LAND USE AND LAND COVER CHANGE, DRIVERS AND ITS IMPACT: A COMPARATIVE STUDY FROM KUHAR MICHAEL AND LENCHE DIMA OF BLUE NILE AND AWASH BASINS OF ETHIOPIA A Thesis Presented to the Faculty of the Graduate School of Cornell University In Partial Fulfillment of the Requirements for the Degree of Master of Professional Studies By Hussien Ali Oumer August 2009 © 2009 Hussien Ali Oumer ABSTRACT Land use and land cover change is driven by human actions and also drives changes that limit availability of products and services for human and livestock, and it can undermine environmental health as well. Therefore, this study was aimed at understanding land use and land cover change in Lenche Dima and Kuhar Michael of Amhara region, Ethiopia. Time-series satellite images that included Landsat MSS, TM, ETM+ and ASTER, which covered the time frame between 1972/3 to 2005, were used. Socio-economic Survey and review of documents was carried out to understand historical trends, collect ground truth and other secondary information required. Analysis of data and other data was accomplished through integrated use of ERDAS imagine (version 9.1), ENVI (version 4.3) and ArcGIS (version 9.2) software packages along with Microsoft office analytical tools. Remote sensing analysis revealed landscape level change of cultivated land to have a net increase in Kuhar Michael, while a decline is found for Lenche Dima. However, socio-economic surveys showed that household level cultivated land has decreased from 1.2ha to 1ha and from 2.2ha to 1.8ha in Kuhar Michael and Lenche Dima respectively, over the last 30years. Major contributing factors included population increase, occurrence of drought, land redistribution, and land degradation. Similarly, average land holding per household has decreased from 1.6ha to 1.5ha and from 2.9ha to 2.2ha in Kuhar Michael and Lenche Dima, respectively. This has jeopardized the capacity of individuals to provide land for their siblings further leading to landlessness, which is becoming a common phenomenon among rural youths. In Kuhar Michael, dense shrub/bush land decreased at an annual rate of -0.1%, while open shrub/bush land increased at a rate of 0.3%. As opposed to this, dense shrub/bush land increased at a rate of 0.2% and open shrub/bush land declined at annual rate of -0.2% in Lenche Dima. Grassland showed a net decrease at a rate of -0.3% in Kuhar Michael due to conversion into cultivated lands, while an increase with annual rate of 0.1% is found in Lenche Dima as a result of implementation of watershed management practices. Along with the observed decrease in vegetation cover, Limited availability and extinction of some tree/shrub species is also reported and research is required to quantify changes and understand the real impacts brought about. Key words: Land Use and land Cover Change, Land degradation, satellite imagery, implication for management, watershed management, Ethiopia, East Africa iii BIOGRAPHICAL SKETCH He was raised and completed High School at HJSSS in Haik town, South Wollo, Ethiopia. He got a Bachelor in Forestry in 2005 from Wondo Genet College of Forestry and Natural Resources with in Hawassa (formerly known as Debub) University. Right after graduation, he was involved in a base line survey conducted jointly by Organization for Rehabilitation and Development in Amhara (ORDA) and German Agro Action (GAA) in his home town. Then after, he joined the natural resources division of Habru district Bureau of Agriculture and Rural Development, in North Wollo Ethiopia and served as a Forestry/Soil and Water Conservation expert for about 8 months. Due to his keen interest to get involved in research endeavor, he was appointed and worked for four months as a junior researcher in the Forestry Research Center of the Ethiopian Institute of Agriculture Research. Later, he moved to the Forestry Research division of Adet Agricultural Research Center of the Amhara Regional Agricultural Research Institute (ARARI) and worked as a junior Researcher for one year and 3 months. Despite his interest to stay longer and involve in research activities, he left ARARI due to his strong desire to join the new Masters program in International Agriculture & Rural Development with concentration on Integrated Watershed Management and Hydrology, offered by Cornell University in Bahir Dar, Ethiopia. Fueled by the inspiration he get from his Professors, Hussien is very interested to pursue a PHD in the areas of climate change impacts and its adaptation. iv Dedicated to My best friend, Sebastian Kaempf v ACKNOWLEDGMENTS My sincere gratitude goes to Dr.Katrien Descheemaeker; my supervisor from ILRI/IWMI, for her continued and unreserved support at every stages of doing this research. The support and guidance I received from Professor Tammo Steenhuis and Dr. Zachary Easton is highly appreciated. Dr. Lisa Robello, Suraj and Ahmed Amdihun and Melakneh Gelete are also appreciated for their supervision and support in remote sensing/GIS issues. This research was conducted with financial support from a collaborative project by IWMI and ILRI entitled “crop-livestock water productivity improvement in Sub-Saharan Africa” and I am very thankful for the financial, technical support, and services I received from the organizations while spending 7 months as a graduate fellow. I am very thankful to Professors Tammo Stenhuis, Dr.Tilahun Amede, and Dr.Katrien Descheemaeker for providing me the fellowship and for their supervision. The administrative staffs also take credit for facilitating logistic support. I would like to extend my earnest appreciation to Sisay Demeku, Getachew Tesfaye, and Samuel Tesfaye, who shared with me important data and experiences. I am very thankful for the hard working and insightful professors who traveled far distance from USA to offer courses, and they also take credit for sharing their experiences and their guidance as well. Dr. Amy Collick is highly appreciated for effective coordination of the program, and for her critical support and guidance during the whole study period including formatting this thesis. I am thankful to all the respondents, development agents and experts in Gubalfto and Fogera district for sharing their experience and facilitation vi during field work. I am very thankful to Jennifer Ward as she has contributed to my work by editing earlier version of the manuscript. My special appreciation goes to Girum Yimer, Nega Ewnetu, Shiferaw Abate and his Wife Hiwot, Habtemariam Assefa, Andinet Ayalew, fellow students, other friends and colleagues who provided me support and encouragement. Words can’t explain the respect and love I have to my best friend, Sebastian Kaempf, who brought me to this track and his continued encouragement and critical financial support enabled me to finish my study. The support, love and care I always get from my beloved parents (Fatie Ebrahim, Workneh Zele, Aster Ali, Tekeste, Minassie and Mekbib Workneh) are always appreciated and remain to have important role for my success. vii TABLE OF CONTENTS BIOGRAPHICAL SKETCH ............................................................................. iii ACKNOWLEDGMENTS .................................................................................. v TABLE OF CONTENTS ................................................................................. vii LIST OF FIGURES .......................................................................................... xi LIST OF TABLES ......................................................................................... xiii LIST OF ABBREVIATIONS ............................................................................ iii CHAPTER ONE ............................................................................................... 1 INTRODUCTION AND BACKGROUND ........................................................ 1 GENERAL OBJECTIVE ............................................................................. 7 SPECIFIC OBJECTIVES ........................................................................... 7 THESIS ORGANIZATION .......................................................................... 7 CHAPTER TWO ............................................................................................... 9 LITERATURE REVIEW ................................................................................. 9 DEFINITIONS AND RATIONALE OF LAND USE AND LAND COVER CHANGE STUDY ...................................................................................... 9 APPLICATION OF REMOTE SENSING FOR LAND USE AND LAND COVER CHANGE .................................................................................... 13 Basics of remote sensing...................................................................... 18 Approaches in image classification ....................................................... 20 APPROACHES IN LAND USE AND LAND COVER CHANGE DETECTION ............................................................................................ 21 CHAPTER THREE ......................................................................................... 24 DESCRIPTION OF THE STUDY AREAS AND IMAGE ANALYSIS ............. 24 viii GEOGRAPHIC LOCATION ..................................................................... 26 FIELD SURVEY AND DATA COLLECTION ............................................ 27 DESCRIPTION OF LAND USE AND LAND COVER CLASSES ............. 28 STEPS IN ANALYSIS OF SATELLITE IMAGES FOR LAND COVER CHANGE ................................................................................................. 31 Image Calibration and its Importance ................................................... 32 Layer Stacking ...................................................................................... 35 Geo-Referencing of Images.................................................................. 37 IMAGE CLASSIFICATION ....................................................................... 39 General Considerations and Decision Rules during unsupervised classification ......................................................................................... 40 Supervised Classification of ASTER Image .......................................... 46 Post Classification Operations .............................................................. 47 CHAPTER FOUR ........................................................................................... 49 RESULTS AND DISCUSSION .................................................................... 49 RESULTS OF REMOTE SENSING AND GIS ANALYSIS OF TIME- SERIES SATELLITE IMAGES ................................................................. 49 Land Use and Land Cover Dynamics in Kuhar Michael ....................... 49 LAND USE AND LAND COVER DYNAMICS IN LENCHE DIMA ......... 53 Accuracy Assessment .......................................................................... 57 DISCUSSION ON LAND USE LAND COVER CHANGE OF THE STUDY AREAS ..................................................................................................... 57 RESULT FROM SOCIO-ECONOMIC SURVEY ...................................... 64 KUHAR MICHAEL ................................................................................... 64 Changes Associated with Livestock Feed Resources .......................... 64 Challenges Related to Grazing Land .................................................... 69 ix Evolution and Challenges of Fertilization Use ...................................... 69 Trends in Crop Production and Associated Shifts ................................ 70 Insight over Shifts in Types of Crops Grown ......................................... 71 Shifts in Crop Production and associated Issues .................................. 72 Causes of Expansion of Cultivated Land .............................................. 72 LENCHE DIMA ........................................................................................ 73 Evolution in Management of Grazing Land and Feed Sources............. 73 Preference Ranking of Different Sources of Animal Feed .................... 75 Emerging Issues, State, and Challenges of Crop Production ............... 76 Overview of Land Fragmentation and Its Impacts ................................ 77 Land Certification Process and Its Impacts .......................................... 78 Evolution of Land Management and Associated Challenges ................ 78 Local Land Administration and Emerging Issues .................................. 80 Rationale and Insight over Area Closure Management ........................ 81 Challenges in Community-Based Management of exclosures.............. 82 Area Closure and Its Influence over Animal Feed Availability............... 84 Misuse of Hillsides and Its Consequences ........................................... 85 Land Use and Land Cover Change and Climatic Variables as a Proxy 86 COMPARISON OF CHANGE IN LAND AND LIVESTOCK RESOURCES Insight over Changes in Household Level Land Holdings ........................ 86 Major changes associated with Livestock resource .............................. 91 CHAPTER FIVE .......................................................................................... 96 CONCLUSION AND RECOMMENDATION ................................................ 96 CONCLUSIONS ...................................................................................... 96 RECOMMENDATIONS .......................................................................... 101 REFERENCES ............................................................................................. 103 x APPENDICES .............................................................................................. 112 APPENDIX I: HOUSEHOLD LEVEL QUESTIONNAIRE ........................................ 112 APPENDIX IIA: NUMERIC VALUE OF THE VARIOUS PARAMETERS UTILIZED FOR AUTOMATED CALIBRATION OF THE DIFFERENT VERSIONS OF LANDSAT IMAGES IN ENVI. ......................................................................................................... 118 APPENDIX IIB: NUMERICAL VALUE OF DIFFERENT CALIBRATION PARAMETERS USED FOR CALIBRATION OF ASTER IMAGE. ................................................... 119 ANNEX IIC: DETAIL CALCULATIONS MADE TO PRODUCE OUT PUT USED FOR ASTER CALIBRATION WIT IN THE BAND MATH OPERATOR OF ENVI .................. 120 APPENDIX IIIA: DETAILS OF GCPS USED FOR GEOREFERENCING OF ASTER IMAGE FOR KUHAR MICHAEL ......................................................................... 121 APPENDIX IIIB: DETAILS OF GCPS USED FOR GEOREFERENCING OF ASTER IMAGE FOR LENCHE DIMA WATERSHED .......................................................... 122 xi LIST OF FIGURES Figure 1: Comparison of Spectral Bands between ASTER and Landsat-7 Thematic Mapper (source: Abrams et al., 2009) ............................................ 18 Figure 2: Location Map of Kuhar Michael Peasant Association (D) From Fogera District (B, Left), and Lenche Dima Watershed (E) From Gubalafto District (C, Right), Amhara Region, Ethiopia. ................................................. 25 Figure 3: General shape of Kuhar Michael (left) PA and Lenche Dima watershed (right) used for sub-setting the time series images ....................... 39 Figure 4: Summary of major steps followed during satellite image analysis for land use and land cover change detection ..................................................... 48 Figure 5: Land use and land cover map of Kuhar Michael in 1973 ................. 50 Figure 6: Land use and land cover map of Kuhar Michael in 1985 ................. 51 Figure 7: Land use and land cover map of Kuhar Michael in 1999 ................. 52 Figure 8: Land use and land cover map of Kuhar Michael in 2005 ................. 53 Figure 9: Land use and land cover map of Lenche Dima in 1972 .................. 54 Figure 10: Land use and land cover map of Lenche Dima in 1986 ................ 55 Figure 11: Land use and land cover map of Lenche Dima in 2000 ................ 56 Figure 12: Land use and land cover map of Lenche Dima in 2005 ................ 57 Figure 13: Summary of land use and land cover change in Kuhar Michael from 1973-2005 ...................................................................................................... 59 Figure 14: Summary of land use and land cover changes in Lenche Dima from 1973-2005 ...................................................................................................... 62 Figure 15: Summary of household level change in average total land holding in the study areas. .............................................................................................. 90 xii Figure 16: Trends in household level change in population of cattle (a) pack animals (b), small ruminants(c), and average total livestock (d) in the study areas. ............................................................................................................. 93 xiii LIST OF TABLES Table 1: Spectral property of various versions of Landsat images (source: Engineering manual, 2003) ............................................................................ 16 Table 2: Land use and land cover classes considered and their description (adapted and modified from Yitaferu, 2007) ................................................... 31 Table 3: Details about the satellite images used for this study ....................... 32 Table 4: Appearance of various surface features in different band combinations (Source: Tucker et al., 2004) .................................................... 45 Table 5: Accuracy assessment report of ASTER image in the two study areas ....................................................................................................................... 58 Table 6: Summary of magnitude and rates of change in land use and land cover of the study areas over the entire study period (from 1972/73 to 2005). ....................................................................................................................... 63 Table 7: Ranking of different animal feed sources/types in Kuhar Michael (% of the respondents) ............................................................................................ 66 Table 8: Ranking of different animal feed sources/types in Lenche Dima ...... 75 Table 9: Distributed exclosure in Lenche Dima watershed (source Laste Gerado kebele Agriculture and Rural Development office)............................. 84 Table 10: Characteristics of interviewed households and their land holding. . 89 Table 11: Summary of the average number of animals of different livestock types based on socio-economic survey .......................................................... 91 iii LIST OF ABBREVIATIONS ASTER Advanced Space Borne Thermal Emission & Reflection Radiometer ERDAS Earth Resources Data Analysis System DEM Digital Elevation Model DN Digital Number EPA Environmental protection authority ER Electromagnetic radiation ESUN Exo-atmospheric spectral irradiance ET Evapotranspiration ETB Ethiopian Birr ETM+ Enhanced Thematic Mapper ENVI Environment for Visualization of Images EOS Earth observing system EROS Earth resources observation and science EVI Enhanced Vegetation Index FAO Food and Agriculture Organization GCPs Ground Control Points GDP Gross Domestic Product GLOBE Global Land One-km Base Elevation EMA Ethiopian Mapping Authority GIS Geographic Information System GPS Geographic Positioning System HDF Hierarchical Data Format ILRI International Livestock Research Institute IWMI International Water Management Institute iv JPEG Joint Photographic Experts Group LUCC Land Use and Land Cover Change LULC Land Use and Land Cover M a s l Meters above sea level MSS Multi Spectral Scanner NDVI Normalized Difference Vegetation Index RMSE Root Mean Square Error RS Remote Sensing SWIR Short wave infrared TIR Thermal Infrared TLU Tropical Livestock Unit TOA Top of Atmosphere TM Thematic Mapper USGS United States Geological Survey UTM Universal Trans Mercator VNIR Visible To Near Infrared WGS World Geodetic System 1 CHAPTER ONE INTRODUCTION AND BACKGROUND Land use and land cover dynamics are widespread, accelerating, and significant processes driven by human actions but also producing changes that impact humans (Agarwal et al., 2002). These dynamics alter the availability of different biophysical resources including soil, vegetation, water, animal feed and others. Consequently, land use and cover changes could lead to a decreased availability of different products and services for human, livestock, agricultural production and damage to the environment as well. In Ethiopia, the availability of natural resources as well as their dynamics and management vary considerably from area to area. For instance, different parts of the Ethiopian highlands receive between 600 and 2700 mm of rainfall annually (Zeleke and Steenhuis, 2005). Besides high rainfall variability, also water shortages are prevalent in the Ethiopian highlands. For the coming decades, it is generally estimated that water withdrawal and consumption will increase as a result of increasing food needs by a growing population (CA, 2007). Attributed to large runoff coefficients, more than 90% of the annual discharge of rivers in the Lake Tana basin is lost during the rainy months, creating serious shortages of water in the dry season for irrigation agriculture and domestic use (Yitaferu, 2007). In Lenche Dima, a relatively small catchment, runoff coefficients were found to be around 5% (McHugh et al., 2007). This watershed is characterized by a rain-fed subsistence mixed farming system with prevailing problems like soil erosion, poor soil fertility, 2 shortage of rainfall, livestock water shortage, poor grazing/dry season feed shortage and shortage of wood for house construction (Gizaw et al., 1999). The traditional management of land-based resources, coupled with a growing interest and reliance on various products and services from those resources poses a challenge for managing the natural resources. Changes in land use and land cover conditions and agricultural water management practices in irrigation could be responsible for the problems associated with hydrological resources of the Lake Tana Basin (Yitaferu, 2007). Inappropriate allocation and utilization, lack of capacity to develop and use poorly accessible water resources, loss of water due to its seasonality and run off are some of the problems associated with the water resources in the basin .Therefore, producing more food under conditions of increasing water scarcity and without creating further environmental degradation is a challenge being faced (CA, 2007). Mixed crop-livestock farming is the main economic activity in Amhara region with cattle kept as the most important livestock for traction and milk production (Mwendera et al., 1997). Livestock provide essential commodities and services to much of the world's population and serve as a strategic reserve that reduces risk and adds stability to farming system. Livestock provide hides and skins, draught power and manure to enhance soil fertility. They form an integral part of the social fabric for many people in Ethiopia, while they serve as a capital reserve available for bad times (Hugo et al., 1997 in Mwendera et al., 1997). In Ethiopia, livestock contribute about 30 to 35 percent of agricultural gross domestic product (GDP) and more than 85 percent of farm 3 cash income (Benin and Pender, 2006). In particular, the Ethiopian highlands support over 90% of the human population, 75% of livestock, and 95% of annual cropland in the country (FAO, 2006). Furthermore, a livestock study in Ethiopia indicated that animal production is a key enterprise supporting and sustaining the livelihoods of the rural dwellers (80% of Ethiopia’s total population) (Amsalu et al. 2006). However, it is likely that variation in the degree of integration and balance between the crop and livestock components in various districts and localities depend on resource availability and biophysical setting. Cattle distribution and average livestock holding per household in Amhara region vary from one district to another depending on human population, resource availability, prevalence of animal diseases, and size of the land holding (BoARD 2003) and total livestock population consists of 49 percent of cattle, which is 80 percent in tropical livestock unit (TLU). Hence, cattle are considered the most