slowdown in the growth of the world meat economy (FAO, 2000b). 44 On land, pastures and rangelands are being grazed at or beyond capacity. Sustainable grazing of livestock is dependent on balancing the scale of use with the capacity of land, water and vegetation to tolerate that use over time. In most regions, grazing is now accompanied by soil erosion and adverse changes in plant species composition. At the same time, the world’s livestock population is growing even more rapidly than its human one. Moreover, rangeland alone can support only a portion of the world’s cattle and sheep, and it is of no use for raising pigs and poultry. Today, 37 percent of the world’s grain is used to feed livestock (World Resources Institute, 1994). 2.3. Conclusion It is now well accepted that, at least over the medium term, there appear to be no major global constraints to expanding world food production at a rate sufficient to match the growth of the effective demand for food. Yet the world still faces a fundamental food security challenge. Despite steadily falling fertility rates and family sizes, the world population is expected to grow. Despite progress on average per capita consumption of food, people in many countries still suffer from food insecurity, and malnutrition will still persist (Pretty and Hine, 2001). The deceleration over time of the effective demand for food contributes materially to this "happy" state of affairs. Such deceleration results from both positive and negative developments from the standpoint of human welfare. The positive ones are the slowdown in population growth due to voluntary reductions in fertility around the world and the fact that an ever growing proportion of 45 world population gradually achieves sufficient levels of nutrition beyond which there is only limited scope for further increases in perperson food demand. The negative aspects are the contributions of higher mortality to the slowing of global population growth, and the role of poverty in depressing demand for food. Even if demand for food is decelerated, the reason is because a significant part of world population with still very inadequate consumption levels lacks purchasing power and has no way to express their need to increase consumption in the form of solvable demand in the marketplace. This is why the problems of food insecurity afflicting many countries and population groups remain as severe as ever, regardless that price trends in world markets indicate once again an overabundance of food relative to effective demand at the global level. World market prices do not reflect adequately the problems of the poor and the food insecure (Alexandratos, 1998). Concerning the environmental and sustainability dimensions of the expansion and further intensification of agriculture, we note that (a) the foreseen land expansion need not be associated with the rapid rates of tropical deforestation observed in the past, though there is no guarantee that this will be so; (b) there will be further increases in the use of agrochemicals (fertilizer, pesticides) in the developing countries, though at declining rates compared with the past; (c) increased use of fertilizer is often indispensable for sustainability (to prevent soil mining); and (d) the need to accept trade-offs between production increases and the environment will continue to exist in the foreseeable future and the policy problem is how to achieve such increases while minimizing adverse impacts on natural resources and the wider environment (Alexandratos, 1998). Considering these 46 consequences, the population monster must be controlled. At the world's current rate of growth it will be impossible to fed the entire world and insure food security for everyone. World population will far outgrow food production. To avoid the harsh outcomes projected for the future, we must stop world population growth, and conserve our land, water, and energy resources that are vital for a sustainable economy and environment (Pimentel and Giampietro, 2001). It is also important to understand that each and every factor, political, gender, population, and others too all intertwine and influence each other. Not only must these factors be overcome but also other factors such as environmental degradation, finances, and geographical factors must be considered. If these factors are not overcome, our world will not be able to feed itself in the twenty-first century (Blundon, 2001). Effective and lasting solutions to problems related to food insecurity must be found in the policies and actions which pay adequate attention to those processes of development that aim primarily toward strengthening the human (Nshimbani, 2004). In the next chapter, we will show how these policies and resource management strategies can be affected to intensify or decline the consequences of population growth. 47 Chapter Three Sustainability: Basic Challenges 3.1. Importance As we enter the new millennium, one of the most challenging questions, if not the most challenging one to be addressed, is how to assess, build, and maintain a sustainable economy that will allow the human society to enjoy a sufficiently high standard of living without destroying its natural and biological support (Andriantiatsaholiniaina, 2001). The first in a new series on population and SD provides new ways of thinking about population trends in the 21st century. While the 20th century was the century of population growth with the world's population increasing from 1.6 to 6.1 billion, the 21st century is likely to see the end of world population growth and become the century of population ageing. At the moment we are at the crossroads of these two different demographic regimes, with some countries still experiencing high population growth, while others are already faced with rapid ageing. As it has become more and more clear that it is technically feasible to provide enough food for projected populations, the concept of sustainability has broadened to cover the environmental impacts of agriculture. This is a welcome development, but it is not enough as there are many ways in which food requirements and demand can be met. Based on general estimation of available natural resources, scientists of the Royal Society and the U.S. National Academy of Sciences have 48 issued a joint statement reinforcing the concern about the growing imbalance between the world's population and the resources that support human lives (RS and NAS, 1992). This is the main reason why sustainability issues have gained substantial importance on the political agenda in the recent years. SD is nowadays the goal, in words at least, of most politicians and decision-makers. Since publication of the Brundtland report in 1987 (WCED, 1987), the concept of sustainability has gained increasing attention among policy-makers and scientists, which culminated during the 1992 Earth Summit held in Rio de Janeiro. Among the results of the Earth Summit, Agenda 21 is a comprehensive list of actions needed to achieve SD (UNCED, 1992). Leaders from over 150 states committed themselves to undertaking actions, which will render future development sustainable but without scientific tools to guide policy-making towards a sustainable path (HMSO, 1994). Decisions leading to SD require a pragmatic approach to assess sustainability based on good science and adequate information. The latter is provided in the form of data about environmental, social, and economical factors known as indicators of sustainability. Sustainable projects and optimal strategies for development necessitate answering four fundamental questions: “why unsustainable development occurs”, “what is sustainability”, “how can it be measured”, and “which factors affect it” (Atkinson et al., 1999). There is evidence that development is currently unsustainable. Ozone depletion, global warming, depletion of aquifers, species extinction, collapse of fisheries, soil erosion, and air pollution are among the obvious signs of ecological distress (Brown et al., 2000). Our society is also showing similar signs due to poverty, illiteracy, AIDS, social 49 and political unrest, and violence (IUCN/UNEP/WWF, 1991; UNEP, 1992). Sustainability is an inherent vague and complex concept. As pointed out in the literature, it is not that sustainability indicators are lacking but their fragmentary and polymorphous nature hampers their direct usefulness in the quest of strategies for SD (Brink et al., 1990; OECD, 1994). Despite the fact that the concept of SD is ill-defined, policymakers should strive towards it and scientists should find scientific ways to assess and improve upon it. The development of sustainable policies is a necessity if we adopt the precautionary approach of development and if we want to avoid nasty surprises species as the ozone depletion or irreversible actions such as species extinction. Of course, even then, no one can guarantee complete avoidance of surprises given our incomplete state of knowledge of an extremely complex environment, but at least we do our best. Although sustainability is a goal for international and national policymakers, there is no measuring yardstick against which to assess practical policy (WCED, 1987). Sustainability is difficult to define or measure because it is an inherently vague and complex concept. Fuzzy logic, due to its capability to emulate skilled humans and its systematic approach in handling vague situations where traditional mathematics are ineffective, seems to be a natural technical tool to assess sustainability. What we need is adequate information that is tailored to quantitative sustainability objectives. Brink (1991) states that such information should: (a) give a clear indication as to whether objectives of sustainability are met, (b) concern the system as a whole, (c) have a quantitative character, (d) be understandable to non-scientists, and (e) 50 contain parameters which can be used for periods of one or more decades. The need for a practical tool to assess sustainability is crucial to policy-makers if they are to secure future development. Since such a tool is not available, management by trial-and-error instead of management by knowledge and prediction is currently the only way used to establish sustainable policies. A deadlock further impeding the measurement of sustainability lies in the fact that scientists are waiting for important political issues to be raised by policy-makers, while policy-makers are waiting for important ecological issues and ecological indicators of sustainability to be defined by scientists (Brink, 1989). Unfortunately, unsustainability may not be easily reversible because the natural and biological support systems of an economy are subject to thermodynamic laws and irreversibility. 3.2. Definitions The fundamental question is: “what is sustainability and how sustainable is an economic system” (Procter and Gamble, 2005). Since SD is a vague continually evolving concept, it is difficult to define it in an appreciate way (McKeown et al., 2002; Pembleton, 2004). The commonly used definition of SD was put forward by the Report of the World Commission on Economy and Environment (the "Brundtland Report") in 1986: "To meet the needs of the present without compromising the ability of future generations to meet their own needs" (WCED, 1987; p. 43). The precise meaning of SD has been widely debated (Wikipedia, 2005). For example, two years after the Brundtland Commission's Report popularised the term, over 140 definitions of SD had been catalogued. Before the widespread use of the term sustainable industries, the terms sustainable economy and SD 51 were prevalent. Their popularization started with the United Nations Conference for Environment and Development (the Earth Summit) in 1992. The conference was prompted by the report Our Common Future (1987, World Commission on Environment and Development, also known as the Brundtland Commission), which called for strategies to strengthen efforts to promote sustainable and environmentally sound development. Caring for natural resources and promoting their sustainable use is an essential response of the world community to ensure its own survival and well-being (Nationmaster, 2005). The issue of intergenerational responsibility also is raised. The Bruntland report does not denounce the depletion of non-renewable resources. Serageldin suggested that "sustainability is to leave future generations as many opportunities as, if not more than, we have had ourselves" (Takle, 2001). This is a very far-reaching principle that admits a wide range of activities to allow residents of the planet present and future - to live fulfilling lives. One definition of SD that appears to have more resonance with the general public is that used by the United Kingdom government: "SD is about ensuring a better quality of life for everyone, now and for generations to come." This focus of SD on improving quality of life is becoming more widely accepted by governments, companies, civil society organizations, and others (Holliday and Pepper, 2005). A quality of life focus makes the concept more aspirational, and it changes the tone and content of the SD debate so that the emphasis is more on solutions than the problems. Another definition of SD, demands that “we seek ways of living, working and being that enable all people of the world to lead healthy, fulfilling, and economically secure lives without destroying the environment and without endangering the future welfare of people 52 and the planet” (Cumber, 2004). Based on this definition, we are constantly required to re-evaluate our values and decisions: for example, the present realities of malnutrition, lack of suitable housing, and lack of safe drinking water, suggest that significant development is needed for the present generation. 3.3. Dimensions A series of seven UN conferences followed on environment and development. The first UN Conference on Environment and Development (UNCED or the Rio Summit) has emphasized this shift of attention in the form of Agenda 21. Since then, countries and regions have been formulating strategies towards the achievement of sustainability in various sectors of national and regional development (GLEAM, 2002). These strategies will have to cover the environmental, social, economic (Hart, 2000), and recently, is more defined by adding the fourth, which is institutional dimension (GLEAM, 2002). SD is also one of the issues addressed by international environmental law and focuses on the overall performance or health of ecosystems. Social sustainability seeks to reduce the vulnerability of various segments of the society, particularly the poor, and maintain the health of social and cultural systems. Economic sustainability aims to maximize the flow of income, while maintaining the stock of assets required for these benefits. The institutional dimension reflects the whole set of norms and beliefs on which personal preferences and attitudes as well as private and public organizations are built. Institutional sustainability links to the availability of mechanisms to implement the other dimensions of sustainability as well as the long-term viability of the 53 institutions in them. In all cases each system's capability to withstand shocks (vulnerability and resilience) is an important aspect of sustainability (Ibid). In other words, the environmental dimensions of sustainability refer to the need to maintain (or restore) the physical resource base so that it endures indefinitely to meet the needs of the present without compromising the capacity of future generations to meet their needs. This also highlights the underlying and fundamental time component inherent in sustainability. Economic sustainability is equally conditional on the use of resources so as to avoid their overexploitation either in terms of their quality or quantity, or the use of resources which results in the generation of waste in excess capacity of the environment's to absorb it effectively. The balance between environmental and economic sustainability is mediated through the institutional arrangements that shape and condition the management and use of the land, and those social norms that influence community values. In effect, the different dimensions of sustainability constitute the key components of the system, and act in concert to either promote or constrain the achievement of sustainability (MAF, 2000). Land use is the visual expression of the interplay among those different dimensions and as such can be an important indicator of the health of the overall ecosystem. Therefore, SD calls for long-term structural change in our economic and social systems, with the aim of reducing the consumption of the environment and resources to a permanently affordable level, while maintaining economic output potential and social cohesion (ARE, 2005). While, environmental sustainability identifies energy, fresh water, and reversing land and soil degradations as priorities, especially for developing countries to protect their natural resource base (EC, 2002; 54 Pembleton, 2004), social sustainability seeks to reduce the vulnerability of various segments of the society, particularly the poor, and maintain the health of social and cultural systems. In general, we can divide all these dimensions in two categories: the technical definition being "a sound balance among the interactions of the impacts (positive and/or negative), or stresses, on the four major quality systems: People, Economic Development, Environment and Availability of Resources," and the none-technical definition being "a sound balance among the interactions designed to create a healthy economic growth, preserve environmental quality, make a wise use of our resources, and enhance social benefits". When there is a need to find a solution to a problem or a concern, a sound solution would be to choose a measure or conduct an action, if possible, which causes reversible damage as opposed to a measure or an action causing an irreversible loss (Earth Government for Earth Community, 2003). 3.4. Modeling problems Modeling sustainable systems has seen considerable improvement in its application and utilization since the Rio Summit. This has been driven partly by the development and negotiation about various Multilateral Environmental Agreements and by the need for scientific assessments in support of these negotiations. The UN Framework Convention for Climate Change has generated a broad variety of integrated assessment models focussing on the climate issue. Some trend setting examples of these models are AIM (NIES, Japan), MARIA (Tokyo University), IMAGE (RIVM, The Netherlands), MESSAGE (IIASA, Austria) and MiniCam (PNNL, USA). Also, the integrated models try to provide a powerful tool to undertake such 55 analyses and evaluations (Morita et al., 2000). Until recently, such assessment modeling efforts have been typically limited in geographical focus and in the sectors and issues covered. Some of these models are gradually being expanded in order to cover the broader perspectives of sustainability, vulnerability, durability and human security, but they are still insufficient. Substantial model development is required to cover the different dimensions of sustainability and the spatial and temporal variety within these dimensions. While some models try to include environmental dimension, they neglect socio-economic dimensions. On the other hand, while socio-economic dimensions are emphasized by the rest, the environmental dimension has been neglected (GLEAM, 2002). In practice, both researchers and policy-makers have found these characteristics of sustainability difficult to handle. There is no good conceptual framework available to untangle financial/environmental links. The fundamental institutional changes that sustainability may require, may themselves need strength of political will not readily secured. Consequently, in most countries, policy efforts have focused on the incorporation of sustainability within existing operational and ideological frameworks. More effort has been spent on issues of intergenerational transfer than to any redistribution of resources within any one-generation. Equally more environmental legislation, determined on traditional lines, has often been substituted for efforts to better integrate environmental values into all decision making. To this extent environmental policies more frequently remain a "tag-on" to existing economic strategies, and production and environmental policies have generally remained polarised and apart, despite rhetoric to the contrary (MAF, 2000). 56 As the concept of sustainability represents a fundamental challenge at the theoretical and methodological levels, reorientation within the social sciences themselves is required, implying (UNESCO, 1996): • Firstly, to give more attention to current vital issues such as land use, social use of natural resources like water or wood, production and consumption patterns, loss of biodiversity, etc.; • Secondly, to expand the problem-oriented cooperation between the social and the natural sciences on issues and questions of sustainability. The crucial importance is that natural and social sciences cooperate on an equal basis, starting from the phase of defining the problems under study; and • Thirdly, to develop and improve interdisciplinary cooperation among the various disciplines. This is necessary to achieve a more integrated and comprehensive understanding of development processes, as well as the relationships between individuals and the environment in their social, political, economic, psychological and cultural aspects. With regard to this, the historical boundaries between the disciplines must be re-examined and methodologies of interdisciplinary research are to be developed. With the increasing attention given to the economic, social and ecological impacts of SD, the need for interdisciplinary approaches in agricultural research is likely to continue growing. However, the experiences of the last few decades of interdisciplinary approaches such as Farming Systems Research (FSR) and Farming Systems Research/Extension (FSRE) are not very promising: disciplinary conflict is common, productive interaction seems difficult to achieve 57 and practical results have been disappointing (Hawkins, 1997). Within the realm of agricultural research, farming systems research (FSR) and land use planning are areas where interdisciplinary is indispensable (Aenis and Nagel, 2000). However, as Hawkins (1997) reports, only few FSR projects have reached interdisciplinary status involving both, technical and social sciences. Sustainability in social, economic, and ecological terms is the central goal of R&D projects in the field of natural resource management. Up until now, impacts seem to be marginal (Pretty, 1995). As a consequence, a change with regard to institutionalizing and organizing participation of all relevant actors is necessary. But which efforts ensure participation? How to manage this process? Which structures support efficient outputs? (Aenis and Nagel, 2000). Some practical experiences have been made in training small teams for interdisciplinary studies and still, there is a growing scientific knowledge on economic and ecological sound planning and management techniques for natural resources. It is also argued that interdisciplinary research requires: a) an appreciation of holistic approaches; b) a common understanding of goals and objectives and skills in planning methods to reach these; and c) effective communication and management procedures to enable expert teams to function efficiently (Hawkins, 1997a). This group of experts help to give better understanding of assessing the SD in the situation. They have a lot of information on dimensions, indicators, and assessment of them (Meadows, 1998). 