CARBON STOCK POTENTIALS OF WOODLANDS AND LAND USE AND LAND COVER CHANGES IN NORTH WESTERN LOWLANDS OF ETHIOPIA
MSc. THESIS
BINYAM ALEMU YOSEF
HAWASSA UNIVERSITY, WONDO GENET COLLEGE OF FORESTRY AND NATURAL RESOURCES, WONDO GENET, ETHIOPIA
OCTOBER, 2012
CARBON STOCK POTENTIALS OF WOODLANDS AND LAND USE AND LAND COVER CHANGES IN NORTH WESTERN LOWLANDS OF ETHIOPIA
BINYAM ALEMU YOSEF
A THESIS SUBMITTED TO SCHOOL OF NATURAL RESOURCES AND ENVIRONMENTAL STUDIES, WONDO GENET COLLEGE OF FORESTRY AND NATURAL RESOURCES HAWASSA UNIVERSITY WONDO GENET, ETHIOPIA
IN PARTIAL FULFILMENT OF THE REQUIREMENTS FOR THE DEGREE OF MASTER OF SCIENCE IN INTEGRATED WATERSHED MANAGEMENT
OCTOBER, 2012
Approval Sheet 1
This is to certify that the thesis entitled “ Carbon Stock Potentials of Woodlands and Land
Use and Land Cover Changes in North Western Lowlands of Ethiopia” submitted in partial fulfillment of the requirement for the degree of Master of Sciences with specialization in Integrated Watershed Management of the Graduate Program of the
School of Natural Resources and Environmental Studies , Wondo Genet College of
Forestry and Natural Resources, is a record of original research carried out by Binyam
Alemu Yosef Id. N o. MSC/IWM/033/10 , under my supervision; and no part of the thesis has been submitted for any other degree or diploma.
The assistance and help received during the courses of this investigation have been duly acknowledged. Therefore I recommended that it be accepted as fulfilling the thesis requirement.
Dr.Zewdu Eshetu
Name of major advisor Signature Date
OR
Dr. Efrem Garedew Name of co-advisor Signature Date
Approval Sheet 2
We, the undersigned, members of the Board of examiners of the final open defense by
Binyam Alemu have read and evaluated his thesis entitled “Carbon Stock Potentials of
Woodlands and Land Use and Land Cover Changes in North Western Lowlands of
Ethiopia ” and examined the candidate. This is therefore to certify that the thesis has been accepted in partial fulfillment of the requirements for the degree of Master of Science.
Dr. Melaku B.
Name of the Chairperson Signature Date
Dr. Zewdu E.
Name of Major Advisor Signature Date
Dr. Tsegaye B.
Name of Internal Examiner Signature Date
Dr. Tesfaye B.
Name of External Examiner Signature Date
Acknowledgements
I would like to express my sincere thanks to my main supervisors Dr. Zewdu Eshetu,
Addis Ababa University, (AAU) for his regular support, gentleness, keen support and close follow up while I am conducting this research from the very beginning of the development of research proposal to the accomplishment of the report. I would also like to acknowledge my co-supervisor Dr. Efrem Garedew, Wondo Genet College of Forestry and Natural
Resources (WGCF and NRs), for his critical comments and ideas with inclusive support in the structure and arrangement of the thesis work at various levels.
My special thanks go to Ato Zewdu Yilma and Temesgen Yohannes (FRC, Addis Ababa) for their technical advice and teaching during the field data collection and Ato Ayenachew
Dagnie for his nice drive to several field sites. I would like also to thank Kafta-Humer,
Metema and Sherkole districts Office of Agriculture for the permission to work in the field. I am grateful to Natural Gum Processing and Marketing Enterprise (NGPME) for its interest in my work and providing valuable data sets relevant for this research. I also acknowledge the support of Forestry Research Centre (FRC) and National Soil Laboratory for determining soil bulk density and soil carbon content, respectively. I also thank Mr.
Tezera and Mulye for their assistance in providing and interpreting the GIS and satellite images, that are valuable components of this thesis. The financial support for this research is obtained from Center for International Forestry Research (CIFOR); and the scientific critique I obtained from Dr. Habtemariam Kassa is highly appreciated. I am very grateful to Wollo University for providing me MSc scholarship to study in Hawassa University.
Finally, my special deepest gratitude goes to my mother and my family for their inspiration, love and support throughout my life.
i
Table of Contents
Acknowledgements ...... i
Dedication ...... vi
List of Tables ...... vii
List of Figures ...... viii
List of Tables in Appendices ...... ix
List of Abbreviations ...... x
Abstract ...... xi
1. INTRODUCTION ...... 1
1.1. Background ...... 1
1.2. Statement of the Problem ...... 3
1.3. Significance of the Study ...... 4
1.4. Objective ...... 5
1.4.1. General Objective ...... 5
1.4.2. Specific Objective ...... 5
1.5. Research Questions ...... 6
2. LITERATURE REVIEW ...... 7
2.1. Carbon Stock Pools ...... 7
2.1.1. Aboveground Biomass Carbon Stock ...... 7
2.1.2. Root Biomass Carbon Stock ...... 9
2.1.3. Dead Wood Biomass Carbon Stock...... 9
2.1.4. Litter Carbon Stock ...... 10
2.1.5. Soil Carbon Stock ...... 11
2.2. The Role of Forest Carbon Stock on Climate Change Mitigation ...... 12
2.3. Gum and Resin Production in Ethiopia ...... 14 ii
2.4. Land Use and Land Covers Change Analysis ...... 15
2.4.1. Image Classification ...... 16
2.4.1.1. Unsupervised Classification ...... 16
2.4.1.2. Supervised Classification ...... 16
2.4.2. Change Detection Methods ...... 17
2.5. Trading Carbon Credits and its Potential to Improve Livelihoods ...... 17
3. MATERIALS AND METHODS ...... 19
3.1. Study Areas Description ...... 19
3.1.1. Kafta Humera District ...... 19
3.1.2. Metema District ...... 21
3.1.3. Sherkole District ...... 22
3.2. Sampling and Data Collection Methods ...... 23
3.2.1. Study Site Selection ...... 23
3.2.2. Tree Sampling ...... 25
3.2.3. Dead Wood Sampling ...... 25
3.2.4. Herb, Litter and Crop Biomass Sampling ...... 25
3.2.5. Soil Sampling ...... 26
3.3. Data Analysis ...... 27
3.3.1. Carbon Stock Estimation ...... 27
3.3.1.1. Aboveground and Belowground Biomass ...... 27
3.3.1.2. Dead Wood Biomass ...... 28
3.3.1.3. Herb, Litter, Annual Crop and Crop Root Biomass Estimation ...... 30
3.3.1.4. Soil Organic Carbon ...... 31
3.3.1.5. The Total Carbon Stock Density Estimation ...... 32
3.3.2. Land Use and Land Cover Change Analysis ...... 32
3.3.2.1. Remote Sensing Data Acquisition ...... 32 iii
3.3.2.2. Image Analysis ...... 33
3.3.2.2.1. Image Pre-processing ...... 33
3.3.2.2.2. Image Classification Analysis ...... 33
3.3.2.2.3. Accuracy Assessment ...... 34
3.3.2.2.4. Matrix of Land Use and Land Cover changes ...... 35
3.3.3. Statistical Analysis of the Various Carbon Stocks ...... 37
4. RESULTS ...... 38
4.1. Vegetation Characteristics ...... 38
4.2. Carbon Stocks in Different Carbon Pools ...... 39
4.2.1. Aboveground Biomass Carbon Stock ...... 39
4.2.2. Belowground Root Biomass Carbon Stock ...... 40
4.2.3. Dead Wood Biomass Carbon Stock...... 41
4.2.4. Herb, Annual Crop and Litter Carbon Stock ...... 42
4.2.5. Soil Carbon Stock ...... 43
4.2.6. Total Carbon Stock Density ...... 47
4.3. Yield of Gum and Resin ...... 47
4.4. Land Use and Land Cover Changes ...... 49
4.4.1. Land Use and Land Cover Mapping ...... 49
4.4.1.1. Land Use and Land Cover of Kafta Humera District ...... 50
4.4.1.1.1. Land Use and Land Cover Change Matrices of Kafta Humera District . 52
4.4.1.1.2. Rate of Land Use and Land Cover Change in Kafta Humera District .... 54
4.4.1.1.3. Accuracy Assessment of Kafta Humera District ...... 54
4.4.1.2. Land Use and Land Cover of Metema District ...... 55
4.4.1.2.1. Land Use and Land Cover Change Matrices of Metema District ...... 57
4.4.1.2.2. Rate of Land Use and Land Cover Change in Metema District ...... 59
4.4.1.2.3. Accuracy Assessment of Metema District ...... 60 iv
4.4.1.3. Land Use and Land Cover of Sherkole District ...... 60
4.4.1.3.1. Land Use and Land Cover Change Matrices of Sherkole District ...... 62
4.4.1.3.2. Rate of Land Use and Land Cover Change in Sherkole District ...... 64
4.4.1.3.3. Accuracy Assessment of Sherkole District ...... 65
5. DISCUSSIONS ...... 66
5.1. Vegetation Characterstics ...... 66
5.2. Carbon Stocks in Different Carbon Pools ...... 67
5.2.1. Aboveground Carbon Stock...... 67
5.2.2. Belowground Carbon Stock ...... 68
5.2.3. Dead Wood Carbon Stock ...... 69
5.2.4. Herbaceous and Litter Carbon Stock ...... 69
5.2.5. Soil Carbon Stock ...... 70
5.2.6. Total Carbon Stock Density ...... 72
5.3. Value and Functions of Gum and Resin ...... 73
5.4. Land Use and Land Cover Change ...... 74
5.4.1. The Causes of Wood Land Cover Change...... 76
6. CONCLUSIONS AND RECOMMENDATIONS...... 79
6.1. Conclusions ...... 79
6.2. Recommendations ...... 80
REFERENCES ...... 82
APPENDICES ...... 98
BIOGRAPHICAL SKETCH ...... 114
v
Dedication
“To my Mother Mulu Kidane and My Family, for their inspiration, love and support throughout my life.”