important livestock species due to their significant contribution to traction power for cultivation, and threshing, and they produce the largest proportion of milk and meat for human consumption as well. However, the livestock population is growing asymptotically and thus failed to keep pace with the rapid human population growth (Ashine 2002 in McCornick et al., 2003). Livestock mortality rates are also very high, with maximum rates up to 40-50% in bad years for sheep and goats, and 10-20% for cattle. Bad years are characterized by severe water and feed shortages and disease outbreaks (Descheemaeker, 2008). The available livestock feed resources, including grazing, natural pastures and crop residues are inadequate with uneven distribution throughout the year. Because of its seasonality in quality and amount, the natural pasture cannot 4 support the livestock’s natural feed requirements, and animals often lose their live weight, especially during the dry season. A survey conducted in Amhara region indicated that livestock productivity is low due to water shortages, weed infestation, dry season feed shortages, shortage of grazing lands, poor marketing conditions, and diseases (Gizaw et al., 1999). Grazing systems offer only limited potential for intensification, and livestock production is becoming increasingly crop-based. Therefore, among other issues, understanding the availability and seasonal variation of animal feed resources is very important. This in turn will assist efforts to improve livestock productivity, and its contribution to the household food economy and income of smallholder farmers. In response to climate change and land degradation, people develop various strategies to adapt to changing conditions. For instance, farmers in the Nile Basin in Ethiopia have developed adaptation strategies to mitigate the negative impacts of climate change (Harrington et al., 2007). These include using more irrigation, changing crop varieties or crops, shifting planting dates, and shifting from crop cultivation to livestock production. The influence of climate change on agricultural production, natural resource management, and subsequent response mechanisms vary greatly from one area to another. Different projects, basin priorities and research agendas for increasing water- use efficiency for food production through better livestock management in the Nile River Basin have been identified (Harrington et al., 2004). These include integrated basin management systems, crop water productivity improvement 5 and building the capacity of the people in the catchments. This needs knowledge on how natural resources evolve over time so as to devise alternative and appropriate strategies for exploiting the resources available. Therefore, this study is intended to understand land use and land cover changes, their drivers and impact on vegetation dynamics and animal feed availability in Kuhar Michael and Lenche Dima watersheds, which are both comprised of mixed crop-livestock production systems with some differences in agro-ecological conditions. For instance, the geology, topography and vegetation found in Lenche Dima, along with low and unreliable rainfall and dry spell, have dictated its hydrological condition creating soil moisture stress and limited availability of fresh river water. The basalts of the mountainous terrain in the area are highly fractured and filled by secondary carbonate minerals, while the valley bottom is filled with silty clay sediments. The very steep topography further contributes to the absence of springs and streams in the area (Gizaw et al., 1999). As a result, there is no practice of irrigation and even very limited potential for irrigation agriculture exists except for those individuals who own land in Alewuha, an adjacent locality with small irrigation schemes and practices near Alewuha River. Hence, crop production is highly dictated by the occurrence of rainfall resulting in rain-fed agriculture that allows producing only one crop per year in most cases. However, Kuhar Michael, forming part of the Fogera plain, has better agricultural potential and irrigation is possible due to the presence of Gumera and other small rivers. As a result, the spread and diversification of cash crops through expansion of irrigation agriculture and the production of up to three crops per year is made possible. Besides the aforementioned 6 differences, other variations in biophysical conditions, socio-economic and cultural practices and land use and land cover conditions are expected, which will be dealt with in brief in this study. This will further help to generate better understanding about local knowledge and coping strategies, which will form the basis to assist and formulate projects geared toward improving inhabitants’ livelihoods by counteracting undesirable changes in land use and land cover and subsequent climate change. Some research has been conducted in Amhara Region, but there is significant variation in the level of analysis performed and purpose and output of the studies. Furthermore, some of the contributing factors and processes are likely to be area specific and evolve over time as well. Hence, some of the outputs and resulting influences may not be applicable in other areas. On top of all these, continuous iterations of such studies by incorporating new approaches and taking advantage of advancements in science are needed to better understand continuously evolving processes. Backing up on-going and directing intended development projects will further help in securing a well informed resource use and management system, minimizing undesirable outcomes from traditional resources management. It is hoped that this study will provide information for decision makers and development practitioners about the magnitude and dimensions of long term land use and land cover changes, their drivers, impacts and community mitigating strategies in the study areas and surrounding. Understanding such changes is critical for formulating effective environmental policies and management strategies (Agarwal et al., 2002). Therefore, based on the 7 information generated, issues that need immediate action will be identified and prioritized. In addition, the study will give an overview of the various ranges of local knowledge and practices within the study areas from which we can plan and formulate future projects. GENERAL OBJECTIVE The study aims to identify and compare changes in land use and land cover, their drivers and impact on vegetation and animal feed dynamics in water abundant (Kuhar Michael) and drought prone (Lenche Dima) areas from the Nile and Awash basins of Ethiopia. SPECIFIC OBJECTIVES This study encompasses several specific research objectives. Among these, the following may be considered to be the primary objectives: 1. To understand changes in land use and land cover occurring in Kuhar Michael and Lenche Dima catchments based on analysis of remotely sensed data 2. To identify and compare drivers of land use and land cover change in the study areas 3. To investigate the impact on vegetation, crop production and animal feed dynamics in the study areas. THESIS ORGANIZATION Complete presentation of the thesis includes five separate but connected chapters including this introductory chapter. Chapter one gives background information about the study. Chapter two introduces the study in the context of 8 existing literature on land use and cover change and its drivers and impact. It explains how the objectives can be achieved using remote sensing as a tool along with the major procedures and software required. It also includes a review of several studies done in the areas along with important findings and gaps of knowledge that need further research, which further guides us to formulate the purpose of the study and central questions to be answered. Chapter three describes the biophysical characteristics of the study areas, sources and methods in collection of primary and secondary data, major approaches followed during analysis of both spatial and socio-economic data and how biophysical results have been related with socio-economic issues and human behavior. Chapter four presents the core findings of the study derived upon analysis of all available data and discusses the major findings of the study with reference to existing knowledge, and relates them to the objectives, central questions and overall framework of the study. Finally, chapter five presents concluding remarks and implications of the major findings in addressing prevailing challenges in the target areas and also presents issues for further research. 9 CHAPTER TWO LITERATURE REVIEW DEFINITIONS AND RATIONALE OF LAND USE AND LAND COVER CHANGE STUDY Land cover is defined by the attributes of the earth’s land surface captured in the distribution of vegetation, water, desert and ice and the immediate subsurface, including biota, soil, topography, surface and groundwater, and it also includes those structures created solely by human activities such as mine exposures and settlement (Lambin et al., 2003; Chrysoulakis et al., 2004; Baulies and Szejwach, 1998). On the other hand, land use is the intended employment of and management strategy placed on the land cover by human agents, or land managers to exploit the land cover and reflects human activities such as industrial zones, residential zones, agricultural fields, grazing, logging, and mining among many others (Zubair, 2006; Chrysoulakis et al., 2004). Land use change is defined to be any physical, biological or chemical change attributable to management, which may include conversion of grazing to cropping, change in fertilizer use, drainage improvements, installation and use of irrigation, plantations, building farm dams, pollution and land degradation, vegetation removal, changed fire regime, spread of weeds and exotic species, and conversion to non-agricultural uses (Quentin et al., 2006). Land use and land cover changes may be grouped into two broad categories as conversion and modification. Conversion refers to changes from one cover or use type to another, while modification involves maintenance of the broad 10 cover or use type in the face of changes in its attributes (Baulies and Szejwach, 1998). According to Lambin (2005) sustainable resource use refers to the use of environmental resources to produce goods and services in such a way that, over the long term, the natural resource base is not damaged so that future human needs can be met. One of the most significant global challenges in this century relates to management of the transformation of the earth’s surface occurring through changes in land use and land cover (Mustard et al., 2004, cited in Daniels et al., 2008). It is estimated that undisturbed (or wilderness) areas represent 46% of the earth’s land surface. Forests covered about 50% of the earth’s land area 8000 years ago, as opposed to 30% today. Agriculture has expanded into forests, savannas, and steppes in all parts of the world to meet the demand for food and fiber (Lambin et al., 2003). Based on data from diverse sources, the Global Forest Resources Assessment 2000 estimated that the world’s natural forests decreased by 16.1 million hectares per year on average during the 1990s, which is a loss of 4.2% of the natural forest that existed in 1990 (Lambin et al., 2003). Land use in East Africa has changed swiftly over the last half-century: expansion of mixed crop-livestock systems into former grazing land and other natural areas and intensification of agriculture are the two largest changes that have been detected (Olson and Maitima, 2006). Accordingly, land cover classification has recently been a hot research topic for a variety of applications (Liang et al., 2002). A great deal of research has 11 been conducted throughout the world in an attempt to understand major shifts in land use and land cover and to relate them to changing environmental conditions. According to Baulies and Szejwach (1998), during the next decades, land-use dynamics will play a major role in driving the changes of the global environment. Hence, global mapping of irrigated and dry land agriculture, semi-natural areas and forest cover, reflecting their dynamics, can contribute to the assessment of the biophysical implications of land use and land cover change within the Earth’s system. Generally, agriculture is found to be the major driver of land cover change in tropical regions (Lambin et al., 2001 cited in Daniels et al., 2008). Over the past 50 years in East Africa, there has been expansion of agriculture at the expense of grazing land (Olson and Maitima, 2006). Before 1950, semi-arid and sub-humid areas were predominantly pastoral with scattered settlement and cultivation but from then onwards, there has been significant transformation of grazing land to mixed crop-livestock agriculture. Understanding the mechanisms leading to land use and land cover changes in the past is crucial to understand the current changes and predict future ones. These changes occurred at different time periods, paces, and degrees of magnitude and with diverse biophysical implications (Baulies and Szejwach, 1998). Therefore, Land use and land cover change (LUCC) research needs to deal with the identification, qualitative description and parameterization of factors which drive changes in land use and land cover, as well as the integration of their consequences and feedbacks (Baulies and Szejwach, 1998). However, one of the major challenges in LUCC analysis is to link behavior of people to biophysical information in the appropriate spatial and temporal scales (Codjoe, 12 2007). But, it is argued that land use and land cover change trends can be easily assessed and linked to population data, if the unit of analysis is the national, regional, district or municipal level. Land use and land cover changes result from various natural and human factors within social, economic and political contexts. Hence, the local human activities expressing the drivers can be determined by measuring the rates and types of changes and analyzing other relevant sources of data like demographic profiles, household characteristics and policies related to land resources administration. To achieve this, it is crucially important to consider multiple sources of information and to acquire temporal, spatial and other non-spatial forms of data. This is due to the fact that land use attributes are complex and the boundaries between different types of data are quite diffuse (Baulies and Szejwach, 1998). LUCC studies have been designed to improve understanding of the human and biophysical forces that shape land use and land cover change. Thus, linking human behavior and social structures to biophysical attributes of the land is a fundamental aspect of LUCC research (Baulies and Szejwach, 1998). Land use and land cover plays an important role in global environmental change and sustainability, including response to climate change, effects on ecosystem structure and function, species and genetic diversity, water and energy balance, and agro-ecological potential (Codjoe, 2007). 13 Land use and land cover mapping is one of the most important and typical applications of remote sensing data (Chrysoulakis et al., 2004). Remotely sensed data are a useful tool and have scientific value for the study of human- environment interactions, especially land use and land cover changes (Dale et al., 1993 cited in Codjoe, 2007). Few studies have been conducted to understand land use and land cover change and other related issues in the proposed study areas. Yitaferu (2007) has done satellite image analysis of the Lake Tana basin between 1985/86 and 2001/03. He found that croplands increased by about 4.2%, which largely occurred at the expense of grasslands and shrub lands. Furthermore, forest cover in the basin was found to have increased by about 0.23% in the same time frame. Analysis of satellite images and aerial photos of 40 years provide evidence of changing land cover and flood levels for Hara Swamp (Mc Hugh et al., 2006). It was found that the vegetation of the wetland catchment was almost gone by 2000, although it was covered by dense woody vegetation in the time between 1964 and 1986. The wetland showed continuous increment and was completely flooded around 2000, whereas it was almost dry around 1964. Furthermore, the numbers of houses in Hara town, near the wetland have greatly increased in this period of analysis. APPLICATION OF REMOTE SENSING FOR LAND USE AND LAND COVER CHANGE There is significant variation between various sensor instruments’ capability and wealth of information captured and also the applicability depends on the objective of the intended study. There is also clear variation in the spatial and 14 spectral properties of satellite images acquired by different versions of a particular sensor instrument. Landsat instruments can be taken as a good example of showing continuous improvement in radiometric and spectral property of images enabling better understanding of land resources. Since 1972, the Landsat satellites have provided repetitive, synoptic, global coverage of high-resolution multispectral imagery. Their long history and reliability have made them a popular source for documenting changes in land cover and use over time (Turner et al., 2003) and their evolution is further marked by the launch of Landsat 7 by the US government in 1999. Multispectral Scanner (MSS) data from the U.S. Geological Survey's (USGS) EROS Data Center (EDC) has provided a historical record of the Earth's land surface from the early 1970s to the early 1990s. The MSS and TM sensors primarily detected reflected radiation from the Earth's surface in the visible and IR wavelengths, but the TM sensor provides more radiometric information than the MSS sensor (http://edc.usgs.gov/guides/landsat_mss.html#mss4). The wavelength range for the TM sensor is from the visible (blue), through the mid-IR, into the thermal-IR portion of the electromagnetic spectrum and it has a spatial resolution of 30 meters for the visible, near-IR, and mid-IR wavelengths and a spatial resolution of 120 meters for the thermal-IR band. Each pixel of Landsat TM images contains a wealth of information about the surface materials that reflect light from that pixel to the satellite sensors. Each band in a TM image represents a separate piece of data whose value ranges from 0 to 255 enabling the whole image to contain 2565 (approximately 1.1 billion) different possible spectral combinations. However, it does not mean 15 that each one of these combinations represents a different type of land cover and most of these variations represent very small and, to us, "unseeable" differences in surface reflectance. ETM+ instrument measures upwelling radiance in the same seven bands as the TM, and has an additional 15 m resolution panchromatic band (Mather, 2004). The spatial resolution of the thermal infrared channel is 60 m rather than the 120 m of the TM thermal band and this instrument has substantially the same operational characteristics as Landsat-4 and Landsat-5. All the Landsat image archives used for this study were acquired from ILIR’s information database. The characteristics of the MSS and TM bands (Table 1) were selected to maximize each band's capabilities for detecting and monitoring different types of land surface cover characteristics. For example, MSS band 1 can be used to detect green reflectance from healthy vegetation, while MSS band 2 is designed for detecting chlorophyll absorption in vegetation. MSS bands 3 and 4 are ideal for recording near-IR reflectance peaks in healthy green vegetation and for detecting water land interfaces. MSS Bands 4, 2, and 1 can be combined to make false-color composite images, where band 4 controls the amount of red, band 2 the amount of green, and band 1 the amount of blue in the composite. This band combination makes vegetation appear as shades of red with brighter reds indicating more vigorously growing vegetation. Soils with sparse or no vegetation will range from white (sand) to green or brown, depending on moisture and organic matter content. Water appears dark blue to black in color, while sediment-laden or shallow waters appear lighter in color. Urban areas appear blue-gray in color. 16 ASTER (Advanced Space borne Thermal Emission and Reflection Radiometer) images were purchased from the USGS Center for Earth Resources Observation and Science (EROS) through the Land Processes Distributed Active Archive Center (LP DAAC) and the data were provided in hierarchical data format (HDF) - Earth Observing System (EOS) (HDF-EOS). The data format for both study sites was ASTER L1B.003, which indicates the level of processing, granule type and algorithm followed in processing the data. Table 1: Spectral property of various versions of Landsat images (source: Engineering manual, 2003) Serial no Name of Satellite Sensor Band number Band wavelengths ( µm) size of Pixels (m) 2 Landsat 4-5 MSS TM 1 2 3 4 1 2 3 4 5 6 7 0.5 to 0.6 0.6 to 0.7 0.7 to 0.8 0.8 to 1.1 0.45 to 0.52 0.52 to 0.6 0.63 to 0.69 0.76 to 0.9 1.55 to 1.75 10.4 to 12.5 2.08 to 2.35 82 82 82 82 30 30 30 30 30 120 30 3 Landsat 7 ETM PAN 1 2 3 4 5 6 7 4 0.45 to 0.52 0.52 to 0.6 0.63 to 0.69 0.76 to 0.9 1.55 to 1.75 10.4 to 12.5 2.08 to 2.35 0.5 to 0.9 30 30 30 30 30 120 30 15 17 The ASTER image purchased for Lenche Dima areas has a cloud cover of about 4 percent but fortunately, cloud was 0% in the watershed boundary of the study area. ASTER consists of three separate instrument subsystems that comprises of 14 bands of different wavelengths and spatial resolution, each operating in a different spectral region (Figure 1, below) that are the visible - near infrared (VNIR), the shortwave infrared (SWIR) and the thermal infrared (TIR). It has three spectral bands in the VNIR, six bands in the SWIR, and five bands in the TIR regions, with 15, 30, and 90 meter ground resolution, respectively (Yamaguchi et al., 1998). The VNIR imagery is used for land cover/use classification and mapping (Chrysoulakis et al., 2004) by creating a pseudo-color composition of VNIR bands that have a spatial resolution of 14m (RGB: 3Nadir-2-1) with a spectral range of 0.52-0.6 µm, 0.63-0.69 µm and 0.78-0.86 µm respectively. High spatial resolution sensors like ASTER are capable of providing accurate land cover maps with precise land use classification results, and it also aids to discriminate a variety of surface materials and reduce problems in some lower resolution data resulting from mixed pixels (Chrysoulakis et al, 2004, Yamaguchi et al., 1998). 18 Figure 1: Comparison of Spectral Bands between ASTER and Landsat-7 Thematic Mapper (source: Abrams et al., 2009) Basics of remote sensing Spectral resolution is the size and number of wavelengths, intervals, or divisions of the spectrum that a system is able to detect. Fine spectral resolution generally means that it is possible to resolve a large number of similarly sized wavelengths, as well as to detect radiation from a variety of regions of the spectrum (Engineering manual, 2003). Digital sensors record the intensity of electromagnetic radiation (ER) from each spot viewed on the Earth’s surface as a digital number (DN) for each spectral band. The exact range of DN that a sensor utilizes depends on its radiometric resolution. For example, a sensor such as Landsat MSS measures radiation on a 0-63 DN scale whilst Landsat TM measures it on a 0-255 scale. Although the DN values recorded by a sensor are proportional to upwelling ER (radiance), the true units are W m-2 ster-1 mm-1 (Anonymous, 2000). 