58 3.5. Conclusion It is fair to say that some clear measures or, at least, indicators of sustainability exist, but the effectiveness of policies towards a goal of sustainability cannot be assessed. Attempts have been made to measure sustainability using economic, ecological, or combined economic- ecological approaches, but the results still lack universal acceptance. Examples of existing sustainability measurements are Pearce and Atkinson, 1993; OECD, 1994; Sherp, 1994; IUCN/IDRC, 1995; Rennings and Wiggering, 1997. For the sake of analysis, researchers have broken down sustainability into a large number of individual components or indicators whose synthesis into one measure appears to be next to impossible. As pointed out in the literature, it is not so much that environmental and socio-economical information is lacking but the fragmentary, often qualitative, and very detailed nature of this information hampers its direct usefulness in policy making (Brink et al., 1990). Not only are there no common units of measurement for the indicators of sustainability but quantitative criteria for certain values are also lacking. A systemic method based on a reliable scientific methodology is needed to combine multidimensional components and assess uncertainty. Such a method should be flexible in the sense that one can add or remove indicators to achieve a better assessment of the system according to the context. In reality the border between sustainability and unsustainability is not sharp but rather fuzzy. This means that it is not possible to determine exact reference values for sustainability and a scientific evaluation of uncertainty must always be considered in the procedure of sustainability assessment. For this reason, the use of natural language and linguistic values based on the 59 fuzzy logic methodology (Munda et al., 1994) seems more suitable to assess sustainability. In the two next following chapters, we will first explain sustainability in rangeland management and some basic concerned challenges, which can be the causes of (dis)equilibrium. Then, in Chapter 5, we will introduce fuzzy logic as a powerful tool to deal with these challenges. 60 Chapter Four Rangeland Management: Basic Challenges and Principles 4.1. A review of literature An accelerating rate of literature is being published on rangeland management, although, the use of rangelands for domestic productions is increasingly questioned on conservation and sustainability grounds in all over the world (Fleischner, 1994, Grigg, 1995). However, much recent literature on grazing management, particularly literature from the arid to sub-humid regions, is at a theoretical level (e.g. Westoby et al., 1989). This is in part because adaptive management research - to determine appropriate management practices for extensively managed rangelands - is now just beginning at the operational scale. Simulation modeling and decision support systems are currently the only way to explore the many possible management alternatives. Grazing managers do not just adopt a new grazing system but rather they adapt a suite of management changes in conjunction with implementing a new grazing system. Savory and Butterfield (1999) are right in stating that grazing management may not be researched in the manner of normal biological research. There is a need to develop principles that can be integrated using simulation modeling into grazing management systems. Business schools have extensively studied management practices of successful businesses by interviewing managers and observing how successful businesses are operated (Peters and 61 Waterman, 1982), not by setting up replicate corporations based on different management challenges and collecting data on them over a number of years. However, even while modeling approaches are being developed and validated, there is still a need now for some practical guidelines for grazing management. Thus for the present and foreseeable future common sense and rule of thumb solutions to grazing management problems may be the most useful. This chapter reviews some practical recommendations for managing grazing on rangelands. In doing so, as most of nomadic people in the world and especially in Iran live in the arid to sub-arid rangeland areas (McMurphy et al., 1990), our focus would be around the conditions of these areas which are faced by socio-economic and ecological challenges, rather than the more humid areas where agronomic challenges are widely survived. The debate in preparing this chapter has been to understand why we propound especial principles and challenges to reach equilibrium in rangeland management. The main reason is, as we have discussed in Chapter 5, a need to develop a new model; using fuzzy logic, which needs some criteria as inputs of the model. This is particularly because over 4,000 papers and books were published on the subject since 1970 to show how we can arrive at the challenges of grazing management in practice (Walker and Hodgkinson, 2000). In the following, at first, we give a discussion to find the concept of rangeland management; if it implicates to art or/and science. Secondly, we describe equilibrium and disequilibrium systems in rangeland management. Then, we explain current challenges in pastoral systems, and finally, we explain basic principles to reach equilibrium. 62 4.2. Rangeland management: Art or science? Although the definition of rangeland management has changed somewhat over the years, the part that states it is both an art and a science has not. In 1943, Stoddart and Smith defined rangeland management as "the science and art of planning and directing rangeland use so as to obtain the maximum livestock production consistent with conservation of the range resources" (Stoddart and Smith, 1943). This definition implies a sustained yield of livestock over a long period of time. The book leaves no doubt that the main objective was to produce livestock. In the second edition (1955) of their book, rangeland management was the science and art of obtaining maximum livestock production from rangeland consistent with the conservation of land resources. This definition asserts rangeland management is closely related to animal husbandry and plant ecology; a movement from livestock production to land management in 12 years. Twenty years later (in the third edition) Stoddart et al., (1975) defined rangeland management as "the science and art of optimizing the returns from rangelands in those combinations most desired by and suitable to society through the manipulation of rangeland ecosystems. The science portion of rangeland management provides a body of knowledge in which principles are developed. Principles should not dictate management, but provide guidelines for management. An artful management is therefore, the skillful and/or ingenious application of scientific principles, experience, and creativity to land management practices. In new definitions, however, rangeland management is a lot more of an art than science, since there is so much we do not know about rangeland ecosystems’ mechanisms and processes that drive rangeland 63 ecosystems. Ecosystems are too complex and too dynamic in both time and space to ever have fully figured out. This makes it easy for interest or user groups from all perspectives to challenge us. We are also challenged because many land management decisions were not made in an artful manner (Miller, 1995). 4.3. Equilibrium and disequilibrium systems in rangeland management For the objectives of this dissertation, we consider equilibrium versus disequilibrium continuum in pastoral systems. Equilibrium systems are those that may change plant species composition in response to grazing, but when grazing pressure is reduced, will return to their former composition. These systems typically have a long evolutionary history of grazing, are not susceptible to invasion by undesirable plants, either woody or herbaceous, and can be managed using simple models of succession. Disequilibrium systems are those that never reach a steady state and are controlled primarily by unpredictable factors such as rainfall and fire. These systems typically have a short evolutionary history of grazing, are prone to invasion by undesirable plants and are best managed using the state and transition model of community dynamics (Walker and Hodgkinson, 2000). Management of equilibrium systems is fairly straightforward because these systems respond to change in total grazing pressure. Outcomes must be described and criteria established to monitor progress. A grazing management plan could be developed based on manipulating livestock species, grazing season, distribution and stocking rate to accomplish goals. For instance, if maximum sustainable production of livestock products is the objective and distribution is not considered a 64 problem, then year-long continuous grazing using sheep and cattle at a variable stocking rate tied to annual rainfall would be appropriate. Decision support systems are currently available to monitor and adjust stocking rate to match animal demand with primary production (Stuth and Lyons, 1993). In equilibrium ecosystems, management to maintain the right rate of stocking is the most important factor of grazing management. This can be done with a simple rotation or even continuous grazing system. Season or timing of use, livestock species and distribution can be adjusted to increase carrying capacity or accomplish other management objectives such as multiple uses. The objectives of multiple uses will complicate the grazing management of equilibrium systems, but in many cases there is commonality of states that meet different objectives. For example, grazing management that retains all palatable plant species will probably meet both biodiversity and sustainable production goals (Walker and Hodgkinson, 2000). Disequilibrium systems are more difficult to manage because even in the absence of livestock they may be unstable, particularly if invasion by undesirable plant and animal species is a problem. Invasion by woody plants is a widespread problem, which although potentially accelerated by grazing, typically will also occur in the absence of grazing (Archer, 1996). Fire is the best tool for controlling woody plant invasion, but the ability to use this tool is dependent upon grazing systems to accumulate and manage grass fuel levels. In these systems grass production will usually need to be managed both as forage and as fuel (Kothmann et al., 1997). Because of the tendencies of disequilibrium systems to cross thresholds to less productive states, grazing systems that have been shown to advance succession will have the greatest potential for maintaining plant communities in the desired 65 vegetation state. Because of the long deferment periods and resultant increased grazing intensity on the grazed units, these types of systems tend to be very sensitive to stocking rate particularly as it affects animal performance. Another advantage of grazing systems with long deferment periods is that they are the ones best adapted to accumulate the grass fuel required for a successful prescribed fire program (Howery et al., 2000). We have purposefully avoided discussing grazing systems per se, because there is no evidence to indicate that they provide any advantage other than the flexibility that cross fencing and grouping of animals provide for managing livestock distribution and timing of grazing. However, as we learn more about how timing may be used to provide either rest or grazing pressure at critical times to accomplish vegetation management goals, this flexibility will be increasingly important. Likewise, when prescribed fire becomes an integral part of the grazing management plan the flexibility of multiple rangelands to manage grass as forage or fuel becomes very important (Walker and Hodgkinson, 2000). 4.4. Current challenges in rangeland management 4.4.1. Overgrazing The most contentious and emotional rangeland use issue now and in the foreseeable future is balancing private rights with the public interest. This debate often terminates to overgrazing (private rights) and sustainability (public rights) (Box, 1995). This issue will surface in such different rangeland use debates as historical preservation, endangered species, location of waste disposal facilities, and public rangeland management. The controversy of overgrazing on public 66 rangelands, therefore, is not just about cows and grass. It is about ownership of the public lands and it is about what rights pastoralists can hold in a common resource (Box, 2002; Deadman, 1999). Whenever pastoralists prefer to focus on their own rights and to neglect the public rights, rangeland degradation will be unavoidable, especially for next generations (Azadi et al., 2003). In general, rangeland degradation falls into two broad categories: that resulting from extended periods of drought, and that resulting from overuse through cultivation or overgrazing (Hiernaux, 1996). With uncontrolled grazing, most rangelands are overgrazed and the vegetation is depleted. Due to overuse of resources, especially overgrazing, and the application of non-suitable management practices such as low recognition of prevalent natural vegetation cycles in grass and thorn bush savannahs without considering long-term degradation processes, the rangeland quality of many rangeland areas has declined (Buss and Nuppenau, 2002). A visible decreasing appearance of natural composition of grass and bush cover, bush encroachment and a decreasing biodiversity indicate lower stocking potentials for domestic livestock on large areas of rangeland, especially in developing countries such as Iran (Mesdaghi, 1995). In this case, rangeland degradation becomes an economical threat to falling pastoralists’ income, a social threat to the continuation of rural-urban immigration, and an environmental threat to desertification. As a response, on one hand, many pastoralists increase their livestock and overgraze the rangelands. On the other hand, economic and political pressures push them to produce red meat for a growing population, especially for urban people. Plantation density, therefore, will decrease and degradation will happen. Sometimes, 67 government policies inadvertently provide incentives that encourage overgrazing. Overgrazing occurs, even though analysis demonstrates that maximum net economic returns will occur below the maximum sustainable level of livestock off-take per unit area (Buxton and Stafford Smith, 1996). Financial crisis is another factor tempting graziers to push stocking to the limit. Pastoral businesses facing insolvency are greatly tempted to increase stocking rates in an effort to sustain their operation until commodity prices or climate take a favourable turn. Carrying capacity and the appropriate stocking rate cannot be determined until the decisions relative to so species of livestock, season of use and distribution have been revised (Taghi Farvar and Jandaghi, 1998; Walker and Hodgkinson, 2000). 4.4.2. Carrying capacity The science of rangeland management adapted carrying capacity concepts to grazing systems on the rangelands. The logic basis for this concern is the concept of rangeland carrying capacity. Carrying capacity is considered to be the average number of animals that a particular rangeland or range can sustain over time. Stocking Rate (SR) is expressed as the number of animal unit months (aum)1 supplied by unit of area (e.g. hectare). An animal unit month is the amount of forage required by an animal unit (au)2 grazing for one month (Kopp, 2004). The responsibility of rangeland managers is to try to balance livestock grazing pressure with the natural regenerative 1. The term aum is commonly used in three ways: (a) Stocking rate, as in "X unit of area per aum"; (b) forage allocations, as in "X aum(s) in Allotment A"; (c) utilization, as in "X aum(s) taken from Unit B (Glossary of Terms Used in Range Management, 2002). 2 . An animal unit (au) is defined as an "average" live body weight equal to 1000 lbs (453.59 kg) (NRCS, 2000). 68 capacity of rangeland plants. The estimations of carrying capacity are usually based on assumptions about the impact of livestock on plants and plant succession. Heavy livestock grazing is thought to lead to a decline in rangeland condition, and reducing or removing grazing pressure assumed plant successional processes would restore the rangeland to its previous condition. By knowing the rangeland condition class1, the proper use factor, or the amount of forage to leave to allow plant nutrients to be restored, and taking into account distance to water, slope steepness, and other factors, carrying capacities for a particular rangeland or pasture could be determined (Miller, 2005). These managerial estimations have usually been used in many countries such as Iran. Before turning to the next section, it should be noted that in most instances where rangelands are being grazed by domestic livestock in an unsustainable manner, the root cause is the lack of well-matched theory on grazing management. However, we believe that the belowdiscussed principles can be understood to alleviate the most of abovediscussed challenges associated with grazing by domestic livestock. Pastoralists can be expected to make rational decisions from their point of view. In general, it should be noticed where grazing management is poor, the theory may be wrong or incomplete, or it may apply to only part of a system being managed. Other factors such as social and economic may override consideration of grazing management. 1 . Rangeland condition class is defined by the percent of climax for the range site, classified as "Poor" (0-25), "Fair" (26-50), "Good" (51-75), and "Excellent" (76-100) (Glossary of Terms Used in Range Management, 2002). 69 4.5. Basic principles in rangeland management The purposes of grazing management is to minimise the adverse impacts of domestic livestock and other herbivores that comprise the total grazing pressure on the natural resources of soil and biota and to maximise the probability that the grazing enterprise will be sustainable. This statement does not imply that grazing is not sustainable or that grazing is not an important ecological force in many rangeland ecosystems (Milchunas et al., 1988). However, at the level of individual plants, defoliation decreases production in the vast majority of instances (Jameson, 1963) and in most instances the purpose of grazing management is to ameliorate this negative effect and maintain a competitive balance among the plant assemblage. There are now multiple desired outcomes for a sustainable grazing enterprise (Walker, 1995), and this trend is likely to increase. The goals of grazing management may include control of undesirable vegetation (Olson and Lacey, 1994), enhancing wildlife habitat (Mosley, 1994), reduction of fire hazard, maintenance of biodiversity (Landsberg et al., 1999), animal traction, manure, banking livestock capital and, of course, for the profitable production of food and fibre. Herbivores, including domestic livestock, highly select the plants and plant parts they graze (Leigh and Mulham, 1966) and the parts of the landscape they favour for grazing (Landsberg and Stol, 1996, Roshier and Nicol, 1998). Plants are the victims of defoliation, overgrazing, and suffer through reduction in the plant's resource capturing ability. Ultimately, such reduction impairs the ability of rangeland plants to survive times of resource scarcity (Hodgkinson, 1996). Ecosystem functioning may also change under the impact of grazing and other anthropomorphic influences. For example, the chance of fire may be 70 greatly reduced (Hodgkinson and Harrington, 1985) and competitive relationships among plant species may be altered because of elevated atmospheric CO2 that has resulted from the burning of fossil fuels (Polley, 1997). Development of a grazing management strategy for a pastoral enterprise requires definition of the desired outcomes and of the inherent constraints. From these, the solution space of the problem can be defined and management can then adjust the stocking rate, grazing patterns, livestock species and distribution of grazing animals in different seasons. It should be recognised that any time domestic livestock are grazed, related decisions to livestock species, timing of grazing and stocking rate have to be made, either consciously or unconsciously. Improving grazing distribution, however, normally requires the investment in capital improvements such as fencing or water development. Stocking rate is usually considered the most important factor as well as the most abused factor in grazing management. The importance of season of grazing is dependent upon the climate of an area and whether cross fencing is in place that would allow the implementation of a grazing management system. Existing conditions in many areas dictate that on a large percentage of grazing lands, season of grazing is not a factor that can be managed readily. However, resting from grazing to foster plant recovery and seed production, and in some cases to prevent accelerated mortality of domestic stock due to forage and water scarcity, is desirable in ecosystems of non-seasonal rainfall. This can be achieved by lease grazing, stock sale or feeding of livestock on forage reserved within the property (Walker and Hodgkinson, 2000). 