vi
List of Tables
Table 1: Landsat Data Used in Land Use and Land Cover Classification ...... 32
Table 2: Yield and Income of Gum Olibanum and Gum Arabic...... 48
Table 3: Description of Land Use and Land Cover Types Identified ...... 49
Table 4: Areas of LULC of Kafta Humera District for the Years 1985, 1995 and 2010 .... 50
Table 5: LULCC Matrices of the Kafta Humera District (1985-1995) ...... 52
Table 6: LULCC Matrices of Kafta Humera District (1995-2010) ...... 53
Table 7: Rate of Changes in LULC Classes (1985-2010) ...... 54
Table 8: Areas of LULC of Metema District for the Years 1985, 1995 and 2010 ...... 55
Table 9: LULCC Matrices of Metema District (1985-1995) ...... 57
Table 10: LULCC Matrices of Metema District (1995-2010) ...... 58
Table 11: Rate of Changes in LULC Classes (1985-2010) ...... 59
Table 12: Areas of LULC of Sherkole District for the Years 1985, 1995 and 2010 ...... 62
Table 13: LCLCC Matrices of Sherkole District (1985-1995) ...... 62
Table 14: LULCC Matrices of Sherkole District (1995-2010)...... 63
Table 15: Rate of Changes in LULC Classes (1985-2010) ...... 65
vii
List of Figures
Figure 1: Map of the Study Sites (Kafta Humera, Metema and Sherkole Districts) ...... 23
Figure 2: Plot Size and Quadrants for Tree, Soil, Herbs, Crop and Litter Sampling ...... 26
Figure 3: The General Framework for the Study of LULCC ...... 36
Figure 4: Diameter Size Class Distribution of the Entire Woodlands ...... 38
Figure 5: Mean Carbon Stock of Different Carbon Pools of Woodland in Adi-Goshu ...... 44
Figure 6: Mean Carbon Stock of Different Carbon Pools of Farmland in Adi-Goshu ...... 44
Figure 7: Mean Carbon Stock of Different Carbon Pools in Lemlem-Terara ...... 45
Figure 8: Mean Carbon Stock of Different Carbon Pools in Gemed...... 45
Figure 9: Mean Carbon Stock of Different Carbon Pools in Untapped Stratum ...... 46
Figure 10: Mean Carbon Stock of Different Carbon Pools in Tapped Stratum ...... 46
Figure 11: Total Carbon Stock Density of Different stratum ...... 47
Figure 12: LULC Map of Kafta Humera District for the Years 1985, 1995 and 2010 ...... 51
Figure 13: LULCC (1985-1995) of Kafta Humera District ...... 52
Figure 14: LULCC (1995-2010) of Kafta Humera District ...... 53
Figure 15: LULC Map of Metema District for the Years 1985, 1995 and 2010 ...... 56
Figure 16: LULCC (1985-1995) of Metema District ...... 57
Figure 17: LULCC (1995-2010) of Metema District ...... 58
Figure 18: LULC Map of Sherkole District for the Years 1985, 1995 and 2010 ...... 61
Figure 19: LULCC (1985-1995) of Sherkole District ...... 63
Figure 20: LULCC (1995-2010) of Sherkole District ...... 64
viii
List of Tables in Appendices
Appendix 1: Height and DBH Ranges of the Vegetation Characterstics ...... 98
Appendix 2: Mean Carbon Stock of the Three Sites (Mean ± SE) ...... 99
Appendix 3: Mean Carbon Stock of Adi Goshu Farm Land (Mean ± SE) ...... 100
Appendix 4: Adi Goshu Woodland ANOVA Results ...... 100
Appendix 5: Adi Goshu Farm Land ANOVA Results ...... 101
Appendix 6: Lemlem Terara Woodland ANOVA Results ...... 103
Appendix 7: Overall Untapped Boswellia papyrifera Woodlands ANOVA Results ...... 105
Appendix 8: Overall Tapped Boswellia papyrifera Woodlands ANOVA Results ...... 106
Appendix 9: The Overall Comparison of Models ...... 107
Appendix 10: Biomass and Carbon Stock Estimation Using WBISPP (2000) ...... 108
Appendix 11: Accuracy Assessment of Kafta Humera ...... 108
Appendix 12: Accuracy Assessment of Metema ...... 109
Appendix 13: Accuracy Assessment of Sherkole ...... 110
Appendix 14: Tree Data Collection Form ...... 112
Appendix 15: Herb, Litter and Soil Sample Form ...... 113
Appendix 16: Dead Wood Data Collection Form ...... 113
ix
List of Abbreviations
CDM Clean Development Mechanism
CSA Central Statistical Agency of Ethiopia
ERDAS Earth Resources Data Analysis System
GHG Green House Gas
GIS Geographic Information System
HBC Herbaceous Biomass Carbon
KHDOA Kafta Humera District Office of Agriculture
LULCC Land Use and Land Covers Change
MDOA Metema District Office of Agriculture
Mg Mega Gram
NGPME Natural Gum Processing and Marketing Enterprise
REDD Reducing Emissions from Deforestation and Forest Degradation
RS Remote Sensing
SAS Statistical Analysis System
SOC Soil Organic Carbon
TU Tapped Boswellia Papyrifera Woodland
UW Untapped Boswellia Papyrifera Woodland
x
Carbon Stock Potentials of Woodlands and Land Use and Land Cover Changes in North Western Lowlands of Ethiopia
Binyam Alemu Yosef Mobile phone: +251-9-12770997 E-mail: [email protected] Abstract A major problem being faced by human society is the rising of global temperature mainly due to human activity that emit carbon dioxide to the atmosphere. The problem of increasing atmospheric carbon dioxide can be addressed in a number of ways. One of such actions is forestry development and forest management undertakings. Sustainable forest management and development is believed to be an asset for increasing societal adaptive capacity to climate anomalies. This paper examined the potential of the dry western woodlands of Ethiopia for carbon sequestration in response to woodland cover changes. The study was based on the assumption that increasing societal adaptive capacity is possible through asset building from financial earnings obtained from carbon trading and non-timber forest products, especially gum and resin from Boswellia papyrifera woodland. GIS and RS were used to determine the LULCC. To estimate the amount of carbon stocked in dry land forests; vegetation inventory including dead wood, litter and herbaceous biomass collection were conducted in the 36 sample plots across the three districts of the study area namely: Kafta-Humera 17 plot, Metema 9 plot and Sherkole 10 plot. The sample plots were taken in transects line method in the two categories of woodlands, untapped and tapped Boswellia papyrifera. A total of 24 species were recorded. The soil samples were taken from 0-30 cm soil depth to determine the potential of soil carbon sequestration. To analyze the total woodland carbon stock, allometry equations were used to determine the aboveground, belowground and dead woods biomasses; litter and herbaceous biomasses were determined using direct harvesting method; and the SOC was estimated using standard methods. The result showed, the estimated mean carbon stocks of the aboveground, belowground and the dead wood biomasses for the UW in the Lemlem Terara site were significantly higher (P < 0.05) than that of the Adi Goshu site. In the Gemed site, the mean HBC stock was 1.2 Mg ha -1, which is significantly highest (P=0.0207) than the other two study sites (Lemlem Terara, 0.42 Mg ha -1 and Adi Goshu, 0.45 Mg ha -1) for the TW. In UW, the mean soil carbon stock of the Lemlem Terara site (58.19 Mg ha -1) was significantly (P=0.0019) higher than that of Adi Goshu (33.61 Mg ha -1). However, no statistical variation (P=0.8884) was observed between the mean soil carbon stocks across sites in the tapped stratum. In the case of the total carbon stocks in UW stratum, for the Adi-Goshu site the carbon stock was estimated to be about 55.26 Mg ha -1 while 96.74 Mg ha -1 for Lemlem Terara. In the TW stratum, however, the total carbon of Adi-Goshu, Lemlem-Terara and Gemed sites were 65.93, 68.77 and 71.01 Mg ha -1 respectively. The results of LULCC analysis showed that in all study sites the classes of agricultural and bare land have been increased at an average rate of 2,322.94 and 726.58 ha/year, respectively; while the woodland coverage in the three district was decreasing at an average rate of 2,833.77 ha/year during the last 25 years (1985-2010). The woodland coverage was converted mainly to agriculture at an average rate of 2,057.9 ha/year. Despite the rapid decline in the woodland coverage, the existing wood land has a huge potential for carbon sequestration, which calls for the promotion of sustainable woodland management in this climate sensitive areas. From the view of points of woodland management in a sustainable manner, the study suggested that the NGPME and the regional government should made fundamental thinking in the policy of woodland management in such a way of promoting carbon trading for additional financial incentive to the local community who are depending on the woodland resource. The data obtained from the current study can be used as a baseline data set of carbon stock to make inferences about the carbon stocking in the areas where the study was conducted. Keywords: woodlands, carbon stock, soil organic carbon, Boswellia papyrifera, land use and land cover change.
xi
1. INTRODUCTION
1.1. Background
Ethiopia is endowed with different vegetation cover in dry land areas. Combretum-
Terminalia woodlands and Acacia-Commiphora woodlands are the two dominant vegetation types that cover large parts of the dry land areas in Ethiopia (Eshete et al .,
2011). Forests provide a great variety of products and services to human kind. The major economic value of forests comes from wood of trees, used or traded as lumber, plywood, fuel wood or charcoal. Other economic importance includes food, medicines, fodder for livestock, natural gums, etc. The latter collectively called non-timber forest products
(NTFPs) (Ros-Tonen et al ., 1995). The dry land woodlands in Ethiopia possess diverse tree species that are known for their valuable Non-Timber Forest Products (NTFPs) of local, national and international significance. One of the well-known species in this regard is Boswellia papyrifera . The species is a deciduous multipurpose tree with the potential for economic development and desertification control (Lemenih and Teketay, 2003, 2004).