19 According to the Modified UNESCO Classifications scheme, there are only about 157 different land cover types and no study site will have all of those different land cover types (GLOBE toolkit, 2003). Hence, it will be necessary to group pixels together into a smaller number of closely related "classes", based on spectral similarity. This is done in a process known as "Classification”. Similar to gray scale, bright regions of color composite images have high reflectance, and dark areas have low reflectance. However, interpretation may get difficult when we combine different bands of data to produce what is known as false-color composites (Engineering manual, 2003). But this can be addressed by using field knowledge of the areas and historical information. The majority of image processing exercises are done based on raw DN values in which actual spectral radiances are not of interest (e.g. when classifying a single satellite image). However, there are problems with this approach as the spectral signature of a habitat is not transferable, if measured in digital numbers. These values are image specific as they are dependent on the viewing geometry of the satellite at the moment the image was taken, like the location of the sun, specific weather conditions, and so on. Accordingly, it is generally far more useful to convert the DN values to a spectral signature with meaningful units as they can be compared from one image to another. This process is called image calibration and is required where the area of study is larger than a single scene or if scenes taken over a period of years are being compared. The relative calibration of satellite imagery will correct for differences in atmospheric path radiance, detector calibration, sun angle, earth-Sun distance, atmospheric attenuation, and phase angle conditions. Radiometric control sets representing temporally invariant features are used to derive gains and offsets for a linear transformation (EPA, 1999). Calibration is 20 band specific and is undertaken on each band of all images with reference to a common radiometric reference, which involves the transformation of digital numbers to physical values of radiance or reflectance. There are numerous approaches to characterizing land cover change. Each one of it has a set of strengths and weaknesses and as a result no single approach is optimal for all types of landscapes and land cover features (EPA, 1999). There are two major approaches of classification of remotely sensed images for various applications. In a supervised classification, the software is “trained" to recognize that certain types of pixels represent specific land cover types. Knowledge of the area and information collected during field work are important inputs, which are used by the software to classify the pixels into similar groups based on sample signatures specified. In an unsupervised classification, or "clustering", the desired number of groups, or "clusters", will be inputs to the software (GLOBE toolkit, 2003). The software then groups the pixels according to similar spectral characteristics in order of decreasing brightness. These groupings are not made on the basis of land cover, but on the similarity in spectral characteristics of the pixels. It is also common to come across a third approach known to be hybrid classification that combines both supervised and unsupervised classification techniques. Approaches in image classification Remote sensing change detection techniques can be broadly classified as either pre- or post-classification change methods. A pre-classification process 21 refers to operations carried out to bring satellite images to the desirable geometric and spectral standard by correcting errors, and it is performed prior to image classification. Whereas, post-classification methods refers to activities done after classification of images like computation of class statistics, accuracy assessment, and map preparation. Pre-classification methods can further be characterized as being spectral or phenology based. Originally, the post-classification approach was considered to be the most reliable approach and was used to evaluate emerging methods (Weismiller et al., 1977 in EPA, 1999). Factors that limit the application of post-classification change detection techniques include cost, consistency, and error propagation (Singh, 1989 in EPA, 1999). Numerous pre-classification change detection approaches have been developed and refined to provide optimal performance over the greatest possible range of ecosystem conditions (Lunetta et al., 2006). The satellite instruments employed some decades ago provided images with coarse resolution. With advancement in remote sensing science, various sensor instruments with improved radiometric, temporal and spatial resolution were being developed. Hence, this allowed the integration of satellite images acquired by various sensor types in order to better understand land resources dynamics. The use of data from different sensors poses a serious challenge to many change analyses, which can be addressed through use of post- classification comparisons (EPA, 1999). APPROACHES IN LAND USE AND LAND COVER CHANGE DETECTION Research evaluating the comparative performance of various land cover change detection methods has indicated that no uniform combination of data 22 types and methods can be applied with equal success across different ecosystems (Lu et al., 2004 cited in Lunetta et al., 2006).Despite this, the two general approaches to change detection are comparative analysis of independently produced classifications, and simultaneous analysis of multi- temporal data. Examples of the simultaneous analysis techniques include image differencing, ratioing, principal component analysis (PCA), and change vector analysis (Singh, 1989 in EPA, 1999). The first approach is straightforward and employs independently classified images being converted to same projections and it has the advantage that it allows compensating for variations in atmospheric and phenological conditions. The method has been criticized as it tends to compound errors that may have occurred in the two initial classifications (Gordon, 1980; Stow et al., 1980; Singh, 1989 in EPA, 1999). On the other hand, simple image differencing is a widely used technique that involves taking the mathematical difference between geo-registered images from two dates (EPA, 1999). Even if the method has often been reported to produce excellent results, it has been suggested that image differencing alone may be too simple a procedure to adequately describe many surface changes (Weismiller et al., 1977; Jensen and Toll, 1982; Sohl, 1999 in EPA, 1999). A major attribute of the landscape is its spatial pattern and structure. It is shown that the detection of land-cover change processes by remote sensing is improved when both spectral and spatial indicators of surface condition like slope and topography are used (Lambin and Strahler 1994 in EPA, 1999). It is further suggested by this author that spectral indicators are more sensitive to 23 fluctuations in primary productivity associated with the inter-annual variability in climatic conditions. Temporal aspects of natural phenomena are important for image interpretation because such factors as vegetation growth and soil moisture vary during the year ,and hence, more positive results can be achieved by obtaining images at several times during the annual growing cycle (Lillesand and Kieffer, 2004). Furthermore, changes in landscape spatial pattern are more likely to reveal long term and long lasting land cover changes. Following image classification as part of the change detection process, accuracy needs to be assessed to evaluate the degree of acceptability of the classification process. A standard accuracy assessment procedure for baseline land cover products involves the use of the error matrix (EPA, 1999) and the standard procedure for one-point-in-time land cover products can be extremely difficult to apply to multi-temporal change analysis products (EPA, 1999). The methods are well established for small areas and single time periods. However, the assessment of accuracies for large areas, past time periods, and change databases can become problematic (Dobson and Bright, 1994 cited in EPA, 1999) as it will be difficult to acquire an adequate database of historical reference materials. Accordingly, accuracy assessments are usually limited to the very recent image that serves as a reference using ground control points (GCPs) collected as part of the data required for the change analysis. 24 CHAPTER THREE DESCRIPTION OF THE STUDY AREAS AND IMAGE ANALYSIS The study focused on two areas in Amhara region characterized by different biophysical composition (Figure 2). Kuhar Michael is one of the 25 rural peasant associations1 (PA) found in Fogera, while Lenche Dima is a catchment forming part of Laste Gerado PA of Gubalafto district. Based on classification of highlands of Amhara region (Desta et al., 2000), Fogera falls in the category of high agriculture potential and high population density, while Gubalafto is classified as an area with low agricultural potential and lower population density. On the other hand, both areas were found to share a common feature, as they both have high access to market (Desta et al., 2000). Furthermore, both areas have similarity in that the prevailing farming systems belong to the mixed crop-livestock farming systems. Kuhar Michael is characterized to be moisture sufficient, even with flooding events in some parts of the Fogera plain; while Lenche Dima is a drought prone area with very little potential for irrigation. There is an on-going joint project by the International Livestock Research Institute (ILRI) and the International Water Management Institute (IWMI), which focuses on improving the livelihoods of smallholder farmers through improving water productivity of mixed crop- livestock systems in sub-Saharan Africa. This particular study focuses on the target areas of this project and is geared towards generating information on the dynamics of land use and land cover change, its impacts and drivers so as 1 refers to local level administrative organ established during the Derg regime with the mandate to handle land related matters in a specified geographic area 25 to support the on-going research project dealing with crop-livestock water productivity improvement. Figure 2: Location Map of Kuhar Michael Peasant Association (D) From Fogera District (B, Left), and Lenche Dima Watershed (E) From Gubalafto District (C, Right), Amhara Region, Ethiopia. A B C D E 26 GEOGRAPHIC LOCATION Kuhar Michael Peasant Association (PA) is located at the southern border of Fogera District (see Figure 2B above) and north of Bahar Dar, with coordinate values that extend from 11°50'37" to 11°53'37" N and from 37°38'10" to 37°42'17" E along the Bahar Dar – Woreta – Gondar highway (Descheemaeker, 2008). It falls within the upper Nile basin with an estimated area of about 2755 ha and annual rainfall of just over 1200mm, and its altitude ranges from 1792 to 1959 m a.s.l. In Kuhar Michael, mean monthly temperature fluctuates around 20 °C the whole year around, with slightly lower temperatures (18 – 19 °C) in winter (December and January) and in July and August and slightly higher temperatures (21 °C) in April and May (Worldclim, Hijmans et al., 2005 in Descheemaeker, 2008). Regarding the soil type of the area, detailed descriptions are lacking, but generally, Eutric Vertisols are found in the plains and Haplic Luvisols in the upland hills (Haileslassie et al., in Descheemaeker, 2008). The area is capable of surplus food production with rice, fish, horticultural crops and livestock being the major agricultural commodities. Lenche Dima watershed is comprised of Kolo Kobo, Lenche Dima, Oromo and Hartibo villages. It forms part of Laste Gerado peasant association in Gubalafto wereda (see Figure 2C above), North Wollo zone and is located at some 20 km distance from the zonal capital, Woldia town. Geographically, the watershed extends from 11°49’13” to 11°51’57" North latitude and from 39°40’07" to 39°44’22" East longitude and its altitude ranges between 1520 to 1890 m a.s.l. The water outflow of the watershed contributes to Alewuha River, which flows to Awash River (Gizaw et al., 1999). The watershed is 27 classified as tepid to cool sub-moist mountains and plateaus (Gizaw et al., 1999), and has a dry tropical climate with a temperature range of 20 to 29ºC (McHugh et al., 2007) and a bimodal rainfall distribution. The short rainy season includes March, April and May, while the second and long rainy season includes July, August and September with an average annual rainfall of 667 mm. The mean daily maximum temperature is 33 °C (June) and the mean daily minimum is 12°C (November) (Gizaw et al., 1999). The major soil types of the watershed are Regosol, Leptosol, Vertisols and Fluvisol (Gizaw et al., 1999). FIELD SURVEY AND DATA COLLECTION The ultimate purpose of the field survey conducted was to collect qualitative and quantitative information to help to better understand, explain, and interpret the land use and land cover change, which is the core issue of this study. Hence, understanding trends in resource dynamics required historical information, which can be achieved using qualitiative and quantitative data collected through interview and group discussion with selected informants believed to have a good understanding of the issues of interest. Accordingly, detailed individual interviews and group discussions were conducted with 50 selected key informants from both study areas to collect the data required. A purposive sampling technique, involving the targeting of individuals who suited the subject and nature of study using predetermined selection criterion, was used to select the participants through consulation with Develeopment Agents (DA) of the respective study areas. A questionnaire (see Appendex I) covering a wide range of topics relevant to the central issue of interest was developed. 28 It was also pre-tested so as to evaluate the understandability of the questions and modifications were made accordingly. Ground Control Points (GCPs) were collected to aid different steps of image processing and classification for change detection. Besides this, field observation was made to have better information about the nature of the various land use and land cover classes prevailing in the area. DESCRIPTION OF LAND USE AND LAND COVER CLASSES Based on field observation and general historical information gained from participants during the survey, it was decided to focus on the following major land use and land cover classes summarized below and in Table 2. CULTIVATED LAND Cultivated land includes most flat areas and also some steep slopes where various food crops are grown, either on a rain-fed basis or using irrigation. Irrigation is commonly practiced in some parts of Kuhar and very few localities of Alewuha, which is located out of Lenche Dima catchment. WETLANDS The category of wetlands encompasses those areas located down in the Fogera plain which experience frequent flooding and immersion in water and are covered by wetland vegetation including grasses. The ultimate use of such land was as a grazing and holding place for livestock. Currently, such areas have been converted to cultivated land for growing rice and other cash crops. 29 GRASSLAND Grazing land refers to those land units allocated as a source of animal feed, including privately and communally owned grazing areas and also those owned by various institutions (church and school). There is variation in their management schemes, from those which are open to access year round, those accessed only during selected months of the year, and those completely closed year round. In the latter feed is made available through the cut and carry system, whereby grazers are expected to cut and haul away the fodder to the animals. SHRUB/BUSH LANDS The category includes areas covered with different species of shrubs and bushes with widely varying density from one locality to another, and often found in hilly areas. In Lenche Dima, those areas that are not closed as part of the watershed management program serve as a communal grazing and free access resource with very few above ground biomasses. In Kuhar Michael, area closure is not practiced, even though the district announces it to be one of the activities they are currently working at. Hence, vegetated areas upslope are not well protected and are being accessed freely to serve as grazing land. In this study, vegetated areas are classified in two general classes, as either open or dense, and comparison of their relative greenness values using the Normalized Difference Vegetation Index (NDVI) value was employed to make a distinction between them. 30 BARE LAND The category of bare land denotes areas that are without any vegetation cover at the time of satellite image acquisition but could be categorized in one of the land use and land cover types mentioned above. They are not covered by any type of crops, grass or any other shrub or tree species. According to a classification scheme developed for woody biomass Inventory and strategic planning project (WBISPP), bare land meant to include those areas with exposed sand/soil, salt flats, and exposed rock as well (WBISPP, 2002). In addition to the land use and land cover classes described above, it is also apparent that there are areas that serve as settlements, most of which are made of up of traditional huts and very few with corrugated metal roofs, scattered around. There was a complete absence of urban built up areas in both study areas. Hence, because it was not possible to easily identify such features in the satellite images and due to the relative small size of the study areas, the decision was made not to consider settlements among the land use classes. Accordingly, in such areas, growing various tree and shrub species as live fences, boundary plants, or woodlots has been observed to be a common practice. It is also apparent that various agricultural crops like cereals, fruit trees, and sometimes vegetables are grown in the homesteads, depending on the availability of water supply. 31 Table 2: Land use and land cover classes considered and their description (adapted and modified from Yitaferu, 2007) Class name Description 1 Cultivated land Plain and slightly undulating landscapes that are intensively cultivated (>75%). Foot-slopes and undulating landscapes are usually under moderately cultivated condition (50-75%).There are other land use and land cover types interspersed with the cultivated lands, but their individual size is negligible as compared to those cultivated. This class also includes those areas left barren temporally. 2 Shrub/bush land Refers to those areas covered with tree, shrub, bushes and some grasses that dominate the foot- slopes and riverine landscapes. There exists variation in vegetation between dense shrub/bush lands with an estimated cover of >50% and open shrub/bush lands with less than 50% cover. The latter are not bare at all, but being degraded from competing use of grazing, cultivation, and deforestation as some of the degraded shrub/ bush lands serve for grazing purposes. 3 Grass land This land cover includes short term flooded flat lands that are usually used for intensive grazing. Many of these lands in Kuhar Michael are periodically flooded. 4 Bare land Refers to those land surface features devoid of any type of vegetation cover. 5 Wetland Represents most plains areas with frequent flooding event during the rainy season. STEPS IN ANALYSIS OF SATELLITE IMAGES FOR LAND COVER CHANGE In order to cover the intended period of study, different type of images originating from different types of sensors were used. Hence, land sat multispectral scanner (MSS), Thematic Mapper(TM) and Enhanced Thematic Mapper Plus (ETM+) from beginning of 1970s, mid 1980s and 2000s respectively along with ASTER image of 2005 were used for a time series of images for the land use/cover change study (Table 3). 32 Image Calibration and its Importance The ultimate purpose of calibration is to convert Landsat (MSS, TM or ETM+) and ASTER digital numbers to radiance or exo-atmospheric reflectance (reflectance above the atmosphere) using published post-launch gains and offsets. Table 3: Details about the satellite images used for this study Image District/ PA Date of acquisition No of bands Band combination Spatial resolution (m) Sensor type ASTE RL1B Kuhar Michael Dec1 ,2005 14 VNIR (1, 2, 3N) 15 ASTER P169r 52 Kuhar Michael Oct 23,1999 7 3,2,1 30 ETM+ P16r5 Kuhar Michael Nov 9,1985 7 3,2,1 30 TM P182r 52 Kuhar Michael Feb 1,1973 4 3,2,1 80 MSS ASTE RL1B Lenche Dima Oct16,2005 14 VNIR ( 1, 2, 3N) 15 ASTER P168r 52 Lenche Dima Dec 5,2000 7 3,2,1 30 ETM+ P168r 52 Lenche Dima Jan 5 ,1986 7 3,2,1 30 TM P180r 52 Lenche Dima Nov 1, 1972 4 3,2,1 80 MSS Calibration is band specific and carried out for each separate band of all land sat sensors and ASTER images independently. The following formulas are the basis of the calibration:  BiasDNGainL  ..............…………….....…………………………… (1)  = ( L* *d 2 )/ (ESUNi*Sin ( ))……..………………….……………. (2) Where λ is the ETM+/TM band number, L=at-satellite radiance, Gain=band specific, provided in the header file sceneid.h1, Bias=band specific, provided in the header file sceneid.h1, ρ= at-satellite reflectance, unit less, d=Earth-sun 33 distance in astronomical unit, ESUN =solar exo-atmospheric spectral irradiance (see Annex IIc), θ= sun elevation angle, provided in the header file sceneid.h1. Calibration of MSS was completed by specifying the satellite type to be landsat-1, the time of acquisition (Date, Month and Year) and the sun elevation angle obtained from the header file of the respective images. Then the calibration process was undertaken through an automatic built-in operator within the ENVI interface. For TM, landsat-5 was selected to be the satellite type and the specific time of acquisition and sun elevation angle were also used. To complete the calibration process of TM images, the presence of minimum and maximum scales of radiance for each of the 6 bands was pertinent. However, review of literature revealed that there are two different sets of band specific values depending on the acquisition time of the image. Therefore, by observing the acquisition times of the two images from their header files, the corresponding values were selected and provided in the dialogue box to accomplish the calibration process. Based on these additional input values, the software automatically assigned gain and bias values to complete the calibration process. Regarding the ETM+ image, the path and row numbers, date of acquisition were used as inputs to generate calibration parameters including sun elevations, minimum and maximum radiance values. This was retrieved from an online database www.ittvis.com/envi_landsat/landsat_gain.asp . All of this 34 information, along with the satellite type, was fed into the calibration interface and the process was completed successfully for each of the six bands except the thermal band, as it is not considered in this study. For ASTER, a different approach was followed, as an automated built-in calibration procedure has not yet been developed for this type of image and does not exist in ENVI version 4.3. Hence, all the required inputs to complete the calibration process were collected from different sources (Smith, 2008 and Mather, 2004) and data from the header file provided along with ASTER image (see Annex II band c). Finally, correcting the values from the working images was accomplished by making use of the Band Math operator in ENVI. Below is the formula used to convert DN to ASTER spectral radiance:   tcoefficienconversionUnitxDNL 1 …….……………………..………. (3) First the DN value of each band was converted to spectral radiance ((the amount of electromagnetic radiation leaving a point on the surface expressed in W/m2)) value using equation 3, above. Then, the output of this step was used in order to calculate the top of atmosphere reflectance (RTOA) of each band in ASTER images, except the thermal infrared reflectance (TIR) bands.     