71 Choice of species of livestock is an under-utilized management tool at the present time. The selection of the species of livestock best adapted to a particular environment is often the simplest solution to grazing management problems. However, many graziers are reluctant to make such changes often for cultural reasons but also because of physical reasons such as predators. In both developed countries with the majority of pastoral business (e.g. the USA) and less developed countries with the majority of nomadic people (e.g. Iran), the impact of grazing on riparian areas, is one of the biggest grazing management problems (Kauffman and Krueger, 1984). Overgrazing of riparian areas affects biodiversity and water quality and is caused primarily by a distribution problem. Whereas ribbon fencing of riparian areas is a solution that has been used in some areas, a much simpler solution is to change from cattle to sheep or goats, which prefer uplands to riparian areas (Glimp and Swanson, 1994). Unfortunately, evaluating the appropriate species of livestock, which should be the first consideration in developing a grazing management plan, is seldom considered. Control of undesirable plants is another example of a problem that often is best solved by changing species of livestock. It is a truism that plants dominating an area are the ones avoided by the dominant herbivores in the system. Whereas determining the dietary preferences of grazing livestock is often difficult, determining the plants that are avoided is typically much easier. If a species of livestock exists that has a preference for the undesirable plant, changing livestock species is an easy solution. Classic examples include the use of sheep or goats to control leafy spurge (Euphorbia eusla L.; Johnson and Peake, 1960) or gorse (Ulex europaeus L.; Radcliffe, 1985). Finally, if the management goal is to maximise 72 profit and minimise risk, then mixed species grazing, which has been shown to accomplish this objective under a wide array of price ratios (Connolly and Nolan, 1976) should be implemented. However, except for a few areas where it is the cultural norm this practice is rarely used (Walker, 1994). Timing of defoliation may adversely affect a plant's ability to replace tissue, reproduce and compete for resources. There are times when a plant can be grazed safely and there are other times when grazing will raise the probability of the plant's dying or its growth being impaired (Blaisdell and Pechanec, 1949). Animal grazing preferences vary seasonally and interact with seasonal plant responses. In temperate climates with strong seasonal differences in temperature and topography, seasonal grazing practices have evolved to meet stock and natural resource requirements. Properties with contrasting soil types, such as floodplains and runoff landscapes, also are a form of seasonal grazing based on flooding patterns. Timing of grazing, for whatever reason, will affect stocking rate because grazing at sensitive times will reduce primary production and community stability unless grazing levels are low. In areas with large topography differences, phenological development may vary by 60 to 90 days between the valley floor and the alpine grasslands. Grazing management accommodates these differences in phenological development as livestock are moved up the mountain following the green (Burkhart, 1996). Similarly, in those rangelands where properties are large and rainfall is irregular in space and time, such as in semi-arid woodlands, livestock may be moved to a paddock(s) where high rainfall from a convective storm has generated a pulse of plant growth (Hodgkinson and Freudenberger, 1997; Martin, 1978). Tactical grazing, particularly 73 in semi-arid areas, is a management option that is under-utilized. It is time to reevaluate the appropriateness of continuous grazing (Holmes, 1996). The economic and ecological implications of grazing selected rangeland areas only during the seasons or years with adequate rainfall for plant requirements need urgent evaluation. Landscape degradation and adverse vegetation change from overgrazing was one of the primary factors leading to the development of the discipline of range science. While the importance of proper stocking rate is well recognised and techniques for assessing the problem are available, overgrazing is still a major problem in many countries such as Iran. However, it is more of a problem of wrong mental models than inadequate technology and there are several factors that tempt pastoralists to exceed the carrying capacity of a property. These factors include the variable and unpredictable nature of rainfall coupled with the overly optimistic attitude of graziers who gamble on the ending of drought. Unfortunately, in many arid areas, probability of a drought is greater than the probability of normal rainfall (Riggio et al., 1987). Sometimes, government policies inadvertently provide incentives that encourage overgrazing (Foran and Stafford Smith, 1991; Holechek and Hess, 1995). Overgrazing occurs even though analysis demonstrates that maximum net economic returns will occur below the maximum sustainable level of livestock per unit of area (Buxton and Stafford Smith, 1996). Financial crisis is another factor tempting graziers to push stocking to the limit. Pastoral businesses facing insolvency are greatly tempted to increase stocking rates in an effort to sustain their operation until commodity prices or climate take a favourable turn. Properties located in the more arid regions with lower carrying capacities tend to be less 74 profitable because of higher fixed cost (Holechek and Hawkes, 1993). Thus, the temptation to overgraze is often greatest in the areas that are the most vulnerable. Carrying capacity and the right rate of stocking cannot be determined until the decisions relative to species of livestock, season of use and distribution have been made. Although, the correct stocking rate is dependent upon these other three factors, the overall success of the system will be dependent upon setting the right rate of stocking. In general, when stocking rate is too high, the system will not be sustainable even if the other components of grazing management are correct. Unsuitable spatial distribution is the cause of many problems associated with grazing livestock (Holechek et al., 1989). Furthermore, uneven distribution is caused by characteristics of both the grazing animal as well as the plant community or landscape. Distribution of livestock occurs at many levels and affects and is affected by livestock species, season of use and stocking rate. Many overgrazing problems are actually species of livestock or season of grazing problems and are in a sense manifested as distribution problems. Examples include overgrazing of riparian areas or grazing prior to range readiness, both of which can cause adverse impacts at very low grazing intensities. Carrying capacity in large pastures with unsuitable livestock distribution is lower than on similar areas with suitable distribution. Unsuitable distribution may buffer the effect of changes in stocking rate on livestock performance compared to the classic response curve where animals are well distributed (Stafford Smith, 1996). Fencing and water development have traditionally been used to improve distribution and the benefits from infrastructure development can be modeled. 75 Grazing systems are based on manipulation of spatial and temporal distribution of livestock grazing pressure. Because grazing systems involve rotating animals among rangelands they actually determine the distribution of livestock. There is a good evidence that any production benefits that result from implementing grazing systems arise from improved distribution and increased carrying capacity (Hart et al., 1993) rather than altered foraging behaviour (Gammon and Roberts, 1978) or increased primary production (Heitschmidt et al., 1987). While most researches on animal distribution have been conducted at the enclosure level, the effect of meta-scale distribution, which involves transportation systems and grazing livestock only during those seasons or years with adequate rainfall, deserves more consideration in disequilibrium areas as a means of achieving sustainable systems. In contrast, when non-pastoral goals such as maintenance of biodiversity become important grazing management objectives, practices that improve livestock distribution may not be implemented because they conflict with the other objectives (Landsberg et al., 1997). Ensuring that parts of enclosures are ungrazed or lightly grazed would meet biodiversity objectives at local scale. The most important challenge and basic principles in rangeland management is, to devise and implement strategies that will strike a balance between the needs of pastoralists to consume rangeland at local level, and the needs of future generations to conserve rangeland biodiversity at national level. The greatest impediment to these trade-offs is the value systems and personality characteristics of pastoralists on one hand, and decision makers, on the other hand (Thompson, 1995). These characteristics, 76 which are important for the successful settlement and development of the current livestock production systems, tend to make people with current tenure of the land very resistant to manage sustainably. However, if one accepts the self-evident statement that "the only sustainable rangeland management is profitable rangeland" (Ainesworth, 1989) and the trends discussed above continue, eventually economics will either transform pastoralists or cause a change in ownership. Then the products and services produced on rangelands will be more congruent with contemporary socio-economic systems. This will also affect the types of information and technology that will be demanded from rangeland specialists, although the management principles can remain constantly (i.e., livestock species, distribution, timing and stocking rate). One of the best examples in this case is the current situation of Iranian nomadic people. Where there is overstocking, this is environmentally unfriendly since overgrazing and subsequent land degradation would result. Iranian pastoralists have to be educated on limiting their livestock according to the carrying capacity of the land somehow overgrazing and degradation do not occur. In this case, a restocking program for the districts is needed. Therefore, it is a challenge to improve the management of livestock rising by introducing new methods of rangeland management including zero grazing (Azadi et al., In revision). 4.6. Conclusion A claim is commonly made that the rangelands of the world are overgrazed and hence producing edible forage and animal produce at less than their potential (Wilson and Macleod, 1991). Globally, 77 rangelands are at risk from numerous pressures (Mitchell et al., 1999). Some of these pressures arise from livestock/rangeland systems. Livestock have been a key factor in the development of civilization, but their role in the future is not clear as well as how the science of rangeland management should change in order to meet the challenges of the future. Carrying capacity is the most important variable in rangeland management (Walker, 1995). At a time when the planet's limited carrying capacity seems increasingly obvious, the rationale and measures of rangelands carrying capacity are increasingly criticized. One of the key elements of rangeland capacity is stocking rate. Calculating stocking rate is relatively simple once the concept and terminology are understood. The ability to calculate stocking rate and make timely management decisions is vital to maximizing net returns from the livestock operation (Redfearn and Bidwell, 2004). If stocking rate is not close to the proper level related to equilibrium rate, then, regardless of other grazing management practices employed objectives will not be met (Roe, 1997). The recent literatures on rangelands disequilibrium call in question any specific measures of carrying capacity, whether the range is stocked or unstocked, managed or mis-managed. The stocking rate of a given area can vary in accordance with management decisions (Kenny, 2004). Stocking rate should be based on average long-term end-of-season standing crop values for an operation to remain productive and sustainable. The procedure for calculating stocking rates can be used on either forests or rangelands. Stocking rates are based on the amount of forage that is standing at the end of the growing season in an ungrazed condition. End-of-season standing crop is not total production because much of the production has been 78 lost to decomposition and insects. Actual forage production is often twice as large as the end-of-season standing crop. Forage production information is useful but is very time consuming to obtain. That is why end-of-season standing crop is used for estimating stocking rate (Redfearn and Bidwell, 2004). The common mathematical process used to estimate stocking rate from herbage biomass is to determine the amount of herbage available during the grazing season, which is the average value of the mean monthly standing herbage biomass values for the grazing-season months. The mean monthly standing herbage biomass should be determined by clipping and weighing the dry herbage from each pasture and averaging the weights over several years. The general monthly herbage values on the herbage weight are the averages of herbage production on well-managed pastures during years with normal precipitation (Manske, 2004). Recommended stocking rates are based upon results from grazing research, local experience and clipped-plot yields (Lacey and Taylor, 2005). The recommended stocking rates for rangelands should also be based on moderate utilization (economic long-term optimum) of the annual forage standing crop and assume uniform grazing distribution. It is also assumed that 50% of the annual peak standing crop can be removed from the ecological site without negatively affecting the plant community relative to species abundance or for beef cattle production. This is the origin of the “take half and leave half” rule-ofthumb that is often used. This is also the source of difference in stocking rate management between rangeland and introduced forages (Redfearn and Bidwell, 2004). The old rule of thumb “take half, leave half” is well publicized, but may not be well understood. This rule 79 applies to average annual forage production. It does not mean that half the forage can be allotted to grazing animals. Part of what is taken will go to the animals, but part will disappear through trampling, decay and insect damage. This disappearance is usually about 25 percent of the average annual production. Therefore, only 25 percent is left for the grazing animal (Lyons and Machen, 2005). Ideally, such objections can be taken into account for any individual carrying capacity estimated by accepting that it has to be determined on a case - by - case basis in the field. Once one knows the size of the grazing and browsing animals, the biomass production of the area, the pattern of range management, and so on, he/she can - so this argument goes - produce a site specific stocking rate estimated for the range area under consideration. But, it cannot pack livestock into a given rangeland, without at some point deteriorating that range demonstrably. Surely, biomass production is going down on rangelands precisely because stocking rate has been exceeded for so long, even taking into account factors such as drought and climate change (Hardesty et al., 1993). The rationale and measures of rangeland carrying capacity are increasingly criticized. It seems that even under environmental conditions of great certainty, the notion of rangeland equilibrium would still be ambiguous and confused. Moreover, since environmental conditions are highly uncertain for the dry rangelands of the world such as Iran, current understanding of rangeland equilibrium turns out to be all the more questionable. There is no workable, practical “equation” for rangeland management in general, and carrying capacity in particular (Roe, 1997). Similar problems exist in other field of sustainable rangeland management. Here, we have 80 observed a number of publications which used fuzzy logic as a valuable tool (Andriantiatsaholiniaina, 2001; Cornelissen et al., 2001; de Kok et al., 2000; Dunn et al., 1995; El-Awad, 1991; Ferraro, et al., 2003; Gowing et al., 1996; Marks et al., 1995; Phillis and Andriantiatsaholiniaina, 2001; Sam-Amoah and Gowing, 2001; Sicat et al., 2005). In these studies, fuzzy logic is used to construct a model for evaluating sustainability in different areas. These models promise to be a valuable tool in evaluating the sustainability in general and equilibrium specifically in this dissertation. The purpose of the next chapter is to design a fuzzy model based on the experts’ knowledge for solving the mis-management of the Iranian rangelands. 81 Chapter Five Application of Fuzzy Logic in Sustainable Rangeland Management Albert Einstein: "So far as the laws of mathematics refer to reality, they are not certain. And so far as they are certain, they do not refer to reality" (Kosko, 1993). 5.1. Fuzzy Logic: A shifting paradigm As we discussed in Chapter 4, there is a conflict that arises between consumption (economic dimension) and conservation (environmental dimension) in rangeland management. In general, interdisciplinary problem-oriented projects have undergone substantial change over the last few decades (Barendse and van der Hoek, 1996) since, social scientists (social dimension) have tried to reconcile such a conflict (Azadi et al., 2003) by creating interdisciplinary experts’ teams, which are constructed by cooperation of technical scientists (Shaner et al., 1982). This approach, which started in 1970s, has labelled differently1. The interdisciplinary experts’ teams usually require not only thorough specialist knowledge about problem sources, mechanisms and options for solution(s), but also the ability to integrate knowledge from different fields (Barendse and van der Hoek, 1996) to introduce new methods of grazing management for reaching sustainability (Azadi et 1 . FSA: Farming System Analysis, FSAR: Farming System Adaptive Research, FSCR: Farming System Component Research, FSBDA: Farming System Base-line Data Analysis, NFSD: New Farming Systems Development, FSRAD: Farming Systems Research and Agricultural Development FSRE: Farming System Research Extension (Sands, 1986). 82 al., 2003). They try to develop a systematic approach for better understanding the complex situation of pastoralists. But as the interdisciplinary team tries to implement various ideas belonging different experts for the same situation (FAO, 2004c; Shaner et al., 1982; Zilberman and Alix, 2005), the members often fail to reach an identical understanding of this complex situation (Shahvali and Azadi, 1999). More clearly, the main problems involved in interdisciplinary teams are as follows: Misunderstandings and misconceptions because of different disciplinary languages (jargons) (Pickett et al., 1999); Overestimating or underestimating the contribution of their own discipline in analyzing and solving problems (Heemskerk et al., 2003); Lack of knowledge about integrating knowledge and insights from different disciplinary fields (Barendse and van der Hoek, 1996); and Lack of knowledge about integrating methods and techniques (Barendse and van der Hoek, 1996; Pickett et al., 1999). Thereby, when presenting their ideas concerning the sustainability in rangeland management, they usually: (Azadi et al., 2003) Select rangeland equilibrium indicators differently; Weigh the indicators unequally; and Assess them in different ways. 83 The above-mentioned problems exist in other field of SD. To solve these problems we use fuzzy sets theory to assess sustainability, and to deal with different experts’ knowledge in rangeland management. Fuzzy logic is one of the fastest growing methodologies in systems engineering (Grint, 1997). In 1965, Lotfi A. Zadeh laid the foundation of fuzzy logic theory. Since then and, especially, after the announcement of the first fuzzy chips in 1987, the literature on both theory and applications of fuzzy logic has been growing (Berkan and Trubatch, 1997). In a broad sense, fuzziness is the opposite of precision. Everything that cannot be defined precisely (that is, according to some broadly accepted criteria or norms of precision) and everything that has no clearly described boundaries in space or time is considered a bearer of fuzziness. In a narrow sense, fuzzy logic relates to the definition of fuzzy sets as proposed by Zadeh (Zimmermann, 1996): sets, the belongingness to which is measured by a membership function whose values are between 1 (full belongingness) and 0 (non- belongingness). According to ‘Principle of Incompatibility’, “as the complexity of a system increases, human ability to make precious and relevant (meaningful) statements about its behaviour diminishes until a threshold is reached beyond which the precision and the relevance become mutually exclusive characteristics” (Zadeh, 1973, p. 29). It is, therefore, that fuzzy statements are the only bearers of meaning and relevance. Zadeh used this principle for extending the applicability of his fuzzy sets theory and fuzzy logic to the analysis of complex systems. It is now realized that complex real-world problems require intelligent systems that combine knowledge, techniques, and methodologies from various sources. These intelligent systems are 84 supposed to do better in changing environment, and explain how they make decisions or take actions (Jang et al., 1997). Ecological studies are known to be complex in nature (Silvert, 1997) and therefore fuzzy logic seems to be an appropriate technique to solve the dichotomy that is inherent in sustainability of natural resources (Andriantiatsaholiniaina, 2001; Cornelissen et al., 2001; Dunn et al., 1995; Marks et al., 1995). 5.2. Foundations of fuzzy logic The mathematics of fuzzy sets and fuzzy logic is discussed in detail in many books (e.g., Lee, 1990; Driankov et al., 1996; Ruspini et al., 1998; Zimmermann, 1996). Here, we provide only a very basic aspect of the mathematics of fuzzy logic. 5.2.1. Crisp models In quantitative sciences where mathematical models are used for analyzing real-world phenomena, (stochastic) variables are introduced having a ‘well-defined’ meaning. During their scientific work, the corresponding scientists apply mathematical tools from calculus, from the theory of differential equations, from discrete mathematics, from (vector) algebra, from numerical methods, from (complex) function theory, and more (van den Berg, 2004). The resulting models offer an ‘idealized’ world, an ‘objective and structured reality’ with, hopefully, rather general validity. Uncertainty is usually described in probabilistic and statistical terms like probabilities on crisp events (i.e. events that do occur or do not occur at all), expected values, statistical tests (that are either rejected or not rejected), interval estimations, et cetera (Zimmermann, 1996). Propositions within these approaches are 85 usually supposed to be either true or false (and sometimes unknown). In line with this way of working as applied in physics, chemistry, econometrics, and other ‘hard sciences’, the first knowledge-based systems developed in the community of Artificial Intelligence were founded on the ‘physical symbol system hypothesis’ expressing that symbols (and only symbols) can represent states of the world and statements about the world. Again, the only ‘epistemological commitments’ allowed for these statements are either true, false or unknown. The physical symbol system hypothesis has still to be proven (van den Berg, 2004). 5.2.2. Boolean vs. Fuzzy Three hundred years B.C., the Greek philosopher, Aristotle came up with binary logic (0,1), which is now the principle foundation of Mathematics. It came down to one law: A or not A, either this or not this. For example, a typical rose is either red or not red. It cannot be red and not red. Every statement or sentence is true or false or has the truth-value 1 or 0. This is Aristotle's law of bivalence and was philosophically correct for over two thousand years (Kosko, 1993). Two centuries before Aristotle, Buddha, had the belief which contradicted the black-and-white world of worlds, which went beyond the bivalent cocoon and see the world as it is, filled with contradictions, with things and not things. He stated that a rose, could be to a certain degree completely red, but at the same time could also be at a certain degree not red. Meaning that it can be red and not red at the same time. Conventional (Boolean) logic states that a glass can be full or not full of water. However, suppose one were to fill the glass only halfway. Then the glass can be half-full and half-not-full. 86 Clearly, this disproves Aristotle's law of bivalence. This concept of certain degree or multivalence is the fundamental concept, which propelled Zadeh at the University of California in 1965 to introduce fuzzy logic. The essential characteristics of fuzzy logic founded by him are as follows (Abdul Aziz, 1996): • In fuzzy logic, exact reasoning is viewed as a limiting case of approximate reasoning, • In fuzzy logic everything is a matter of degree, • Any logical system can be fuzzified, • In fuzzy logic, knowledge is interpreted as a collection of elastic or, equivalently, fuzzy constraint on a collection of variables, and • Inference is viewed as a process of propagation of elastic constraints. The third statement hence, defines Boolean logic as a subset of Fuzzy logic. 