This species is found in the Combretum-Terminalia (broad-leaved) deciduous woodland and wooded grassland usually dominant on steep rocky slopes, lava flows or sandy valleys, within the altitudinal range of 950-1800 m a.s.l. altitude (Eshete et al ., 2005).
In Ethiopia, Boswellia papyrifera provides the widely known and traded frankincense that accounts more than 80% of the export revenues that the country is earning from gum and resin resources (Eshete et al ., 2011; Eshete et al ., 2012). Ethiopia is the leading producer and exporter of frankincense and with significant local and national economic benefits
(Eshete et al ., 2012). Boswellia papyrifera provides several goods and services such as poles, timber, fodder, nectar and gum, which is useful for traditional medicine, religious ritual and income generation (Lemenih and Teketay, 2003).
1
In addition to the aforementioned benefits the presence of woodlands in the dry lands of
Ethiopia could serve as a sink for atmospheric CO 2 and have potential contributions to climate change mitigation and adaptation locally as well as globally. Globally, forest ecosystems account for approximately 90% of the annual carbon flux between the atmosphere and terrestrial ecosystems (Dixon et al ., 1994). Carbon storage in forest ecosystems involves numerous components including above-ground and below-ground biomass, deadwood, litter and soil carbon.
Forests are relevant to climate change mitigation through their potentials in mitigation
GHGs, particularly carbon sequestration. Although there is no doubt growing trees function as an active carbon sink, large emissions from dead organic matter and soil would count as a reduction in the amount of sequestrated carbon (Takahashi et al. , 2010).
However, dead organic matter and soil carbon stock are influenced by vegetation, site conditions and forest management practices. More biomass results in increased production of aboveground litter and belowground root activities. Some research undertaking also indicated that by adding trees in grassland or pasture systems, the SOC content can be increased considerably (Reyes-Reyes et al ., 2002; Yelenik et al ., 2004). Therefore, managing the woodland ecosystems is a cost-effective carbon storing/sequestering effort towards absorbing carbon dioxide from the atmosphere (Smith et al ., 1993; Dixon et al .,
1994; Lal, 2005) and plays an important role in global carbon cycle.
Related to the effects of human activity in terrestrial ecosystems, land-use category is a key factor for determining the equilibrium level of carbon stock in the soil (Paul et al ., 2002).
In Ethiopia, more than half of the country’s land area is located in such dry areas and associated tropical dry forest (NCSS, 1993). The biomass carbon densities expressed as mass per unit area (Mg/ha) for different forest types were one of the important components
2
for assessing the contribution of forestlands to the global carbon cycle (Haripriya, 2002).
In this respect, the local farmers or forest dependent communities should be eligible for payments from carbon credits/market for managing the woodland resources, and this will be strong additional incentive for promoting sustainable woodland development and management in dry lands of Ethiopia. Therefore, this study was undertaken to estimate carbon stock in the north western woodlands of Ethiopia using direct field measurement, laboratory analyses, RS and GIS techniques in relation to historical woodland cover changes.
Furthermore, remote sensing provides a reliable source of data of land use and land cover change that can be extracted efficiently and cheaply in order to assess and monitor spatial and temporal changes in land use and land cover. Thus change detection has become a major application of remotely sensed data because of the repetitive coverage at short intervals and consistent image quality. Land use and land cover map can be a powerful tool to compare the changes of an area over time. It is impossible to cover a large area in short time through manual survey but with remote sensing (land use and land cover map), it is relatively an easier task (Billah and Rahman, 2004).
1.2. Statement of the Problem
In Ethiopia, woodlands are covering large areas and their carbon stock is much higher than high forests which are 1,263.13 million tons of carbon per 29.55 million hectare in woodland and 434.19 million tons of carbon per 4.07 million hectare in the high forest
(Moges et al ., 2010). But this resource as an alternative source of income for household is not yet well quantified. Woodlands are under heavy pressure: they are cleared for fire wood, expansion of cash crops and new settlements and apparently are shrinking overtime.
Household income from various woodland products (gum and resin, fire wood, etc) are not
3
compared through cost benefit analyses/opportunity costs. Additionally, dry land plantation development by introducing locally adaptive species to buffer the gum and resin bearing woodland species is not well studied and practiced. Revenues from carbon sequestration also have never been utilized for dry lands.
Yet much effort on promoting the diversity of livelihoods through forest carbon development is given to land/forest rehabilitation and reduction of deforestation and degradation in the highland areas e.g. in the South-Eastern Ethiopian highlands, the Bale
Eco-Region REDD project covers 700,000 hectares of forest over fourteen districts
(BERSMP, 2010). Thus, there is a need in extending this experience to the dry land areas so that the livelihoods of the woodland community could be diversified and climate resilient by managing these resources for their multiple uses which amongst others are carbon sequestration, gum and resin production, and other associated values. This led to propose and conduct this study in three selected western low land sites that are owned under concession by the Natural Gum Processing and Marketing Enterprise.
In addition, it is important to assess the status of the wood land area coverage using remote sensing data helps to analyze the land use and land cover change in the study area. Because as in other areas of Ethiopia, wood land exploitation is largely unregulated with communities rapidly deforesting to meet their livelihood needs and the new settlers
(BERSMP, 2010).
1.3. Significance of the Study
Currently, the importance of forest ecosystems are consideration in the context of climate change mitigation because they can act as the sinks of CO 2 (Lyngbaek et al. , 2001). If the woodlands is to be used in carbon sequestration schemes such as the CDM/REDD, it is a better options in developing countries for the dual objective of reducing greenhouse gas
4
emissions and contributing to sustainable woodland development and management
(Roshetko et al. , 2002). As a result, carbon determination may provide clear indications of the possibilities of promoting dry woodland development and management for climate change mitigation through soil and vegetation carbon sequestration and opportunities for economic benefit through carbon trading to farmers (Lal, 2009). In addition to this the woodlands have non-timber products (Gum and Resin) and the local farmers can benefit as an additional income. Moreover, a combined use of RS/GIS technology can be invaluable to address a wide variety of resource management problems including land use and land cover changes (Tekle and Hedlund, 2000). The study provide insights into the rate of woodland resource changes, associated changes in carbon stocks, potential income generation from carbon trading and Gum and Resin marketing, that would assist in developing sustainable dry land forest management towards building climate resilient livelihoods in dry land areas of Ethiopia.
1.4. Objective 1.4.1. General Objective
The general objective of this study is to investigate the potential contribution of carbon trading to livelihoods options and promoting woodland development and management in the dry land areas of Ethiopia.
1.4.2. Specific Objective
• To estimate the total carbon stock density of the study areas
• To assess the production and income of gum and resin products and the widely
open opportunity to local farmers
• To quantify the trends of land use and land cover changes of the three study sites
using different dating remote sensing data
• Make recommendations on sustainable woodland management and development
5
1.5. Research Questions
• How much C is stored in the western woodlands of Ethiopia?
• What is the contribution of non-wood products to households?
• What is the magnitude of woodland cover changes?
• What driving forces are contributing to the woodland cover changes?
6
2. LITERATURE REVIEW
2.1. Carbon Stock Pools
2.1.1. Aboveground Biomass Carbon Stock
Carbon sequestration can be defined as the removal of CO 2 from the atmosphere and store into green plant biomass (sink) where it can be stored indefinitely through the process of photosynthesis (Watson et al ., 2000). These sinks can be above ground biomass (trees), living biomass below the ground in the soil (roots and micro organisms) or in the deeper sub-surface environments (Nair et al ., 2009). Forests are major contributors to terrestrial carbon sink, mitigating climate change and associated economic benefits (Waston et al .,
2000; FAO, 2005; Sheikh et al ., 2009). As a leading tree based system, especially in the tropics, agroforestry, afforestation and reforestation has been suggested as one of the most appropriate land management systems for mitigating the atmospheric carbon increase
(Dixon, 1995; Albrecht and Kandji, 2003; Montagnini and Nair, 2004). The estimation of the total global carbon sequestration potential for afforestation and reforestation activities for the period 1995-2050 was between 1.1-1.6 Gt carbon per year and of which 70% will be in the tropics (IPCC, 2000). Even though the climate protection role of forests is apparent, it is complex to determine how much of the forest carbon sink and reservoir can be managed to mitigate atmospheric CO 2 and in what way to buildup. Four major strategies are available to mitigate carbon emissions through forestry activities: (i) increase forest land area through reforestation and afforestaton, (ii) increase the carbon density of existing forests at both stand and landscape scales, (iii) expand the use of forest products that sustainably replace fossil-fuel, and (iv) reduce emissions from deforestation and degradation (Canadell and Raupach, 2008).