zCOSxESUNLdR iTOA 2 …………………………………………………. (4) Where, Pi=3.14159, RTOA is planetary reflectance, Lrad is the spectral radiance at the sensor’s aperture; ESUNi is the mean solar exoatmospheric irradiance of each band, i;z is the solar zenith angle(Zenith angle=90-solar elevation angle), which is within the ASTER header file; and d is the earth-sun 35 distance,in astronomical units, which is calculated using the following equation (Achard and D’souza 1994; Eva and Lambin, 1998 in Smith 2008): ))))4(9856.0((01672.01(  JulianDayRADIANCECOSd ………………. (5) Major inputs required to compute the earth-sun distance (d) were radiance, and Julian date of the image acquisition. Different approaches of computing the band specific ESUNi value are available. From the alternatives available, the one computed by Smith (2008), interpolating the ASTER spectral response functions to 1nm and convolving2 it with the 1nm step with world resource center (WRC) data (Smith, 2008), was used. The input values for the various versions of Landsat and ASTER images were retrieved from various secondary sources of data. Detail numerical value of the different parameters considered as inputs during the calibration process are provided at the end of this report as a series of Annexes (IIa, IIb and IIc). After calibration is performed in each band of the different images, color composite images were made whose reflectance value ranges from 0 to 1. Layer Stacking During layer stacking, the Universal Traverse Mercator (UTM) system with WGS84 as a datum was assigned as a preference as far as projection is 2 A pixel based spatial filter applied on an image based on weighted average of coefficients 36 concerned. All four bands of MSS and five bands of TM and ETM+, excluding the thermal band, were considered for Layers stacking. The nature of these different bands had to be considered to make a decision as to which three band combination would be most helpful for classification and visual interpretation. The Band 4 reflective infrared wavelength (0.76-0.90 μm) is absorbed by water (appearing dark) and reflected by vegetation (appearing bright), while mid-infrared bands 5 (1.55-1.75 µm) and 7 (2.08-2.35 µm) contrast well, revealing differences in types and conditions of vegetation and soil (McHugh, 2006). Furthermore, the green (0.5-0.6 µm) and red (0.6-0.7 µm) wavelengths of Bands 1 and 2 in MSS imagery are absorbed by vegetation, showing differences in vegetation health. By taking all these facts into consideration, a false color composite for ETM+ (Kuhar Michael 1999 and Lenche Dima 2000) were created using band combinations of band 4 in the green, band 5 in the red and band 7 in the blue; while band 4 in the green, band 5 in the red and band 7 in the blue wavelengths were used to create a false color composite from TM images (Lenche Dima 1986 and Kuhar Michael 1985). False color composites of MSS images (Lenche Dima1973 and Kuhar Michael 1972) were created using a combination of band 4 in the Green, band 2 in the Red, and band 1 in the Blue. After layer stacking, all the scenes were re-projected to UTM Zone 37 North using WGS 84 as a datum. The default pixel size of the stacked image was 73.4 m for MSS, 36.7 meters for TM and 34 meters for ETM+. However, to standardize pixel size and to minimize possible error that could arise due to variation from pixel resolution, output cell size for all landsat composite images was set to be 14m, which was the default size of ASTER image. 37 Geo-Referencing of Images When the ground control points (GCPs) were overlaid on the color composite of the ASTER images, it was observed that the points did not fit exactly with the corresponding ground features represented on the image. This triggered the need to undertake geo-referencing of ASTER, which was accomplished through the image-to-image geo-referencing operation of ENVI that requires use of GCPs. Image-to-image geo-referencing allows interactive selection of GCPs. Orthorectified Landsat ETM+ images whose geographical accuracy is checked with the GCPs served as a base image and ASTER as a warp image. Warp image refers to the one to be corrected using a geometrically corrected image, which will serve as a base image. At least four points are required for defining a warp polynomial so as to predict the corresponding locations of the selected GCPs in the warp image. However, twenty points were selected to improve the accuracy of registration. These included easily identifiable features that exist in both the base and warp images. Both the base and warp images were displayed side to side and effort has been made to minimize the overall registration error by relocating the position of each of the GCPs with in the particular ground feature selected as ENVI provides the flexibility to do so. The first order polynomial was selected as a warping method as it enables to achieve more accurate results. Furthermore, a cubic Convolution was selected for image resampling as it was the method being used by the data providers. 38 Hence, Georeferencing of ASTER image for Kuhar Michael was accomplished with a total mean square error (RMSE) value of 2.05 (See Appendix IIIA). The same procedure was applied to correct the ASTER image representing Lenche Dima. However, prevalence of cloud cover in ASTER image of Lenche Dima made the identification of GCPs a bit difficult. Despite this, effort was made to improve registration accuracy by relocating position of GCPs and the geo-referencing process was completed with an over all RMSE value of 1.86 (see Appendix IIIB). Finally, positional accuracy of both corrected ASTER (warped) images was checked visually by linking it to the corresponding Landsat ETM+ scenes that served as base image and using Google earth as well. Imagery of the study areas were clipped out using the extract by mask operator of the ArcGIS spatial analyst tool box (Figure 3). To achieve this, pre- defined shape files of the two study areas were used. Triangulation with nearest neighborhood method that employs the value of the closest pixel to assign a value to the output pixel was preferred as a method for re-sampling during the calibration and layer stacking operations as well. The nearest neighbor method takes the value of the pixel in the raw (untransformed) image that is closest to the computed (c, r) coordinates (Mather, 2004). It has two advantages; it is fast and its use ensures that the pixel values in the output image are ‘real’ in that they are copied directly from the raw image, while bilinear and cubic convolution resampling methods use interpolation algorithms to ‘fabricate’ out put coordinates. 39 Histogram equalization is a non-linear stretch that redistributes pixel values to make sure that there are approximately the same numbers of pixels within each value within a range. Hence, it was carried out so as to increase the visual interpretability of the images. Figure 3: General shape of Kuhar Michael (left) PA and Lenche Dima watershed (right) used for sub-setting the time series images IMAGE CLASSIFICATION ASTER image was the very recent image available for study areas and served as a reference image. Hence, it was possible to undertake field visit and collect GCPs and supervised classification was preferred. The three landsat images; MSS, TM and ETM+ images were also included to meet the preferred time horizon of study. But, due to lack of fine details, unsupervised classification was selected to classify landsat images. Meanwhile, it must be noted that effort has been made to integrate few historical information acquired from surveys to minimize complete reliance on spectral information 40 and solve mystery of spectral similarity of different land cover classes in order to improve classification accuracy. General Considerations and Decision Rules during unsupervised classification Image interpretation process is like the work of a detective trying to put all the pieces of evidence together to solve a mystery (Lillesand and Kieffer, 2004). Success in solving the mystery is a function of the level of detail contained in images, experience and knowledge about situations and history of the study area, and expertise of the interpreter as well. Despite general variations in the specific field application of remote sensing studies, most image analysts are recommended to pay attention to characters of ground features like shape, size, pattern, tone (or hue); texture, shadows, site, association, and resolution of an image (Olson, 1960 cited in Lillesand and Kieffer, 2004), and some of them are applied in this particular study. Furthermore, an interpreter is advised to consider wealth of information available from maps and other sources at different levels of systematically examining image. For instance, from imagery alone, it is not possible to make a distinction between cropland and grassland generally with a high degree of accuracy and uniformity (Hardy et al., 1971 in Anderson et al., 1976). Hence, field observation, measurements, and survey results can provide important information to aid the process. Accordingly, some of the communal grasslands and shrub/bush lands observed during fieldwork and interview has been serving for same purpose since 1970’s, except for the reported shrink and deterioration in quality. Hence, such historical information was also used 41 during unsupervised classification of images as location of these features is marked with GPS. Integration of such historical information in the decision process makes the classification approach hybrid rather than purely unsupervised. Furthermore, an interpreter has to take into account the probability of certain ground cover types occurring on certain dates at certain places and knowledge of the crop development stages or calendars (Lillesand and Kieffer, 2004). In the true color composite (R-3, G-2 and B-1); bare land appears white color due to its high reflectance value (Tucker et al., 2004). But this might not be enough information to make decision about information classes as cloud/snows may exhibit same character, which might lead to errors. Hence, the interpreter needs to be aware of this and other similar tricks and needs to employ additional techniques and information to make a distinction through an informed decision. Accordingly, knowledge about pattern (spatial arrangement of objects) and associations (occurrence of certain features in relation to others) will be important evidences. For instance, wetlands frequently are associated with topographic lows, even in mountainous regions (Anderson et al., 1976) and one might expect flooding experiences to occur in relatively plain areas but not in steep areas. Furthermore, tone, known appearance of certain ground features in a different composite image, and level of details expected from an image with particular resolution, along with knowledge on historical trends of an area provide important clues to aid classification process. Figure 4 at the end of this section summarizes the processes used during the image analysis and classification of this study. 42 The first image of Gubalafto was acquired in November 1972. According to evidence from respondents; bare land was expected during the reign of Hailesilassie due to relatively low human population and less shortage of land. However, it was not possible to identify it from the images, which might be due to the coarse resolution of the image that did not allow depicting such details and due to possible coverage of such areas with grasses and herbs in November. The second image was taken in January 1986, when crops were harvested and significant areas in the landscape lack cover. This made it difficult to identify bare land precisely as most cultivated lands appear bare. It was possible to identify bare lands in Kuhar Michael during the first two year of analysis (February 1973 and November 1985). On the other hand, despite use of a relatively finer resolution image for the reference year 2005, a general decrease of fallowing practice in recent years is reported due to the growing need and shortage of cultivated. Hence, due to this and specific acquisition dates of the images, bare land was not recognized in this particular year as well. Hence, due to their particular nature of association, most bare lands are assumed to be included with cultivated lands. Furthermore, as suggested by Anderson et al. (1976), land may appear barren because of man’s activities and when it may reasonably be inferred from the data source that the land will be returned to its former use, it is not included in the Barren category but classified on the basis of its site and situation. Generally, the relative small size of the study areas, very course resolution of the images and the lack of series of images to discriminate the inter-seasonal differences within the same year and crop calendar made it difficult to identify 43 bare lands in general , and also formed part of the rigor faced during image interpretation process. Despite lack of detail and specific historical information about previous land use and land cover in the area, inhabitants’ reflection about the overall trends served as a guide during image classification, change detection process and result interpretation. Because of this, unsupervised classification was preferred for classification of all landsat images, except the ASTER image. During unsupervised classification, patterns inherent in the spectral data drive the process. Accordingly, a set of spectral classes will be formed that requires interpreter’s input to visually correlate map patterns to meaningful ground categories or land cover classes. This requires knowledge about major classes expected in the scene, ground truthing, and skill to visually correlate map patterns to their ground counterparts. After determination of the major land use and land cover classes to be included, a natural cluster comprised of twenty classes was created using the unsupervised classification operator of ERDAS imagine. This software uses the ISODATA (Iterative Self-Organizing Data Analysis Technique) algorithm to perform an unsupervised classification. It performs an entire classification (with a thematic raster layer as an output) repeatedly and recalculates statistics by locating the clusters that are inherent in the data using the minimum spectral distance formula to form clusters. For cluster formation, either arbitrary cluster means or means of an existing signature set are available and the latter was used for this study. 44 The maximum iteration was set to be ten. In each iteration, means of clusters were shifted and the new cluster means were used for the next iteration. Output natural clusters files were set to be displayed as an approximate true color composite. Finally, historical information, knowledge3 of the area acquired from field work and knowledge about color representation of important features in different band combinations (see Table 4) were employed to associate those arbitrarily formed clusters/classes with their corresponding meanings among the ground features. Ground features covered with vegetation (trees and shrubs) appeared to be red in the 4, 3, 2-RGB false color composite (FCC) of Landsat images (Tucker et al., 2004). Furthermore, cultivated lands are displayed as pink to red and wetland vegetation appears dark red; whereas in the true color composite (3, 2, 1-RGB), wetlands look dark green to black, and bare soil appears to be white to light due to its high reflectance value. On the other hand, another band combination, including short wave infrared (7, 4, 2-RGB) is also utilized to aid in the categorizing process of arbitrarily formed natural clusters to their ground meaning (Tucker et al., 2004). Particularly, this was of importance for the TM and ETM+ Landsat images. In 3 Included information about state of certain ground features like communal grazing and shrub lands that didn’t show much change during the study period, based on information from elders. 45 this color composite, bare soil is displayed as magenta or pale pink while water bodies appear black or dark blue. Table 4: Appearance of various surface features in different band combinations (Source: Tucker et al., 2004) Furthermore, trees, shrubs, and crops appear as shades of green, creating confusion when attempting to discriminate. However, during formation of natural clusters comprising of 20 classes, better distinction between these land covers was found with different intensities of greenness. Furthermore, as well described in previous section, knowledge on pattern and association of certain ground features was also employed and provided important clues in differentiating certain ground features. TRUE COLOR RED: BAND 3 GREEN: BAND 2 BLUE: BAND 1 FALSE COLOR RED: BAND 4 GREEN: BAND 3 BLUE: BAND 2 SWIR RED: BAND 7 GREEN: BAND 4 BLUE: BAND 2 Trees and Bushes Olive green Red Shades of green Crops Medium to light green Pink to red Shades of green Wetland vegetation Dark green to black Dark red Shades of green Water Shades of blue and green Shades of blue to gray Black to dark blue Urban areas White to light blue Blue to gray Lavender Bare soil White to light gray Blue to gray Magenta, lavender, or pale pink Clouds white White White-pink- lavender Snow/Ice white Light green-blue Medium blue 46 Supervised Classification of ASTER Image Once geo-referencing is accomplished and the positional accuracies of features on images from different time are checked, supervised image classification of the ASTER image was done. Hence, selected ground control points that include the major land use and land cover classes were sampled to create a signature file to help train the software to classify the entire study areas. Care was taken to minimize error by avoiding mixed pixels, and an effort was made to include areas relatively uniform in spectral pattern. Moreover, attention was also paid during collection of GCPs in the field by taking GPS measures at the center of relatively large and uniform sampling unit representing each land cover class type. As opposed to unsupervised classification, decision as to the number of classes expected in that particular area is the first important step to be made. Following that, as part of the interpretation process, information was provided by taking sample signature points so as to supervise and help the computer accomplish the classification process in the whole study area based on samples provided. Particularly, parametric decision rule with maximum likelihood classifier was selected as it allows classification of all the pixels in scenes, and avoids the probability of assigning a single pixel in more than one class. Moreover the maximum likelihood decision rule is based on the probability that a pixel belongs to a particular class and its basic equation assumes that these probabilities are equal for all classes, and that the input bands have normal distributions. 47 Based on comparison of the different algorithms of supervised classification to monitor landscape changes in Abuja, findings revealed that the maximum likelihood algorithm performed better than the others (Ojigi, 2006 cited in Omo- Irabora and Obuyemib, 2007). Post Classification Operations Finally, class statistics of each classified image were computed in ENVI. Following this, a majority/minority filter with an operating window size of 3 by 3 was run in ENVI. A majority filter is a logical filter applied on a classified image that consists of labels rather than quantised counts and its simplest form involves the use of a filter window, usually measuring 3 rows by 3 columns, centered on the pixel of interest (Mather, 2004). Then, the number of pixels allocated to each of the k classes is counted and, if the centre pixel is not a member of the majority class (containing five or more pixels within the window) it is given the label of the majority class. The effect of this algorithm is to smooth the classified image by weeding-out isolated pixels, which were initially given labels that were dissimilar to the labels assigned to the surrounding pixels (Mather 2004). Accuracy of classification and Kappa coefficient error matrix were also determined based on classification result of ASTER images. Finally; the classified images were exported to ArcGIS for map preparation. 48 TM ETM+ Majority/minority Filter MSS Image Classification Map Preparation ASTER Image calibration Georeferencing Pre processing Post classification Land use and Land Cover Map Figure 4: Summary of major steps followed during satellite image analysis for land use and land cover change detection 49 CHAPTER FOUR RESULTS AND DISCUSSION RESULTS OF REMOTE SENSING AND GIS ANALYSIS OF TIME-SERIES SATELLITE IMAGES Land Use and Land Cover Dynamics in Kuhar Michael Based on information from local inhabitants, cultivated, fallow, communal and private grazing, and shrub/bush land were the major land use and land cover classes during the study periods. In Ethiopia, before 1974, the relationship between land users and owners was based on a feudal system (Desta et al., 2000) under which the ownership of land was limited to a few individuals. Most inhabitants were only eligible to get access to farmland through share cropping. During this period, it was highly likely that a portion of the land was left abandoned. The human population was relatively low with low pressure on the land and associated resources in general and with prevalence of fallow lands. An image of February 1973 was acquired, when crop harvesting had already started, and farmlands appear bare. Regarding vegetation, there were also relatively undisturbed areas that had been serving as a home for some wild animals with varying levels of density, ground cover and disturbance, according to respondents. Some of these areas were accessed and served as a source of wood and other products used for house construction, fuel wood, farm implements and fencing. Hence, for this particular study, such areas were broadly categorized as open and closed shrub/bush lands, depending on their level of ground cover. Analysis of the 1973 Image revealed that cropland 50 constituted the largest proportion of land in Kuhar Michael with a value of 31%, followed by grassland which accounts for 22 % (Figure 5). Dense shrub/bush land and wetland constituted 15% and 12% respectively, while both bare and open shrub lands showed similar coverage values of 10% each. Figure 5: Land use and land cover map of Kuhar Michael in 1973 As can be seen in Figure 6 for the 1985 image of Kuhar Michael, the proportion of land allocated for cultivation increased to 54%. Furthermore, dense shrub land expanded and covered 20% of the landscape. However, the proportions of grassland, open shrub/bush and bare land have decreased to 14%, 7% and 4 % respectively. The wetland land cover that appeared in the 51 previous classification is not included as it could not be identified due to possible dry up as the image was acquired in November. Figure 6: Land use and land cover map of Kuhar Michael in 1985 In accordance with the trends from the past, the cultivated land expanded again from 1985 to 1999. Hence, it can be seen in Figure 7 that cropland was again the major land use class covering 62% of the landscape. Similarly, open shrub/bush land increased to 13 %. As a result of the increase in agricultural (8%) and open shrub lands (6%), grassland and dense shrub/bush lands decreased to 9% and 6% respectively between1985 and 1999. 52 Figure 7: Land use and land cover map of Kuhar Michael in 1999 Due to the ever increasing shortage of and need for more cultivated land, as well as the high likelihood of not getting barren areas in October when the ground is often covered with grasses, herbs and crops; bare land is not observed as an important land cover class and was excluded accordingly. 53 Figure 8: Land use and land cover map of Kuhar Michael in 2005 The 2005 image analysis of Kuhar Michael showed that cultivated land accounts for 57%, while the closed shrub/bush, open shrub/bush and grasslands covered 13%, 19% and 11% respectively (Figure 8). LAND USE AND LAND COVER DYNAMICS IN LENCHE DIMA For the 1972 image (Figure 9), cultivated land accounts for 43 % of the landscape followed by open shrub/bush land with a value of 25 %. Dense shrub/bush land and grasslands had an almost equal coverage of 16% each. 54 Figure 9: Land use and land cover map of Lenche Dima in 1972 In 1986 (Figure 10), cropland increased in coverage to 52 %, while both dense and open shrub lands declined to 14% and 17 % respectively. Grassland increased slightly17%. 55 Figure 10: Land use and land cover map of Lenche Dima in 1986 Contrary to the trend observed in Kuhar Michael, cultivated land in Lenche Dima declined from 52% in 1986 to 48% in 2000. This could be related to the expansion of big gullies that are currently prevalent in the area. Cultivated and grass lands decreased slightly to 48% and 16 % respectively, while dense shrub/bush stayed unchanged. Finally, open shrub/bush land increased in coverage from 17% in 1986 to 22% in 2000 due to increased pressure on natural vegetation from the inhabitants (Figure 11). 