5.2.3. Towards soft computing It is clear from history that the hard sciences have been and still are quite successful in many areas. Based on this success, they have obtained a strong and predominant position and many scientists working in this field seem to believe that their approach of crisp, twovalued logical, precise mathematical modeling where uncertainty is modeled within a probabilistic, statistical framework, is the one and only true, applicable approach (van den Berg, 2004). 87 For several reasons, however, like ‘irrelevance’ and ‘complexity’, one may doubt whether hard computing is always the right tool. Considering ‘Principle of Incompatibility’, we try to answer the following series of questions related to problems with increasing complexity. They may convince you of the validity of Zadeh’s principle: (1) How sustainable is a rangeland? (2) Which measures are important for a sustainable rangeland management? (3) How much is the stocking rate of a sustainable rangeland? (4) How much is its plantation density? (5) How much annual rainfall is needed in a sustainable rangeland? (6) How many pastoralists’ families can live in? Another, every-day, example related to the irrelevance of information may also be illustrative. This example shows that offering some short, incomplete information can be much more relevant than communicating an extensive and precise message: if you are sauntering on a road and a car is approaching you with high speed, the warning “A car of length 4.65 m and height 1.54 m with mass 1348.7 kg is approaching you with a speed of 56.645 kmph” is probably much less relevant than the quite fuzzy cry: “Look Out!!”. From the context, the saunterer will probably immediately understand the meaning of this message and undertake appropriate action (by running away). In the world of Artificial Intelligence, similar lessons have been learned. While, implementing systems based on ideas like the abovementioned physical symbol hypothesis, problems arose related to the modeling of the likeliness of a certain conclusion and to the lack of robustness and flexibility. Apparently, the tools as made available by the hard sciences also have their limitations when trying to apply 88 ‘intelligent techniques’. In several cases, it has been shown that alternative approaches with fuzzy or other ‘vague’ ingredients work better. Successful fuzzy modeling projects exist since 1975 on topics like automatic control, printed character recognition, target selection for marketing purposes, financial modeling, SD, and more (van den Berg, 2004). 5.2.4. Towards fuzzy sets Let U be a collection of objects u which can be discrete or continuous. U is called the universe of discourse and u represents an element of U. A classical (crisp) subset C in a universe U can be denoted in several ways like, in the discrete case, by enumeration of its elements: C = {u1, u2 ,… , uP} with ∀i: ui ∈ U. Another way to define C (both in the discrete and the continuous case) is by using the characteristic function χF: U→{0, 1} according to χF (u) = 1 if u ∈ C, and χF (u) = 0 if u ∉ C. The latter type of definition can be generalized in order to define fuzzy sets. A fuzzy set F in a universe of discourse U is characterized by a membership function µF which takes values in the interval [0, 1] namely, µF: U→[0, 1]. 5.2.5. Operators on fuzzy sets Let A and B be two fuzzy sets in U with membership functions µA and µB, respectively. The fuzzy set resulting from operations of union, intersection, etc. of fuzzy sets are defined using their membership functions. Generally, several choices are possible: 89 Union: The membership function µ A∪ B of the union A∪B can be defined by ∀u : µ A∪B = max{µ A (u), µ B (u)} or by ∀u : µ A∪B = µ A (u) + µ B (u) − µ A (u) µ B (u). Intersection: The membership function µ A∩ B of the union for all A∩B can be defined by ∀u : µ A∩ B = min{µ A (u ), µ B (u )} or by ∀u : µ A∩B = µ A (u). µ B (u). Complement: The membership function of the complementary fuzzy set Ac of A is usually defined by ∀u : µ Ac (u) = 1 − µ A (u) . 5.2.6. Linguistic variables Fuzzy logic enables the modeling of expert knowledge. The key notion to do so is that of a linguistic variable (instead of a quantitative variable) which takes linguistic values (instead of numerical ones). For example, if stocking rate (SR) in a rangeland is interpreted as linguistic variable, then its linguistic values could be one from the socalled term-set T(SR) = {low, medium, high} where each term in T(SR) is characterized by a fuzzy set in the universe of discourse, here, e.g., U = [0, 5]. We might interpret low as a “stocking rate of less than approximately 1.5 aum/ha”, medium as a “stocking rate close to 2 aum/ha”, and high as a “stocking rate of roughly more than 2.5 aum/ha” where the class boundaries are fuzzy. These linguistic values are characterized by fuzzy sets whose membership functions are shown in Fig. 5.1. 90 µi(SR) 1 low medium high SR (aum/ha) 0 1 2 3 5 Fig. 5.1. Diagrammatic representation of the linguistic variable stocking rate in a rangeland having linguistic values low, medium, and high defined by a corresponding membership function. 5.2.7. Knowledge representation by fuzzy IF-Then rules Fuzzy logic is a scientific tool that permits simulation of the dynamics of a system without a detailed mathematical description. In an expertdriven approach, knowledge is represented by fuzzy IF-THEN linguistic rules having the general form If x1 is A1 AND x2 is A2 Λ AND xm is Am THEN y is B, where x1, … , xm are linguistic input variables with linguistic values A1, … , Am respectively and where y is the linguistic output variable with linguistic value B. To illuminate we consider animal unit and plantation density as the principal factors for having equilibrium. Then the relevant fuzzy rules could be: - IF amount of animal unit is low AND plantation density is poor THEN equilibrium is medium. - IF amount of animal unit is medium AND plantation density is poor THEN equilibrium is weak. - IF amount of animal unit is high AND plantation density is poor THEN equilibrium is very weak. 91 5.2.8. Architecture of fuzzy systems Fuzzy Inference Systems or, shortly, Fuzzy Systems (FSs) usually implement a crisp input-output (IO) mapping consisting of basically four units, namely • A Fuzzifier transforming crisp inputs into the fuzzy domain, • A Rule Base of fuzzy IF-THEN rules, • An Inference Engine implementing fuzzy reasoning by combining the fuzzified input with the rules of the Rule Base, • A Defuzzifier transforming the fuzzy output of the Inference Engine to a crisp value (Fig. 5.2). Fig. 5.2. Building blocks of a Fuzzy Inference System (FIS). In some practical systems, the Fuzzifier or the Defuzzifier may be absent namely in cases where fuzzy input data are available or the fuzzy system output can be interpreted directly in linguistic terms. Corresponding "approximate reasoning techniques" are available (see, e.g., Jang et al., 1997). 5.2.9. Fuzzy reasoning Probably, the hardest part to understand is the precise way fuzzy reasoning can be implemented. An extensive discussion of this topic is outside the scope of this dissertation so we limit ourselves here to 92 present just the basic idea. Classical logic is our starting point using the classical reasoning pattern ‘modus ponens’: Given fact “x is A” and rule “IF x is A, THEN y is B”, we conclude “y is B”. “Applying fuzzy reasoning, classical modus ponens can be generalized to an ‘approximate reasoning’ scheme of type.” Given fact “x is A' ” and rule “IF x is A, THEN y is B”, we conclude that “y is B' ”. Here, the assumption made is that the closer A' to A, the closer will B' be to B. It turns out that especial combinations of operations on fuzzy sets like ‘max-min’ and ‘max-product’ composition can fulfill this requirement. The complete fuzzy reasoning in a FS can be set up as follows: 1. The fuzzification module calculates the so-called ‘firing rate’ (or degree of fulfillment) of each rule by taking into account the similarity between the actual input A' defined by membership function µA'(x) (and in case of a crisp input xp defined by the value µA(xp) and the input A of each rule defined by membership function µA(x). 2. Using the firing-rates calculated, the inference engine determines the fuzzy output B' for each rule, defined by membership function µB'(y). 3. The inference engine combines all fuzzy outputs B' into one overall fuzzy output defined by membership function µ(y). 4. The defuzzification module calculates the crisp output yp using a defuzzification operation like ‘centroïd of gravity (area)’. 93 For a treatment in depth on FSs, its construction and corresponding reasoning schemes (including the most popular systems like Mamdani (Mamdani and Gaines, 1981) and Tagaki-Sugeno Fuzzy Models (Tagaki and Sugeno, 1985), we refer to the above-mentioned textbooks. 5.3. Theoretical frameworks 5.3.1. Architecture the EAFL model Considering sustainability in rangeland management, we develop a model called Equilibrium Assessment by Fuzzy Logic (EAFL) which provides a mechanism to approach eqilibrium by assessing Right Rate of Stocking. The scheme of the EAFL model applying approximate reasoning to come up with RRS is shown in Fig. 5.3. In general, the following basic steps should be done to construct a fuzzy model (van den Berg, 2004): 1. Determining the relevant input and output variables; 2. Defining linguistic values; 3. Constructing membership function; 4. Determining the fuzzy rules; 5. Computing degree of membership of crisp inputs; 6. Determining approximate reasoning; 7. Computing crisp output (defuzzify); and 8. Assessing the model performance. 94 Determining inputs Defining linguistic values Input1 L.V.(I1) Step 2 IF – THEN rules Step 4 µi(I1p), µj(I2p), µk(I3p) Step 5 Given: µi(I1p), µj(I2p), µk(I3p) ; derive:µi(RRSp) Step 6 RRSp = COG µ(RRS) Step 7 Assessing & fine-tuning Step 8 Computing degree of membership of crisp inputs Assessing the model performance L.V.(I3) Step 1 Step 3 Determining fuzzy rules Computing crisp output (defuzzy) L.V.(I2) Input3 1 µi(I1), µj(I2), µk(I3) Constructing membership functions Determining approximate reasoning Input2 Fig. 5.3. Scheme of development of the EAFL model applying approximate reasoning to assess the Right Rate of Stocking (RRSp) based on the inputs values (I1p, I2p, and I3p)2 (adapted from Cornelissen, 2003, p.51). 1. To simplify the model, we have shown three inputs, however, the number can be higher or lower. 2. The grey color notifies the fuzzy parts of the model. 95 5.3.2. Architecture of multi-fuzzy model While fuzzy specialists usually use homogeneous experts’ knowledge to construct fuzzy models, it is much more difficult to deal with knowledge elicited from a heterogeneous group of experts. They usually try to elicit and deal with homogenous experts’ knowledge and hardly refer to heterogeneous experts. Experts’ knowledge, however, is influenced by individual perspectives and goals (Ford and Sterman, 1998). This issue especially holds in the area of the sustainable rangeland management. One way to deal with the diversity of opinions is to develop a fuzzy system for all experts and to combine all these so-called primary systems into one multi-fuzzy model. When constructing a fuzzy model, an important consideration is how to deal with differences in personal experience. The effect of these differences is assumed to be smaller in a homogeneous (e.g. only pastoralists) than in a heterogeneous group (e.g. different experts). As experts have graduated in different disciplines, they may come to a different evaluation of sustainable rangeland management than, for example, pastoralists. Such differences, however, are not necessarily disadvantageous. A heterogeneous group of experts, can be an advantage over a homogeneous group through considering all knowledge and, compensating for dissenting points of view by more liberal ones (Cornelissen, 2003). However, the heterogeneity in experts’ knowledge makes unclear decisions in rangeland management for reaching equilibrium in practice. Here, based on three Mamdani-type of fuzzy models (three EAFL models), we have designed a specific multi-fuzzy model (Fig. 5.4) to assess the final Right Rate of Stocking (RRSf). The following basic steps are also suggested to enable to calculate the final crisp output: 96 1. Constructing several EAFL models (described in the last section); based on the experts’ knowledge resulting from the semi-structured interviews (described in the next chapter); 2. Computing, for several typical cases, the crisp primary outputs of the models and comparing them; and 3. Combining different outputs using a voting process and calculating the final crisp output. The architecture of the multi-fuzzy model is depicted in Fig. 5.4. 97 Ii Ii Rules Base (ni) Inference Engine 2 Rules Base (ni) Inference Engine 3 W1 RRSi W2 RRSi Voting Process Defuzzifier 1 Inference Engine1 Defuzzifier 2 Ii Rules Base (ni) Defuzzifier 3 Ii Fuzzifier 1 Ii Ii Fuzzifier 2 Model 2 Ii Ii Fuzzifier 3 Model 1 Ii Model 3 Mamdani fuzzy systems W3 RRSi Fig. 5.4. Architecture of the multi-fuzzy model to deal with different experts’ knowledge. The primary outputs RRSi (i = 1,2,3,…n) of three Mamdani fuzzy systems having different input variables and different number of rules base (ni), are combined into output value RRSf using a weighted voting process to calculate final output. 98 3 RRSf = ∑ w RRS i i =1 i To apply the above holistic approach, we conducted a multiple case study in three different areas of the Fars province in Southwest Iran. In the next chapter, we give some details concerning the research methodology of this study. 99 Chapter Six Research Method 6.1. The population of study It can be stated that the majority of the Iranian people were engaged in pastoral subsistence and animal breeding for a long time (Sogol, No date). One third of the total area in Iran (164 million hectares) is unusable for any purpose other than pastoralism (Emadi, 2003). Today, there are over 1.5 million nomads in Iran. Many of nomad tribes such as the Kurds, Bakhtiyaris (Bactrians), Lors, Guilaks, and the Baloochs are descendants of the original invaders who came from Central Asia to settle in the Iranian plateau (Iran Rozaneh, 2003). Most of the tribes of Central Iran are pure Aryan, while others such as Khoozestan and Khorasan's Arabs, Qashqai, Turkaman (decendants of Mongols), Shahsavan and Afshar's tribes in Azarbaijan had ancestors who passed through Iran (Travel Explorations, 2004). By 1920, nomadic pastoral tribes constituted over a quarter of the Iranian population. Their number declined sharply as a result of forced settlement between 1920s and 1930s. Continued pressures as well as the lure of the cities and settled life have resulted in a further sharp decline since 1960s (Iran Rozaneh, 2003). A public census conducted on the Iranian nomadic tribes in 1987 puts the number of tribes in 96 tribes and 547 clans consisting of 180,223 families or 1,152,099 people (597,774 men and 554,325 women). Holding the results of next public census, some 96 independent tribes 100 and 547 clans have been registered in Iran in 1988. Iranian tribesmen have been scattered throughout the country, mainly in the provinces of East and West Azarbaijan, Kordestan, Gorgan and Gonbad, Lorestan, Fars, Kerman, Khorasan, and Sistan and Balouchestan. In 1991, public census on the Iranian population does not mention to the tribesmen. It rather mentions urban and rural residents and non-resident people. It is not known whether the non-residents, as mentioned in 1991, are the same as tribesmen (Namey-e-Otaq-e-Bazargani, 1996). According to the Statistic Centre of Iran in 1996, the number of nonresidents is approximately 2,110,406. Though the tribes and clans are scattered in distinct areas of the country, this in itself denotes the influence of the central rule on such realms. At times, due to political reasons, tribes were compelled to migrate to other regions. Such an example can be Kordestan's Kurds who migrated to the territory in Northern Khorasan. But it can be stated that each tribe withheld its own cultural and social traditions wherever they resided; such as Shahsavan in Northern Azarbayjan and Kordestan's Kurds, even so, between Qashqaie and Bakhtiyari's tribes. Historical surveys reveal that some of the Iranian tribes have a common ancestor. A large portion of the tribes in Central and Western Iran has Lor dialect. These are divided in two groups; Lor-e-Bozorg (Greater Lors) and Lor-e-Koochak (Smaller Lors). Branches of these tribes were decamped to the mountainous regions in Central Iran. Tribes such as the Bakhtiyari, Kohgilooyeh, Mamasani and Boyer Ahmad are of this group, and yet are completely distinct from each other. During the Safavid era, Afshar's tribes were decamped from Khorasan to Azarbayjan, and still another group to Kohgilooyeh and Khoozestan. By the Fars conquest tribes leaded Aqa Mohammad Khan Qajar in 101 1206 A.H., 12,000 families that proved rebellious were decamped from Shiraz to Tehran. During Nasereddin Shah's regime, Hezareh's tribe was decamped to Khorasan, but due to unrest and turmoil, was compelled to scatter in smaller groups. Formerly, this dispersion depended solely on the acquirement of pastoral vicinities. But gradually this gained a political aspect, thus conserving limits and distinctions as to the jurisdiction of tribes. Currently, the tribes are dispersed in the following regions in Iran (Sogol, No date): • North and Northwestern; the provinces of Golestan and Khorasan - comprising of various clans such as the Turkaman tribes. • Northwestern; Shahsavan, Arasbaran, Afshar-e-Qezelbash, Garahgozloo and various clans of Khamseh's tribe. These are within the limits of Eastern and Western Azarbayjan, Hamadan, Ardabil and Zanjan. • Western; comprising of those having a Kurdish dialect, Kalhor, Sanjabi, Gurkani and... They reside in the provinces of Kermanshah, West Azarbayjan and Kordestan. • Southwest and Southern; comprising of various clans such as Khamseh, Qashqaie, Arab and Lor-e-Koochak which are settled in the provinces of Fars, Khoozestan and Lorestan. • Southeastern; comprising Baloochi tribes residing in the province of Sistan and Baloochestan. • Center; these are namely Bakhtiyari, Boyer Ahmad, Doshman Ziyari, Charam, Bavi, Bahmehyi, Tayebi, Mokran and... which reside within the limits of the provinces of Chahar Mahal and Bakhtiyari, Khoozestan, Kohgilooyeh and Kerman. 102 • Eastern and Northeastern; which comprising of various clans settled in the province of Khorasan. As noticed, in South Iran, especially the Fars province is rich in pastoral groups usually specialised in sheep and these are described in a number of monographs (e.g. Barfield, 1981; Barth, 1961; Bates, 1973; Black-Michaud, 1986; Irons, 1975)1. For centuries, the Fars province has been a multi-ethnic region, in which tribal and pastoral nomadic groups compose a large part of the population as follows (Qashqai.net., 2003): Qashqaie's tribe consists of the following clans: Darreh Shouri (5,265 families), Kashkooli Bozorg (4,862 families), Shesh Boluki (4,350 families), Kashkooli Kuchak (650 families), Qaracheh (430 families), Safi Khani (235 families), Rahimi (370 families), Farsi Madan (1,505 families), and Amaleh (5,397 families). The above-mentioned pastoral nomads move - with their herds of sheep and goats - between summer rangelands in the higher elevations of “Zagros” in North of Shiraz and winter rangelands at low elevations in South of Shiraz. Their migration routes are considered to be among the longest and most difficult in all Iranian pastoral tribes (Oldcarpet, 2005). 6.2. The area of study In this study, we have focused on three different regions of the Fars province in Southwest Iran: first, Cheshme-Anjir from Shiraz county which covers 2575 hectares, 3200 livestock and 12 pastoral families; 1 . It should also be noted that a significant body of literature on pastoral nomadism such as the works of Barth (1961); Beck (1986); Digard (1990); Fazel (1971); Irons and Dyson-Hudson (1972); Mortensen (1993); Swee (1981) and Tapper (1979, 1983 and 1997) contributed to the development of better understanding of the socio-economic organization and migratory patterns of the Iranian tribal society. 103 second, Morzion from Sepidan county having 2000 hectares, 1570 livestock and 19 pastoral families; and third, Kheshti from Lamerd county with 6900 hectares, 3804 livestock and 20 pastoral families. The regions have different climatic and geographical conditions (Table 6.1). 104 Table 6.1. General information of the three regions of the study. Region Morzion Cheshme-Anjir Kheshti Area (ha) 2000 2575 6900 No. of Livestock (aum) 1570 3200 3804 No. of Pastoral Families 19 12 20 Average Temperature (°C) 11.9 15.0 22.8 105 Average Annual Rainfall (mm) 750 317 244 Soil Texture Sandy Loam Sandy Loam Sandy Loam Clay Height (m) (from see level) 2300 1400 770 The main reason for selecting these regions was the management activities done by the Natural Resources Administration of Fars Province (NRAFP) to balance livestock number and rangeland conditions in these regions. 6.3. Research method There is a growing body of opinion that argues that qualitative research, including the case study, has an important place among the variety of research methods available to the researcher. ‘Yet the traditional case study still remains firmly within the domain of the qualitative researcher’ (Tesch, 1990; p. 69). Burns (1994) has argued that the case study, because of its intense nature and its ability to generate rich subjective data, may generate more intensive research. The case study allows for in-depth probing of phenomena, as Burns stated: ...typically involves the observation of an individual unit... to qualify as a case study, it must be a bounded system, an entity in itself. A case study should focus on a bounded subject/unit that is either very representative or extremely atypical (Burns, 1994; p. 312). Many approaches have been taken in terms of the development of case studies. Tesch (1990) argued that a case study is an ‘intensive and detailed study of one individual or a group as an entity, through observation, self-reports, and any other means’ (Tesch, 1990; p. 39). Others see the case study in more elaborate terms. The case study is judged to be appropriate for this study somehow it would allow in-depth interviews of issues that are important in sustainable rangeland management. At the same time, it would impose structure on the research and allow for judgment (Niblo and Jackson, 106 1999). Generally, case studies are the preferred strategy when "how" or ‘why" questions are being posed, when the investigator has little control over events, and when focus is on a contemporary phenomenon within some real-life context. In more precise definition, a case study is an empirical inquiry that: (1) investigates a contemporary phenomenon within its real-life context; when (2) the boundaries between phenomenon and context are not clearly evident; and in which (3) multiple sources of evidence are used (Yin, 1989; p. 13). This research method explores the problems for the qualitative researcher in trying to explore a hidden phenomenon. The questions should be asked in depth and be probing (Niblo and Jackson, 1999). ‘These “how” and “why” questions, capturing what researcher is really interested in answering, lead him\her to the case study as the appropriate strategy...’ (Yin, 1989; p. 30). That is to say, the researcher has to leave room for material and information to emerge during the course of the research project. Nevertheless, Yin also argued that every exploration still should have a purpose. Some basic skills that need to be mastered by case studier are (Yin, 1989): • being able to ask questions, • interpreting the answers, • being a good "listener", • being adaptive and flexible, • having a firm grasp of the issues being studied, and • being unbiased by preconceived notions. 107 6.3.1. Multiple-case study The specific approach implemented in this study is a multiple-case study in the framework of an exploratory research design to find sustainability indicators in rangeland management in the views of selected experts. For each region of the research1, a case study involving both observation and interviews with pastoralist’ experts was conducted. The implementation of multiple-case studies allowed us to gather more evidences and to make comparisons between regions with different socio-economic and environmental conditions. The approach is open-ended in that it allows researchers to go as far as possible before exhausting the sources of information. Yin has discussed the difference between ‘replication, not sampling logic, for multiple-case studies’ (Yin, 1989; p. 53). Burns (1994; p. 316) also elaborated on the concept of replication: “A collection of case studies, which is called the multiplecase study, is not based on the sampling logic of multiple subjects in an experimental design. If the cases are not aggregated it is convenient to apply the term case study to such an investigation.” For each region, then, an extensive case study was conducted which provided the materials of the present research. The first step involved the preparation of the detailed case studies through procedures of direct observation and extensive in-depth interviewing and second, abstracting from observations and interview materials in each of these which are drawn upon for the explication of material relevant to the questions which guided this research. 1 . Since, we will finally estimate the RRS for each "pasture" in the regions, the pasture can be considered as the case or the unit of analysis of this study. 108 6.4. Sampling method Since fuzzy models usually use the expert knowledge, it is important to identify an expert properly. For example, there are some differences between stakeholders and experts (Cornelissen, 2003) and also differences between experts and those people who are selected to interview based on a few personal contacts, or on the basis of availability during a short-term period (Davis and Wagner, 2003). In this dissertation, an expert is defined as a person whose knowledge in a specific domain (e.g. equilibrium in a rangeland) is obtained gradually through a period of learning and experience (Bromme, 1992 and Turban, 1995 in Cornelissen, 2003). To call a holistic approach, both homogeneous and heterogeneous experts were selected based on the ‘socio-metric’ method. According on this method, concerned information is obtained directly from key informant experts who are nominated by the majority of stakeholders (Ortega, 2002). Totally, in this study, based on several discussions with the stakeholders of NRAFP at the first round of study, 9 experts in three different categories nominated by them were interviewed (Table 6.2). Table 6.2. Some personal characteristics of 9 nominated experts at the first round of study. Pastoralist experts Expert no. Age (yr.) Education level Discipline/Job Experience (yr.) 1 55 Diploma Pastoralist All his life 2 63 Reading & Writing Pastoralist All his life 3 65 Reading & Writing Pastoralist All his life Administrative experts Expert no. Age (yr.) Education level Discipline Experience (yr.) 1 33 M.Sc. Desert Management 8 2 45 M.Sc. Animal Husbandry 18 3 48 M.Sc. Rural Development 10 4 47 M.Sc. Admin. Management 16 Researcher experts Expert no. Age (yr.) Education level Discipline Experience (yr.) 1 42 Ph.D. Animal Husbandry 8 2 48 Ph.D. Rage management 10 109 During three rounds of data collection in a period of 18 months, by conducting 9 main interviews and 21 follow-ups, we were led to remove one administrative expert and both two researchers by discounting others or even themselves1. In other words, they eliminated others or even themselves and pointed to the rest who are more expertized in this field. It should be noted again that the experts who have different personal characteristics are expertized differently within and between groups according to their last and present background. NRAFP, for instance, has hired a large number of experts who hold in bachelor and master degrees in different disciplines of rangeland management. 6.5. Data collection and applied techniques In order to construct a multi-fuzzy model, several semi-structured interviews were held with elites2 who are called, in fuzzy studies, experts. As Jones (1985) described, a semi-structured interview is: • a social interaction between two people (the researcher and one of his\her experts); • in which the interviewer (researcher) initiates and varyingly controls the exchange with the respondents (the experts); • for the purpose of obtaining quantifiable and comparable information (defining sustainability indicators); and • relevant to an emerging or stated hypothesis (if-then rules for making the balance between the different levels of the indicators). 1 . See appendix 1. . Elite interview conducts with prominent people. This technique provides valuable information which can be obtained because the position of interviewees; Elite also can provide overall view his/her (Marshall and Rossman, 1995). 2 110 The entire script was written ahead of time, with an eye to an almost total standardization of the interview from one expert to the next. The standardized, open-ended interview was used when it was important to minimize the variation of the questions posed to interviewees. This helps to reduce the bias that can occur from having different interviews with different experts (Patton, 1987). The open-ended questionnaire was used to conduct interviews included a set of questions which were carefully worded and arranged for the purpose of taking each expert through the same sequence and asking him the same questions with essentially the same words (Gamble, 1989). More specifically, the main questions were conducting the experts i) to introduce the main indicators of sustainability in range management, ii) to define the labels (linguistic values), iii) to determine the fuzzy ranges of each value label, and iv) to express the fuzzy if-then rules. 6.5.1. Data analysis The qualitative techniques were used to analyze the elicited knowledge of experts. The main techniques included open and axial coding to find the main indicators (Strauss and Corbin, 1990). The Matlab Fuzzy Toolbox (version 7) was used to construct the fuzzy model. Finally, to deal with different experts’ knowledge, we used Microsoft Excel 2003. More details and extensive discussions regarding data analysis are given in the next chapter. 111 Chapter Seven Fuzzy Analysis and Discussion This chapter presents the analysis of data using fuzzy logic and an extensive related discussion. To come up with a holistic approach in the area of sustainable rangeland management, based on the theoretical frameworks presented in Chapter 5, we will first construct a fuzzy model called EAFL and explain its eight steps, and second, to deal with different experts’ knowledge, we will develop a multi-fuzzy model and introduce voting methods including (un)supervised techniques. 7.1. Development the EAFL model 7.1.1. Determining the relevant input and output variables Equilibrium between stocking rate and plantation density is difficult to define but many experts recognize that it is a function of three major components (inputs) which are: Stocking Rate in a pasture (SR), the amount of Plantation Density per hectare (PD), and the Number of Pastoralists in a pasture (NP). 7.1.2. Defining linguistic values In the EAFL model, the linguistic values of each variable are recapitulated in Table 7.1. 112 Table 7.1. Linguistic values used in the EAFL model. Variable Linguistic values Stocking Rate (SR) low, medium, high Plantation Density (PD) poor, acceptable, rich Number of Pastoralists (NP) low, medium, high 7.1.3. Constructing membership functions Based on Pedrycz (1994; 2001) the experts' experience and knowledge can be considered as a guide to define all membership functions which in this study none of them are expressed continuous values. Therefore, both triangular and shoulder shapes were suggested. The suggestion was based on special range of values, which are stated by the experts for each linguistic value. All experts proposed the left-shoulder shape for the smallest and the right-shoulder for the largest linguistic values and the triangular for the rest (Fig. 7.1). 113 µi(SR) a low 1 medium high SR (aum/km2) 50 0 µj(PD) 70 b poor 1 55 acceptable rich PD (tn/km2) 0 10 30 µk(NP) 1 50 c low medium high NP (p/km2) 0 0.1 0.3 0.5 Fig. 7.1. Membership functions for a) Stocking Rate, b) Plantation Density, and c) Number of Pastoralists 1 7.1.4. Determining the fuzzy rules In this study, the rules are expressions of the role of interdependencies among factors of equilibrium, which were elicited from pastoralists’ 1 . While most of the ranges were elicited by interview, the rest were calculated by means. 114 experts by interviews. They state different dimensions of sustainability in range management. To determine the overall equilibrium, the rule base needs 33 = 27 rules since we have 3 linguistic values and 3 linguistic variables (SR, PD and NP), which are stated by pastoralists’ experts. The complete rules base used to construct the overall experts’ knowledge base are summarized in Table 7.2 for different linguistic values. All rules-base was elicited by interviews and all pastoralists’ experts were agreed at the end with several follow-ups. Table 7.2. The complete rules base (33 = 27) used to construct the overall experts’ knowledge base. Rule r 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 1 if Stocking Rate is low low low low low low low low low medium medium medium medium medium medium medium medium medium high high high high high high high high high and Plantation Density Is Poor Poor Poor Acceptable Acceptable Acceptable Rich Rich Rich Poor Poor Poor Acceptable Acceptable Acceptable Rich Rich Rich Poor Poor Poor Acceptable Acceptable Acceptable Rich Rich Rich and Number of Pastoralists is low medium high low medium high low medium high low medium high low medium high low medium high low medium high low medium high low medium high Then Stocking Rate must be1 Low Low low low low low medium medium medium low low low medium low low medium medium medium low low low medium medium low medium medium medium . Considering long-term consequences which is certainly back to "overgrazing", none of pastoralists express "high" as the linguistic value of the output variable (RRS). 115 7.1.5. Computing degree of membership of crisp inputs We present a numerical example illustrating how the EAFL model can compute degree of membership of crisp inputs. Suppose that information concerning the input variables is expressed numerically as follows: SR = 75 (aum/km2) (Fig. 7.2a), PD = 35 (tn/km2) (Fig. 7.2b), and NP = 0.4 (p/km2) (Fig. 7.2c). µi(SR) a low 1 medium high SR (aum/km2) 50 0 55 µj(PD) 1 70 75 b poor acceptable rich 0.75 0.5 0.25 PD (tn/km2) 0 10 30 µk(NP) 1 35 40 50 c low medium high 0.5 NP (p/km2) 0 0.1 0.3 0.4 0.5 Fig. 7.2. Linguistic values and fuzzification of crisp inputs for a) Stocking Rate, b) Plantation Density, and c) Number of Pastoralists. 116 Fuzzification yields the following inputs for the inference engine: Input 1: SR is high with membership grade µh(SR) = µh(75) = 1; Input 2: PD is acceptable with membership grade µa(PD) = µa(35) = 0.75 and rich with membership grade µr(PD) = µr(35) = 0.25; Input 3: NP is medium with membership grade µm(NP) = µm(0.4) = 0.5 and high with membership grade µh(NP) = µh(0.4) = 0.5. 7.1.6. Detemining approximate reasoning Now, we compute the degree to which each rule is applicable to the input. The only consistent rules are those in which SR is high, PD is either acceptable or rich, and NP is either medium or high. These are rules 23, 24, 26, and 27 of Table 7.2. The conclusions of these rules are expressed as follows: Rule 23: If SR is high with membership grade 1 and PD is acceptable with membership grade 0.75 and NP is medium with membership grade 0.5, then the output SR must be low with membership grade: µPREMISE23 = min ({1, 0.75, 0.5}) = 0.5 With the same calculation: µPREMISE24 = min ({1, 0.75, 0.5}) = 0.5 µPREMISE26 = min ({1, 0.25, 0.5}) = 0.25 µPREMISE27 = min ({1, 0.25, 0.5}) = 0.25 117 For the remaining rules of the rule base, we have µPREMISEr = 0. We observe that rules 23, 24 and 27 assign the same linguistic value low to SR with membership degree 0.5, 0.5 and 0.25 respectively. Now, based on degree of membership of inputs value, the fuzzy outputs µB'(RRS) of each rule are calculated and combined into one fuzzy output µ(RRS) (Fig. 7.3). 7.1.7. Computing crisp output (defuzzify) Finally, we use “Center Of Gravity” (COG) for defuzzifying (Jang et al., 1997; Zimmermann, 1996) yielding the RRS. In the example, the RRS was assessed by using the Matlab Fuzzy Toolbox (version 7) yielding RRS = 32.9 (aum/km2) = 0.329 (aum/ha) (Fig. 7.3). Rule 23 Rule 24 Rule 26 Rule 27 Fig. 7.3. Graphical illustration of the EAFL model for approximate reasoning and defuzzification. Approximate reasoning starts with a two-steps process comprising the implication process and the aggregation process yielding the overall fuzzy output µ i(RRSp) based on the fuzzy conclusions of the inputs (SRp, PDp and NPp) for each rule. Finally, the center of gravity method divides the area under curve into two equal sub-areas hereby determining the crisp output value: RRSp = 32.9 (Fuzzytoolbox in Matlab 7). 118 7.1.8. Assessing the model performance Having available a large set of input-output data, the performance of the system can be evaluated and parameters of the system can be finetuned in order to achieve a low ‘generalization error’. In such a datarich situation, a training set is used to fit the models, a validation set is used to estimate the prediction error for model selection and a test set is used for assessing the generalization error of the final model chosen (Hastie et al., 2001). If, like in our case, no large data sets are available, the best way to assess model performance and fine-tune the system is based on experts’ judgements (Davis and Wagner, 2003). By using different real inputs and observing crisp outputs, judgement is possible by experts. They can assess several scenarios and conclude whether the performance of the model is (not) reasonable. In our case, a small set of real input-output data appeared to be available. This data set was used to describe the behaviour of the EAFL model (Table 7.3). Table 7.3. Assessing the performance of the EAFL model by using real data. Output: Real Inputs Active RRS 1 Area NP SR PD 2 2 2 Rules (aum/km2) (aum/km ) (tn/km ) (p/km ) Cheshme-Anjir Morzion Kheshti 124 94 55 18 12 28 0.4 0.9 0.3 20, 21, 23, 24 20, 23 11, 14 32.1 26.1 26.1 ∆SR: (RRS – SR) (aum/km2) – 91.9 – 67.9 – 28.9 Table 7.3 shows three different outputs (RRSs) corresponding to three real input data. By comparing the Right Rate of Stocking (RRS) with the current Stocking Rate (SR) for each area, it becomes clear that the current SR values are considerably higher than the RRS values: RRS = 1 . Using Membership Function Editor of Fuzzytoolbox to automatically define membership functions of the inputs, the outputs (RRSs) would be 44.2, 16.3, and 21.8 for the studied areas respectively that also confirm "overgrazing" in these areas. 119 32.1, 26.1 and 26.1 when SR = 124, 94 and 55, respectively. The negative ∆SRs (– 91.9, – 67.9 and – 28.9 respectively) exhibit the exceeding rate of SR compared to that of RRS and therefore suggest general overgrazing in the three-prototypical areas in Southwest Iran. All pastoralists’ experts in the study areas, on one hand, agree with this result. They believe that most important issue nowadays they are challenging with, is overgrazing and this is actually the reason why they did not consider the “high” value for RRS as the output of the model (Table 7.2). They are afraid that, by choosing this value, even the favourite conditions (e.g. rules 8 and 16), overgrazing will continue to happen and further be encouraged in the future. By comparing the current SR to the RRS, on the other hand, the correct decision can easily be made by pastoralists. In all three areas, to return to an equilibrium state, pastoralists should decrease the following amounts of their livestock per square kilometer: ∆SR1 = RRS – SR = 32.1 – 124 = – 91.9 (aum/km2) = – 0.9 (aum/ha) ∆SR2 = RRS – SR = 26.1 – 94 = – 67.9 (aum/km2) = – 0.6 (aum/ha) ∆SR3 = RRS – SR = 26.1 – 55 = – 28.9 (aum/km2) = – 0.2 (aum/ha) Such a decision has a lot of consequenses. In this case, for example, as they would loose a part of their major income, they usually do not consider this option (decrease). Consequently, pastoralists will experience unbalance and unavoidable degradation in their rangelands. For making money to return to balance (without any overgrazing), the Natural Resourse Administrations in Iran have started to offer them the other jobs since 2000. These are included handcrafts, horticulture, agronomy, animal husbandry and other jobs related to agriculture. These jobs can be varied due to different 120 climaticand geographical conditions as well as the pastoralists’ experiences. The current state of the majority of the Iranian rangelands, namely overgrazing, of course, may change in the future. In fact, a pastoral system is a dynamic system, where socio-economic conditions change over the time. Therefore, it may be needed to change or add input variables to the model or to redefine the membership functions, yielding in different output rates. If so, “overgrazing” may change to “normal grazing” or even, “undergrazing” based on different socioeconomic and dynamic ecological conditions in the future. 7.2. Development the multi-fuzzy model Due to heterogeneous experts’ knowledge as collected in the semistructured interviews, we have constructed three different Mamdanitypes of fuzzy models. Each model has its own specific inputs, linguistic values, fuzzy range and if-then rules (Table 7.4). 121 Table 7.4. Inputs, linguistic values and fuzzy range of each expert. Expert His discipline Inputs (Ii) Linguistic values Fuzzy range Unit PA Verylow,Low,Medium,High,Veryhigh 500,2500,5000,10000,20000 ha Desert 1 AR Verylow,Low,Medium,High,Veryhigh 100,200,350,600,800 mm Management PD Verylow,Low,Medium,High,Veryhigh 10,25,35,55,70 % PA Low,Medium,High 500,1000,2000 ha FTP Negative,Stable,Positive 20,50,70 % Rural 2 SP Deserts,Mountains 5,12 % Development TP Low,Medium,High 100,450,1700 m2 AR Verylow,Low,Medium,High 50,100,250,500 mm FTP Negative,Stable,Positive 0,50,100 % Animal 3 CC Low,Medium,High,Veryhigh 0.5,2,5,6 aum Husbandry PS Poor,Normal,Good,Rich 25,50,75,100 % Abbreviations: PA: Pasture Area, AR: Annual Rainfall, PD: Plantation Density, FTP: Future Trend of a Pasture, SP: Slope of Pasture, TP: Topography of Pasture, CC: Carrying Capacity, PS: Pastoralists Situation. 122 Considering the second column with the third of Table 7.4 makes clear that the different administrative experts, graduated in different disciplines, have different knowledge concerning the indicators influencing the balance between ‘consumption’ and ‘conservation’ in sustainable rangeland management. They usually introduce those indicators which are most related to their own discipline. Thereby, based on the experts’ knowledge, we constructed three fuzzy models. The inputs of the first model are Pasture Area, Annual Rainfall, and Plantation Density, where the second model holds Pasture Area, Future Trend of a Pasture, Slope of Pasture, Topography of a Pasture, and Annual Rainfall and the third model includes Future Trend of a Pasture, Carrying Capacity and Pastoralists Situation as the inputs (Table 7.4). The administrative experts introduced different linguistic values for their inputs. The first expert, for example, considered five levels (Very low, Low, Medium, High, Very high) for all three of his own inputs, while, the second expert considered two, three and four levels for his own indicators. Also, the third expert deemed three and four levels for his own inputs. Table 7.4 also shows that for the same inputs (e.g. variable AR and FTP), experts may have different ideas regarding the linguistic values to be used. We also asked the experts to define the fuzzy ranges of the linguistic values, i.e., the membership functions that define the linguistic values selected. All experts considered the trapezoidal shape for the smallest and the largest linguistic values and the triangular for the rest. In this way, we defined both triangular and trapezoidal membership functions based on administrative experts’ knowledge. 123 The experts were also asked to express their own knowledge in a set of ‘fuzzy IF-THEN rules’ while offering all possibilities. We prepared all combinations of the inputs and asked the experts to fill out the output column. Thereby, the number of if-then rules was determined based on the number of inputs variables and their linguistic values1: The number of if-then rules (Model 1) = 5 * 5 * 5 = 125 The number of if-then rules (Model 2) = 3 * 3 * 2 * 3 * 4 = 216 The number of if-then rules (Model 3) = 3 * 4 * 4 = 48. Although the inputs of the three fuzzy models are different due to quantity and linguistic values, the Right Rate of Stocking (RRS) is chosen as the only output of the model (Table 7.5). Table 7.5. Characteristics of the output (RRS) for three fuzzy models. Model Linguistic values Fuzzy ranges 1 Verylow,Low,Medium,High,Veryhigh 0.5,1.0,2.5,4.0,6.0 2 Low,Medium,High 0.5,0.6,1.0 3 Low,Medium,High 1.0,2.0,3.0 Unit aum/ha aum/ha aum/ha 7.2.1. Computing the crisp primary outputs Before being able to compute a final crisp output value of the multi-fuzzy model, we must first calculate the primary outputs, i.e., the output of each fuzzy model. To do so, we need a representative set of data. We have been able to collect the inputs data of five prototypical cases for each region of our study. Next, we computed the primary output values RRSi of each the three fuzzy models developed. The results of this procedure are summarized in Tables 7.6, 7.7 and 7.8. 