7
Deforestation and the burning of forests release CO2 to the atmosphere. Indeed, land use and land cover change especially deforestation is responsible for about 25% of all greenhouse emissions to the atmosphere (Waston et al ., 2000). On the other hand, forest ecosystems could also help reduce greenhouse gas concentrations by absorbing carbon from the atmosphere through the process called photosynthesis. Of all the global forests, tropical forests have the greatest potential to sequester carbon primarily through reforestation, agroforestry and conservation of existing forests (Brown et al ., 1996).
Forests are also producing renewable materials in order to substitute fossil fuel (Watson et al ., 2000).
Forests operate both as vehicles for capturing additional carbon and as carbon reservoirs. A young forest, when growing rapidly, can sequester relatively large volumes of additional carbon roughly proportional to the forest’s growth in biomass. An old-growth forest acts as a reservoir, holding large volumes of carbon even if it is not experiencing net growth.
Thus, a young forest holds less carbon, but it is sequestering additional carbon over time.
An old forest may not be capturing any new carbon but can continue to hold large volumes of carbon in its biomass over long periods of time. Managed forests offer the opportunity for influencing forest growth rates and providing for full stocking, both allow for more carbon sequestration. Forest management for carbon sequestration would have associated with it a relative increase in stock of carbon held captive in the forest ecosystem over what would have occurred in the absence of such focused management. Increases in the stock of carbon could be accomplished as the result of an increase in the forest biomass and as a result of an increase in forest soil carbon directly. Finally, if the stock of long-lived wood products increases, the carbon held captive in wood products stock would also increase
(Sedjo, 2001).
8
2.1.2. Root Biomass Carbon Stock
Roots are an important part of the carbon balance, because they transfer large amounts of carbon into the soil. More than half of the carbon assimilated by the plant is eventually transported below-ground via root growth and turnover, root exudates (of organic substances) and litter deposition. Depending on rooting depth, a considerable amount of carbon is stored below the plow layer and better protected from disturbances, which leads to longer residence times in the soil. With some trees having rooting depths of greater than
60 m, root carbon inputs can be substantial, although the amount declines sharply with soil depth (Cairns et al ., 1997). Root biomass in ecosystems is often estimated from root-to- shoot ratios. The ratio ranges from 0.18 to 0.30, with tropical forests in the lower range and the temperate and boreal forests in the higher range (Cairns et al ., 1997).
Roots make a significant contribution to SOC (Strand et al ., 2008). About 50% of the carbon fixed in photosynthesis is transported belowground and partitioned among root growth, rhizosphere respiration, and assimilation to soil organic matter (Lynch and
Whipps, 1990; Nguyen, 2003). Roots help in accumulation of SOC by their decomposition and supply carbon to soil through the process known as rhizodeposition (Rees et al ., 2005;
Weintraub et al ., 2007). Increased production and turnover rates of roots lead to increased
SOC accumulation following root decomposition (Matamala et al ., 2003).
2.1.3. Dead Wood Biomass Carbon Stock
Dead organic matter is composed of litter and dead-wood and generally divided into course and fine, with the breakpoint set at 10 cm diameter (Harmon and Sexton, 1996; Takahashi et al. , 2010). Although logged dead wood, standing and lie down on the ground, is often a significant component of forest ecosystems, often accounting for 10-20% of the aboveground biomass in mature forests but it tends to be ignored in many forest carbon
9
budgets (Delaney et al ., 1998). The quantity of dead wood does not generally correlate with any index of stand structure (Harmon and Sexton, 1996). The primary method for assessing carbon stock in the dead wood pool is to sample and assess the wet-to-dry weight ratio, with the large pieces of dead wood measured volumetrically as cylinders and converted to biomass on the basis of wood density, and standing trees measured as live trees but adjusted for losses in branches (<20%) and leaves (<2-3%) (MacDicken,
1997).
Dead trees serve many key functions in the ecosystems (Franklin et al ., 1987). Since dead trees may persist for centuries, they can influence ecosystems as long as living trees.
Woody detritus reduces erosion, they are a major source of energy and nutrients, serves as a seedbed for plants and they are a major habitat for microbes, invertebrates and vertebrates (Harmon et al . 1986).
2.1.4. Litter Carbon Stock
Carbon is stored in trees (stem, branches, leaves and root), understory, forest litter and forest soils. The mechanism of species driven carbon sequestration in soil is influenced by two major activities, aboveground litter decomposition and belowground root activity
(Lemma et al ., 2007). Litter decomposition is one of the major sources of SOC and the quality of litter is very important in this regard (Mafongoya et al ., 1998; Issac and Nair,
2006; Lemma et al ., 2007). In the systems with high plant diversity, litters are present with different degrees of chemical resistance, creating the possibility of longer residence of carbon through slower decomposition of litters from some species. Lignin in litter is highly resistant to decomposition and therefore, litter with high lignin content would have slower decomposition rate (Mafongoya et al ., 1998). In contrast, litter with low lignin, phenols, and high nitrogen content would have faster rate of decomposition.
10
2.1.5. Soil Carbon Stock
The term ‘‘soil carbon sequestration’’ implies the removal of atmospheric CO 2 by plants and storage of fixed carbon as soil organic matter. The strategy is to increase SOC density in the soil, improve depth distribution of SOC and stabilize SOC within stable micro aggregates, so that carbon is protected from microbial processes or as recalcitrant carbon with long turnover time. Soil carbon sequestration also increases SOC stocks through judicious land use and recommended management practices. The potential of soil carbon sink capacity in managed ecosystems is approximately equals to the cumulative historic carbon loss estimated. The attainable soil carbon sink capacity is only 50-66% to the potential capacity. The strategy of soil carbon sequestration is cost-effective and environmentally friendly (Lal, 2004).
Soils are the largest carbon reservoirs of the terrestrial carbon cycle, 1500–1550 Gt, of organic soil carbon and soil inorganic carbon approximately 750 Gt both to 1 m depth.
About three times more carbon is contained in soils than in the global vegetation (560 Gt) and soils hold double the amount of carbon that is present in the atmosphere (720 Gt) (Post et al ., 2001; Lal, 2004). Soils play a key role in the global carbon budget and greenhouse gas effect. Soils contain 3.5% of the earth's carbon reserves, compared with 1.7% in the atmosphere, 8.9% in fossil fuels, 1.0% in biota and 84.9% in the oceans (Lal, 2004).
The Soil Science Society of America recognizes that carbon is sequestered in the soils directly and indirectly (SSSA, 2001). Direct soil carbon sequestration occurs by inorganic chemical reactions that convert CO 2 into soil inorganic carbon compounds such as calcium and magnesium carbonates. Indirect plant carbon sequestration occurs as plants photosynthesize atmospheric CO 2 into plant biomass. Some of this plant biomass is indirectly sequestered as SOC during decomposition processes. The amount of carbon
11
sequestered at a site reflects the long-term balance between carbon uptake and release mechanisms. Because those flux rates are large, changes such as shifts in land use and land cover practices that affect pools and fluxes of SOC have large implications for the carbon cycle and the earth’s climate system (Lal and Bruce, 1999; Lal, 2008).
Forest soils are one of the major carbon sinks on earth, because of their higher organic matter content. Soils can act as sinks or as a source for carbon in the atmosphere depending on the changes happening to soil organic matter. Equilibrium between the rate of decomposition and rate of supply of organic matter is disturbed when forests are cleared and land use and land cover is changed (Lal, 2004). Soil organic matter can also increase or decrease depending on numerous factors, including climate, vegetation type, nutrient availability, disturbance, and land use and management practice. About 75% of the total terrestrial carbon is stored in the global soils and 40% of it resides in forest ecosystem
(Dixon et al. , 1994; Six and Jastrow, 2002; Baker, 2007).
2.2. The Role of Forest Carbon Stock on Climate Change Mitigation
Forest ecosystems are the largest terrestrial ecosystem comprised 4.1 billion ha (Dixon et al ., 1994; Brown et al ., 2002) and are critical in reducing the rate of CO 2 build-up in the atmosphere responsible for climate change (Streck and Scholz, 2006). Forests account for
80%-90% of the total global carbon reservoir in the living biomass (Dixon et al ., 1994), cover 30%-40% of the vegetated area of the earth and exchange carbon with the atmosphere through photosynthesis and respiration (Malhi et al ., 1999), thus playing an important role in the global carbon cycle. Forest ecosystems accumulate carbon through the photosynthetic assimilation of atmospheric CO 2 and the subsequent storage in the form of biomass (trunks, branches, foliage, roots, etc.) (Brown et al ., 1996; Malhi et al ., 2002;
12
Houghton, 2005), litter, woody debris, soil organic matter and forest products (Malhi et al .,
2002), and organic carbon in the soil (Houghton, 2005).
Carbon cycle and budget have been intensively studied to quantify different processes, with a view to mitigate climate change (Tupek et al ., 2010). Forest carbon cycle is an important part of global carbon cycle and a primary focus of climate change mitigation policy (White et al ., 2005). So, it is necessary to estimate forest carbon cycle and carbon budget. Land use and land cover change and forestry is one of the sector for which a national inventory of sources and sinks of GHGs must be developed. With reference to forests, the inventory must include estimates of carbon emissions and removals caused by changes in forest biomass stocks due to forest management, harvesting, plantation establishment, abandonment of lands that are re-grow to forests, and forest conversion to non-forest use (Brown, 2002).