56 Figure 11: Land use and land cover map of Lenche Dima in 2000 In 2005, cultivated land accounted for about 36%, followed by dense shrub land (24%) and grassland (21%). On the other hand, the percentage of open shrub land has shown a slight decrease to 19% in 2005 (Figure 12).This improvement in vegetation cover could be attributed to the implementation of a watershed management program through coordinated efforts of AMAREW project along with the Bureau of Agriculture and other partner institutes and the involvement of the community at large. 57 Figure 12: Land use and land cover map of Lenche Dima in 2005 Accuracy Assessment The study focused on relatively small geographic units as compared to basin wise studies. Hence, much variability is not expected within a short distance and some of changes might not be captured due to image resolution. There is three year difference between time of image acquisition and collection of GCPs and few changes in land use or cover might be expected as individuals constantly make decisions to adapt to changes situations. Generally, classification accuracy could be affected by lack of fine details; resolutions of images used, due the need to make generalizations, and errors are always expected accordingly. To assure wise use of land cover maps and 58 accompanying statistics derived from remote sensing analysis, the errors must be quantitatively explained (Sherefa, 2006). The most common and typical method used by researchers to assess classification accuracy is the use of an error matrix (Congalton et al. 1999 in Sherefa 2006). Therefore, Table 5 shows details of the accuracy assessment done for the two study areas. Table 5: Accuracy assessment report of ASTER image in the two study areas Reference C la s s if ie d I m a g e Class name Study area Cultivate d land Dense Shrub Open Shrub Grass land Total Users Accuracy (%) Cultivated land Kuhar.M 6 2 8 75 Lenche.D 7 2 1 10 70 Dense shrub Kuhar.M 1 2 1 4 50 Lenche.D 1 5 2 8 63 Open shrub Kuhar.M 2 1 2 5 40 Lenche.D 2 4 6 67 grassland Kuhar.M 2 2 6 10 60 Lenche.D 3 1 2 5 11 45 Total Kuhar.M 11 3 4 9 27 Lenche.D 13 6 10 6 35 Producers Accuracy (%) Kuhar. M 55 67 50 67 Lenche.D 54 83 40 83 Users accuracy=number correct/classified total Producers accuracy=number correct/reference total Based on assessment made, producer’s accuracy is found to be 68% for Kuhar Michael and 72% for Lenche Dima (Table 5). Likewise, user’s accuracy is found to be 69% for Lenche Dima and 65% for Kuhar Michael. The overall classification accuracy is revealed to be 59% for Kuhar Michael and 60% for Lenche Dima. 59 DISCUSSION ON LAND USE LAND COVER CHANGE OF THE STUDY AREAS Cultivated land area doubled in Kuhar Michael during the period of observation from 1973 to 1999 (Figure 13). In particular, it was 31% in 1973 and increased to 54 % in 1985, and then to 62% in 1999. This is an annual increase in cropland of approximately 1.7% (Table 6). Figure 13: Summary of land use and land cover change in Kuhar Michael from 1973-2005 The ASTER image indicated that the cultivated land covered 57% of the area which is 5% less than in 1999. Since the grass land increase by approximately the same amount, the difference with 1999 might have been caused by spectral similarity of crops and grassland because the image is acquired during peak growing season. This makes the overall annual rate of expansion 0 10 20 30 40 50 60 70 C u lt iv a te d la n d D e n s e s h ru b /b u s h O p e n s h ru b /b u s h G ra s s la n d B a re l a n d w e tl a n d Land use land cover type P e rc e n t C o v e ra g e (% ) 1972 1985 1999 2005 60 of cultivated to be 0.8% (Table 6), considering the whole study period (1973- 2005). Historically, most of the plains in Fogera, which include the lower part of Kuhar Michael, served as a grazing land and as a place for maintaining animals for longer time as there was no tradition to maintain animals at home during the reign of Hailesilaasie. Hence, in the 1973 image, 12% of the landscape in Kuhar Michael constituted wetlands, serving as both grazing spots and homes for livestock. However, following the Derg’s efforts to improve upon the agricultural systems, through the establishment of cooperatives, the introduction of rice production in the marshy plains, and the cultivation of cash crops, these areas have been converted into cultivated lands. Just after the introduction of rice around 1995 (Descheemaeker, 2008), there were some challenges to convincing farmers to take to the track of growing rice. As a result of extension efforts, and due to attractive returns gained by innovative farmers, rice cultivation has grown to be an attractive business nowadays. It was found to be 12% in 1973, and 9% in 1999, while it was not recognized in 1985 and 2005. However, as described earlier, this percentage cover of wetlands is not merely allocated as a site for grazing and included those land parcels growing rice, the private and communal grazing areas in the plain. According classification scheme developed by Anderson et al., 1976, cultivated wetlands such as the flooded fields associated with rice production are classified as agricultural land. Expansion of cultivated land was observed in Lenche Dima only during the first period (1972-1986), when it increased from 43% to 52% at an average 61 rate of increase of 0.6% per annum. However, the cover decreased continuously to 48% and 36% during the second (1986-2000) and third (2000- 2005) periods of analysis with an estimated rate of decline of 0.3 and 2.4% per year, respectively (Table 8). On the average the cultivated land has decreased by 0.2% in the period from 1972 to 2005. Soil erosion, which has been occurring in Lenche Dima watershed since the first rainy season after the 1973 drought, is the major problem in the watershed damaging the low lying cropland and hilly areas (Gizaw et al., 1999). Dense shrub/bush land in Lenche Dima occupied 16% of the landscape in 1972 and this value decreased to 14% by 1986 and stayed unchanged until 2000. Later in 2005, it showed an improvement and accounted for 24% of the landscape, based on ASTER image. Open shrub/bush land in the area was 25 % in 1972, and this value declined to 17 % by 1986 and then increased again to 22% in 2000, and finally dropped down to 19% in 2005. Finally, the percentage of grassland remained almost unchanged between 1972 and 2000. Later in 2005, the value rose to 21%. The improvement in coverage of shrub lands and grasslands (Figure 12) found during the third period (2000- 2005) of analysis could be largely due to introduction and expansion of exclosures introduced by AMAREW project undertaken for five years from 2003 to 2007. Other studies corroborate our findings for Kuhar Michael. Yitaferu (2007) showed that in the Lake Tana basin during the period from around 1985 to 2003 cropland increased from 46.6% to 50.8%; forest land from 0.8% to 1.04%, shrub land decreased from 7.6 to 5%, wetlands (water and 62 swampland) remained nearly constant at 22.5%, and grassland decreased slightly from 21% to 20%. Further analysis by this author in Wej Awuramba4 , a smaller area within the Lake Tana Basin adjacent to Kuhar Michael, indicted that the cropland expanded from 53% to 61%, grassland decreased from 21.2 to 12.6% and shrub land increased from 20.4 to 22.4% in the periods from 1982 to 2003. Figure 14: Summary of land use and land cover changes in Lenche Dima from 1973-2005 4 It was one of the case study location selected by Yitaferu’s study entitled “Land Degradation and Options for Sustainable Land Management in the Lake Tana Basin (LTB), Amhara Region, Ethiopia”. The area is found in the fogera district, very close to one of the focus area of this study (Kuhar Michael Peasant association). 0 10 20 30 40 50 60 C u lt iv a te d la n d D e n s e s h u b /b u s h la n d O p e n s h ru b /b u s h la n d G ra s s la n d land use land cover type P e rc e n t c o v e ra g e (% ) 1973 1986 2000 2005 63 Table 6: Summary of magnitude and rates of change in land use and land cover of the study areas over the entire study period (from 1972/73 to 2005). T im e f ra m e N o . o f y e a rs Temporal pattern of percent and rate [1] of change in land use and land cover change Cultivated land Dense shrub/bush KM [2] LD [3] KM LD % ∆ Rate % ∆ Rate % ∆ Rate % ∆ Rate 1972-1985 13 23 1.7 6 0.5 1985-1999 14 9 0.6 -11 -0.8 1999-2005 6 -6 -1 3 0.5 1972-2005 33 26 0.8 - 2 0.1 1972-1986 14 9 0.6 -2 -0.2 1986-2000 14 -4 -0.3 0 0 2000-2005 5 -12 -2.4 10 2 1972-2005 33 -7 - 0.2 8 0.2 Open shrub/bush Grassland KM LD KM LD % ∆ Rate % ∆ Rate % ∆ Rate % ∆ Rate 1972-1985 13 -2 -0.2 -8 -0.6 1985-1999 14 6 0.4 -8 -0.6 1999-2005 6 6 1.1 5 0.8 1972-2005 33 9 0.3 -11 -0.3 1972-1986 14 -7 -0.5 1 0 1986-2000 14 5 0.4 -1 -0.1 2000-2005 5 -3 -0.6 5 1 1972-2005 33 - 5 -0.2 5 0.1 Bare land Wetlands KM LD KM LD % ∆ Rate % ∆ Rate % ∆ Rate % ∆ Rate 1972-1985 13 -6 -0.4 -12 -1 1985-1999 14 -4 -0.3 9 0.6 1999-2005 6 0 0 -9 -1.5 1972-2005 33 - 10 - 0.3 - 12 -0.4 1972-1986 14 0 0 0 0 1986-2000 14 0 0 0 0 2000-2005 5 0 0 0 0 1972-2005 33 0 0 0 0 [1] Computed per annum basis by dividing % difference in LULC between selected periods by the time difference. [2] Kuhar Michael Peasant association of Fogera District of South Gonder in Amhara Region, Ethiopia. [3] Lenche Dima watershed from Gubalafto district of North Wollo Zone in Amhara Region, Ethiopia. 64 RESULT FROM SOCIO-ECONOMIC SURVEY KUHAR MICHAEL Changes Associated with Livestock Feed Resources A socio-economic survey was conducted from August 2009 to November 2009 and it involved interview of selected households and group discussion to generate information on household level change in land and livestock holding, and to get insight into various political, social, and environmental factors that influence decision on land use and land cover at household and landscape level. The time horizon considered for the trend analysis using the socio- economic study corresponded with the time horizon considered above for the remote sensing analysis. During the reign of Hailesilassie, the major sources of animal feed were communal holdings that included communal grasslands and shrub/forest lands. Other sources of animal feed that are used nowadays like rice, teff and maize straw were not even collected as they were left in the field and burned to be further incorporated as part of the land preparation activity for next cropping. According to respondents, communal grazing, locally termed Nur in Kuhar Michael, served as the sole animal feed source until the Derg regime came to power in 1974. Following village establishment during the Derg administration, which was made effective around 1987, collection of crop straw started due to the conversion of communal grazing areas in the flood plain to crop production. It was also remarked by respondents that, in addition to this land’s role as a feed source, such areas also served as a holding place for animals during the reign of Hailesilassie, even during periods when most feed resources were depleted. However, the increased risk of disease transfer 65 between animals as they were densely maintained, sometimes beyond the carrying capacity of the areas, was a major disadvantage of such grazing systems. The presence of the tse-tse fly was also another source of threat for animals. Current feeding strategies in the area consist of maintaining the animals around farm boundaries from July to September and allowing animals to graze freely following crop harvest from October to January. Crop byproducts including millet and rice straws, have become major feed sources (Table 7) and complement grazing, particularly during the period between January and July when other feed resources are scarce. Despite variation in locality, seasonal factors and variation in crop type, a generalized comparison was made of the major feed sources available in the study area. Hence, as survey results revealed, rice straw, millet straw, communal grazing, private grazing and maize stover ranked as the top five livestock feed sources in Kuhar Michael (Table 7). More particularly, rice is a primary source for those who have a rice field in the plains. On the other hand, millet grows in the upper parts of Kuhar Michael so that it is the primary source of animal feed for farmers who have farmland in those areas. As can be seen in Table 7, about 63 % of the respondents ranked rice as a primary feed source, while 27% and 14 % of respondents gave it second and third rankings, respectively. This could be attributed to differences between individuals in terms of preference, access, and ownership of the various feed sources, and individual’s preference ranking did show a variation. 66 Table 7: Ranking of different animal feed sources/types in Kuhar Michael (% of the respondents) Sixty-four percent of the respondents in Kuhar Michael reported that they own private grazing land, which expanded from 0.11ha to 0.14ha per household over the last thirty years (Table 10). Furthermore, respondents in Kuhar Michael indicated that 18 % of communal grazing land was converted to cultivated land during the last 30 years. Hay is one of the feed sources for livestock in Kuhar Michael, but it can only be produced by farmers who own relatively large tracts of land and allocate part of them for grassland. Hence, hay is used less as a source of animal feed compared to other sources, as most individuals prefer to cultivate crops on most of their land. Hay can also be produced by institutions like schools or Feed type/source Rank (based on % value) 1 2 3 4 5 Rice straw 63 27 14 Millet stover 25 67 28 Communal grazing 6 29 14 8 Grazing natural private pasture 6 14 29 Maize green leaves 8 Maize straw 7 37 Aftermath grazing 12 29 8 Grazing on reserves for dry season 12 8 Hay from private pasture 12 14 25 Tree fodders 0 33 Hay from private pasture 14 14 8 Total 100 100 100 100 100 67 churches, and sold to individuals to complement feeding whenever there is shortage of feed. Kuaya is a type of grass harvested from private grazing reserves and serves in house construction. But, it is now becoming extinct and is found only in the schools’ premises due to prevalence of free grazing, the substitution of corrugated roofing and sesame stalk (yenug geleba) for hut roof construction, along with conversion to cultivated land. Some respondents pointed out that, among rural households, it was traditionally preferred to have greater number of children and animals for various reasons. To accommodate the latter, they tend to allocate private grazing land from areas of cultivated land, despite the prevalence of a serious shortage of cultivated land. By so doing, they have been managing to strike a balance between crop and livestock production so as to sustain a mixed system that gives relatively better food security than either one of them singly. It is also found that economic returns to land in mixed crop-livestock systems are often higher than for pastoral livestock systems alone (Olson and Maitima, 2006). According to respondents, recent shifts in animal feeding strategies have included efforts to avoid free gazing and the promotion of controlled feeding systems that employ the collection of hay and straws. Furthermore, after 1995, a rule has been enforced to limit grazing by locality, disallowing free grazing away from one’s locality. This has created an increased shortage of grazing area and feed in some areas. 68 Those grazing lands in the plain get flooded and majority of feed in the areas is lost, and individuals in some localities have started using cut and carry system so as to maximize feed production. In these areas, grazing is not between January and May, and also from the start of July to September. However, the area will be open for grazing from September to January, and during the month of May. But, the actual dates in September and May could be changed depending on the beginning and end of the rainfall season. Cut and carry system is generally applied during the rainy seasons, and during the crop growing periods. On the other hand, communal grazing is also managed in a shifting manner that allows animals to graze in some parts, while the remaining is protected for growth. However, this is only possible in localities that have alternative sources of grazing land. According to an elder man in Kuhar Michael, the availability of surplus animal feed in earlier times enhanced frequent breeding of animals. Up to Village establishment, seasonal mobility and free grazing was possible. Later, it was claimed by the then officials that such practices had to be abandoned, in order to minimize degradation. Nowadays, in addition to the serious shortage of animal feed, seasonal mobility from one locality to another in search of grazing land is not possible as feed shortages prevail everywhere, impacting most individuals. Accordingly, seasonal mobility to areas like Dera5, served to provide good feed resources like grass to support the livestock. Nowadays, 5 It is an adjacent district to fogera, which is characterized with availability of well drained areas, covered with scattered vegetation that served as a seasonal grazing for animals from nearby districts. 69 such areas are cultivated by the inhabitants due to a shortage of available cultivable land as a result of population increment. Challenges Related to Grazing Land Prevalence of invasive weeds is one of the problems in Kuhar Michael, and farmers are involved in campaigns to destroy thorny weeds, locally termed Amekela, which invade communal grazing lands creating obstacles for the cattle to graze freely. This problem is more pronounced in the lower plain areas, following continuous grazing of cattle in the wetlands. The problem is minimal in upslope areas that have better drainage. Evolution and Challenges of Fertilization Use Village establishment (Mender Misreta) was made around 1987 during the Derg administration. After that, fertilizer was introduced and provided on loan basis as one of the strategies to improve agricultural productivity. Regardless of farmer’s interest, officials pressured farmers in using fertilizers through established cooperatives. People around Kuhar Michael became dependent on the use of fertilizers. The cost of fertilizer has been steadily rising to a market price of about 640 ETB per 100 kg at the time of the interview (September 2008). According to respondents, coupled with shortage of supply, the increase in price has made it difficult to acquire fertilizer despite most farmers’ high dependency on them. Some farmers remarked that it is becoming almost impossible to produce crops without the use of fertilizer. However, the government is promoting organic based farming and advising individuals to switch to compost, manure and other organic sources of soil fertilization. But, despite the government’s directives and push to get farmers 70 to switch to organic sources of fertilizer, technical shortcomings and subsequent unsatisfactory results upon application of compost, farmers still prefer inorganic fertilizers. However, local administrators’ continued warnings about the limited supply of artificial fertilizers, and even their intention to stop the supply, has created frustration and tension among farmers. Trends in Crop Production and Associated Shifts According to the respondents, the productivity of cropland has deteriorated over time due to continuous cultivation, requiring the use of fertilizers. In the past, traditional practices were undertaken to improve the fertility of farm land, known locally as Tiget Mewgat and Chikcheka. Tiget Mewgat is a periodic process of plowing land every two months starting from July or August for a period of one year to provide rest and is usually practiced by individuals who own more than one parcel of land. Likewise, individuals who own a relatively small tract of land practice crop rotation as a strategy to regain land fertility. Chikcheka is a traditional practice of improving land fertility that involves maintaining livestock in the farmland to accumulate manure and incorporating it into the soil through the movement of the animals. In addition, Chikcheka is also practiced on grazing land to accelerate the growth of grass. It involves the formation of groups of people living in the same locality and the process will be carried out in a shifting basis and will cover the land of all members. It was practiced during Hailesilassie’s time from October to January. The availability of relatively large land holdings and large numbers of livestock had favored 71 these traditional fertility maintenance practices. Tiget Mewgat was usually followed by Chikcheka. Decline in land holdings and livestock population, along with the growing use of animal dung as energy source, are major factors responsible for the decline of these traditional practices. Insight over Shifts in Types of Crops Grown Teff, Guizotia abyssinica (noug), and Millet were some of the major crops during Hailesilassie regime. A trend towards replacing them with rice and maize is reported by respondents. In addition, tomato has become an important cash crop and is widely cultivated in recent years but tomato fields are susceptible to erosion due to low biomass covering the ground. Some respondents pointed out that they started growing tomatoes from 2003 onwards and found an attractive income to fill various income gaps. One farmer noted that he produced tomatoes are worth about 9000 ETB from 0.1ha land. Due to the allure of cash, more cash crops are grown in the area. In an effort to increase household level food and income security, fruit growing and selling are also becoming common following extension services. Garlic was produced during the Kremt rainy season but it now substituted by red onion and tomatoes. However, onion is easily susceptible to disease and its market price has fluctuated due to overproduction in some years. Before rice was introduced in the area, finger millet, noug, and different teff varieties like Netch Laba, Mure, Bukri, and Bohri, were major crops grown. The introduction of rice into the area during the Derg regime was welcomed by the farmers because the land was depleted from continuously growing similar crops like teff, chickpeas and finger millet only. In the uplands rice varieties are 72 grown adapted to the well drained areas but they produce generally less than the lowland varieties. According to the respondents, rice is considered the most productive crop followed by millet and maize. Some rice producers are able to generate substantial savings, with some of them being able to buy residence areas in the town of Woreta. Shifts in Crop Production and associated Issues During the Derg regime, mofer zemet6 was not allowed. This practice was reinstalled following overthrow of the Derg and farmers were able to cultivate land far away from their residence locality. Farmers have been exchanging land among themselves through various arrangements to be able to grow food in proximity to their residence. In Kuhar Michael two or even three crops can be grown per year where irrigation is available. In the upper part of Kuhar Michael only one crop of, maize, millet or teff can be grown. As a result of the increasing population pressure marginal lands are being increasingly used for growing crops, which is revealed by several stony features. Soil erosion is becoming much pronounced. Causes of Expansion of Cultivated Land The expansion of cultivated land as shown on the satellite images in Kuhar Michael was confirmed by interviews with inhabitants that indicated that the land conversion was caused by the high numbers of the youths, who were 6 It is a practice of plowing land located far away from one’s locality. 