1 . See appendix 2. 124 Table 7.6. Computing the outputs of the first model with 5 cases for each region. Model: 1 First region: Morzion Real Inputs Active Output Case PA (ha) Rules RRS AR (mm) PD (%) 1 (aum/ha) 1 100 750 12 16,17,21,22 0.68 2 100 750 11 16,17,21,22 0.56 3 200 750 12 16,17,21,22 0.68 4 50 750 13 16,17,21,22 0.77 5 250 750 12 16,17,21,22 0.68 Second region: Cheshme-Anjir Real Inputs Active Output Case PA (ha) Rules RRS1 (aum/ha) AR (mm) PD (%) 6 100 315 18 6,7,11,12 1.07 7 200 315 19 6,7,11,12 1.11 8 300 315 17 6,7,11,12 1.02 9 250 315 18 6,7,11,12 1.07 10 300 315 16 6,7,11,12 0.97 Third region: Kheshti Real Inputs Active Output Case PA (ha) Rules RRS1 (aum/ha) AR (mm) PD (%) 11 200 240 20 6,7,11,12 1.15 12 500 240 22 6,7,11,12 1.23 13 200 240 19 6,7,11,12 1.11 14 300 240 20 6,7,11,12 1.15 15 700 240 20 6,7,11,12 1.15 Abbreviations: PA: Pasture Area, AR: Annual Rainfall, PD: Plantation Density, and RRS: Right Rate of Stocking. 125 Table 7.7. Computing the outputs of the second model with 5 cases for each region. Model: 2 First region: Morzion Real Inputs Output Active Case PA RRS2 FTP SP TP AR Rules (ha) (%) 1 2 3 4 5 100 100 200 50 250 50 50 50 50 50 Case PA (%) (mm) 15.7 100 750 40 15.7 120 750 40,44 15.7 115 750 40,44 15.7 130 750 40,44 15.7 145 750 40,44 Second region: Cheshme-Anjir Real Inputs Active FTP SP TP AR Rules (ha) (degree) (%) 6 7 8 9 10 100 200 300 250 300 55 55 55 55 55 13.4 13.4 13.4 13.4 13.4 Case PA (ha) (m2) (m2) (mm) 180 315 39,40,43,44,63,64,67,68 160 315 39,40,43,44,63,64,67,68 200 315 39,40,43,44,63,64,67,68 185 315 39,40,43,44,63,64,67,68 170 315 39,40,43,44,63,64,67,68 Third region: Kheshti Real Inputs Active FTP SP TP AR Rules (degree) (%) (m2) (mm) (aum/ha) 0.70 0.85 0.82 0.91 0.98 Output RRS2 (aum/ha) 1.12 1.12 1.13 1.12 1.12 Output RRS2 (aum/ha) 11 200 20 8.6 30 240 2,3,14,15 0.45 12 500 20 8.6 20 240 2,3,14,15 0.45 13 200 20 8.6 35 240 2,3,14,15 0.45 14 300 20 8.6 40 240 2,3,14,15 0.45 20 8.6 25 240 2,3,14,15,74,75,86,87 0.45 15 700 Abbreviations: PA: Pasture Area, FTP: Future Trend of a Pasture, SP: Slope of Pasture, TP: Topography of Pasture, AR: Annual Rainfall, and RRS: Right Rate of Stocking. 126 Table 7.8. Computing the outputs of the third model with 5 cases for each region. Model: 3 First region: Morzion Real Inputs Active Output Case Rules RRS3 (aum/ha) FTP (%) CC (aum) PS (%) 1 50 0.27 60 18,19 0.84 2 50 0.25 65 18,19 0.84 3 50 0.22 60 18,19 0.84 4 50 0.26 55 18,19 0.80 5 50 0.28 65 18,19 0.84 Second region: Cheshme-Anjir Active Output Real Inputs Case Rules RRS3 (aum/ha) FTP (%) CC (aum) PS (%) 6 50 0.36 65 18,19 0.84 7 50 0.34 50 18 0.76 8 50 0.36 70 18,19 0.80 9 50 0.38 60 18,19 0.84 10 50 0.35 60 18,19 0.84 Third region: Kheshti Real Inputs Active Output Case Rules RRS3 (aum/ha) FTP (%) CC (aum) PS (%) 11 20 0.42 30 1,2,17,18 0.84 12 20 0.40 35 1,2,17,18 0.84 13 20 0.43 40 1,2,17,18 0.84 14 20 0.44 30 1,2,17,18 0.84 15 20 0.41 35 1,17 0.84 Abbreviations: FTP: Future Trend of a Pasture, CC: Carrying Capacity, PS: Pastoralists Situation, and RRS: Right Rate of Stocking. Table 7.6, 7.7 and 7.8 show that, for equal cases, the primary outputs RRSi (i = 1, 2, 3) are usually different. It clarifies that our decisions to select the best final output as an estimation of the RRS is not a trivial task. Actually, we need to find a solution for dealing with the differences among the primary outputs. More formally, we should find an 'optimal' way to combine the primary outputs RRSi(I), based on a given input vector I, in order to calculate one final crisp output value RRSf(I) of our multi-fuzzy model. This combining process is sometimes, especially in environments of supervised learning, termed ‘voting’1 (Hastie et al., 2001). 1 . Voting and rating activities require the gathering of participants' opinions from large distances, and therefore, they are closely connected to network issues and distributed processing (Kovács and Micsik, 2005). 127 7.2.2. Implementing voting In the next sections, we introduce and discuss several ways to implement voting, i.e., to calculate the weights for combining the primary outputs. 7.2.2.1. Method 1: Calculating the mean of outputs Table 7.9 shows the primary outputs of models 1, 2 and 3, and the final output RRSf of 15 cases (c = 1,2,…,15) in three different study regions. In this approach, all final crisp outputs are simply equal to the arithmetic mean of the primary outputs of the three models, i.e., RRS f ( I c ) = Mean = 1 3 ∑ RRSi ( I c ), where Ic represents the input vector for 3 i =1 the cth case. Table 7.9. Finding the final outputs by calculating the mean of primary outputs. First region: Morzion Primary outputs: RRSi Mean Case SD (RRSf) Model 1 Model 2 Model 3 1 0.68 0.70 0.84 0.08 0.74 2 0.56 0.85 0.84 0.16 0.75 3 0.68 0.82 0.84 0.08 0.78 4 0.77 0.91 0.80 0.07 0.82 5 0.68 0.98 0.84 0.15 0.83 Second region: Cheshme-Anjir Primary outputs: RRSi Mean Case SD (RRSf) Model 1 Model 2 Model 3 6 1.07 1.12 0.84 0.14 1.01 7 1.11 1.12 0.76 0.20 0.99 8 1.02 1.13 0.80 0.16 0.98 9 1.07 1.12 0.84 0.14 1.01 10 0.97 1.12 0.84 0.14 0.97 Third region: Kheshti Primary outputs: RRSi Mean Case SD (RRSf) Model 1 Model 2 Model 3 11 1.15 0.45 0.84 0.35 0.81 12 1.23 0.45 0.84 0.39 0.84 13 1.11 0.45 0.84 0.33 0.80 14 1.15 0.45 0.84 0.35 0.81 15 1.15 0.45 0.84 0.35 0.81 128 As table 7.9 shows, different regions have different RRSf as the means of the primary outputs. Since, the second region (Cheshme-Anjir) gains the highest and the first region (Morzion) has the lowest means, the third region (Kheshti) stands between them. Therefore, according to these estimations, the second region can hold the most aum/ha, while the capacity of the first region is the least. The standard deviations of RRSi in the three regions have also been calculated. Table 7.9 shows that the highest deviations of the primary RRSi are found in the third and the lowest in the first region. The method of calculating the mean of the primary outputs has some strengths and weaknesses. It has some strengths because it concerns a simple calculation and it covers all three primary outputs. It has some weaknesses, as it uses all data in our calculations with equal weights. By doing so, outliers are equally important as points close to the expected value of the output. Therefore, we think it would be better to look for voting methods where the primary outputs are calculated as a weighted mean, i.e., N RRS f ( I c ) = ∑ wi RRS i ( I c ) (1) i =1 Here, in a more general setting, N equals the number of primary models, where the weights wi are subject to the following constraints: N ∑w i = 1 and ∀i : wi ≥ 0. The underlying assumption of this weighted i =1 approach is that ‘each expert has something to say’ and, in addition, that ‘certain experts have something more to say than others’. We now discuss several methods for calculating ‘optimal’ weight values wi. 129 7.2.2.2. Method 2: Minimizing the sum of squared errors Where a training set of C input-output cases (Ic,RRSc), (c = 1, 2, …, C) is available, we can calculate optimal weights wi by choosing the weights wi such that the following sum of squared errors SSE is minimized: C C N c =1 c =1 i =1 SSE = ∑ ( RRS f ( I c ) − RRS c ) 2 = ∑ (∑ wi RRS i ( I c ) − RRS c ) 2 . (2) This approach of supervised learning can be considered as a regression method where the final outputs RRSf(Ic) of the multi-fuzzy model are as much as possible equated to the correct output values RRSc. If desired and needed, even more sophisticated supervised methods from predictive data mining like ‘bagging’ and ‘boosting’ (Hastie et al., 2001; Ishibuchi et al., 1999) can be considered for implementing optimal voting schemes. Unfortunately, like in our case, the above-given methods of supervised learning are not applicable where a set of correct inputoutput values is not available. Therefore, we will not further discuss method 2 in this study. Instead, we are challenged to come up with an unsupervised method for implementing voting, i.e., a method where we try to optimize the ‘consistency’ of the final output values of the system by harmonizing the values of the primary outputs. 7.2.2.3. Method 3: Minimizing an approximation of the sum of squared errors One might wonder whether we can approximate the approach of method 2 by using an approximation RRS′c of the correct output values RRSc. Knowing the primary outputs RRSi, we can use as an 130 approximation RRS ' c = Mean = 1 N N ∑ RRS ( I i c ) and next try to minimize the i =1 approximation of the sum of squared errors SSE' defined as C C N SSE ' = ∑ ( RRS f ( I c ) − RRS ' c ) 2 = ∑ (∑ wi RRS i ( I c ) − c =1 c =1 i =1 1 N N ∑ RRS i ( I c )) 2 , (3) i =1 N where the above-given constraints ∀i : wi ≥ 0 and ∑w i = 1 still hold. i =1 Unfortunately, this method does not work since equation (3) has a trivial minimum equal to zero, namely, in case ∀i : wi = 1 / N resulting into Method 1 from section 4.3.1, therefore, we should look for another approach. Below we shall look for ‘harmonizing methods’ where the dissimilarities between the primary outputs RRSi are minimized. 7.2.2.4. Method 4: Harmonizing the primary outputs A natural approach for harmonizing existing differences in the primary outputs is to put less emphasis on outliers. By doing so, we hope to find a more unbiased estimation RRSf of the right rate of stocking. In addition, a smaller standard deviation SD of the weighted primary outputs wi RRS i is expected to be found simply because the primary output values close to the mean get more weight in the voting process. The above-given idea can be formalized as follows. Given input vector Ic, let ∆ i , j ( I c ) =| RRS i ( I c ) − RRS j ( I c ) | represent the absolute value of the difference between the primary outputs of model i and j. Using all input vectors Ic available, we can calculate the sum ∆i of the absolute differences between primary ouput RRSi and all other primary outputs RRSj, j ≠ i, defined by 131 C C ∆ i = ∑∑ ∆ i , j ( I c ) = ∑ ( ∆ i ,1 ( I c ) + ∆ i , 2 ( I c ) + Λ + ∆ i ,i −1 ( I c ) + ∆ i ,i +1 ( I c ) + Λ ) . c =1 j ≠i (4) c =1 If ∆ i > ∆ j , this means that model i generates, on average, more outlying output values than model j and therefore, in our approach, should get a lower weight. This can be implemented by giving model i a weight which equals the normalized inverse of ∆i or, more precisely, wi = 1/ ∆i 1/ ∆i = ∑ 1 / ∆ j 1 / ∆1 + 1 / ∆ 2 + 1 / ∆ 3 + Λ (5) j It should be clear that by providing the primary outputs RRSi the weights wi as defined by equation (5), the above-mentioned constraints ∀i : wi ≥ 0 and ∑w i = 1 are automatically fulfilled. i =1 Having determined the weights of the primary outputs, the standard deviation of the weighted primary outputs can be calculated. Since for Method 4 outliers have received less weight, these standard deviations are expected to be smaller than in case of using Method 1 where the primary outputs have equal weights. Furthermore, the final output of the multi-fuzzy model can be calculated using equation (1). We have done these calculations using the available field data from each of the different regions of study. The results found are summarized in Table 7.10. 132 Table 7.10. Estimating the final output RRSf by calculating the sum of weighted outputs for separated regions according to Method 4. Case 1 2 3 4 5 Primary outputs (RRSi) Model 1 Model 2 Model 3 0.68 0.70 0.84 0.56 0.85 0.84 0.68 0.82 0.84 0.77 0.91 0.80 0.68 0.98 0.84 ∆1-2 0.02 0.29 0.14 0.14 0.3 Delta’s ∆2-3 0.14 0.01 0.02 0.11 0.14 ∆1-3 0.16 0.28 0.16 0.03 0.16 Sum Inverse Weights(wi) Case 6 7 8 9 10 Primary outputs (RRSi) Model 1 Model 2 Model 3 1.07 1.12 0.84 1.11 1.12 0.76 1.02 1.13 0.80 1.07 1.12 0.84 0.97 1.12 0.84 ∆1-2 0.05 0.01 0.11 0.05 0.15 Delta’s ∆2-3 0.28 0.36 0.33 0.28 0.28 ∆1-3 0.23 0.35 0.22 0.23 0.13 Sum Inverse Weights(wi) Case 11 12 13 14 15 Primary outputs (RRSi) Model 1 Model 2 Model 3 1.15 0.45 0.84 1.23 0.45 0.84 1.11 0.45 0.84 1.15 0.45 0.84 1.15 0.45 0.84 ∆1-2 0.7 0.78 0.66 0.7 0.7 Delta’s ∆2-3 0.39 0.39 0.39 0.39 0.39 ∆1-3 0.31 0.39 0.27 0.31 0.31 Sum Inverse Weights(wi) First region: Morzion Sum of Deltas ∆1 ∆2 ∆3 0.18 0.16 0.3 0.57 0.3 0.29 0.3 0.16 0.18 0.17 0.25 0.14 0.46 0.44 0.3 1.68 1.31 1.21 0.59 0.76 0.82 0.27 0.35 0.38 Second region: Cheshme-Anjir Sum of Deltas ∆1 ∆2 ∆3 0.28 0.33 0.51 0.36 0.37 0.71 0.33 0.44 0.55 0.28 0.33 0.51 0.28 0.43 0.41 1.53 1.9 2.69 0.65 0.52 0.37 0.42 0.34 0.24 Third region: Kheshti Sum of Deltas ∆1 ∆2 ∆3 1.01 1.09 0.7 1.17 1.17 0.78 0.93 1.05 0.66 1.01 1.09 0.7 1.01 1.09 0.7 5.13 5.49 3.54 0.19 0.18 0.28 0.29 0.28 0.43 133 Sum 4.2 2.18 1.00 Sum 6.12 1.55 1.00 Sum 14.16 0.65 1.00 Weighted outputs w1RRS1 w2RRS2 w3RRS3 0.18 0.24 0.31 0.15 0.29 0.31 0.18 0.28 0.31 0.21 0.31 0.30 0.18 0.34 0.31 Weighted outputs w1RRS1 w2RRS2 w3RRS3 0.45 0.38 0.20 0.46 0.38 0.18 0.43 0.383 0.19 0.45 0.38 0.20 0.40 0.38 0.20 Weighted outputs w1RRS1 w2RRS2 w3RRS3 0.34 0.124 0.36 0.36 0.124 0.36 0.32 0.124 0.36 0.34 0.124 0.36 0.34 0.124 0.36 SD 0.06 0.09 0.06 0.05 0.08 SD 0.12 0.14 0.12 0.12 0.11 SD 0.13 0.13 0.12 0.13 0.13 (RRSf) Summation 0.74 0.76 0.78 0.83 0.84 (RRSf) Summation 1.03 1.02 1.00 1.03 0.98 (RRSf) Summation 0.82 0.84 0.81 0.82 0.82 As Table 7.10 shows, there are different weights for each region. While the weights of the model 1, 2 and 3 for the first region are 0.27, 0.34, and 0.37, the weights for these models for the second region are 0.42, 0.33 and 0.23 and for the third region are 0.29, 0.27 and 0.42 respectively. In other words, when the first model gets the highest weight in Cheshme-Anjir (w1 = 0.42), the second model gives the highest weight in Morzion and Cheshme-Anjir (w2 = 0.34 and 0.33) respectively, and the third model gains the highest weight in Kheshti (w3 = 0.42). So, we have different weights in different regions showing that the expertize of the experts in the various regions seems to be different. Now, by calculating the sum of weighted outputs, we can easily estimate the final outputs: 3 RRS f = ∑ wi RRS i (6) i =1 Based on equation (6), again, we estimated the final output RRSf for separated regions. As Table 7.10 shows, the estimations for various regions are different. While the highest amount of RRSf (more than 1) is estimated for the second region and the lowest for the first (less than 0.8), the third region has got more than 0.8. On the other hand, although the estimations are different for the ‘between groups’, they are approximately similar for the ‘within groups’. Comparing Table 7.9 and 7.10, the standard deviation has indeed been reduced, as expected. 7.2.3. Comparison of Method 1 and Method 4 Based on the two applied methods (Mean and Harmonized), we estimated the final Right Rate of Stocking (RRSf). In the first method 134 (Table 7.9), we estimated RRSf by calculating the average of the primary outputs. The method considers the same weights for all outputs of the three models and therefore, it treats outliers and ‘normal’ data equally. This approach may introduce some bias in our calculations. To decrease this weakness, we introduced other voting methods. As discussed, a good voting method is provided by a ‘supervised learning’ algorithm where optimal weights are calculated by minimizing the sum of squared errors of the output values. A necessary precondition for applying this method is the availability of a representative data set. Otherwise, we can use a voting procedure based on ‘unsupervised learning’. Using this method (the harmonized method) and the first method (the mean method), we are able to compare the assessments of the RRSf for all three regions of our study (Fig. 7.4). Fig. 7.4. Comparison of the RRSf for the harmonized method and mean method. 135 As Fig. 7.4 demonstrates both methods estimate the highest RRSf for the second region and the least for the first where the third region stands between them. Also, while the RRSf for the second region is considerably different from the two other regions, the RRSf for the first and third region are quite close. We observe here that differences, even small ones, between the values of the Right Rate of Stocking have serious consequences because of ‘scaling factor’. In other words, since the total area of each region is very large (e.g. 2000 hectares in Morzion, 2575 hectares in Cheshme-Anjir and 6900 hectares in Kheshti), small differences become big ones if we multiply the estimated values of the RRSf by the amount of pasture area. A primary validation shows that all experts confirm the abovementioned outcomes. Considering our results, they believe that the second region (Cheshme-Anjir) has the best conditions to reach sustainability. These conditions include social, geographical and environmental circumstances and are supposed to be strongly related to the values of the input variables (indicators) of the fuzzy models discussed above. The social problems are supped to be less in Cheshme-Anjir because the region benefits by a good manager who solves many of their problems, especially, those related to the usual bureaucratic problems in the different Iranian administrations. The region also benefits by higher education level of its expert. In addition, the administrative experts believe that the second region has better strategic conditions as it falls between the two main roads which are near to Shiraz, the capital of the Fars province. Finally, as the weather in this region is not very cold (like the first region in Morzion) and not very warm (like the third region in Kheshti), the temperate weather makes better environmental conditions in 136 Cheshme-Anjir. Thus, these conditions make the second region the best prototypical case in comparison to the two other regions and can explain the high differences between outcomes for the second area and the rest. The administrative experts also agree upon estimations of the RRSf for the two other regions. As vegetation period in the third region (Kheshti) is longer than the first region (Morzion), they expect a lower and a higher RRSf for the first and third region respectively. They also believe that the warm weather of Kheshti creates a longer time period of grazing. In contrast, the cold weather of Morzion declines its capacity to hold livestock to graze. The experts add, however, if we neglect the time period of grazing, the capacity of the third region for holding the RRS should decrease (as we have seen for the second model). This can also be a good evidence to show that the experts were differently expertized in the various regions. In addition, since the experts consider 2-3 aum/ha as the medium range of RRS (see Table 7.5), they believe that currently the three regions of our study have a much smaller grazing capacity than might be possible. This outcome confirms the general believe that many pastoral regions in Iran are currently facing overgrazing and are exacerbating by the unsustainable situation. By taking appropriate measures, circumstances for SD may be improved in the future. Consequently, the experts’ knowledge will change according to new conditions and higher estimates are likely to be found. As has been shown in Tables 7.4 and 7.5, the nominated experts in Iran choose different indicators to estimate the Right Rate of Stocking, even in the cases where the same social, environmental, and geographical conditions hold. They come up with sustainability in 137 rangeland management by different environmental (e.g. Slope of Pasture) and socio-economic indicators (e.g. Pastoralist Situations). Therefore, our study supports the reality that sustainability in range management is a multi-dimensional vague concept. We also wish to emphasize here that the supervised voting process (Method 2) should be tried out in future research, after we have collected a set of real input-output data from the field. The collection of this data is, however, a time consuming process, which may take several years. 138 Chapter Eight Summary and Conclusions 8.1. Summary A claim is commonly made that the rangelands around the world are overgrazed and hence producing edible forage and animal products at less than their potential Wilson and Macleod, 1991). Globally, rangelands are at risk from numerous pressures (Mitchell et al., 1999). Some of these pressures arise from livestock/rangeland systems. Livestock have been a key factor in the development of civilization, but their role in the future is not clear as well as how the science of rangeland management should change in order to meet future challenges in rangelands. Carrying capacity is the most important issue in range management (Walker, 1995). At a time, when the planet's limited carrying capacity seems increasingly obvious, the rationale and measures of rangelands carrying capacity are increasingly criticized. One of the key elements of rangeland capacity is stocking rate. If stocking rate is not close to the proper level related to equilibrium rate, then, regardless of other grazing management practices employed equilibrium will not be met (Roe, 1997). This applies to many countries, including Iran. It is a regular topic of books, articles and symposia (Conference on Sustainable Range Management, 2004; Iranian Nomadic Organization, 1992), and a common justification for further research. 