The carbon balance of a forest ecosystem [net ecosystem production (NEP)] is the net result of carbon acquisition through photosynthesis and carbon losses through autotrophic and heterotrophic respiration (Malhi et al ., 1999). In other words, whether a forest ecosystem is a carbon sink or source depends on the balance of photosynthetic uptake and respiratory release of CO 2 (Malhi et al ., 1999; Janisch and Harmon, 2002). The NEP is an important indicator for estimating carbon sink or source in terrestrial ecosystems and is influenced by land use and management through a variety of anthropogenic actions such as deforestation, afforestation, fertilization, irrigation, harvest, and species choice (IPCC,
2005). Disturbances (e.g., harvesting, conversion to non-forest uses, wildfires, etc.), can convert a forest from a sink to a source for atmospheric carbon when NEP and net biome production (NBP) become negative (Janisch and Harmon, 2002). On the other hand, an
13
area can become a carbon sink if the forest is allowed to regenerate after a disturbance when NEP and NBP become positive (Brown et al ., 1996).
2.3. Gum and Resin Production in Ethiopia
Frankincense or olibanum is an oleo-gum resin obtained by tapping trees and shrubs of the genus Boswellia , family Burseraceae. This resin has been in use since at least 1700 BC
(Howes, 1950), and is still widely used for incense burning, for example in church ceremonies worldwide (Coppen, 1995; Demissew, 1996). Internationally it has and important industrial uses, including pharmaceutical, perfumery and food flavoring uses. In north-east Africa, local communities use large quantities of resin for traditional medicine, in coffee ceremonies and for chewing (Coppen, 1995; Lemenih and Teketay, 2003;
Lemenih et al ., 2003).
At present, about half of the 20 Boswellia species known for their frankincense yield,
Boswellia papyrifera is one of them (Tucker, 1986; Coppen, 1995). It is claimed that this species is the true source of ancient frankincense (Tucker, 1986). This species is a deciduous, multipurpose tree and the largest and longest frankincense tree reaching a height of up to 12 m. The species also provides wood for timber, leaves for fodder and flowers for bee keeping purposes (Gebrehiwot et al ., 2002).
In addition to Ethiopia, Boswellia papyrifera grows in the dry areas of Africa, mainly in the Sudanian and Sahelian regions where it occurs in Nigeria, Cameroon, Central African
Republic, Chad, Sudan, Uganda and Eritrea (Demissew, 1993). In Ethiopia, it is distributed in the northern, western and central parts of the country, including Tigray, Gonder, Gojam,
Welega, Wollo and Shewa regions (Bekele et al ., 1993; Taddese et al ., 2002). It is found in the arid regions mainly on steep, rocky slopes in hilly areas, often on lava flows or sandy river valleys and shallow soils (Bekele et al ., 1993; Taddese et al ., 2002).
14
This species has an economic importance of gums and resins, this income as a source of valuable foreign currency and employment opportunities have been generated throughout the year by the Boswellia papyrifera products include tapping and collection, transportation, processing (cleaning, sorting and grading), marketing of frankincense and guarding of storage facilities (Taddese et al ., 2002). At household level, studies carried out in one region of Ethiopia have shown that the gum and resins business provides income about three times greater than the contribution of crop farming (Gebrehiwet et al ., 2002;
Lemenih et al ., 2003).
2.4. Land Use and Land Covers Change Analysis
Every parcel of land on the Earth’s surface is unique in the cover and it possesses.
According to FAO (2000), “land cover is the observed biophysical cover on the earth’s surface”. The same document also defines land use as the arrangements, activities and inputs that people under take on a certain land cover type. According to this definition, land cover corresponds to the physical condition of the ground surface, like forest and grassland, while land use reflects human activities such as the use of the land such as industrial zones, residential zones, and agricultural fields.
Land use and land cover change has become a central focal point in the current strategies for managing natural resources and monitoring environmental change. Since the late
1960’s, the rapid development of the concept of vegetation mapping has lead to increased studies of land use and land cover change worldwide. Providing an accurate assessment of the extent and health of the global forest, grassland, and agricultural resources has become an important priority. The land use and land cover pattern of a region is an outcome of natural and socioeconomic factors and their utilization by man in time and space. Land is becoming a scarce resource due to immense agricultural and demographic pressure. Hence,
15
information on land use and land cover and possibilities for their optimal use is essential for the selection, planning and implementation of land use and land cover schemes to meet the increasing demands for basic human needs and welfare. This information also assists in monitoring the dynamics of land use and land cover resulting from the changing demands of the increasing population (Ahadnejad, 1999).
2.4.1. Image Classification
According to Jensen (1996), digital image classification is the process of assigning or sorting pixels into a finite number of individual classes, or categories of data, based on their data file values. Usually, each pixel is treated as an individual unit composed of values in several spectral bands. By comparing pixels to one another and to pixels of known identity, it is possible to assemble groups of similar pixels into classes that match to the informational categories of interest to users of remotely sensed data. Digital image classification is divided into two supervised and unsupervised classification.
2.4.1.1. Unsupervised Classification
Unsupervised classification is an approach that uses statistical clustering techniques to combine pixels into groups (classes) based on the degree of similarities of their brightness value in each spectral band. The analyst then combines spectral classes into real land cover type using maps and field based knowledge. The analyst should understand the spectral characteristics of the terrain in the area of interest well enough to properly label certain clusters into a specific information class (land cover type). In this process many spectral classes can be assigned to a few land cover types (Jensen, 1996).
2.4.1.2. Supervised Classification
Supervised classification is the process of grouping pixels using a known identity of specific sites through a combination of fieldwork, analysis of aerial photography, maps and
16
personal experience using remotely sensed data, which represent homogenous examples of land use and land cover types to classify the remainder of the image. These areas are commonly referred to as training sites (Jensen, 1996).
2.4.2. Change Detection Methods
Change detection is the process of identifying differences in the state of an object or phenomenon by observing it at different times. Essentially, it involves the ability to quantify temporal effects using multi-temporal data sets. Post classification is among the most widely applied techniques for change detection purpose (Singh, 1989). Numerous studies have been carried out using post classification approach. In post classification approach two images from different dates are classified and labeled. The area of change is then extracted through the direct comparison of the classified results (Lunetta, 1999). This method needs both images to be individually rectified and classified before they can be compared pixel by pixel (Jensen, 1996). This method provides information and results in a base map that can be used for the subsequent year. Therefore, it identifies where and how much change has occurred.
2.5. Trading Carbon Credits and its Potential to Improve Livelihoods
Carbon sequestration plant biomass has a direct market value. Here, carbon sequestration defined as the amount of carbon that can be additionally stored in forested agro-ecosystems
(Henry et al ., 2009). Indeed, trading carbon credits offers a new hope to resource poor and small land holder farmers of the region that are prone to climate change and variability by creating another important income stream that would make the local livelihoods resilient to climate change. The choice of strategies to mitigate carbon abundance from the atmosphere is also linked to carbon trading, which is an emerging global market opportunity (Kandlikar, 1996; Tucker, 2001; Persson et al ., 2006).
17
The economic gains, through increased biomass productivity, and hence, carbon sequestration, have a potential to positively impact the economy and the environment
(Johnson and Heinen, 2004; Bohannon, 2007). The Kyoto Protocol, Clean Development
Mechanism (CDM) and REDD +, open up new opportunities for using the terrestrial biosphere as a carbon sink (Schlamadinger and Marland, 1998; Olsson and Ardo, 2002;
Diakoulaki et al ., 2007). Tree planting and sustainable woodland management can be economically beneficial for the livelihoods of the local farmers in the case of carbon credits payment to protect and manage the reforested areas. This could bring financial incentives from the carbon credit payments in principle. In addition to this, the local small scale and marginal farmers engaged in woodland development could be able to generate higher rate of financial return from their lands as well as create self employment opportunities to their family members (Diakoulaki et al ., 2007). Thus, managing woodland resources would enhance climate resilient livelihoods.
18
3. MATERIALS AND METHODS
3.1. Study Areas Description
The study sites for the present study were located in three selected districts/woredas (Kafta
Humera, Metema and Sherkole) in the western lowlands of Ethiopia. The sites were gum and resin tapping concession sites owned by the Natural Gum Processing and Marketing
Enterprise.
3.1.1. Kafta Humera District
Kafta Humera district is located in the north-western Ethiopia and in the western part of
Tigray Regional State (Figure 1) and 991 km away from Addis Ababa. Kafta Humera is bordered with “Tsegede” on the south and with Sudan in the west. In the north, the Tekeze
River separates the district from Eritrea, in the east “Tahtay-Adiyabo” bordered with the district and in the southeast with “Wolqayt”. The district administrative center is Humera town. The geographical location of this study area was ranging from 36 0 27’ 4.70’’ to 370
33’ 7.12’’ E and from 130 39’ 46.47’’ to 14 0 26’ 34.87’’N with in an altitude range of 560-
1849 m a.s.l.
The district covers an area of 632,877.75 ha which is about 23.6 percent of the western zone of Tigray. The area is located in semi-arid agro-climatic zone. The geological classifications of the district are mainly dominated by early tertiary volcanic and Pre-
Cambrian rocks and also the dominant soil types in the study area are Chromic Eutric and
Calcic Combisols; Chromic and Orthic Luvisols and Chromic and Pellic Vertisols (EMA,
1988). The vegetation communities in the districts include Acacia-Commiphora,
Combretum-Terminalia and dry evergreen woodlands (Eshete et al ., 2011). The mean total rainfall ranges from 400-650 mm. The mean maximum temperature varied between 33 0C in April and 41.7 0C in May, while the mean minimum temperature is between 17.5 0C in
19
August and 22.2 0C in July. The rainy season of the study area is from June to September.