73 looking for land because no employment opportunities outside agriculture were available. Currently triggered by the serious shortage of land, Kuhar Michael church is renting out about 8 ha land to landless youngsters from the15ha that was delineated for grass production. Selling grass for hay was the major source of income for the priests and other servants. The land is rented for 300 ETB per kedema (0.25ha) for a year. Further land redistribution to address landlessness is difficult due to very small size of farm land holdings and enforcement of land certification. According to a survey made in Amhara region, 12.8% of the rural households have no access to land (Adenew and Abdi, 2005). Hence, it will be imperative to look for other options of addressing the growing shortage of land and landlessness. For instance, as suggested by one author, access to farmlands can be improved through development of land rental markets (Benin et al., 2002). Consequently, the district land management office is considering whether other areas owned by churches and schools can be rented out to the landless youngsters. Recently it was decided that the 30 kedema (7.5ha) of grassland owned by Nora Elementary school, will be rented out as cultivated land to landless people in the locality. This land was one of the major sources of hay to the community. LENCHE DIMA Evolution in Management of Grazing Land and Feed Sources Temporal analysis of feed resources dating 30 years back revealed that the major sources of feed were communal and private grazing. During this period, farmers were not interested to collect crop straws/stover and used to leave it on their farmland as they had enough feed resources. As a result, other than 74 occasional consumption of crop byproducts in crop fields, most of it had no other use than being decomposed and incorporated into soil. Up to 1974, two years after Derg came to power, there was no restriction of grazing by locality and most communal grazing lands were open for every resident in the peasant association. From the beginning of July to September, hillside grazing areas were closed. From October to December, the areas were open for free grazing. During harvesting time (November to January), animals were taken to the farmlands to feed on the leftovers as there was no practice of collecting straw. After Derg came to power, the community was taught to decrease the number of animals they owned and to minimize free grazing. Following this, hillsides were closed and rehabilitated and, consequently a shortage of grazing land led to a decline in the number of livestock. As a result, farmers started collecting straw and other byproducts to complement animal feeding and to overcome feed shortages (Table 8). Likewise, nowadays, free grazing is prohibited and people are forced to practice controlled feeding. Animals are maintained at home from July to November and are fed grass collected from closed areas, weeds from crops, and green leaves from crops. From November to the next rainy season, animals are allowed to graze freely, and the few individuals who have Belg7 crops look after their crops. 7 It is a short production season which covers the period between February to April 75 Table 8: Ranking of different animal feed sources/types in Lenche Dima Feed type/source Rank (%) 1 2 3 4 5 Teff Straw 89 8 Sorghum green 5 12 Maize stover 8 Sorghum straw 5 46 25 Maize green leaves 20 42 33 Aftermath grazing 37.5 14 11 Hay from exclosures 12.5 22 Communal grazing 12.5 Grazing near homestead 12.5 14 11 Grazing on private pasture 14 Grazing on waterways 14 Grazing on wetlands 4 22 Total 100 100 100 100 100 Preference Ranking of Different Sources of Animal Feed In accordance with remote sensing analysis, the socio-economic survey showed that availability of communal and private grazing areas have been continuously declining over the last three decades. Only 5% of the interviewed households said that they owned private grazing land three decades ago. Presently, none of the respondents owned private grazing land. To make up for the decline in grazing lands, crop residues and hay from exclosures are being used as feed. Crop residues are being either carried to the house or grazed on the cultivated lands after the harvest. Farmers rank teff straw as most important feed followed by sorghum stover, maize straw, green sorghum, and maize green leaves based on their availability (Table 8). Other coping strategies to shortages in animal feed include purchasing hay from nearby 76 towns like Kobo, migrating to adjacent shrub lands like Hora, found in the Habru district, and use of tree/shrub species that have a forage quality as well. Emerging Issues, State, and Challenges of Crop Production Generally, contrasting views are reflected by respondents from Lenche Dima regarding the state of cultivated land over time. Some of them argued that the net change in cultivated land is negative as the amount of additional land allocated for cultivation is less than the amount of land abandoned due to land degradation and expansion of settlements. On the other hand, respondents agreed that cultivated land has shown an increase due to the encroachment into communally owned lands like grassland, shrub/bush lands, and the close down of fallowing practice triggered by land shortage problem. Furthermore, the analysis of satellite images revealed that cultivated land increased only during the first period followed by a continuous decline in the preceding two periods, making the net change negative. Following conversion of grazing lands to cultivated land, maize and different fruit trees are grown around homesteads depending on water availability, as a strategy to feed a growing family size. To mitigate the serious moisture stress, attempts were made by AMAREW project to introduce technologies like dome- shaped harvesting structures and drip irrigation, which could help in addressing such problems. However, most of the water harvesting structures are not working well, and those structures with water are not properly used to meet the intended objectives. It is also noted that various weeds could affect crop productivity; these include a Janjur invasion in 1974, Kinche or Chebchabe in 1985 and Kinjit occurred recently in the closed areas. According 77 to Gizaw et al. (1999), crop yield was high before 1973. In 1973 and 1984/85 there was no crop production due to drought and after that it has been declining from year to year, and the total type of crops grown has decreased from nine to four. In an effort to increase crop productivity, the use of fertilizer was tried but it was not promising, as the crops burned due to the hot temperature condition and lack of water. In the valley bottom, a significant proportion of land is lost due to gully formation and deposition of sand is another problem hampering potential of land for production. Hence, soil moisture stress from unreliable rain fall, land degradation, limited use of organic fertilizers and little success from use of synthetic fertilizers, low crop diversity, decline in size of household level land holding and landlessness appear to be major challenges facing the crop production. Overview of Land Fragmentation and Its Impacts A respondent in Lenche Dima pointed out that an area which was owned by one individual in the time between 1982 and 1991 is now divided between 7 households. The average size of non-irrigable land holdings has shown a decline from 2 to 1.3ha per households (Table 10) in Lenche Dima over 30 years. Furthermore, the average homestead size in Lenche Dima was about 0.73ha almost thirty years back and this has fallen down currently to 0.23 ha per household (Table 10). Describing how a settlement has been expanding, a respondent pointed out that an area which hosted about 10 to 15 household heads during the 1970s now accommodates about 50 households. Among the various factors responsible for land fragmentation at the household level, inheritance from parents to sons, individuals equally sharing resources 78 including land when divorcing, and land redistribution play significant roles. However, attributed to the shortage of land and other economical factors, newlyweds are often forced to stay with their parents up to 10 years until they can find a way or an enabling environment to start their own family life. Land Certification Process and Its Impacts Prior to the introduction of the land certification process, individuals were allowed to report to the concerned body whenever their land was degraded or lost due to several natural factors. But, it has been found not as easy to do so, following the land certification process, which entitles farmers with the right to use the land and the obligation to protect it as well. As a means to stop the expansion of land degradation, strict enforcement of rules that obligate land owners to put forth maximum effort in conserving their land is imposed. Therefore, complaints and letters provided are not any more eligible with very few exceptions to claim compensation provided that the owner went through all the measure he can achieve by his own up on occurrence of natural hazards. Furthermore, as part of an effort to control problems that affect land quality, agreement was reached in 1998 to penalize individuals who do not or cannot eradicate Kinche (an aggressive weed affecting crop production) from land units they own. Evolution of Land Management and Associated Challenges In the feudal system, land was only owned by a few landlords. But, following inception of the Derg administration, a land proclamation that entitled individuals the right to own and manage land was made operational. However, due to the serious drought in 1985, some of the farmland was left abandoned 79 until 1991, as the quality of the land had shown deterioration in quality and as some of the residents were settled8 in other areas. In 1991, upon progressing to Addis Ababa from North Ethiopia, the former Tigray people Libration Front (TPLF) now called the Ethiopian Peoples’ Revolutionary Democratic Front (EPRDF), gained control over various localities around North Wollo including Lenche Dima, while the Derg administration was still in power in most parts of the country. Hence, land redistribution was carried out by EPRDF so as to solve problems of the community associated with land and to gain popular support from the community as well. Hence, most communally owned lands were divided among individuals as a cultivated land by maintaining part of it for shrub lands and grazing purpose. For instance, land has been distributed to about 251 new couples/newlyweds between 1998 and 2000, and the sources include ‘Yemote keda’ land (whose owners passed away), shrub lands, and cultivation of abandoned land. Another form of land ownership was also introduced later by the EPRDF with the intention of promoting proper land utilization and development by allowing private investment. Since 1994, part of shrub lands has been provided for three individuals for investment with each one having 15-18, 9 and 11ha. According to information from district officials, the third investor did not work long, and the area was distributed in 1998 among landless newlyweds and land registration was carried out later in 2004. However, residents are not happy about land allocated for investment, and have frequently asked local 8 Settlement was made to south west Ethiopia as there was significant amount of cultivable land due to the existing very low population density, and undisturbed ecosystem that can well support agricultural production. 80 administrators to get back the land so as to allocate it for other alternative uses. Justifications for the proposed change include the fact that the investors are producing similar crops to those already grown by the local inhabitants, and crop production is undertaken with traditional production methods. The original plan of the investors was to undertake fattening, fruit, vegetable and oil crop production has not been realized due to the high cost required to pump water for irrigation. Hence, the community strongly needs the investors to engage in the originally proposed activities that could enable them to learn new production techniques and to access some of the produces with relatively minimum cost in local markets. Due to all these factors, the majority of the interviewed individuals prefer to get back the land allocated for investments and to distribute it among the community. Local Land Administration and Emerging Issues Land is a very important asset in the rural economy and has been the cause of several forms of conflicts and claims among community members. The Environmental Protection, Land Administration and Use Authority (EPLAUA) is primarily responsible for issues associated with land resources management at the district level, and there is also a team of individuals responsible for local land administration unit at the peasant association level. However, the works of this latter team is not subjected to evaluation by the community, and this has created opportunities for these leaders to get involved in various activities of corruption in return for favors from individuals. This takes various forms and includes provision of hillsides for cultivation and failing to take measures pertaining to illegally acquired lands. This situation has led the community to lack trust and confidence in the local land administration committees and the 81 performance of their duties and responsibilities. Decisions about the use and management of communally owned lands are the responsibility of the Kebele land administration committee. Hence, individuals need to get permission whenever they want to gather tree branches and related products for house construction and other cultural ceremonies. All of the follow up and regulation in such regards is done by cooperation between local police and kebele administration. Rationale and Insight over Area Closure Management During the beginning of the 1970s, most hills in Lenche Dima like Sebensa and Dolamba were covered with various trees and shrubs. But starting from 1975, the natural vegetation started degrading due to the removal of trees. During the middle of the Derg’s administration, the area close to Urenew (including Likuas, Kocha Maseria and Tit Kebele) had a good and intact forest cover. But from then onwards, until AMAREW started a project, the area was under a serious threat from land degradation. Organized efforts and commitments were made during the Derg government to rehabilitate and protect degraded/vegetated areas, and encroachment of communal holdings in general, was minimal. However, due to population growth, expansion of land degradation, soil fertility decline and subsequent production decline and instability created during the transition period from Derg to EPRDF government, people destructed significant areas of protected land covered with vegetation during the instable transition period from Derg to EPRDF government. This was triggered by the need to cultivate more land and negative attitudes developed from deprivation of access to use resources, while the areas were being protected by power. Furthermore, due to the need 82 to get more wood biomass for energy, construction, grazing pressure, tools and agricultural implements and the continued need to feed the growing population, communal areas, allocated as a grazing land and covered with vegetation, have been converted to cultivated land and suffer from severe encroachment. Simplifying natural ecosystems with agricultural uses initiates processes that lead to biomass and species losses and substitutions, affecting abundance of biodiversity (Blake and Nicholson, 2004). Achievements brought about as a result of AMAREW’s intervention in Lenche Dima include the safeguarding of cultivated lands as a result of upstream treatment, treatment of gullies, rehabilitation of degraded hills, and improved technical skills in Gabion production. As can be seen in Table 9, about 210 hectare of land was closed by AMAREW project in different localities of Lenche Dima, and subsequently distributed among 520 individuals between 2006 and 2008. Apart from efforts made to rehabilitate and reforest degraded areas, challenges due to moisture stress are reported to have negatively affected the anticipated success. For instance, two of the respondents explained that they planted 300 and 1500 eucalyptus and Gravilea seedlings respectively in the gullies being divided; however, most of them failed due to the serious soil moisture stress prevalent in the area. Challenges in Community-Based Management of exclosures As noted by AMAREW project, exclosures have become social exclosures with no armed guards or fences, but only by an agreement reached among the 83 community members to exclude animals from the protected exclosures and to avoid cutting of vegetation (Gebrekidan et al., 2007). All the member individuals have been involved in protecting the closed areas from illegal intrusions9 and penalizing wrong doers. The penalty ranges from 3 to 100 ETB depending on frequency and type of illegal activity, and this is enforced through cooperation between local police and administrators. An agreement is already made to direct individuals to the next level (social justice system) upon showing resistance to accept the penalty imposed on them. Most people involved in those prohibited activities are from adjacent localities. Key informants argued that the major reason for this is that some residents from the nearby areas feel jealous when they see the area rehabilitated. However, realizing the various forms of benefits drawn from the rehabilitated area like protection of cultivated land from erosion, grass and wood products for fuel, construction and farm implements, members continue to safeguard the area. Despite such organized efforts to maintain and benefit from area exclosures, there are some challenges being faced. For instance, various hillslopes in Oromo including Debiso, Gudguad Kebele and Dishike were distributed for 65 individuals (Table 9). However, only Dishike is well protected while the others have been destroyed. According to respondents, there is still pressure to eliminate the rehabilitated catchment in Dishike, which put sustainability of some of the exclosure under question. 9 This includes several prohibited activities like cutting grass/trees, sending animals in restricted areas, whether the violations are intentional or not and penalty applies to members and non-members. 84 Despite some resistance to maintain closed areas, tendency of expanding the rehabilitation and closure of degraded areas was reported by realizing their importance to adjacent localities. As a result, Gerado and Oromo (Kille Gora) have been protected starting from 2006. Fifty seven percent of respondents in Lenche Dima reported that they are involved in area closure management with an average holding of 0.12ha, which is acquired over the last five years. As a strategy to solve the land shortage problems and promote afforestation in Amhara, a land directive was issued in 1998 to distribute degraded communal lands to landless individuals and groups for private development of forest, fruit and fodder production (Desta et al., 2000). Hence, land holding of households has slightly increased in Lenche Dima due to the involvement of the community in area closure. Table 9: Distributed exclosure in Lenche Dima watershed (source Laste Gerado kebele Agriculture and Rural Development office) Name of the hill/ catchment Identification no. of the locality (got) Year of distribution Area (ha) Number of farmers Male Female Total Kolo kobo 3 2006 88.5 142 39 181 Minchu Gora 3 2004 29 39 5 44 Begido 3 2006 33 93 39 132 Oromo 2 2007 24.2 58 7 65 Dilabma 3 2007 15 27 17 44 Kundiw 2 2008 20 36 18 54 Summary 209.7 395 125 520 Area Closure and Its Influence over Animal Feed Availability Some individuals argued that livestock resources are their banks offering better economic returns as compared to what they get from crop cultivation due to higher uncertainty. However, the expansion of the exclosures practice 85 on various hillslopes in Lenche Dima has negatively affected the livestock component due to the exclusion of animals from such areas. This is especially true for goats, which highly depend on browsing and are the major source of cash. In fact, one has to note that exclosures also complement animal feed through the cut and carry of hay for cattle. Furthermore, the provision of communal grazing areas locally termed yekorma sar to investors negatively affected benefits expected from livestock resources by limiting the availability of feed. According to a survey in Amhara, communal resources have encouraged farmers to keep a large stock of animals (Desta et al., 2000).This author speculated that, if grazing areas were managed privately, farmers would be forced to reduce their herds as they cannot afford to pay for grazing. Misuse of Hillsides and Its Consequences There are cases where land is not allocated for its best alternative use. For instance, hillslopes in the Minchute area, like Yeshikurkubi-Gudguad, that need rehabilitation are distributed for individuals for use as cultivated land. This has resulted in flooding hazard that affected farmland owned by 40 households in Oromo locality. Following reports made by the flood victims to the local administrators and development agents, efforts were exerted to mitigate the damage. However, the problem has not yet been addressed as upstream areas are still under cultivation. It is also argued by informants that the problem may not be arrested by the district bureau of agriculture and rural development due to its scope. Furthermore, the illegal settlement of people near the locality during the last three years has resulted in the destruction of seedlings and physical conservation structures established as part of land rehabilitation efforts. 86 Land Use and Land Cover Change and Climatic Variables as a Proxy According to respondents, following transformation in land use and land cover over the study areas, the climatic condition has changed significantly. They reported a trend towards a shorter rainy season that starts later and finishes earlier with a relatively less predictable pattern as compared to the old times. Particularly, respondents from Lenche Dima remarked that there was a more regular and predictable rainfall pattern three decades back with a positive influence on the various farming activities. It is of interest to note that the community has evolved towards other mechanisms to predict possible outcomes of their farming practices using various components of climate variables as a proxy. For instance, in Lenche Dima, the communities have acquired a traditional experience that if there is sufficient rainfall in the period from Pagume 5th (end of 13th month in Ethiopian calendar, 10th of September) to 15th of September, a period locally termed Tatit or Rufael, there is a high likelihood of getting a good crop harvest provided that other factors are kept constant. With regard to this, there is a local saying in Amharic that goes “Rufael Talelish engdih gotashin abishi” to indicate that if it rains in this period, one should not put extra effort to save resources available at home as there are strong indicators that the harvest will be good. COMPARISON OF CHANGE IN LAND AND LIVESTOCK RESOURCES Insight over Changes in Household Level Land Holdings The remote sensing analysis had shown (Table 6) contrasting trends regarding landscape level change of cultivated land with net expansion in 87 Kuhar Michael, and net decline in Lenche Dima. However, socio-economic surveys revealed a net decrease in the average size of household level cultivated land, which changed from 2.2ha to 1.2ha and from 1.24ha to 1ha per household in Lenche Dima and Kuhar Michael, respectively (Table 10). Major contributing factors include population increase, land redistribution, and land degradation. Unlike expansion of small scale irrigation scheme and landscape level increase in cultivated land, size of household level Irrigated and non-irrigated lands in Kuhar Michael has declined from 0.45ha to 0.37 ha and from 0.79ha to 0.63ha respectively during the last thirty years (Table 10). This could be attributed to population increase and subsequent land redistribution, transfer through inheritance, and due to land degradation. In Lenche Dima, total cultivated land decreased during the last 30 years. As opposed to this, size of irrigated land has increased from 0.22ha to 0.49ha (Table 10). This is attributed to a medium scale irrigation scheme established in 1991 by CoSAERAR (Commission for Sustainable Agriculture and Environment Rehabilitation for Amhara Region) through diversion of Alewuha River (Bekele, 2008). However, discussions with Gubalafto district officials revealed that farmers have not yet fully exploited the potential of the irrigation scheme despite serious moisture stress in the area. This is due to the long distance between farmers’ residence and irrigable fields that requires frequent travel and loss of human labor and draft power, less follow up efforts and attack of matured crops by wild animal. Hence, efforts are underway to address this problem by establishing a village in the nearby area to facilitate 88 settlement of model famers from Lenche Dima, who have irrigated land in Alewuha area. Land degradation in the form of gully erosion is reported to be a problem in both study areas with greater magnitude in Lenche Dima, which brought significant area of cultivated land out of production. Unsustainable and improper land use and land cover changes are the major causes of land degradation (Bossio et al. 2007 cited in Descheemaeker et al., 2009). Plantation of trees around homesteads, in the form of woodlots, as a farm boundary, in conservation structures and as a means to stabilize gullies has been tried in both study areas. However, survival of tree seedlings is barely possible in Lenche Dima except in closed areas and it has shown little success due to serious soil moisture stress. On the other hand, tree planting in Kuhar Michael has shown improvement, as plantation area increased from 0.01ha to 0.02ha of woodlot per household over thirty years being a growing practice and important source of income, energy and construction material for the inhabitants. Similarly, as revealed by interview with the selected households in Lenche Dima, size of average exclosure per household is found to be 0.21ha (Table 10). But, this value is found to be 0.4ha when computed based on secondary data about exclosures distribution from Laste Gerado Kebele Office of agriculture and rural development. 