139 Iran has approximately 90 million hectares of rangeland, 9.3 million hectares of which are considered in ‘good’ conditions while the remaining in ‘fair’ or ‘poor’ conditions. The country’s rangelands in a normal year produce around 10 million tons of dry matter (dm), of which 5.8 million tons may be available for grazing. The remaining amount is the minimum required for reproduction and soil conservation. The later amount of dm can support 38.5 million animal units (au) for duration of 8 months. At the moment there is 115.5 million of au in Iran and only 16.5 of them are fed from other sources including by agricultural products. The above figures prove that the rangelands are being utilized at three times more than their peak capacities in a non-drought year. This results in severe degradation of the rangelands and accelerates soil erosion. As the rangeland is considered by its users as "free resource" it is subject to heavy abuse, which further exacerbates the drought (FAO, 2004a; Iranian Nomadic Organization, 1992; Mesdaghi, 1995; UNCT, 2001). The recent literature on rangelands disequilibrium calls in question any specific measures of carrying capacity, whether the range is stocked or unstocked, and managed or mis-managed. Ideally, such objections can be taken into account for any individual carrying capacity estimated by accepting that it has to be determined on a case - by - case basis in the field. Once one knows the size of the grazing and browsing animals, the biomass production of the area, the pattern of range management, and so on, she/he can - so this argument goes produce a site specific stocking rate estimated for the range area under consideration. But, it cannot pack livestock into a given rangeland, without at some point deteriorating that range demonstrably. Surely, biomass production is going down on rangelands precisely because 140 stocking rate has been exceeded for so long, even taking into account factors such as drought and climate changes (Hardesty et al., 1993). However, the rationale and measures of rangeland carrying capacity are increasingly criticized. It seems that even under environmental conditions of great certainty, the notion of rangeland equilibrium would still be ambiguous and confused. Moreover, since environmental conditions are highly uncertain for the dry rangelands of the world such as Iran, current understanding of rangeland equilibrium turns out to be all the more questionable. There is no workable and practical “equation” for rangeland management in general, and carrying capacity in particular (Roe, 1997). Similar problems exist in other field of SD. In most studies, Fuzzy Logic is used to construct a model for evaluating sustainability in different areas. These models promise to be a valuable tool in evaluating the sustainability in general and equilibrium specifically in this study. The main purpose of this dissertation was to design a fuzzy model based on the experts’ knowledge for solving the mis-management of the Fars rangelands in Southwest Iran. 8.2. Conclusions Fuzzy logic appears to be well suited to provide quantitative answers pertaining to sustainability in rangeland management. In this study, we introduced a fuzzy logic-based model for sustainable rangeland management where the RRS is assessed as the output of the model from three input variables. All experts agreed with these three inputs which are: Stocking Rate, Plantation Density and Number of Pastoralists. By using the EAFL model, there will be three different options: 141 i) increasing the Stocking Rate (pastoralists usually like to do so), ii) unchanging the Stocking Rate (the rangeland is already in equilibrium), and iii) decreasing the Stocking Rate (more probable option). Evaluating the EAFL model presented in this dissertation, we conclude that it exhibits three important characteristics: • First, it permits the combination of various aspects of sustainability with different units of measurement, • Second, it overcomes the difficulty of assessing certain attributes or indicators of sustainability without precise quantitative criteria, and • Third, the methodology is easy to use and interpret. The model has, therefore, the potential to become a practical tool to policy-makers and scientists. It is important to note that the model is open for improvement, based on our better understanding of realities in the future. For example, one may construct different fuzzy rules (the number of indicators used to evaluate each linguistic variable of sustainability may be changed according to need or the membership functions of certain linguistic values can be redefined), or may enter some other Boolean measurements (which are normally used by rangelands’ scientists) in fuzzy analyses to emprove the validity of the model. We are aware that our EAFL model is just the first step. The flexibility of our model is one of its advantages over existing static methods. Considering the sequence "crisp input – fuzzifier – inference 142 engine – defuzzifier – crisp output" illustrates the uncertainity that exists in such a complex vague concept as sustainable rangeland management. It also well adjusts to usual ambigious linguistic statements of individuals. Furthermore, fuzzy logic operations compensate for the lack of full knowledge of our system. Uncertainty is ubiquitous in sustainability problems, since we never have complete knowledge of the complex interrelationship of ecological systems and the human thinking. Therefore, the EAFL model is expected to provide a new useful tool for policymakers in order to manage and to predict the overall sustainability in rangelands. Considering a multi-fuzzy model, which is developed in this dissertation, we have come up with other conclusions. While fuzzy specialists usually use homogeneous experts’ knowledge to construct fuzzy models, it is generally much more difficult to deal with knowledge elicited from a heterogeneous group of experts, especially in the area of sustainable rangeland management. In this study, we proposed a multi-fuzzy model to cope with the muti-dimensional vagueness of sustainability in the field of range management. To deal with the heterogeneity of experts’ knowledge, which should be considered either as a reality or necessity, we introduced several voting methods for estimating the Right Rate of Stocking as the final output of several fuzzy models. The first method simply uses the average of the primary outputs as the final right rate of stocking. We also introduced a supervised voting method, which is applicable where a real-world data set is available. In the absence of such a set, like in our study, an unsupervised voting method can be applied which estimates the weights of the primary right rate of stocking using a harmonizing approach. Since this method puts less emphasis on 143 outliers, the harmonizing approach is supposed to result into a more unbiased estimation. In addition, it turns out that the standard deviation of the harmonized weighted primary outputs is smaller that the standard deviation of equally weighted primary outputs. The harmonized method is expected to provide a new useful tool for policymakers in order to deal with heterogeneous experts’ knowledge. By constructing the three fuzzy models based on the heterogeneous knowledge and using some harmonized methods, our study tried to show the multi-dimensional vaguenesses which generally exist in rangeland management, and solve the conflict that especially exists in economic and conservation views in the Iranian rangeland management. Finally, by comparing the estimated Right Rate of Stocking, which elicited from both experts' opinions and Matlab Fuzzytoolbox Editor, with its medium range, the models verified overgrazing in the three regions of the Fars province in Southwest Iran. 144 Bibliography Abdul Aziz, S., 1996. You fuzzyin' with me? Available on: http://www.doc.ic.ac.uk/~nd/surprise_96/journal/vol1/sbaa/articl e1.html Aenis, T. and Nagel, U.J., 2000. Participatory management of large interdisciplinary research teams - First experiences from a rural research and development project in Germany. Available on: http://www.zalf.de/home_zalf/sites/grano/publikation/aenisnagel2000.pdf Ainesworth, E., 1989. LISA men have called you. Farm Journal, 113: 1. Alexandratos, N., 1998. World Food and Agriculture: The Outlook for the Medium and Longer Term. Rome: Global Perspective Studies Unit, FAO. Alexandratos, N., 1999. Plants and population: Is there time? Available on: http://www.nap.edu/books/0309064279/html/5908.html Andriantiatsaholiniaina, L.A., 2001. Sustainability Assessment Using Fuzzy Logic. Ph.D. Dissertation. Chania: Department of Production Engineering and Management, Technical University of Crete, Greece. Archer, S., 1996. Assessing and interpreting grass - woody plant dynamics. In: J. Hodgson, & A.W. Illius (Eds.), The Ecology and Management of Grazing Systems (101-134). Wallingford: CAB International. ARE., 2005. Sustainable development – Definition and constitutional status in Switzerland. Available on: http://www.are.admin.ch/are/en/nachhaltig/definition/ Atkinson, G., Dubourg, R., Hamilton K., Munashinge, M., Pearce, D. and Young, C. (Eds.), 1999. Measuring Sustainable Development: Macroeconomics and the Environment (Second Edition). Northampton: Edward Elgar. Azadi, H., Shahvali, M. and Faghih, N., 2003. Using fuzzy logic in solving conflict between sustainability and pasture. The Proceedings of the 4th Seminar on Fuzzy Sets and its Applications, 1-15. May 28-29, 2003, University of Mazandaran, Babolsar, Iran. Azadi, H., Shahvali, M. and van den Berg, J., (In revision). Sustainable rangeland management using fuzzy Logic: A case study in Southwest Iran. Journal of Ecological Modeling. 145 Barendse, G.W.J. and van der Hoek, D., 1996. Interdisciplinary project teams on agriculture and environment: A problemoriented approach. Journal of Agricultural Extension and Education 1(3), 7. [online] Available on: http://library.wur.nl/ejae/v1n3-7.html Barfield, T., 1981. The Central Asian Arabs of Afghanistan. Austin: University of Texas Press. Barth, F., 1961. Nomads of South Persia: The Basseri Tribe of the Khamseh Confederacy. London: Allen. Bates, D.G., 1973. Nomads and Farmers: A Study of the Yörük of Southeastern Turkey. University of Michigan: Ann Arbor. Beck, L., 1986. The Qashqa’i of Iran. New Haven & London: Yale University Press. Berkan, R.C. and Trubatch, S.L., 1997. Fuzzy Systems Design Principles: Building Fuzzy IF-THEN Rule Bases. New York: IEEE Press. Black-Michaud, J., 1986. Sheep and Land. The Economics of Power in a Tribal Society. Cambridge: Cambridge University Press. Blaisdell, J.P. & Pechanec, J.F., 1949. Effects of herbage removal or various dates on vigor of Bluebunch Wheatgrass and Arrowleaf Balsamroot. Ecology, 30, 298-305. Blundon, A., 2001. The world food prize. Available on: http://www.worldfoodprize.org/Youthinstitute/2001YIProceeding s/papers-html/Hoover%20High%20School.htm Box, T.W., 1995. Private Property Rights and Sustainable Rangeland Use. University of California: Center for Forestry, UC Berkeley. Box, T.W., 2002. Future deserts and sustainable communities. Available on: http://www.rala.is/rade/ralareport/Box.pdf Brink, B.J.E.T., 1989. Systematically in Search of Sustainable Development. Memorandum GWWS 89.006, Tidal water division, The Hague. Brink, B.J.E.T., Husper, S.H. and Colijn, F., 1990. A quantitative method for description and assessment of ecological systems: The AMOEBA approach. Proceedings of the International Conference on the Environmental Management of Enclosed Coastal Seas. Marine Pollution Bulletin, Kobe, Japan. Brink, B.J.E.T., 1991. The AMOEBA approach as a useful tool for establishing sustainable development. In: O., Kuik, H., Verbruggen, (Eds.), Search of Indicators of Sustainable Development. Kluwer Academic Publishers, The Netherlands. 146 Brown, L.R., 1997. Can we raise grain yields fast enough? Worldwatch, 10(4), 9-17. Brown, L.R., Abramovitz, J.N., and Starke, L. (Eds.), 2000. State of the World: A Worldwatch Institute Report on Progress Toward a Sustainable Society. Publisher: W. W. Norton and Company, Inc. Burkhart, J.W., 1996. Herbivory in the Intermountain West: An Overview of Evolutionary History, Historic Cultural Impacts and Lessons from the Past. Station Bull. #58. Idaho Forest, Wildlife and Range Exp. Sta., Moscow Idaho. Burns, R., 1994. Introduction to Research Methods. Melbourne: Longman. Buss, H.J. and Nuppenau, E.A., 2002. More sustainable range use in semi-arid eco-systems through adapted management: A case study on Namibian farms based on bio-economic models. Contributed Paper and Poster, Deutscher Tropentag 2002, Witzenhausen, 9. bis 11. October. Buxton, R. and Stafford Smith, M., 1996. Managing drought in Australia's rangelands: Four weddings and a funeral. Rangeland Journal, 18, 292-308. Conference on Sustainable Range Management, 2004. Proceedings of the Conference on Sustainable Range Management, January 5 – 8. New Orleans, Louisiana. Connolly, J. & Nolan, T., 1976. Design and analysis of mixed grazing experiments. Animal Production, 23, 63-71. Cornelissen, A.M.G., 2003. The Two Faces of Sustainability. Fuzzy Evaluation of Sustainable Development. Ph.D. Thesis. Netherlands: Wageningen University. Cornelissen, A.M.G., van den Berg, J., Koops, W.J., Grossman, M. and Udo, H.M.J., 2001. Assessment of the contribution of sustainability indicators to sustainable development: A novel approach using fuzzy set theory. Agriculture, Ecosystems and Environment, 86, 173-185. Coulter, J., 2003. Sustainable population Australia. Available on: http://www.population.org.au/pressrm/newslet/nl200303.pdf Council for Biotechnology Information, 2002. Growing more food. How biotechnology in a seed can enhance global food security. Available on: http://www.whybiotech.com/index.asp?id=1821 Cumber, I.K., 2004. Chalmers. Annual Meeting of the Alliance for Global Sustainability at Chalmers Workshop on Learning for Sustainable Development, 22-24 March 2004. 147 Dasgupta, P., 1996. The economics of the environment. The British Academy. Proceedings of the British Academy, 90, 165-221. Dasgupta, P. and Maler, K., 1995. Handbook of Development Economics. Amsterdam: North Holland. Davis, A. and Wagner, J.R., 2003. Who knows? On the importance of identifying “experts” when researching local ecological knowledge. Human Ecology, 31(3), 463-489. Deadman, P.J., 1999. Modeling individual behavior and group performance in an intelligent agent based simulation of the tragedy of the commons. Journal of Environmental Management, 56,159-172 Dean, D., No date. World population. Available on: http://www.ctm.org/arti/gentop/wdp.htm de Kok, J.L., Titus, M. and Wind, H.G., 2000. Application of fuzzy sets and cognitive maps to incorporate social science scenarios in integrated assessment models: A case study of urbanization in Ujung Pandang, Indonesia. Integrated Assessment, 1, 177-188. Digard, J-P., 1990. The Evolution of Nomadic Pastoralism in Iran 1960-1978: Problems and Interpretations. In: C. Salzman, & J.G. Galaty (Eds.), Nomads in a Changing World (165-175). Naples: Istituto Universitario Orientale. Driankov, D., Hellendoorn, H. and Reinfrank, M., 1996. An Introduction to Fuzzy Control. New York: Springer. Dunn, E.G., Keller, J.M., Marks, L.A., Ikerd, J.E., Fader, P.D. and Godsey, L.D., 1995. Extending the application of fuzzy sets to the problem of agricultural sustainability. IEEE Proceedings of ISUMA-NAFIPS ’95, Missouri-Columbia, USA. Earth Government for Earth Community, 2003. Earth government for earth community: A grassroots process. Available on: http://members.shaw.ca/earthgov/ECGlobalConcepts.htm EC., 2002. Strategic Guidelines and Priorities for Interventions to be Financed in 2002 and 2003 under Regulation. No. 2493/2000 of the European Parliament and of the Council on Measures to Promote Environment in Developing Countries. European Commission, DG Development. DEV/2790/2002, Budget Line B7 – 6200. EDC News, 2003. Population & land: Arable land per capita diminishing serious implications. Available on: http://www.edcnews.se/Research/Population-Landhunger.html El-Awad, O.M.A., 1991. Multicriterion Approach to the Evaluation of Irrigation Systems Performance. Ph.D. Thesis. Greece: 148 University of Newcastle upon Tyne, Department of Civil Engineering. Emadi, M.H., 2003. Bridging Epistemologies and Integrating Indigenous Knowledge with Scientific Knowledge for Sustainable Development: A Case Study. Research Deputy of Rural Research Center, (RRC), Tehran, Iran. Einstein, D.E. 1998. Intensification: Can agriculture keep up with population growth? Available on: http://darwin.bio.uci.edu/~sustain/global/sensem/einstein298.htm l EuropaWorld, 2003. EuropaWorld. Available on: http://www.europaworld.org/week119/crowthinfood7303.htm Falconer, J., 1990. The Major Significance of ’Minor’ Forest Products. Rome: FAO. Falconer, J. and Arnold, J.E.M., 1989. Household Food Security and Forestry: An Analysis of Socio-Economic Issues. Rome: FAO. FAO, 1999. Towards a hunger-free century. Available on: http://www.fao.org/dg/1999/millen-e.htm FAO, 2000a. FAO report: Enough food in the future--without genetically engineered crops. Available on: http://www.ucsusa.org/food_and_environment/biotechnology/pa ge.cfm?pageID=331 FAO, 2000b. Food and population: FAO looks ahead. Available on: http://www.fao.org/news/2000/000704-e.htm FAO, 2002. World agriculture 2030: Main findings. Available on: http://www.fao.org/english/newsroom/news/2002/7833-en.html FAO, 2003. World Agriculture: Towards 2015/2030. An FAO Perspective. London: Earthscan Publications Ltd. FAO, 2004a. Grassland and pasture crops in the Islamic Republic of Iran. Available on: http://www.fao.org/WAICENT/FAOINFO/AGRICULT/AGP/AGP C/doc/Counprof/Iran/Iran.htm FAO, 2004b. Food topsource. Available on: http://www.t21.ca/food/tp.htm FAO, 2004c. Community Forestry. Rapid Appraisal. Note Number 3. FAO Corporate Document Repository. Forestry Department. Fazel, G., 1971. Economic Organisation and Change among the Boyr Ahmad: A Nomadic Pastoral Tribe in Southwest Iran. Ph.D. Thesis. Berkeley: University of California. 149 Foran, B.D. & Stafford Smith, D.M., 1991. Risk, biology and drought management strategies for cattle stations in Central Australia. Journal of Environmental Management, 33: 17-33. Ford, D.N. and Sterman, J.D., 1998. Expert knowledge elicitation to improve mental and formal models. System Dynamics Review, 14(4), 309-30. Ferraro, D.O., Ghersa, C.M. & Sznaider, G.A., 2003. Evaluation of environmental impact indicators using fuzzy logic to assess the mixed cropping systems of the Inland Pampa, Argentina. Agriculture Ecosystems & Environment, 96, 1-18. Gamble, D., 1989. Two Techniques for Gathering Data and Building Rich Pictures in Rural and Community Development Action Research Projects. Hawkesbury: Hawkesbury Agricultural College. Gammon, D. M. & Roberts, B. R., 1978. Patterns of defoliation during continuous and rotational grazing of the Matopos Sandveld of Rhodesia. 1. Selectivity of grazing. Rhodesian Journal of Agricultural Research, 16, 117-31. Genesis, 2005. Online parallel Bible. Available on: http://bible.cc/genesis/6-5.htm GLEAM., 2002. A Forum on sustainability modeling and assessment: What's new? Available on: http://www.unu.edu/env/Gleam/Concept.html Glimp, H.A. & Swanson, S.R., 1994. Sheep grazing and riparian and watershed management. Sheep Research Journal, Special Issue, 65-71. Glossary of Terms Used in Range Management, 2002. Range sites of Florida. Glossary of terms used in range management. Available on: http://wfrec.ifas.ufl.edu/range/rangelands/glossary.htm Gowing, J., Tarimo, A. and El-Awad, O., 1996. A rational methods for assessing irrigation performance at farm level with the aid of fuzzy set theory. Irrigation and Drainage Systems, 10, 319-339. Grigg, G., 1995. Kangaroo harvesting for conservation of rangelands, kangaroos, and graziers. In: G.C. Grigg, P.T. Hale and Lunney D. (Eds.), Conservation Through Sustainable Use of Wildlife (161-165). Centre for Conservation Biology:University of Queensland. Grint, K., 1997. Fuzzy Management. Oxford: Oxford University Press. Hardesty, L., Lawrence, J.H., Gill, S.J. and Chapman, R.C., 1993. Private forest landowner’s perceptions of forest grazing in Washington State. Journal of Range Management, 46, 49-55. 150 Harris, J.M., 2001. Agriculture in a global perspective. Global Development and Environment Institute. Working Paper No. 0104. Available on: http://ase.tufts.edu/gdae/publications/working_papers/agric4.wo rkingpaper.PDF Hart, M., 2000. An introduction to sustainability. Available on: http://www.sustainablemeasures.com/Sustainability Hart, R.H., Bissio, J., Samuel, M.J. & Waggoner, J.W., 1993. Grazing systems, pasture size, and cattle grazing behavior, distribution and gains. Journal of Range Management, 46, 81-87. Hastie, T., Tibshirani, R. and Friedman, J., 2001. The Elements of Statistical Learning. Data Mining, Inference and Prediction. New York: Springer. Hawkins, R., 1997. Training in interdisciplinary team research for agricultural development: The experience of ICRA. European Journal of Agricultural Education and Extension, 4(1), 49-66. Heitschmidt, R.K., Dowhower, S.L. & Walker, J.W., 1987. 14- Vs. 42-Paddock rotational grazing: Above ground biomass dynamics, forage production, and harvest efficiency. Journal of Range Management, 40, 216-223. Hiernaux, P., 1996. The Crisis of Sahelian Pastoralism: Ecological or Economic? ODI Pastoral Development Network Paper 39a. 20p. London: Overseas Development Institute. Heemskerk, M., Wilson, K. & Pavao-Zuckerman, M., 2003. Conceptual models as tools for communication across disciplines. Conservation Ecology, 7(3), 8. [online] Available on: http://www.consecol.org/vol7/iss3/art8/ HMSO, 1994. Sustainable Development: The UK Strategy. Cm 2429. HMSO, London. Hodgkinson, K.C., 1996. A model for perennial grass mortality under grazing. In: N.E. West (Ed.), Rangelands in a Sustainable Biosphere. Proceedings Vth International Rangeland Congress, Vol. I., pp. 240-241. Denver: Society for Range Management. Hodgkinson, K.C. & Freudenberger, D.O., 1997. Production pulses and flow-ons in rangeland landscapes. In: J.A. Ludwig, D.J. Tongway, D.O. Freudenberger, J.C. Noble & K.C. Hodgkinson (Eds.), Landscape Function, Ecology and Management: Principles from Australia's Rangelands (23-34). Melbourne: CSIRO. 151 Hodgkinson, K.C. & Harrington, G.N., 1985. The case for prescribed burning to control shrubs in Eastern semi-arid woodlands. Australian Rangeland Journal, 7, 64-74. Holechek, J.L., & Hawkes, J., 1993. Desert and prairie ranching profitability. Rangelands, 15, 104-109. Holechek, J.L. & Hess, K., 1995. The emergency feed program. Rangelands, 17, 133-136. Holechek, J.L., Pieper, R.D. & Herbel, C.H., 1989. Range Management Principles and Practices. Englewood Cliffs: Regents/Prentice Hall. Holliday, C. and Pepper, J., 2005. Sustainability through the market: Our opportunity. Available on: http://www1.dupont.com/dupontglobal/corp/documents/US/en_U S/news/speeches/wbcsdreport.pdf Holmes, J.H., 1996. Diversity and change in Australia's rangeland regions: Translating resource values into regional benefits. Rangeland Journal, 19, 3-25. Howery, L.D., Sprinkle, J.E., and Bowns, J.E., 2000. A summary of livestock grazing systems used on rangelands in the Western United States and Canada. Cooperative Extension, College of Agriculture & Life Sciences, The University of Arizona. Iranian Nomadic Organization, 1992. Proceedings of Development Strategy of Iranian Nomadic Life. Iran: Ashayeri Publications. Iran Rozaneh, 2003. Nomads in Iran. Available on: http://www.rozanehmagazine.com/janfeb2003/aculture.html Irons, W. and Dyson-Hudson, N. (Eds.), 1972. Perspectives on Nomadism. Leiden: Brill. Irons, W., 1975. The Yomut Turkmen: A Study of Social Organization among a Central Asian Turkicspeaking Population. Ann Arbor: University of Michigan. ISAS, 2001. Biotech ‘Tailor-Made’ for Africa, researcher tells tufts conference. Council for Biotechnology Information, Washington, D.C., November 19, 2001. Available on: <index.asp?id=1156&redirect=con1309mid17%2Ehtml> Ishibuchi, H., Nakashima, T. and Morisawa, T., 1999. Voting in fuzzy rule-based systems for pattern classification problems. Fuzzy Sets and Systems, 103(1), 223-238. IUCN/IDRC (International Union for the Conservation of Nature/International Development Research Center), 1995. Assessing progress towards sustainability: a new approach. In: 152 Thadeus, Trzuna (Eds.), A Sustainable World: Defining and Measuring Sustainable Development (172-182). Sacramento. IUCN/UNEP/WWF (International Union for the Conservation of Nature/United Nations Environment Program/WorldWide Fund for Nature), 1991. Caring for the Earth: A Strategy for Sustainable Living. IUCN, Gland, Switzerland. Jameson, D.A., 1963. Responses of individual plants to harvesting. The Botanical Review, 29, 532-594. Jang, J.-S.R., Sun, C.