The remaining 8-9 months between October and May/June is dry and hot (KHDOA, 2011).
The land use system was characterized by mixed farming system dominated by open crop cultivation; and this included cereals (31.24%), pulses (5.94%), oilseeds (60.87%) and vegetables (1.95%). Many farmers (68.8%) are practicing a mix of cereal-livestock farming, while 27.97% cultivating annual crops and 3.23% livestock raring. Regarding land ownership and land distribution/holding about 74.74% of farmers possessed own land,
25.09% were renting and the rest 0.17% was other form of land holding (CSA, 2007). The economy of the district is mainly centered on the production of sesame, but after 1996 it was replaced by cotton as the primary cash crop. Sesame is high value edible oil that is exported to Israel, Turkey, the Middle East, Japan and China. Over 400 large scale investors cultivate an average of 600 hectares of sesame, while local farmers cultivate up to 12 hectares/head. In the district, investors cultivate 58% of the cultivated land while local farmers use the remaining 42%. Another important crop in the district is sorghum, and it is grown both by agricultural investors and local farmers for cash and consumption
(KHDLR, 2007).
Population in the district has increased from 48,690 in 1994 to 92,144 (47,899 are men and
44,245 are women) in 2007 suggesting a population growth rate of 3.6% per year.
According to (CSA, 1994, 2007), the majority of the people (about 67.2 %) are living in rural areas, while about 32.8% are urban inhabitants. The men and women population a total of about 47,899 and 44,245 respectively is almost at equal proportion. Kafta Humera with an area of 6,328.78 square kilometers and has a population density of 14.56/km 2, which is less than the zone average 28.9 and the national average 77.72. A total of 23,449
20
households were counted in this district and the average household size is estimated to be
3.93 persons.
3.1.2. Metema District
Metema district is located north western Ethiopia (350 45’ 21.34’’ to 36 0 45’ 31.31’’ E and
12 0 17’ 33.63’’ to 13 0 5’ 52.52’’ N) in the Amhara regional state. Metema is one of the 105 districts, and part of the North Gondar Zone of the Regional State and 889 km far away from Addis Ababa. Metema is bordered in the south with “Qwara” district, in the west with
Sudan, in the north with “Sanja” district and in the east with “Chilga” district. The total annual rainfall ranges from 700-900 mm with minimum annual temperature ranged between 22 0C and 28 0C and mean maximum annual temperature between 35 and 45 0C.
The agroecological zone of the district is classified as semi-arid and the elevation is ranges between 550 and 1600 meters above sea level (MDOA, 2011).
The major land use and land cover types observed in the district were arable/crop land
(42.21%), pasture (7%), forest and shrub land (41.11%) and the remaining 9.68% is considered as a degraded or other land. Teff, corn, sorghum, cotton and sesame are the important cash crops. The town of Metema serves as an important trade gateway between
Sudan and the Amhara Regional state. The economy of the district is predominantly agricultural production based commodity. Gum and incense are important cash crops in the district and the principal species which produce incense is Boswellia papyrifera , while
Acacia seyal and Acacia polyacantha are harvested for their gum production. The geological classifications of the district are manly dominated by tertiary and younger sediments and early tertiary volcanic rocks. The soils in flat land areas are dominantly
Chromic and Pellic vertisols, while the soils in hillsides are Chromic and Orthic Luvisols
(EMA, 1988).
21
The site is one of the natural ecosystems of Boswellia papyrifera where its population is found in a good stock and commercial tapping of incense is widely practiced (Eshete et al .,
2012).
The district has a total population of 110,231 and increase of 100.78% over the 1994 census. Out of which 58,734 are men and 51,497 women and 26.93% of them are urban inhabitants (CSA, 2007). With an area of 380,677.32 ha, Metema has a population density of 28.96 persons per square kilometer, which is less than the Zone average of 63.76 persons per square kilometer. A total of 29,378 households were counted in the district and the average size of the household was 3.75 persons.
3.1.3. Sherkole District
Sherkole district is located in Benshangul Gumuz regional state, western Ethiopia (Figure
1). This study area is one of the 21 districts in the regional state, which is 710 km far away from Addis Ababa. The district is part of the Asosa Zone, in the south bordered with
“Menge” district and in the southwest with “Kormuk” distric. In the northwest, the district has an international border with Sudan and in the northeast with Abay River which separates the district from “Metekel” Zone. In the southeast “Dabus” River on separates the district from the “Kamashi” and “Oda Godere” Zone. The geographical location of the district is 34 0 28’ 48.29’’ to 35 0 13’ 3.14’’ E and 10 0 26’ 18.98’’ to 11 0 14’ 25.65’’ N with in an altitude range between 500 and 1000 m a.s.l. The mean total rainfall varies from 900-
1200 mm with mean annual temperature ranging in between 10.8-42 0C. The geological classification of the district is mainly intrusive rock and the soil type of this study area is dominated by Dystic Nitosols and Calcaric and Eutric Fluvisols (EMA, 1988). The agroecological zone of the district is semi-arid (AWBISPP, 2003).
22
The total land cover of Sherkole district is 351,857.07 ha, out of which 6.89 % cultivated land, 55.25% wooded and shrub land, 18.66% grass land and 19.20% other land uses.
The total population of the district is 19,992 (9,931 men and 10,061 women) whereas
4,237 households (3,837 men and 400 female) were counted. Sherkole has a population density of 5.68 persons per square kilometer which is less than the Zone average of 19.95 and the national average 77.72 (CSA, 2007).
Figure 1: Map of the Study Sites (Kafta Humera, Metema and Sherkole Districts)
3.2. Sampling and Data Collection Methods
3.2.1. Study Site Selection The sample plots for this study were located at a site that have been allocated for gum and resin collection by Natural Gum Processing and Marketing Enterprise. There were left rest
23
for 9 years in Lemlem Terara site and while it is not tapped from the beginning (here after referred as untapped trees) in Adi Goshu site. The other one is tapping were going on
(hereafter referred as tapped trees). The sample plots were located at all study sites.
The study sites were represented by different climatic conditions, and hence the effect of climate gradient on biomass production and carbon stocking were captured. For example there is a gradient of rainfall and temperature north west to south west lowlands, although there is no much difference in elevation gradients. The total annual rainfall at Sherkole
(900-1200 mm) is higher than the rain fall at Metema (700-900 mm) and Kafta-Humera
(400-650 mm). The mean annual temperature at Metema (22-45 0C) is higher than Sherkole
(10.8-42 0C) and Kafta-Humera (17.5-41.7 0C).
In the study sites, vegetation and soil samples were taken from the sampling plots (Figure
2) (Sheikh et al ., 2009). A total of 36 plots were distributed for carbon stock estimation and vegetation records.
a) Adi-Goshu Site:
Of which in Adi-Goshu site, six plots were allocated for each untapped and tapped
Boswellia papyrifera woodlands, two plots for the crop lands within the untapped
Boswellia papyrifera and three plots for the crop lands within the tapped Boswellia papyrifera .
b) Lemlem Terara Site:
In Lemlem Terara site three plots were allocated for the tapped and six plots for the untapped Boswellia papyrifera .
c) Gemed Site:
In Gemed site the allocated plots were 10 for tapped Boswellia papyrifera woodlands.
24
3.2.2. Tree Sampling Trees were inventoried from the larger circular plots with 15 m radius (Figure 2). For the aboveground biomass estimation, field transect of 200 m distance of circular sampling method was applied (Hairiah et al ., 2001). All tree diameters in the larger plot were measured at breast height (DBH) (1.3 m aboveground level) and at stump height (DSH)
(30 cm aboveground level) (Mac Dicken, 1997). In addition, the total tree heights (to the top of the crown) were measured using Hypsometer. In the plot, local names of trees were recorded and later scientific names were identified from “Useful Trees and Shrubs for
Ethiopia” (Bekele, 2007).
3.2.3. Dead Wood Sampling Within the 15 m circular plots where live trees were recorded, dead trees were also measured. Dead woods were classified as standing, stump and felled woods. Standing dead woods are dead trees with stem, branches and twigs; and their DBH, height and standing status were recorded. From stump and felled dead trees height (length) and mid diameter of the woods were measured (Persson et al ., 2005).
3.2.4. Herb, Litter and Crop Biomass Sampling From all study sites, the biomass of herbs was collected during the end of the rainy season, the peak growth period. All the herbaceous vegetation emerging within the quadrant areas
(1m x 1m) (Figure 2) were cut at the ground level/closed to the mineral soil, weighed, and a composite sample was obtained from each subplot for oven-dry mass determination in the laboratory (Roshetko et al. , 2002; Dossa et al. , 2008; Jina et al ., 2008). Similarly, surface litter was sampled from the quadrants and composite litter was collected. In crop land, the whole plant material of maize, sorghum and sesame grown under the trees were collected within 1m x 1m (Figure 2) quadrant sample plot located with circular plot were
25
collected and all the stalks and roots were taken for oven-dry biomass determination in the laboratory.