8 9 Table 10: Characteristics of interviewed households and their land holding. Household characteristics Average Land holding per Household S tu d y a re a Y e a r N u m b e r o f H H in te rv ie w e d % M a le % F e m a le A v e r. A g e Cultivated P ri v a te .g ra z in g H o m e s te a d A b a n d o n e d G u lly W o o d lo ts A re a c lo s u re T o ta l Ir ri g a te d N o n - ir ri g a te d T o ta l K u h a r M ic h a e l Now 28 54 44 51 0.37 0.63 1 0.14 0.26 0.02 0.02 0.08 - 1.48 30 year ago - - - - 0.45 0.79 1.24 0.11 0.26 0.01 0.01 0.01 - 1.59 L e n c h e D im a Now 22 59 41 51 0.49 1.3 1.8 0 0.23 0 0.02 - 0.12 2.15 30 year ago - - - - 0.22 2 2.2 0.01 0.73 0 0.01 - 0 2.93 90 The total average land holding per household in both study areas has shown a decline from 2.92ha to 2.15ha for Lenche Dima and from 1.59ha to 1.48ha in Kuhar Michael over the past 30 years (Table 10 and Figure 15). Figure 15: Summary of changes in average household level land holding in the study areas. Particularly, a greater magnitude of change was found in Lenche Dima than in Kuhar Michael (Figure 15), which could be due to lower land productivity requiring cultivating more land and establishment of exclosures as a response to serious land degradation in Lenche Dima. In Kuhar Michael, the land size remained around 1.5 ha which is likely the minimum amount to make a living. In this location, Bahir Dar town is relatively close and seasonal migration to Metema area is also another potential source of job for those to seek employment that have not sufficient land. 0.00 0.50 1.00 1.50 2.00 2.50 3.00 3.50 30yr 20yr 10yr 5yr Now Time (year) S iz e o f la n d (h a ) Lenche Dima Kuhar Michal 91 Major changes associated with Livestock resource Livestock plays a critical role for supporting communities involved in the mixed crop-livestock production system through its products and services. However, animal keeping tends to affected by changing socio-economic, biophysical, political, land cover cover and land use. Insight over trends in household level change of livestock holding (cattle, small ruminants and pack animal) population over thirty years period is revealed by the socio-economic survey (Table 11 and Figure 16). Table 11: Summary of the average number of animals of different livestock types based on socio-economic survey As revealed by the interviews, the number of cattle holding in Kuhar Michael remained the same while it decreased in Lenche Dima over the last thirty years (Table 11 and Figure 16a). The average number of cattle holding in Lenche Dima 30 years ago was by far greater than Kuhar Michael (Table 11and Figure 16a). A sharp decrease was observed in Lenche Dima between 30 and 20 years ago, while a slight improvement was found in Kuhar Michael due to changes made in land ownership policy and distribution of communal grazing areas for cultivation. As a result, cattle holding in both areas Average Livestock holding per household(number of animals) Study area Cattle Pack animals Small ruminants Total livestock 30 years ago Now 30 years ago Now 30 years ago Now 30 years ago Now Kuhar Michael 3.43 3.43 0.25 0.75 0.18 0.79 3.86 4.97 Lenche Dima 5.82 4.09 0.41 0.82 5.91 4.55 12.14 9.46 92 converged to an average value of four animals per household 20 years ago. From that time onwards, population of cattle in Kuhar Michael remained unchanged starting from the point of convergence up to 5 years ago and showed only a slight decrease up to recent years (Figure 16a). Despite periodic variations, the over all trend of cattle holding decreased more in Lenche Dima than in Kuhar Michael. This result corresponds well with the reported shrink and complete conversion of most grassland in Lenche Dima. Despite varations in number of pack animal,in the study area over the last 30 years (Figure 16b), a net increase was found for both study areas (table 11), but with higher magnitude in Kuhar Michael than in Lenche Dima. This increase in both study areas is a consequence of the growing need for transportation means. According to respondents, there was no camel in ten years ago and following its introduction, it kept growing and 32% of the respondents said to have camels now. This is largely due to the comparative advantage of camels due to its diversified feeding ability , capablity to withstand drought (water and feed shortages), capability to carry heavy loads, enabling owners to offer rental service during distribution of food aid, crop harvest, and they also facilitate the selling of charcoal and fuel wood to markets. Hence, owners are able to generate significant income, which form part of the mechanisms employed to assures household income and food security. An adverse impact on the local environment is also being reported due to introduction and growing population of camels in Lenche Dima. 93 Figure 16: Trends in average household level change in holding of cattle (a) pack animals (b), small ruminants(c), and total livestock (d) in the study areas. Over the past thirty years, the number of small ruminants was greater in Lenche Dima than Kuhar Michael. The number decreased despite variations in Lenche Dima and increase slightly in Kuhar Michael (figure 16c & table 13). From the households surveyed, about 23% and 3.5% of the respondents reported to own sheep in Lenche Dima and Kuhar Michael respectively. Furthermore, 67% of the respondents in Lenche Dima said to have had up to 25 goats per household five years ago, of which 43% sold all goats following closure of hillsides by AMAREW project . Initially, goats were kept and 0.00 2.00 4.00 6.00 8.00 10.00 30yr 20yr 10yr 5yr Now N o m b e r o f r u m in a n t s Time(year) c 0.00 1.00 2.00 3.00 4.00 5.00 6.00 7.00 30yr 20yr 10yr 5yr Now N u m b e r o f c a t t le Time(year)a 0.00 0.10 0.20 0.30 0.40 0.50 0.60 0.70 0.80 0.90 30yr 20yr 10yr 5yr Now N u m b e r o f p a c k a n im a ls Time(year)b 0.00 2.00 4.00 6.00 8.00 10.00 12.00 14.00 16.00 30yr 20yr 10yr 5yr Now T o ta l n u m b e r o f l iv e st o ck Time horizon(year)d 94 additional was distributed to selected households by the project as a strategy to create income and improving the livelihood of the community. After watershed management (area closure and gully rehablitation) was implemented and some areas were closed from human and animal contact. Besides its protection, members were also granted accesss of grass through cut and carry system, which became a highly accepted activity by the majority of the watershed communities as income producing activity (Gebrekidan et al., 2006). However, unlike speculations made during earlier stages of the project, farmers are now complaining about the difficulty of rearing goats as their sole feed sources are being protected and mobility is restricted. According to Descheemaeker (2006), apart from its positive and fast effect on vegetation recovery, exclosures can also cause increased pressure on the remaining grazing lands and natural forests and they do not yield many tangible and material benefits. Hence, the author further remarked that the practice may not always get support by the local farmers due to the opportunity cost from economic loss of other activities that disappear upon establishment of exclosures, which should not be ignored. This requires creating arrangements to allow farmers get access to woody biomass and helping them to benefit from non-wood forest products. According to survey made in Gojam area of Amhara region, goat represented 10% of farm capital invested in livestock yet accounted for about 40% of cash income from livestock (CEDEP, 1999 in Desta et al., 2000) reflecting the critical role of goats as source of cash compared to other animal types. Therefore, the observed decline of goat population in recent years could affect 95 household income, which could be translated as an impaired ability to escape bad times contingencies. Despite the periodic fluctuations of the various livestock types, the net change over the last three decades appears to be an increase for Kuhar Michael and a decrease for Lenche Dima (Table 11 and Figure 16d). However, it was also found to have relatively better stability in Kuhar Michael, while it has undergone a continuous change in Lenche Dima due to the serious decrease and shortage of private and communal grasslands being created following land redistribution, and area closure interventions. This trend in livestock population agrees well with general patterns of grassland as revealed by the remote sensing analysis. 96 CHAPTER FIVE CONCLUSION AND RECOMMENDATION CONCLUSIONS Remote sensing results showed that the area of cultivated land in both study areas has undergone a change, but with differing magnitude and rates. In Kuhar Michael, an expansion was found during the first two periods with annual rates of 1.7% and 0.6% , while a decline was revealed during the third period at annual rate of -1% . Despite the periodic variations in change of cultivated land, the overall rate of change during the whole period of analysis (1972-2005) is found to be 0.8% per annum, which resulted in a net increase. However, in Lenche Dima, cultivated land showed a declining trend both at household and landscape level during most analysis periods largely due to land degradation. Expansion of cultivated land was encountered only during the first period (1972-1986) with annual rate of 0.6%. During the second and third periods, a decline was found at annual rate of -0.3 to -2.4%, making the over all rate of decline -0.2% per annum. Furthermore, abandonment of traditional soil fertility improvement practices is reported in both study areas due to the declining size of land holding at household level from redistribution and inheritance, the decrease in livestock population and subsequent lower allocation of manure, straw and other crop byproducts. Consequently, strong dependence on use of inorganic fertilizer prevails in Kuhar Michael to maintain soil fertility. In Lenche Dima, inorganic fertilizer was also tried but did not gave a satisfactory result due to the prevailing serious soil moisture stress and as crops get burnt up on application 97 of fertilizer. Furthermore, along with the expansion of cultivated land, introduction of new crops like rice, onion, tomatoe took place in Kuhar Michael as availability of water made it possible to develop small scale irrigation schemes. This has contributed significantly towards improved household income in Kuhar Michael from production and sell of cash crops. On the other hand, crop production in Lenche Dima has been constrained by land degradation, unreliable rainfall and soil moisture stress, abandonment of traditional soil fertility improvement practices, and limited success upon use of inorganic fertilizers. Added to this, the number of crop types grown is reported to have decreased from 9 to 4 (Gizaw et al., 1999) leading to limited production possibility. Generally, communal grazing used to be the major source of animal feed in the study areas during the reign of Hailesilassie. However, later during the Derg administration, part of the communal grazing lands was distributed for cultivation and this created shortage in feed. Hence, crop byproducts and leftovers were introduced as alternative sources of animal feed. Preference ranking revealed that rice, millet, and maize straw/stover in Kuhar Michael and teff, sorghum, and maize green/stover to be important feed sources currently, besides to grasslands. As far as private grassland is concerned, all respondents in Lenche Dima replied that they do not have any due to serious shortage of cultivated land, while 64% of the respondents in Kuhar Michael claimed to own it. As revealed from the socio-economic survey and confirmed by RS/GIS analysis of satellite images, grasslands have shown changes that extend from 98 a decline in cover to complete extinction in some localities of the study areas. In particular, the shrinkage was found to be more pronounced in Lenche Dima than in Kuhar Michael. However, ASTER image analysis had shown a recent improvement in grasslands cover due to watershed management and exclosure interventions made by AMAREW project, which contributed in addressing feed shortage created upon distribution of such areas to community members. Consequently, an overall improvement of grassland at annual rate of 0.1% is revealed. The interventions made by AMAREW project have brought about several benefits and evaluations made during earlier stages of the projected indicated that most individuals have accepted the intervention, which could be due to direct benefits and influence by others. However, recent surveys showed that shortcomings are emerging in the long run after phase out of the project as it is not so easy to immediately evaluate the overall acceptance and success of an intervention. Hence, a detailed evaluation of the impacts brought about as a result of AMAREW’s intervention is highly required. In Kuhar Michael, grasslands owned communally and those that belong to institutions are under continuous conversion to cultivated lands. Grasslands declined in the first two periods at a rate of -0.6% each, followed by a slight improvement at annual rate of 0.8% in the third period, which makes the annual rate of net change to be -0.3%. Unlike this, there are some individuals who still allocate part of their land for private grazing unlike the case in Lenche Dima, as crop production is supported with irrigation and orientation towards cash crops enabled farmers to get a better production per unit area, which, as a result, allowed farmers to allocate part of their land as grassland. Despite 99 landscape level decline in grassland as revealed by remote sensing image analysis, respondents in Kuhar Michael confirmed that farmers are still able to allocate small parcels for private grassland to complement animal feed. The grassland presented by the remote sensing analysis meant to represent most of the communal holdings as those owned privately may not be captured by the remote sensor instruments due to disparity between the relative small size of privately owned grasslands and spatial representation (scale) of the remote sensor instruments. Based on temporal analysis of livestock population in the study areas, the net population is found to grow slightly in Kuhar Michael, while much decline is revealed in Lenche Dima over the last thirty years. Particularly in Kuhar Michael, pack animals and small ruminants increased while cattle holding stayed same. On the other hand, in Lenche Dima, pack animals showed an increase and the household holding of cattle and small ruminants declined. The observed pattern of net change in livestock holding corresponds well with the reported shrink and complete conversion of most grassland in Lenche Dima. This decline in household level livestock holding along with challenges posed on the crop cultivation implies jeopardy on livelihood of Lenche Dima inhabitants. The state of dense and open shrub/bush land has also shown a contrasting trend in the study areas. Particularly, dense shrub land has shown an overall improvement in Lenche Dima with a rate of 0.2%, while open shrub/bush land declined at annual rate of -0.2%. This could be associated with watershed management practices implemented by AMAREW project. In Kuhar Michael, 100 Dense shrub/bush land has shown a net decrease at a rate of -0.1%, while open shrub/bush land expanded with annual average of 0.3% over the study period. This could be due to lack of protection of communally owned shrub/bush lands and access by animals and damage from grazing, collection of fuel wood that facilitated conversion of dense to open shrub/bush land. Apart from the decline in both open and dense shrub/bush lands as revealed by the remote sensing analysis, respondents testified the limited availability of some tree/shrub species and some are also reported to be under extinction. For instance, introduction and population increase of camels is reported to have brought about adverse impacts on the local environments of Lenche Dima as camels browse agressively with a very wide range of feed preferences including most shrub, bush and tree species in the landscape. This has disrupted some ecological functions through the pressure created on regeneration and dynamics of scattered shrubs and trees in farmlands, near homesteads and in unprotected communal areas. One farmer even remarked that ”all the shrub, bush and tree species currently found scattered in the farmlands are those that finished their seedling/sapling stage and grew up to be a tree prior to the arrival of camels in the area”. This means that succession is seriously undermined and the regeneration ability of some species like Ziziphus spina-christi that are well adapted in Lenche Dima is being affected. Similarly, due to the fact that most shrub/bush lands in Kuhar Michael are open access for human and animals and consequent pressure created due to fuel wood collection and grazing,regenerative capacity, distribution and status of some tree/shrub species is hampered. 101 RECOMMENDATIONS 1. In the face of the growing household level land shortage and growing number of landless youths, it will be imperative to create and strengthen non-farm/off-farm income generating activities due to limited capacity of agriculture to accommodate additional population. 2. Forest resources development, protection and use strategies need to be devised to counteract the deteriorating shrub/bushes and avoid further extinction of important shrub/tree species and migration of associated wild animals. 3. Further study is required to quantify reported species extinction and underlying factors responsible for that. Impacts brought about on the local environment of Lenche Dima due to introduction of camel need further research. 4. AMAREW project, encompassing watershed management, microfinance, research and extension activities, has brought several positive benefits. However, negative impacts are also reported few years after termination of the project due to some of the interventions. Hence, further study is required to better understand current challenges and evaluate overall impacts so as to make immediate corrections and draw important lessons to further scale up the experience to other areas. 5. As revealed by the socio-economic survey, land provided for investors in Lenche Dima is done without consent of the community and this led to tension and lack of trust. According to Gubalafto District expert, effort is underway to provide additional area to provide for new investors. The new land rental service opened up for landless youths by converting 102 institutionally owned grasslands to cultivated land is an appreciable effort, which came out of the collaborative effort of Fogera District experts, the local level land administration committee, and community members. Therefore, participation of the community and getting their approval is highly required prior to providing land for investors and making similar decisions in order to avoid conflicts. 6. Designating interventions/activities that can be synchronized with exclosure management is very important, as an effort to improve the livelihood of the community in Lenche Dima and to increase economic importance of such areas. In Kuhar Michael, community participation is lacking in protection of communally owned vegetated areas. Hence, rehabilitation, closure and distribution of degraded hills or other conservation practices has to be launched to minimize land degradation. 103 REFERENCES Abrams, M., Hook, S., Ramachandran, B. 2009, ASTER User Handbook, Version 2. Jet Propulsion Laboratory, California Institute of Technology, USA. Adenew, B. and Abdi, F., 2001. Land Registration in Amhara Region, Ethiopia Research Report 4. London: International Institute for Environment and Development. Agarwal, C., Green, G. M., Grove, J. M., Evans, T. P., and Schweik, C. M., 2002. A Review and Assessment of Land-Use Change Models: Dynamics of Space, Time, and Human Choice. General Technical Report NE-297. Newtown Square, Pennsylvania: U.S. Department of Agriculture, Forest Service, Northeastern Research Station. 61 pp. Amede, T., Mengistu S., and Roothaert R., 2005. Intensification of Livestock Feed Production in Ethiopian Highlands: Potential and Experiences of the African Highlands Initiative, paper presented at the 19th EVA Annual conference, 8 June 2005, ECA, Addis Ababa. Amsalu. A., Stroosnijder, Graaff, J. D., 2006. Long-term dynamics in land resource use and the driving forces in the Beressa watershed, highlands of Ethiopia. Journal of Environmental Management. 83, 448- 459. Anderson, J. R., Hardy, E. E., Roach, J. T., and Witmer, R.E., 1976. A Land Use and Land Cover Classification System for Use with Remote Sensor Data; Geological Survey Professional Paper 964. United States Government Printing Office, Washington. 104 Anonymous, 2000. Radiometric correction of satellite images: when and why radiometric correction is necessary in Applications of satellite and airborne image data to coastal management. Baulies, X., and Szejwach, G., 1998. LUCC data requirements workshop survey of needs, gaps and priorities on data for land-use/land-cover change research organized by IGBP/IHDP-LUCC AND IGBP-DIS, Barcelona, Spain, 11-14 November 1997 LUCC report series no. 3. Bekele, M., 2008. Integrating Livestock Production into Water Resources Development: Assessment on Livelihood Resilience and Livestock Water Productivity at Alewuha and Golina Rivers. M.Sc. Thesis. Awassa, Ethiopia. Hawassa University, Benin, S., Ehui, S., and Pender, J., 2006. Policies for livestock development in the Ethiopian highlands. In Strategies for sustainable land management in the East African highlands. (ed) Pender, J., F. Place, and S. Ehui, pp 141, Washington, D.C.: International Food Policy Research Institute. Benin S. and Pender J., 2002. Impacts of land redistribution on land management and productivity in the Ethiopian highlands. Socio- Economic and Policy Research Working Paper 43. ILRI (International Livestock Research Institute), Addis Ababa, Ethiopia. 26pp. Blake, R. W., and C.F.Nicholson. , 2004. Livestock, land use change and environmental outcomes in the developing world. Pages 133-153 (chapter 9) in Responding to the Livestock revolution-the role of globalization and implications for poverty alleviation. (Owen et al., eds). Nottingham University Press, UK. Bureau of Agriculture and Rural Development (BoARD)., 2003. Rural Households Socio-Economic Baseline Survey of 56 Weredas in the 105 Amhara Region (Phase I). Volume VIII, Main Document. Federal Republic of Ethiopia, Amhara National Regional State, Bahir Dar. CA (Comprehensive Assessment of Water Management in Agriculture), 2007. Water for Food, water for Life: A Comprehensive Assessment of Water Management in Agriculture. Earthscan, London, UK and International Water Management Institute, Colombo, Sri Lanka. Chrysoulakis, N., Kamarianakis, Y., Farsari, Y., Diamandakis, M. and P.Prastacos, 2004.Combining Satellite and Socioeconomic data for Land Use Models estimation. In Goossens, R.