-T. and Mizutani, E., 1997. Neuro-Fuzzy and Soft Computing, A Computational Approach to Learning and Machine Intelligence. Prentice Hall. Johnston, A. and Peake, R.W., 1960. Effect of selective grazing by sheep on the control of leafy spurge (Euphorbia esula L.). Journal of Range Management, 13,192-195. Jones, R.A., 1985. Research Methods in the Social and Behavioral Sciences. Massachusetts: Sinauer Association Inc. Kanwar, R.S., 2003. Environmental degradation and water security for the world population. Available on: http://www.ucowr.siu.edu/proc/T9B.pdf Kautza, T. and Gronski, R., 2003. Water ethic to renew the earth. Catholic Rural Life Magazine, 45(2). Available on: http://www.ncrlc.com/crl-magazine-articles/vol45no2/GronskiKautza.pdf Kauffman, J.B. & Krueger, W.C., 1984. Livestock impacts on riparian ecosystems and streamside management implications: a review. Journal of Range Management, 37, 430-438. Kenny, G., 2004. Report to the productivity commission. Cost and stocking rate data for the Murweh Shire. Available on: http://www.pc.gov.au/inquiry/nativevegetation/consultancy/murw eh1/murweh1.pdf Kopp, J., 2004. Stocking Rates and AUM- frequently Asked Questions. Alberta government. Department of Agriculture, Food and Rural Development. Kosko, B., 1993. Fuzzy Thinking. New York: Hyperion. Kothmann, M.M., Hinnant, R.T. & Taylor, C.A., 1997. The role of grazing management in overcoming juniper. Proc. Juniper Symposium. Texas Agric. Exp. Stn. Tech. Rep. Kovács, L. and Micsik, A., 2005. The design of voting and rating services within Web4Groups. Available on: http://dsd.sztaki.hu/publications/cscwid97bangkok/BankokCikkFinal.htm 153 Landsberg, J.J., James, C. and Morton, S., 1997. Assessing the effects of grazing on biodiversity in Australia's rangelands. Australian Biologist, 10, 153-162. Landsberg, J.J., O'Connor, T. G. and Freudenberger, D.O., 1999. The impacts of livestock grazing on biodiversity in natural ecosystems. In: Proceedings Vth International Symposium on the Nutrition of Herbivores. Landsberg, J.J. and Stol, J., 1996. Spatial distribution of sheep, feral goats and kangaroos in woody rangeland paddocks. Rangeland Journal, 18, 270-291. Larsen, J., 2002. Earth times. Population growth leading to land hunger. Available on: http://www.sdearthtimes.com/et0403/et0403s9.html Larsen, J., 2003. Population growth leading to land hunger. Earth Policy Institute. Available on: http://www.earthpolicy.org/Updates/Update21.htm Lacey, J. and Taylor, J., 2005. Montana guide to range site, condition and initial stocking rates. Available on: http://www.montana.edu/wwwpb/pubs/mt8515.html Lee, C.C., 1990. Fuzzy logic in control systems: Fuzzy logic controller–part I and II. IEEE Trans. Syst. Man Cyber., 20, 404435. Leigh, J. H. & Mulham, W.E., 1966. Selection of diet by sheep grazing semi-arid pastures of the Riverine Plain. 1. A bladder saltbush (Atriplex vesicaria) - cotton bush (Kochia aphylla) community. Australian Journal of Experimental Agriculture and Animal Husbandry, 6, 460-467. Lusigi, W. and Acquay, H., 1999. Challenges facing pastoralists in the drylands of subSaharan Africa. VIth International Rangeland Congress Proceedings, 2, 83-84. Lyons, R.K. and Machen, R.V., 2005. Stocking rate: The key grazing management decision. Available on: http://rangeweb.tamu.edu/extension/rangedetect/l5400_sr.pdf MAF, 2000. Impediments to optimising the economic and environmental performance of agriculture. Available on: http://www.maf.govt.nz/mafnet/rural-nz/profitability-andeconomics/performance/impediments-to-optimumperformance/impopt-05.htm Mamdani, E.H. and Gaines, B.R., 1981. Fuzzy Reasoning and its Applications. London: Academic Press. 154 Manske, L.L., 2004. A method of determining stocking rate based on monthly standing herbage biomass. Available on: http://www.ag.ndsu.nodak.edu/dickinso/research/2003/range03s. htm Marks, L.A. Dunn, E.G. Keller, J.M. and Godsey, L.D., 1995. Multiple Criteria Decision Making (MCDM) using fuzzy logic: An innovative approach to sustainable agriculture. IEEE Proceedings of ISUMA-NAFIPS ’95, Missouri-Columbia, USA. Marshall, C. and Rossman G.B., 1995. Designing qualitative research. Available on: http://www.geocities.com/zulkardi/turmudi.html Martin, S.C., 1978. Grazing systems -- What can they accomplish? Rangeman's Journal, 5, 14-16. McKeown, R., Hopkins, C., Rizzi, R. and Chrystalbridge, M., 2002. Education for Sustainable Development Toolkit. Energy, Environment and Resources Center, University of Tennessee. McMurphy, W.E., Gillen, R.L., Engle, D.M. & McCollum, F.T., 1990. The philosophical difference between range and pasture management in Oklahoma. Rangelands, 12, 197-200. Meadows, D., 1998. Indicators and Information Systems for Sustainable Development. A Report to the Balaton Group. Hartland: The Sustainability Institute. Mesbahi, M., 2004. Water wars. Available on: http://www.countercurrents.org/en-mesbahi251104.htm Mesdaghi, M., 1995. Management of Iran’s Rangelands. Mashhad: Imam Reza University. Milchunas, D. G., Sala, O.E., and Lauenroth, W.K., 1988. A generalized model of the effects of grazing by large herbivores on grassland community structure. The American Naturalist, 132, 87-106. Miller, D.J., 2005. Rangeland management in the Hindu KushHimalayas. International Centre for Integrated Mountain Development (ICIMOD). Available on: http://www.icimod.org/focus/rangelands/range3.htm Miller, R., 1995. Range Management: The Art and Science. The Pacific Northwest Section, Society for Range Management Newsletter. Milne, R., 1995. World food issues: Past and present. A Christian perspective on “population alarmism”. Available on: http://www.agron.iastate.edu/courses/agron342/milne.html Milne, R., 2002. World population. Available on: http://www.leaderu.com/orgs/probe/docs/worldpop.html 155 Mitchell, J.D., 1998. Before the next doubling. Available on: http://www.worldwatch.org/may/1998/98-1.html Mitchell, J.E., Joyce, L.A. and Bryant, L.D., 1999. Applicability of Montreal process criteria and indicators to rangelands. VIth International Rangeland Congress Proceedings, 1, 183-184. Morita, T., Nakicenovic, N. and Robinson, J., 2000. The Relationship between Technological Development Path and the Stabilization of Atmospheric Greenhouse Gas Concentrations in Global Emissions Scenarios. Center of Global Environmental Research (CGER). National Institute for Environmental Studies. Mortensen, I.D., 1993. Nomads of Luristan. History, Material Culture and Pastoralism in Western Iran. London: Thames & Hudson / Copenhagen: Rhodos. Mosley, J.C., 1994. Prescribed sheep grazing to enhance wildlife habitat on North American rangelands. Sheep Research Journal, Special Issue, 79-91. Munda, G., Nijkamp, P. and Rietveld, P., 1994. Qualitative multicriteria evaluation for environmental management. Ecological Economics, 10, 97-112. Namey-e-Otaq-e-Bazargani, 1996. The Life and Economy of Iranian Tribesmen in Iran. Journal of Chamber of Commerce, 3, 63-66. (The publication of the Chamber of Commerce, Industries and Mines of the Islamic Republic of Iran). [online]: http://www.kkhec.ac.ir/Iran%20information/life_and_economy_o f_iranian_trib.htm National Research Council. 1993. Sustainable Agriculture and the Environment in the Humid Tropics. Washington, DC: National Academy Press. Nationmaster, 2005. Encyclopedia: Sustainable development. Available on: http://www.nationmaster.com/encyclopedia/sustainabledevelopment Niblo, D.M. and Jackson, M.S., 1999. Qualitative data & private security in the hospitality industry: The hidden agenda. Available on: http://www.latrobe.edu.au/aqr/offer/papers/MNiblo.htm NGO Forum for Food Sovereignty, 2002. Policies for mobile pastoralism, rangeland management, and arid lands. Available on: http://www.iucn.org/themes/ceep/Wkg_grp/SL/pastoralism%20p olicies.pdf 156 NRCS, 2000. Animal unit. Delaware Nutrient Management Notes, 1(2). [online] Available on: http://www.state.de.us/deptagri/nutrients/newsltr/feb_00.shtml Nshimbani, J., 2004. Zamculture for human and social development. Culture and Rural Development Research. Available on: http://africa.eduprojects.net/agriculture/index.php?action%5B% 5D=IArticleShow::showArticle('352') OECD (Organization for Economic Cooperation and Development), 1994. Environmental Indicators. OECD, Paris. Oldcarpet, 2005. Qashqai. Available on: http://www.oldcarpet.com/qashqai.htm Olson, B.E. & Lacey, J.R., 1994. Sheep: A method for controlling rangeland weeds. Sheep Research Journal, Special Issue, 105112. Ortega, R., 2002. Studies in the subject of opinion leadership. Available on: http://www.ciadvertising.org/student_account/spring_02/adv382j /ortega/Theories.htm Owens, S., 2001. Salt of the earth. European Molecular Biology Organization. Available on: http://www.monsanto.co.uk/news/ukshowlib.phtml?uid=5890 PAI, 2005. Conserving land. Population and sustainable food production. Committee for the National Institute for the Environment. Available on: http://www.cnie.org/pop/conserving/landuse.htm Patton, M.Q., 1987. How to Use Qualitative Methods in Evaluation. London: SAGE Publications. Pearce, D.W., Atkinson, G.D., 1993. Capital theory and the measurement of sustainable development: an indicator of weak sustainability. Ecological Economics, 8, 103-108. Pedrycz, W., 1994. Why triangular membership function? Fuzzy Sets and Systems, 64, 21-30. Pedrycz, W., 2001. Fuzzy equalization in the construction of fuzzy sets. Fuzzy Sets and Systems, 119, 329-335. Pembleton, P.N., 2004. Sustainability and Impacts. Technical Analysis Report. California Department of Transportation, Transportation Energy Efficiency and Conservation Project. Pender, J., 1999. Rural population growth, agricultural change and natural resource management in developing countries: A review of hypotheses and some evidence from Honduras. International Food Policy Research Institute. Environment and Production 157 Technology Division. Available on: http://www.ifpri.org/divs/eptd/dp/papers/eptdp48.pdf Peters, T.J. & Waterman, R.H., 1982. In Search of Excellence Lessons from America's Best-Run Companies. New York: New York Harper & Row, Publishers. Peterson, E., 1998. What earth day means to minority citizens. Available on: http://www.nationalcenter.org/NPA193.html Phillis, Y.A. and Andriantiatsaholiniania, L., 2001. Sustainability: An ill defined concept and its assessment using fuzzy logic. Ecological Economics, 39, 435-456. Pickett, S.T.A., Burch, W.R., and Grove, J.M., 1999. Interdisciplinary research: maintaining the constructive impulse in a culture of criticism. Ecosystems, 2, 302-307. Pimentel, D. and Wilson, A., 2005. World population, agriculture, and malnutrition. Available on: http://www.energybulletin.net/3834.html Pimentel, D. and Giampietro, M., 2001. US population growth factsheet. Available on: http://www.balance.org/articles/factsheet2001.html Pimentel, D., Huang, X., Cordova, A., and Pimentel, M. 1996. Impact of population growth on food supplies and environment. Available on: http://dieoff.org/page57.htm Pinstrup-Andersen, P. and Cohen, M. J., 1998. Aid to developing country agriculture: Investing in poverty reduction and new export opportunities. International Food Policy Research Institute. Available on: <http://www.ifpri.org/2020/briefs/number56.htm> Pinstrup-Andersen, P. and Pandya-Lorch, R., 1999. Meeting food needs in the 21st century. How many and who will be at risk? Available on: http://www.ifpri.org/pubs/jhu/fed21century/fed21century_chapte r01.pdf Polley, H.W., 1997. Implications of rising atmospheric carbon dioxide concentration for rangelands. Journal of Range Management, 50, 561-577. Pretty, J.N., 1995. Regenerating Agriculture – Policies and Practice for Sustainability and Self-Reliance. Washington D.C.: Joseph Henry Press. Pretty, J. and Hine, R., 2001. Reducing food poverty with sustainable agriculture: A summary of new evidence. SAFE-World Research Project. University of Essex, Centre for Environment and 158 Society. Available on: http://www2.essex.ac.uk/ces/ResearchProgrammes/SAFEWexecs ummfinalreport.htm Qashqai.net., 2003. A short historical overview (A meta-narrative?). Available on: http://www.qashqai.net/history.htm Radcliffe, J.E., 1985. Grazing management of goats and sheep for gorse control. New Zealand Journal of Experimental Agriculture, 13, 181-190. Rao, L., 1999. Future harvest: Agriculture & overpopulation. Magazine of the Cousteau Society, October. Available on: http://www.futureharvest.org/news/calypso.shtml Raven, P.H., 2004. Feeding a hungry world: The moral imperatives of biotechnology. The Pontifical Gregorian University, Rome. Available on: http://vatican.usembassy.it/policy/events/Biotech2004/Raven.pdf Redfearn, D.D. and Bidwell, T.G., 2004. Stocking rate: The key to successful livestock production. Available on: http://www.osuextra.com Reich, P.F., Numbem, S.T., Almaraz, R.A. and Eswaran, H., 2001. Land resource stresses and desertification in Africa. In: E.M. Bridges, I.D. Hannam, L.R. Oldeman, F.W.T. Pening de Vries, S.J. Scherr, and S. Sompatpaint (Eds.), Responses to Land Degradation. Proc. 2nd. International Conference on Land Degradation and Desertification, Khon Kaen, Thailand. Oxford Press, New Delhi, India. Rennings, K. and Wiggering, H., 1997. Step towards indicators of sustainable development: Linking economic and ecological concepts. Ecological Economics, 20, 25-36. Richards, J.F., 1990. Land Transformation. In: B.L. Turner (Ed.), The Earth as Transformed by Human Action. New York: Cambridge with Clark University. Rigby, D., Woodhouse, P., Young, T. and Burton, M., 2001. Constructing a farm level indicator of sustainable agricultural practice. Ecological Economics, 39, 463-478. Riggio, R., Bomar, G. and Larkin, T., 1987. Texas Drought: Its Recent History (1931-1985). Texas Water Commission, Austin Texas LP 87-04. 74 pp. Roe, E.M., 1997. Viewpoint: On rangeland carrying capacity. Journal of Range Management, 50(5), 467-472. Roshier, D.A. and Nicol, H.I., 1998. Implications of spatio-temporal variation in forage production and utililisation for animal 159 productivity in extensive grazingsystems. Rangeland Journal, 20, 3-25. RS and NAS, 1992. The Royal Society and the National Academy of Sciences on Population Growth and Sustainability. Population and Development Review, 18 (2), 375-378. Ruspini, E.H., Bonissone, P.P. and Pedrycz, W. (Eds.), 1998. Handbook of Fuzzy Computation. Bristol and Philadelphia: Institute of Physics Publishing. Sam-Amoah L.K. and Gowing, J.W., 2001. Assessing the performance of irrigation schemes with minimum data on water deliveries. Irrigation and Drainage Systems, 50, 31-39. Savory, A. and Butterfield, J., 1999. Holistic Management: A New Framework for Decision Making. Washington D.C.: Island Press. Seers, D. 1982. The meaning of development. In: G.E. Jones and M.J. Rolls (Eds.), Progress in Rural Extension and Community Development, 1, 7-24. John Wiley and Sons, Ltd. Sands, D.M., 1986. Farming systems research: clarification of terms and concepts. Expl. Agric. 22, 87-104. Serf Publishing Inc., 2001. Food & Agriculture...A Growing World Crisis! The U.S. Dept. of Agriculture, the former Ambassador Agricultural Research Department, Agresource Management Associates, Agrotech Research Group. Shah, M. and Strong, M., 1999. Food in the 21st century: From science to sustainable. Consultative Group on International Agricultural Research (CGIAR). Available on: http://www.worldbank.org/html/cgiar/publications/shahbook/sha hbook.pdf Shahvali, M. and Azadi, H., 1999. Designing of extension research on the basis of paradisciplinary. A Quarterly Journal of Ightesade Kishavarzi Va Towse'e 2(4), 21-56. Shaner, W.W., Philip, P.F. and Schmehl, W.R., 1982. Farming Systems Research and Development: Guidelines for Developing Countries. Westview Press. Sherp, J., 1994. What does an economist need to know about the environment? Directorate General for Economics and Financial Affairs. Economic Papers of the European Commission No. 107. Sicat, R.S., Carranza, E.J.M. and Nidumolu, U.B., 2005. Fuzzy modeling of farmers’ knowledge for land suitability classification. Agricultural Systems 83, 49-75. Silvert, W., 1997. Ecological impact classification with fuzzy sets. Ecological Modelling, 96, 1-10. 160 Sogol, No date. Tribes of Iran. Available on: http://www.sogol.com/WHP/IINT/IRTRIBE.htm Stafford Smith, M., 1996. Management of rangelands: Paradigms at their limits. In: J. Hodgson and A.W. Illius (Eds.), The Ecology and Management of Grazing Systems (325-357). Wallingford: CAB International. Stoddart, L.A. and Smith, A.W., 1943. Range Management (First Edition). McGraw-Hill, New York. Stoddart, L.A. and Smith, A.W., 1955. Range Management (Second Edition). McGraw-Hill, New York. Stoddart, L.A., Smith, A.W. and Box, T., 1975. Range Management (Third Edition). McGraw-Hill, New York. Strauss, A. and Corbin, J., 1990. Basic Qualitative Research Grounded Theory Procedures and Techniques. London: SAGE. Stuth, J.W. and Lyons, B.G. (Eds.), 1993. Decision Support Systems for the Management of Grazing Lands: Emerging Issues. Carnforth: The Parthenon PublishingGroup Inc. Swee, G.M., 1981. Sedentarization: Change and Adaptation among the Kordshuli Pastoral Nomads of Southwestern Iran. Ph.D. Thesis. Michigan: Michigan State University. Taghi Farvar, M. and Jandaghi, M., 1998. Thematic Programme Network III for Regional Co-operation under the Convention to Combat Desertification (CCD). Tehran: The Government of the Islamic Republic of Iran and the Interim Secretariat of the Convention to Combat Desertification. Tagaki, T. and Sugeno, M., 1985. Fuzzy identification of systems and its applications to modeling and control. IEEE Transactions on Systems, Man and Cybernetics, 15(1), 116–132. Takle, E.S., 2001. Sustainable development. Available on: http://www.iitap.iastate.edu/gccourse/develop/sust_develop_lectu re.html Tapper, R., 1979. Pasture and Politics: Economics, Conflict and Ritual among the Shahsevan Nomads of Northwestern Iran. London: Academic Press. Tapper, R. (Ed.), 1983. The Conflict of Tribe and State in Iran and Afghanistan. London: Croom Helm. Tapper, R., 1997. Frontier Nomads of Iran. Cambridge: Cambridge University Press. Tesch, R., 1990. Qualitative Research: Analysis Types and Software Tools. New York: The Falmer Press. 161 Thompson, P.B., 1995. The Spirit of the Soil: Agriculture and Environmental Ethics. New York: Routledge. The Online Newsletter of the Bahá'í International Community, 1996. The importance of agriculture. The Online Newsletter of the Bahá'í International Community, 8(2). [online] Available on: http://www.onecountry.org/oc82/oc8202as.html Travel Explorations, 2004. Trekking through the lands of the Iranians nomads – Part 1 of 2. Available on: http://www.travelexplorations.com/cparticle175488-18553a.html Truelsen, S., 2003. Agriculture Faces Huge Task in Scenario for the Future. Available on: http://www.fb.org/views/focus/fo2003/fo0113.html UN, 2000. Sustainable agriculture and rural development. Available on: http://www.un.org/documents/ecosoc/cn17/2000/ecn1720007add3.htm UNCED (United Nations Conference on Environment and Development), 1992. Agenda 21. New York: UNCED. UNCT, 2001. United Nations inter-agency assessment report on the extreme drought in the Islamic Republic of Iran. Available on: http://www.reliefweb.int/w/Rwb.nsf/0/08951820c13c2bc5c1256a 85005081e0?OpenDocument UNDP, 2001. Human development report. Available on: http://hdr.undp.org/reports/global/2001/en/pdf/pr2.pdf UNDP, 2003. Earth policy news – Population growth & land hunger. Available on: http://www.undp.org.vn/mlist/envirovlc/012003/post74.htm UNEP, 1992. Caring for the Earth: A Learner’s Guide to Sustainable Living. New York: United Nations. UNEP, 2000. Atmosphere. Available on: http://www.unep.org/geo2000/english/0048.htm UNESCO, 1996. Enhancing Social Scientific Understanding of Sustainability. Declaration on the occasion of the conference "Sustainability as a Social Science Concept" within the UNESCO-MOST Programme, held in Frankfurt, November 1996. Available on: http://www.unesco.org/most/sustdecl.htm UNFPA, 2003. State of World Population 2003. Making 1 Billion Count: Investing in Adolescents’ Health and Rights. United Nations Population Fund. Available on: http://www.unfpa.org/swp/2003/english/ch1/ UNFPA, 2005. Population & demographic dynamics. Available on: http://www.unfpa.org/sustainable/demographics.htm 162 UNPD, 2003. World Population Prospects. The 2002 Revision (Vol. I: Comprehensive Tables). New York: UN Population Division. van den Berg, J., 2004. Fuzzy methodologies for evaluating sustainable development. ICMCS Conference, Nigeria: Lagos. van der Werff, T.J., 1998. Threats and opportunities in a new century. Available on: http://www.globalfuture.com/9512.htm Waggoner, P., 1994. How Much Land Can Ten Billion People Spare for Nature? Council for Science and Technology, Ames, Iowa. Walker, J.W., 1994. Multi-species grazing: the ecological advantage. Sheep Research Journal, Special Issue, 52-64. Walker, J.W., 1995. Viewpoint: Grazing management and research now and in the next millennium. Journal of Range Management, 48(4), 350-357. Walker, J.W. and Hodgkinson, K.C., 2000. Grazing management: New technologies for old problems. http://sanangelo.tamu.edu/walker/incpaper.htm WCED (World Commission on Environment and Development), 1987. Our Common Future. Oxford University Press, Oxford. Westoby, M., Walker, B.H. & Noy-Meir, I., 1989. Opportunistic management for rangelands not at equilibrium. Journal of Range Management, 42, 266-274. Wikipedia, 2005. Sustainable development. Available on: http://en.wikipedia.org/wiki/Sustainable_development Williams, J.K., 2000. Churchpeople, socialism, and capitalism. Available on: http://www.libertyhaven.com/noneoftheabove/religionandchristia ns/churchpeople.html Wilson, E., 1999. One third of world's population 2.7 billion people will experience severe water scarcity by 2025, says new study. Future Harvest. Available on: http://www.futureharvest.org/news/03171999.shtml Wilson, A.D. and Macleod N.D., 1991. Overgrazing: Present or absent? Journal of Range Management, 44(5), 475-482. World Bank, 1998. Population growth rate. Available on: http://www.worldbank.org/depweb/english/modules/social/pgr/in dex03.html World Resources Institute, 1994. World Resources 1994-95. New York: Oxford. Yin, R.K., 1989. Case study research design and methods. Available on: http://www.geocities.com/zulkardi/turmudi.html 163 Zadeh, L.A., 1973. Outline of a new approach to the analysis of complex systems and decision processes. IEEE Transactions on Systems, Man, and Cybernetics, 1 (1), 28–44. Zadeh, L.H., 1965. Fuzzy sets. Information and Control, 8, 338–353. Zilberman, D. and Alix, J., 2005. Principles of environmental policymaking. Department of Agricultural and Resource Economics University of California at Berkeley. Available on: http://are.berkeley.edu/courses/ARE253/2004/handouts/envpolic ymaking.pdf Zimmermann, H.J., 1996. Fuzzy Set Theory and its Applications (Third Edition). Boston: Kluwer Academic Publishers. 164 APPENDIX 1 General information of data collection Number of studied area 3 regions Number of data collection 3 rounds Spent time of data collection 18 months Number of conducted major interviews 9 sessions Number of conducted minor interviews 21 follow-ups Average time of major interviews at first round 3 hours Average time of major interviews at second round 3 hours Average time of major interviews at third round 5 hours Average time of minor interviews 2 hours 165 APPENDIX 2 If–Then Rules EAFL Model: 166 Multi Fuzzy Model: Model 1 167 168 169 Model 2 170 171 172 173 174 175 Model 3 176 177