3.2.5. Soil Sampling The soil samples were collected for the bulk density and soil carbon stock analysis. Soil samples were taken from quadrants (1 m2) allocated in the four directions (north, south, east and west) of the circular sample plots (Figure 2). Soil samples for the determination of soil carbon density were collected from 30 cm soil depth after the herbs and litter samples were taken. Within 1 x 1 m quadrant four soil samples were taken by pressing an auger to a depth of 30 cm, and the four soil samples were composited (Roshetko et al. , 2002;
Takimoto et al. , 2008). In addition, from the same quadrants, soil samples for soil bulk density determination were collected from the surface soil using 10 cm length and 3.4 cm diameter core sampler carefully driven into the soil to avoid compaction (Roshetko et al. ,
2002).
1m
1m
5m
15m
Figure 2: Plot Size and Quadrants for Tree, Soil, Herbs, Crop and Litter Sampling
Quadrants on the radius 5 m are for litter, herbaceous plant and soil sample collection, while 15 m radius circular plots are for measuring trees (Persson et al ., 2005).
26
3.3. Data Analysis
3.3.1. Carbon Stock Estimation
3.3.1.1. Aboveground and Belowground Biomass
The usual methods for determining of the aboveground biomass (AGB) of forests are the combination of forest inventories with allometric tree biomass regression models
(Houghton et al ., 2001; Brown, 2002; Houghton, 2005). This estimation of AGB in the forests is based on plot inventories that involve in the following three steps (Brown et al .,
1989; Houghton et al ., 2001; Chave et al ., 2005):
1. The selection and application of an allometric biomass function for the
estimation of individual tree biomass,
2. The summation of individual tree AGB to estimate plot AGB, and
3. The calculation of an across-plot average to hectare based.
In this study, allometric equations given by Brown (1997) and WBISPP (2000) were used.
The equations were developed based on the aboveground tree biomass for different tropical climatic/rainfall regions. The total annual rainfall obtained from a weather station close to the study sites were used to determine the allometric equation. The carbon stock estimated by the two equations was statistically compared. The following allometric equation was used to calculate the AGB of Brown (1997):
AGB = 0.139 DBH 2.32 (Brown, 1997)…………………………………..Equation (1)
Where, AGB = Above Ground Biomass (kg/tree) and
DBH = Diameter at Breast Height
The equations developed by Woody Biomass Inventory and Strategic Planning Project
WBISPP (2000) for all woody species in Dry Kolla and Moist Kolla agro-ecological zone
27
of Ethiopia were also used for estimating tree carbon stock and later to comparing with the results estimated according to Browns equation. The following equation was used:
AGB = (0.4861*DSH) + (0.1659*(DSH exp 2.2)) for Dry Kolla……...Equation (2)
AGB = (1.4277*DSH) + (0.0088*(DSH exp 3.0)) for Moist Kolla…....Equation (3)
Where, AGB = Above Ground Biomass (kg/tree) and
DSH = Diameter at Stump Height (cm)
Tree biomass was converted to carbon stock (MacDicken, 1997):
AGB Carbon Stock = AGB * 0.5 …………………………………...... Equation (4)
Where, AGB = Above Ground Biomass (kg/tree)
Root biomass is often estimated from root-shoot ratios (R/S) by taking 25% of above- ground biomass (Cairns et al ., 1997). The equation presented in Cairns et al. (1997) was used to make estimates of root biomass in a standard manner for forests based on the knowledge of the above-ground biomass. The root biomass was then converted in to root carbon stock by taking 50% of the root biomass.
Root Biomass = AGB x 0.25 ………………………………………..….Equation (5)
Where, AGB = Above Ground Biomass (kg/tree)
3.3.1.2. Dead Wood Biomass
For standing dead wood, which has branches was estimated in a similar manner using the allometric equation of above ground biomass. As the standing dead wood do not have leaves, needs to subtract 5-6 percent for conifer species while 2-3 percent for broadleaved species (Pearson et al ., 2005). In the study, most of the existing species were broadleaved, and hence 2.5 percent reduction was recommended from the total above ground biomass of each standing dead tree.
28
= 2.32 − BSDW 1 0.139 DBH 5.2 % ...... Equation (6)
Where, BSDW 1 = Biomass of Standing Dead Wood in kg
DBH = Diameter at Breast Height of Standing Dead Wood (cm)
In addition, to determine the amount of biomass in the standing stump of dead woods was used the recommended allometric equation from REDD (2009):
1 D 2 BSDW = * ∏* * H * S ...... Equation (7) 2 3 200
Where, BSDW2 = Biomass of Standing Dead Wood (kg)
H = Height of Standing Dead Wood (m)
D = Basal Diameter of Standing Dead Wood (cm)
S = Mean Wood Density of Dead Wood (g cm -3)
The specific density is estimated at 0.5 g cm -3 as default value, but can be used 0.8 for dense hard woods and 0.3 for very scattered species in tropical regions (Hairish et al.,
2001).
The volume of felled dead trees was calculated using the midpoint diameter and height measurements. It is then estimated as the volume of a truncated cylinder.
= Huber’s Formula: V gm L …………………………………………….Equation (8)
Where, V = Volume of the Log
gm = Cross-Sectional Area at Log Mid Point
L = Log Length
Volume was converted to dry biomass using wood density available in REDD (2009):
BDLDW = V x S...... Equation (9)
29
Where, BDLDW = Biomass of Down Lying Dead Wood (kg)
V = Volume of the Dead Wood (m 3)
S = Mean Wood Density of the Dead Wood (g cm -3)
The total biomass of the dead wood was estimated by summing up of the standing, logged and felled dead wood as follow:
TBDW = BSDW 1+ BSDW 2 + BDLDW...... Equation (10)
Where, TBDW = Total Biomass of Dead Wood in a Given Plot
BSDW 1 = Biomass of Standing Dead Wood which have Branches
BSDW 2 = Biomass of Standing Dead Wood which haven’t Branches
BDLDW = Biomass of Down Lying Dead Wood
The total carbon stock in dead wood was computed by multiplying the total biomass of the dead wood by 0.5 (Persson et al ., 2005).
3.3.1.3. Herb, Litter, Annual Crop and Crop Root Biomass Estimation
The collected sample from herbs, crop and litter were oven dried at 70°C till constant weight (Jina et al ., 2008). The root samples were extracted from soil by soaking in water in a shallow dish and followed by two successive washings of the remaining soil and sieving through a 2 mm screen. The roots were rinsed in distilled water and separate from organic debris using forceps. Root samples per subplot were oven dried at 70 oC to constant weight and root weight per hectare was calculated based on the area of the sampling plots (Dossa et al. , 2008).
For the forest floor (understory wood vegetation, herbs, grass, and litter), the amount of biomass per unit area was given by Persson et al . (2005):
Dry Mass (kg) = Sub Sample Dry Mass (kg) * Field Mass (kg)...... Equation (11) Sub Sample Fresh Mass (kg)
30
The biomass density (Mega gram per hectare) was calculated by multiplying the dry mass by an expansion factor calculated from the sample frame or plot size. The expansion factor was calculated as the area of a hectare in square meters divided by the area of the sample in square meters, that is:
Expansion Factor = 10,000 m 2 ……….……………………..Equation (12) Area of the Plot (m 2) The carbon content in herbaceous biomass was calculated by multiplying herbaceous biomass by 0.5 (Persson et al ., 2005).
3.3.1.4. Soil Organic Carbon
The soil samples for soil carbon analysis were air-dried, well mixed and sieved through a 2 mm mesh size sieve. Soil samples were also analyzed following Walkley & Black (1934) and were conducted at the National Soil Testing Laboratory. Bulk density was determined in the Forest Research Center, Addis Ababa, after drying the core samples of soil at 105 oC and the weight of the soil was divided by the volume of the core sampler. The weight of the gravel above 2 mm diameter was subtracted to determine the bulk density of the soil samples. The soil organic carbon stock pool was calculated using the formula (Pearson et al. , 2005):
SOC = BD * d * %C ………………………………………………….Equation (13)
Where, SOC = Soil Organic Carbon [Mg ha -1]
BD = Bulk Density [g cm -3]
d = Depth of the Soil Sample [cm]
% C = Carbon Concentration [%]
31
3.3.1.5. The Total Carbon Stock Density Estimation
The total carbon stock density was calculated by summing all the carbon pool stock densities using the following formula:
C Density = C AGTB + C BGTB + C DWB + C HB + CLB + SOC ………………...Equation (14)
-1 Where, C Density = Carbon Stock Density [Mg ha ]
-1 CAGTB = Carbon Stock in Above Ground Tree Biomass [Mg ha ]
-1 CBGB = Carbon Stock in Below Ground Tree Biomass [Mg ha ]
-1 CDWB = Carbon Stock in Dead Wood Biomass [Mg ha ]
-1 CHB = Carbon Stock in Herb Biomass [Mg ha ]
-1 CLB = Carbon Stock in Litter Biomass [Mg ha ]
SOC = Soil Organic Carbon [Mg ha -1]
3.3.2. Land Use and Land Cover Change Analysis 3.3.2.1. Remote Sensing Data Acquisition
For this study, three dates (1985, 1995 and 2010) of Landsat satellite images were acquired
(Source: http://landcover.org ). A brief description of them is given in Table 1. The downloaded satellite images were in tiff format and were stacked in ERDAS 9.1 software and developing function in it to stack each layer to produce one single layer composing of each band. Then from the stacked band the study area was extracted.