(Editor), Proc.Of 3rd Workshop of EARSeL Special Interest Group on Remote Sensing for Developing Coutries(in press). Codjoe, S.N.A., 2007. Integrating Remote Sensing, GIS, Census, and Socioeconomic Data in Studying the Population–Land Use/Cover Nexus in Ghana: A Literature Update. Africa Development, Vol. XXXII, No. 2, pp. 197–212. Dale, V. H., O’Neill, R. V., Pedlowski, M. and Southworth, F., 1993, ‘Causes and Effects of Land Use Change in Central Rondonia, Brazil’, Photogrametric-Engineering and Remote Sensing, vol. 59, no. 6, pp. 997-1005. Daniels, A.E., Painter, K., Southworth, J., 2008 .Milpa imprint on the tropical dry forest landscape in Yucatan, Mexico: Remote sensing & field measurement of edge vegetation. Elsevier, Amsterdam, The Netherlands, pp.293-304. Descheemaeker, K., 2006. Pedological and hydrological effects of vegetation restoration in exclosures established on degraded hillslopes in the highlands of Northern Ethiopia. PhD thesis, K.U.Leuven, Belgium, 364p 106 Descheemaeker, K., 2008. Baseline data report for the Ethiopian study sites: Fogera: Quhar Michael and Gubalafto: Lenche Dima, BMZ project in Improving Water Productivity of Crop-Livestock Systems of Sub- Saharan Africa. IWMI (International Water Management Institute) and ILRI (International Livestock Research Institute), Addis Ababa, Ethiopia. Descheemaeker, K.; Amede, T. Haileslassie, A., 2009. Livestock and water interactions in mixed crop-livestock farming systems of sub-Saharan Africa: Interventions for improved productivity. Colombo, Sri Lanka: International Water Management Institute. 44p. (IWMI Working Paper 133). Desta, L., Kassie, M., Benin, S., and Pender, J., 2000. Land degradation and strategies for sustainable development in the Ethiopian highlands: Amhara Region. Socio-economics and policy research working paper 32.ILRI (International Livestock Research Institute), Nairobi, Kenya.122 pp. Engineering Manual., 2003. Remote sensing, Engineer Manual No. 1110-2- 2907, Department of the Army, US Army Corps of Engineers Washington, DC. FAO., 2006. Environmental Impact Assessment, Volume II of Livestock Production Originated by Agriculture and Consumer Protection. Gebrekidan.B., Alemayehu, N., , Ashine. Y., Bayafers, G., Kebede, S., 2007. AMAREW Project Terminal Report July 2002 – December 2007; Amhara Micro enterprise development, Agricultural Research, Extension, and Watershed management (AMAREW) Project, Tebikew Baile Building, Bahir Dar, Ethiopia. 107 Gizaw, S., Mekonnen, K., Desta, L., 1999. Lenche Dima Integrated Watershed Development Project feasibility study report. Amhara National Regional State, Bahir Dar, Ethiopia. GLOBE toolkit, 2003 in Landsat Multispectral Scanner (MSS) and Thematic Mapper (TM). (Available online http://iic.gis.umn.edu/finfo/land/ landsat2.htm). (Accessed on December 2008.) Hardy, Ernest E., Belcher, Donald B., and Phillips, Elmer S., 1971, Land use classification with simulated satellite photography: U.S. Dept. of Agriculture, Econ. Research Service, Agr. Inf. Bull., 352 p. Harrington, L.W., Douthwaite, B., Leon, C.D., and Woolley, J., 2007.Most Significant Change (MSC): Stories from the field, CGIAR challenge Program on Water and Food (CPWF). Harrington, L.W., Gichuki, F., Gaheb, K., and Woolley, J., 2004.Changinnig the way we managed water for food, livelihoods health and the environment, CGIAR challenge program on Water and food. Jähne, B., 2002. Digital Image Processing 5th revised and extended edition, Springer-Verlag Berlin Heidelberg, Germany. Lambin, E.F., Turner II, B.L., Geist, H., Agbola, S., Angelsen, A., Bruce, J.W., Coomes, O., Dirzo, R., Fischer, G., Folke, C., George, P.S., Homewood, K., Imbernon, J., Leemans, R., Li, X., Moran, E.F., Mortimore, M., Ramakrishnan, P.S., Richards, J.F., Ska°nes, H., Steffen, W., Stone, G.D., Svedin, U., Veldkamp, T., Vogel, C., Xu, J., 2001. Our emerging understanding of the causes of land-use and - cover change. Global Environ. Change 11, 261–269. 108 Lambin, E.F., Geist H.J., and lepers, E., 2003. Dynamics of land-use and land- cover change in tropical regions, annual review of environmental resources, 28:205–41. Lambin, E.F., 2005. Conditions for sustainability of human–environment systems: Information, motivation, and capacity, Global Environmental Change 15 (2005) 177–180, Elsevier. Lambin, E.F. and Strahler, A.H., 1994. Change-vector analysis in multispectral space: A tool to detect and categorize land-cover change processes using high temporal-resolution satellite data. Remote Sens. Environ., 48, 231-244. Landsat Multispectral Scanner (MSS) and Thematic Mapper (TM) accessed online from http://iic.gis.umn.edu/finfo/land/landsat2.htm on December 2008. Liang, S., Fang, H., Morisette, J.T, Chen, M., Shuey, C.J., Walthall C.L., Daughtry, C.S, 2002. Atmospheric Correction of Landsat ETM+ Land Surface Imagery: II. Validation and Applications. IEEE Transactions on Geoscience and Remote Sensing, Lillesand, T.M., and Kieffer, R.W., 2004. Remote Sensing and Image Interpretation; 4th ed., Willey, New York, 469pp. Lu, D., Mausel, P., Brondizio, E., and Moran, E., 2004. Change detection techniques. Int. J. Remote Sensing, 20(12), 2365-2407. Lunetta, R.L., Knight, F.K, Ediriwickrema, J., Lyon, J.G., and Worthy, L.D., 2006. Land cover change detection using multi-temporal MODIS NDVI data. Remote Sensing of Environment, 105, 142-154. Mather, P. M., 2004 .Computer Processing of Remotely-Sensed Images: an Introduction, 3rd Edition. John Wiley & Sons Ltd, Chichester, England. 109 McCornick, P.G., Kamara, A.B., and Tadesse, G., (eds). 2003. Integrated water and land management research and capacity building priorities for Ethiopia. Proceedings of a MoWR/EARO/IWMI/ILRI international workshop held at ILRI, Addis Ababa, Ethiopia, 2–4 December 2002. IWMI (International Water Management Institute), Colombo, Sri Lanka, and ILRI (International Livestock Research Institute), Nairobi, Kenya. 267 pp. McHugh, O., 2006. Integrated water resources assessment and management in a drought-prone watershed in the Ethiopian Highlands. PhD thesis, Cornell University, US. McHugh, O. V., McHugh, N.P., Eloundou-Enyegue, M., and Steenhuis, T. S., 2007. Integrated Qualitative Assessment of Wetland Hydrological and Land Cover Changes in a Data Scarce Dry Ethiopian Highland Watershed, Land Degradation and Development, Wiley Inter Science. Mustard, J.F., Defries, R.S., Fisher, T., Moran, E.F., 2004. Land-use and land- cover change pathways and impacts. In: Gutman, G., Janetos, A.C., Justice, C.O., Moran, E.F., Mustard, J.F., Rindfuss, R.R., Skole, D., Turner, II, B.L., Cochrane, M.A. (Eds.), Land Change Science: Observing, Monitoring and Understanding Trajectories of Change on the Earth’s Surface. Kluwer, Dordrecht, The Netherlands, pp. 411–429 Mwendera, E. J., Mohamed Saleem, M. A., and Dibabe, A., 1997. The effect of livestock grazing on surface runoff and soil erosion from sloping pasture lands in the Ethiopian highlands, Australian Journal of Experimental Agriculture, No-37, PP 421–30. Olson, J.M., and Maitima, J.M., 2006. Sustainable Intensification of Mixed Crop-Livestock Systems, land use change impacts and dynamics 110 (LUCID) Policy Brief # 1, International livestock research institute, Nairobi, Kenya. Omo-Irabora, O. O., and Oduyemib, K., 2007. A Hybrid Image Classification Approach for the Systematic Analysis of Land Cover (LC) Changes in the Niger Delta Region. Built and Natural Environment, School of Contemporary Sciences, University of Abertay, Scotland, UK. Quentin F. B., Jim, C., Julia, C., Carole, H., and Andrew, S., 2006. Drivers of land use change, Final report: Matching opportunities to motivations, ESAI project 05116, Department of Sustainability and Environment and primary industries, Royal Melbourne Institute of Technology. Australia. Richards, J. A., 1999. Remote Sensing Digital Image Analysis, Springer- Verlag, Berlin, p. 240. Sherefa, B., 2006, Remote Sensing & GIS for Land Cover/ Land Use Change Detection and Analysis in the Semi-Natural Ecosystems and Agriculture Landscapes of the Central Ethiopian Rift Valley. PhD thesis, Institute of Photogrammetry and Remote Sensing, Technology University of Dresden, Germany. Singh, A., 1989. Digital change detection techniques using remotely-sensed data. Int. J.Remote Sensing, 10(6), 989-1003. Smith, A.M.S. How to convert ASTER radiance values to reflectance: an online guide. University of Idaho accessed on November, 17, 2008 from www.cnrhome.uidaho.edu/default.aspx?pid=85984. Tucker, C.J., Grant, D.M., and Dykstra, J.D. 2004. NASA's Global Orthorectified Landsat Data Set: an Introductory Landsat Tutorial for Users of GeoCover Data in Photogrammetric Engineering and Remote Sensing, Volume 70, Number 3, pp. 313-322. 111 U.S EPA., 1999. A Summary of Models for Assessing the Effects of Community Growth and Change on Land-Use Patterns. EPA/600/R- 00/098. U.S. Environmental Protection Agency, Office of Research and Development, Cincinnati, Ohio. 260 pp. Weismiller, R.A., Kristof, S.J., Scholtz, D.K., Anulta, P.E., and Momin, S.A., 1977. Change detection in coastal zone environments. Photogrammetric Engineering and Remote Sensing, 43(12), 1533- 1539. Woody Biomass Inventory and Strategic Planning Project (WBISPP), 2002. Report on natural grazing lands and livestock feed resources, Amhara National Regional State. Addis Ababa. Yamaguchi, Y., Kahle, A., Tsu, H., Kawakami, T., and Pniel, M., 1998. Overview of Advanced Space borne Thermal Emission and Reflection Radiometer (ASTER) IEEE trans. Geosci. Remote sensing, vol. 36, pp 1062-1071. Yitaferu, B., 2007, Land Degradation and Options for Sustainable Land Management in the Lake Tana Basin (LTB), Amhara Region, Ethiopia. PhD thesis, Centre for Development and Environment, University of Bern, Switzerland. Zeleke, G. and Steenhuis .T.S., 2005. The Broken Link in the Agricultural Production System in Ethiopia and its Impacts: Natural Resources Engineering and Management, Cornell University, Ithaca, USA. Zubair, A. O., 2006. Change detection in land use and Land cover using remote sensing data and GIS (A case study of Ilorin and its environs in Kwara State), Msc Thesis, University of Ibadan, Nigeria. 112 APPENDICES APPENDIX I: Household level questionnaire This questionnaire is prepared to collect data about land use and land cover changes, their drivers, impacts and implications for management from two contrasting sites of Amhara region. It is expected to generate and provide helpful information for policy makers and development practitioners about magnitude and trends of land use and land cover change and its impact on availability of livestock feed. Therefore, your inputs as a stakeholder to fill this questionnaire is highly appreciated and information provided will stay confidential and your right to involve or not is also respected. Part I-General background information 1. Date of interview…………………. 2. Start time ……...End time ……… time elapsed…… 3. Name of interviewee………………………....Age……..Position in the HH…………… 4. District ……………… Kebele …………….. Got/PA…………….. 5. Total number of family members ………. F........ M.......... Members able to read and write........... 6. Total land10 owned ………(‘Kada’)………..(ha) Type of land Approximate Size of each parcels of land Underlying reasons Now 5years ago 10years ago 20year s ago Homestead irrigated cropland Non-irrig cropland Private grazing Communal grazing Closed area/hill Abandoned Gully Plantation 7. GPS measure( coordinate ) N………………E…..………………elevation (m)...….precision (m)……. 10 refers to amount of land owned by the interviewed household during the time of interview 113 Part II-Soil and Water related issues 8. What are the major problems associated with water resources in your locality? …………………………………………………………………………………………………. …………………………………………………………………………………………………. …………………………………………………………………………………………………. 9. Is land degradation a problem in your locality? Yes  No  What type of land (land cover, topography, and soil type) is vulnerable to land degradation (in order of vulnerability)? Soil erosion: …………………………………………………………………........................ Gully formation: ………………………………………………………………........................ Vegetation decline: ………………………………………………………………….............. Soil fertility decline: …………………………………………………………………............ Water stress: …………………………………………………………………........................ Others, specify: ………………………………………………………………….................... 10. How do you evaluate trend of land degradation over time? Now/2000 5 years ago/95 10 years ago/1990 20 years ago/1980 Severity of land degradation 1 Extent of land degradation 2 Signs of land degradation 3 1 1: light; 2: moderate; 3: severe; 4: very severe 2 1: absent; 2: present on vulnerable land units; 3: widespread everywhere 3: 1: soil erosion; 2: gully formation; 3: vegetation degradation; 4: soil fertility degradation; 5: water stress; 6: others (specify) 11. What soil and water conservation practices are present in your locality and which ones are your preferences? …………………………………………………………………………………………………. …………………………………………………………………………………………………. Trees, shrubs, animal feed, Vegetation availability and dynamics 12. List the 5 major sources of animal feed based on their order of importance & availability in season? Feed types Now 5 years ago 10 years ago 20 years ago Teff straw Wheat straw Barley straw Rice straw Sorghum stover Sorghum green leaves Maize stover Maize green leaves Millet stover Millet green leaves Aftermath grazing 114 Hay from private pastures Hay from private, improved pastures Hay from exclosures Hay from communal land Tree fodder Tree pods Grazing natural private pastures Grazing improved private pastures Grazing communal grazing land Grazing near homestead Grazing waterways, gullies Grazing wetlands Transhumance Grazing reserves for dry season 13. What is the frequency of feed shortage (frequency, period) and the most used coping strategies? Now 5 years ago 10 years ago 20 years ago frequency Critical months Particular years Coping strategies External support 14. list the major trees/shrub species found in your locality S/N Scientific name Local name Niche Major purpose status 1 2 3 4 5 6 7 8 9 10 11 115 12 (e.g. 1: construction; 2: firewood; 3: fodder; 4: shelter; 5: fruits; 6: conservation; ….) 15. What are the major factors that determine productivity of crops in your locality? ……………………………………………………………………………………………… ……………………………………………………………………………………………… ……………………………………………………………………………………………… Part V–Land use change, their drivers and impact 16. What major shift in land use occurred in your locality in the last 20 years?(provide qualitative description; +, - & No change) 5 years ago to now 5-10 years ago 10-20 years ago Area Quality Area Quality Area Quality Cropland – rainfed Cropland – irrigated Grazing land – private Grazing land – communal Natural vegetation Plantation forest Fallow Exclosure Abandoned 17. What are the major factors that affect your decision related to land use or management in order of importance (+explain)? And what is the difference in these factors between dry/wet/normal years? ………………………………………………………………………………………………… …………………………………………………………………………………………………. …………………………………………………………………………………………………. 18. Describe land lost11 or additional land gained12 during the last 20 years and associated factors? …………………………………………………………………………………………………. …………………………………………………………………………………………………. …………………………………………………………………………………………………. 19. Describe new practices & regulations that influence land management in your locality at different points in time and their impact? Last 5 years: ………………………………………………………………………………… Between 5 and 10 years ago: ……………………………………………………………. Between 10 and 20 years ago: ………………………………………………………… 11 it may include land provided to new weds, taken by government, abandoned due to degradation or others 12 it may include those gained through redistribution, conversion of hillsides, transferred from relatives 116 Other……………………………………………………………………………………………. 20. What are the major changes in land use (area + quality) and management you noted in communal properties over the last 25 years and the institutional changes that go along with these …………………………………………………………………………………………………. …………………………………………………………………………………………………. …………………………………………………………………………………………………. 21. Are there external factors that are out of your control? Describe and explain while differentiating between: Natural factors: …………………………………………………………………………………………. Demographic factors: ……………………………………………………………………………................. Institutional factors, laws: …………………………………………………………………………............. Political factors, policies: ………………………………………………………………………………………. 22. Livestock and crop production dynamics (for crops: indicate hectare or timad) over the last 25 years 23. What are your strategies during drought seasons or in case production failure? ………………………………………………………………………………………………… …………………………………………………………………………………………………. 24. Do you receive food aid and how much kilograms, if any? ………………………………………………………………………………………………… …………………………………………………………………………………………………. 25. List major production determinants in your area according to their importance? (+ how was the ranking of the determinants 5, 10 and 20 years ago) Now………………………………………………….………………………………………… 5years ago……………..……………………………………………………………………… 10 years ago…………………………………………………………………………………… S/N Item Amount owned remark Now/2000 5years ago/95 10years ago/90 20years ago/80 %chang e 1 Sheep 2 Camel 3 Cow 4 Mule 5 Goat 6 Horse 7 Ox 117 20 years ago…………………………………………………………………………………… Part IV - Institutional issues 26. Which organizations are working towards management of various land- based resources in your locality? …………………………………………………………………………………………………. …………………………………………………………………………………………………. 27. How do you evaluate the efforts made? What’s not achieved so far and what could have been done differently? …………………………………………………………………………………………………. …………………………………………………………………………………………………. …………………………………………………………………………………………………. 28. What are the most priority issues in your locality that needs intervention and please forward your suggestion to address it? …………………………………………………………………………………………………. …………………………………………………………………………………………………. …………………………………………………………………………………………………. Part VI-Miscellaneous 29. Do you have additional issues to forward pertaining points discussed? …………………………………………………………………………………………………. …………………………………………………………………………………………………. 30. Comments of the interviewed person regarding the information provided/ Special remarks of the interviewer: ………………………………………………………………………………………………… 1 1 8 APPENDIX IIA: Numeric Value of the Various Parameters Utilized for Automated Calibration of the Different Versions of Landsat Images in ENVI. District Path & raw Acquis date Satellite Sun elevation Sun Azimuth MSS Fogera 182/52 1-Feb-73 landsat 1 45.07 129.6 Gubalafto 180/52 1-Nov-72 landsat 1 52.73 132.58 TM Fogera 169/52 9-Nov-85 landsat 5 50.53 134.82 Gubalafto 168/52 5-Jan-86 landsat 5 42.53 136.2 EMT Fogera 169/52 23-Oct-99 Landsat7 57.5814 134.011 Gubalafto 168/52 5-Dec-00 Landsat7 47.913 143.8351 Radiometric gain and bias of bands Path & raw Acquis. date Satellite 1 2 3 4 5 7 gain bias gain bias gain bias gain bias gain bias gain bias MSS Fogera 182/52 1-Feb-73 landsat 1 0.9725 0 0.7843 0 0.6902 0 Gubalafto 180/52 1-Nov-72 landsat 1 0.9725 0 0.7843 0 0.6902 0 0.5229 TM Fogera 169/52 9-Nov-85 landsat 5 0.6024 -1.52 1.1751 -2.84 0.8058 -1.17 0.8145 -1.51 0.1081 -0.37 0.057 -0.15 Gubalafto 168/52 5-Jan-86 landsat 5 0.6024 -1.52 1.1751 -2.84 0.8058 -1.17 0.8145 -1.51 0.1081 -0.37 0.057 -0.15 1 1 9 APPENDIX IIB: Numerical value of different calibration parameters used for calibration of ASTER image. Aqcuisition Date Study area Sun elevation Sun Azimuth 1-Dec-05 Fogera 52.6099 152.8832 16-Oct-05 Gubalafto 62.9217 139.2751 Radiometric gain and bias values band 1 band 2 band 3N band 4 band 5 band 6 band 7 band 8 band 9 1-Dec-05 Fogera 1.688 1.415 0.862 0.2174 0.0696 0.0625 0.0597 0.0417 0.0318 16-Oct-05 Gubalafto 0.676 0.708 0.862 0.2174 0.0696 0.0625 0.0597 0.0417 0.0318 Key to gain coefficient for ASTER Images Fogera Nor Nor Nor Nor Nor Nor Nor Nor Nor Gubalafto high high Nor Nor Nor Nor Nor Nor Nor 1 2 0 ANNEX IIC: Detail calculations made to produce out put used for ASTER calibration wit in the Band math operator of ENVI Band Sun elevation angle Unit conversion Julian day d d^2 type number Fog Guba Fog Guba Fog Guba Fog Guba Fog Guba VNIR 1 52.6099 62.9217 1.688 0.676 335 289 0.9861 0.9968 0.972394 0.9937 VNIR 2 52.6099 62.9217 1.415 0.708 335 289 0.9861 0.9968 0.972394 0.9937 VNIR 3n 52.6099 62.9217 0.862 0.862 335 289 0.9861 0.9968 0.972394 0.9937 SWIR 4 52.6099 62.9217 0.2174 0.2174 335 289 0.9861 0.9968 0.972394 0.9937 SWIR 5 52.6099 62.9217 0.0696 0.0696 335 289 0.9861 0.9968 0.972394 0.9937 SWIR 6 52.6099 62.9217 0.0625 0.0625 335 289 0.9861 0.9968 0.972394 0.9937 SWIR 7 52.6099 62.9217 0.0597 0.0597 335 289 0.9861 0.9968 0.972394 0.9937 SWIR 8 52.6099 62.9217 0.0417 0.0417 335 289 0.9861 0.9968 0.972394 0.9937 SWIR 9 52.6099 62.9217 0.0318 0.0318 335 289 0.9861 0.9968 0.972394 0.9937 Band Pi*d^2 ESUNi Cos(Z) ESUNi*cos(z) RTOA type number Fog Guba (smith) Fog Guba Fog Guba Fog Guba VNIR 1 3.0549 3.1218 1845.99 0.7945 0.89 1467 1643.6 0.00208 0.0019 VNIR 2 3.0549 3.1218 1555.74 0.7945 0.89 1236 1385.2 0.00247 0.00225 VNIR 3n 3.0549 3.1218 1119.47 0.7945 0.89 889 996.76 0.00343 0.00313 SWIR 4 3.0549 3.1218 231.25 0.7945 0.89 184 205.9 0.01663 0.01516 SWIR 5 3.0549 3.1218 79.81 0.7945 0.89 63.4 71.062 0.04818 0.04393 SWIR 6 3.0549 3.1218 74.99 0.7945 0.89 59.6 66.77 0.05127 0.04675 SWIR 7 3.0549 3.1218 68.66 0.7945 0.89 54.6 61.134 0.056 0.05106 SWIR 8 3.0549 3.1218 59.74 0.7945 0.89 47.5 53.192 0.06436 0.05869 SWIR 9 3.0549 3.1218 56.92 0.7945 0.89 45.2 50.681 0.06755 0.0616 121 APPENDIX IIIA: Details of GCPs used for Georeferencing of ASTER image for Kuhar Michael Base x Base y Warp x Warp y Predict x Predict y Error x Error y RMSE 3235.62 2909.98 1990.09 2196.35 1992.25 2197.37 2.16 1.02 2.39 3184.97 2810.66 1872 2012.43 1871.76 2011.24 -0.24 -1.19 1.21 3180.03 2785.17 1859.81 1963.05 1860.22 1963.1 0.41 0.05 0.42 3141.5 2805.5 1767.33 2003 1766.22 2002.25 -1.11 -0.75 1.34 3121.31 2772.23 1717.5 1939.63 1717.49 1939.35 -0.01 -0.28 0.28 3128.41 2799.5 1736.85 1991.54 1734.48 1991.08 -2.37 -0.46 2.42 3139.41 2784.55 1761.69 1962.38 1761.41 1962.51 -0.28 0.13 0.31 3095.91 2780.5 1654.92 1955.08 1655.6 1955.42 0.68 0.34 0.76 3074.45 2775.73 1605.15 1945.85 1603.39 1946.62 -1.76 0.77 1.92 3037.73 2608.27 1513.08 1625.62 1513.77 1626.54 0.69 0.92 1.15 3032.82 2616 1501.54 1641.15 1501.72 1641.27 0.18 0.12 0.22 3023.77 2626.55 1477.77 1661.38 1479.53 1661.38 1.76 0 1.76 2998.77 2611.95 1415 1630.77 1418.06 1633.09 3.06 2.32 3.84 2981.91 2611.14 1377.62 1632.69 1376.63 1631.32 -0.99 -1.37 1.69 2961.05 2634.32 1327 1675.46 1325.66 1675.71 -1.34 0.25 1.37 2911.45 2561.5 1203.38 1535.62 1202.59 1534.44 -0.79 -1.18 1.43 2865.67 2460 1083.67 1337 1087.25 1336.49 3.58 -0.51 3.62 2856.27 2359.82 1064.38 1141.15 1061.32 1141.55 -3.06 0.4 3.08 3145.67 2695 1777.46 1792.77 1777.84 1792.51 0.38 -0.26 0.46 3208.66 2692.78 1932.85 1788.62 1931.92 1788.29 -0.93 -0.33 0.99 122 APPENDIX IIIB: Details of GCPs used for Georeferencing of ASTER image for Lenche Dima watershed Base x Base y Warp x Warp y Predict x Predict y Error x Error y RMSE 1423.67 2784.67 2574 1062 2573.15 1063.79 -0.85 1.79 1.98 1223 2930 2083 1341 2084.23 1339.36 1.23 -1.64 2.05 1169 3071 1951 1608 1952.56 1606.99 1.56 -1.01 1.86 1042 3137 1643 1733 1642.79 1732.87 -0.21 -0.13 0.25 1210 2890 2053 1262 2052.54 1263.52 -0.46 1.52 1.59 1189 2811 2001 1118 2001.37 1113.69 0.37 -4.31 4.33 1125.5 2917.5 1849 1313 1846.57 1315.83 -2.43 2.83 3.73 1133 2997 1865 1466 1864.81 1466.7 -0.19 0.7 0.72 1524 2695 2818 893 2817.4 894.17 -0.6 1.17 1.31 1550.2 2634.4 2880 779.25 2881.07 779.59 1.07 0.34 1.12 1625.33 2594.33 3064.25 705.75 3063.92 704.03 -0.33 -1.72 1.75 1649 2458 3121.33 448.33 3121.13 446.44 -0.2 -1.89 1.9 1584 2522.5 2963.33 565.67 2963.08 568.07 -0.25 2.4 2.41 1680 2673.8 3194.5 855 3197.33 854.36 2.83 -0.64 2.9 1714.57 2656.29 3281 821 3281.46 821.35 0.46 0.35 0.58 1761.67 2637.67 3395 785 3396.11 786.29 1.11 1.29 1.7 1839.2 2566.8 3584.5 653 3584.61 652.74 0.11 -0.26 0.28 1880.71 2588 3686.75 692 3685.78 692.88 -0.97 0.88 1.31 1857.48 2721.95 3631.92 947.54 3629.84 945.55 -2.08 -1.99 2.88 1906.43 2767.29 3749.5 1030.8 3749.32 1031.12 -0.18 0.32 0.37