Table 1: Landsat Data Used in Land Use and Land Cover Classification Sensors Study Path and Row Bands Pixel Size/Ground Observation Area Resolution (m) Date Landsat Kafta p170 r50 & 51 7 30*30 1985 ETM Humera Landsat Metema p170 & 171 r51 7 30*30 1995 ETM Landsat Sherkole p171 r52 &53 7 30*30 2010 ETM
32
3.3.2.2. Image Analysis
3.3.2.2.1. Image Pre-processing
Preprocessing involve those operations that are normally required prior to the main data analysis and extraction of information. Selecting appropriate satellite imagery is the first task in image data processing.
a) Geometric Correction
In this work, the processing of these images have been geometrically corrected with road and river intersection on the images themselves and the topographic map of the study area with scale of 1:250,000. After the raw data are georeferenced, they have been clipped with the boundary of the study area for further processing.
b) Image Enhancement
The goal of image enhancement is to improve the visual interpretability of an image by increasing the apparent distinction between features in the scene (Lillesand and Kiefer,
2000). If the image is enhanced the distinct of features are more clear so that image analysis, classification and interpretation is better. In addition, Image enhancement is used to increase the details of the image by assigning the image maximum and minimum brightness values to maximum and minimum display values, it is done on pixel values, and this makes visual interpretation easier and assists the human analyst. The original low dynamic range of the image is stretched to full dynamic range which is from 0 to 256 by using histogram equalization. Moreover, spatial enhancement of convolution of Kernel 5 by 5 of high pass filtering has been done on the images of the respective years.
3.3.2.2.2. Image Classification Analysis
In this study, unsupervised and supervised classification methods were used. Supervised image classification was a method in which the analyst defines small training sites on the
33
image, which were representative of each desired land cover category. The delineation of training areas representative of a cover type was most effective when an image analyst has knowledge of the geography of a region and experience with the spectral properties of the cover classes. However the unsupervised classification technique is performed when there was little or no knowledge to the geography of the region where classification is under taken. Therefore, first the satellite image was classified in the unsupervised classification for identification of the features in a pixel form. Then by observing and recording identifiable coordinate points of features in the Google Earth were perform the supervise classification using the training points.
For this study a simple classification scheme comprising six land use and land cover types was developed for the purpose of mapping. A combination of information collected from the field and a satellite image were effectively used in the preparation of the legend.
Identification of some of the land use and land cover classes were required a number of field visits and discussions with farmers, to have not only a clear understanding of the main land use and land cover types but also to establish what types of changes are expected over time. What stage and type of land use and land cover is to be expected in what season of the year should be properly established to enable interpretation of the satellite images.
Categorization of land use and land cover types were culminated in the production of the land use and land cover legend, establishment of its characteristics, and identification and mapping of the various land use and land cover types (Tegene, 2002).
3.3.2.2.3. Accuracy Assessment
Land use and land cover maps derived from remote sensing always contain some sort of errors due to several factors, which range from classification technique to method of satellite data capture. In order to wisely use the maps the errors must be quantitatively
34
evaluated in terms of classification accuracy and intended to produce information that describes reality. Therefore, an accuracy classification assessment was carried out to verify to what extent the produced classification is compatible with what actually exists on the ground (Congalton, 1991). It involves the production of references (samples) that evaluate the produced classification. These references were produced from Google Earth and GPS points during field work, which were independent of the ground truths used in the classification. Using this process error matrix was produced for each image of the three districts.
3.3.2.2.4. Matrix of Land Use and Land Cover changes
The change matrixes were determined by overlaying two land use and land cover maps at a time in ERDAS imagine software. The areas which were converted from each of the classes to any of the other classes were computed (Gautam et al ., 2003).
35
Landsat Image Landsat Image Landsat Image 1985 1995 2010
Geometric Correction
Image Enhancement
Image Classification
LULC Map LULC Map LULC Map (1985) (1995) (2010)
Ground Verification
Accuracy Accuracy Accuracy Assessment Assessment Assessment (1985-Map) (1995-Map) (2010-Map)
LULCC Matrix (1985 & 1995)
and (1995 & 2010)
Figure 3: The General Framework for the Study of LULCC
36
The rate of change was also calculated for each land use and land cover using the following formula:
Rate of change (ha/year) = (A-B)/C …………………………………..Equation (15)
Where, A = Recent area of the land use and land cover in ha
B = Previous area of the land use and land cover in ha
C = Time interval between A and B in years
It should be noted that the negative values indicate the magnitude of decline in that particular land use and land cover type.
3.3.3. Statistical Analysis of the Various Carbon Stocks
The effect of ongoing tapping and tapping rest period as well as effect of site variation on carbon stock was tested using two ways ANOVA. Honesty Significance Difference (HSD) tests were also performed to separate means when ANOVA results indicated the presence of significant differences in mean differences between stratum and study sites on carbon stock of the biomass as well as the soil organic carbon. All statistical tests were performed with SAS 9.0 and statistical mean differences were considered significant when P-value is less than 0.05. A fixed effect model was used to estimate the effects of tapping and sites on the carbon stock of woodlands following the formula:
Yij = µ + αi + βj + eij ………………………………………………...Equation (16)
Where, Yij = carbon stock density
µ = the overall mean of carbon stock
th αi = the i tapping effect
th βj = the j site effect
eij = the error term
37
4. RESULTS
4.1. Vegetation Characteristics
A total of 24 species were recorded across the three study sites. A range of different plant characteristics such as DBH and height were also identified (Appendix 1). The average number of stems per hectare across the sites was 276 and the average basal area was estimated to be 9.33 m 2 ha -1. The highest number of plant species (15), number of stems
(359 ha -1) and BA (11.32 m 2 ha -1) were identified and estimated for the Lemlem Terara site and the lowest for Adi Goshu (number of species, 13; number of stems, 224 ha -1 and BA, 7 m2 ha -1). As compared with other two sites, the average DBH (19.28 cm) and height (7.12 m) of the woody species were larger at Gemed site than Adi Goshu (18.81cm and 6.22 m) and Lemlem Terara site (18.77cm and 6.49 m). In Lemlem terara site, the computed average above ground biomass, using the equation developed by Brown (1997), was the highest (54.94 t ha -1) than in Gemed (51.73 t ha -1) and in Adi Goshu site (33.43 t ha -1).
140 -1 120 100 Adi-Goshu Lemlem-Terara 80 Gemed 60 40 20 Number of Individual ha Individual of Number 0 1 2 3 4 5 6 7 8 9 DBH Class Figure 4: Diameter Size Class Distribution of the Entire Woodlands
Diameter class in cm: class 1=5-10, 2=10-15, 3=15-20, 4=20-25, 5=25-30, 6=30-35, 7=35- 40, 8=40-45 and 9>=45.
38
The population structure of the entire woodland showed higher stem densities in the middle diameter classes and progressively declining stem densities with increasing diameter classes at both sites of Adi-Goshu and Lemlem-Terara (Figure 4).
Based on their population structures, the woodland species can be categorized into two diameter class distribution patterns. Group I contained species with a progressively declining numbers of trees with increasing diameter in Gemed site. Group II was comprised of both Adi-Goshu and Lemlem-Terara sites with a bell-shaped or irregular distribution.
4.2. Carbon Stocks in Different Carbon Pools
4.2.1. Aboveground Biomass Carbon Stock
The results showed that overall aboveground biomass carbon density were varied significantly between the two methods (P<0.0001) (Appendix 9 and 10). The calculated carbon density using the Brown (1997) function appeared to be the average while the
WBISPP method underestimates the carbon density. This underestimation might occur due to the formulation of the function relied on the diameter at stump height.
As a result, the Brown (1997) allometric function developed for dry tropical forest was adopted for conventionally estimating the aboveground biomass. The calculated mean aboveground biomass carbon stock of Kafta Humera (Adi Goshu site) in untapped and tapped Boswellia papyrifera woodlands was 16.71 and 19.29 Mg ha -1 (Figure 5) respectively. There was no significant differences within the site at (P=0.6207). 78.84% and 67.69% were sharing the Boswellia papyrifera of untapped and tapped, respectively. In the farmland the above ground biomass carbon was estimated to be 8.92 and 16.77 Mg ha -1
(Figure 6), respectively in untapped and tapped Boswellia papyrifera woodlands.
Similarly, there was no significant differences found between these strata (P=0.3732). This
39
also occupied 50% and 33.33% of the carbon stock was by Boswellia papyrifera, respectively for untapped and tapped stratum.
In Metema (Lemlem Terara site) no significant differences (P=0.7695) were found in the mean carbon stock between untapped (27.91) and tapped (26.59) Mg ha -1 (Figure 7)
Boswellia papyrifera woodlands. 65.73% and 38.42% of the carbon stock were share by
Boswellia papyrifera for the untapped and tapped stratum, respectively. In Sherkole
(Gemed site), the estimated mean aboveground carbon stock was 25.87 Mg ha -1 (Figure 8) for the tapped Boswellia papyrifera woodland. The Boswellia papyrifera contributes
16.01% of the mean carbon stock density.
The estimated mean aboveground carbon stock in Lemlem Terara for the untapped
Boswellia papyrifera woodland 27.91 Mg ha -1 (Figure 9). Whereas, in Adi Goshu it was estimated 16.71 Mg ha -1 and a significant variation (P=0.0074) was observed between these two sites. On the contrary, in the tapped Boswellia papyrifera woodlands, the estimated mean carbon stock of the aboveground biomass showed no significant differences (P=0.4747) among Adi Goshu (19.29 Mg ha-1), Lemlem Terara (26.59 Mg ha -
1) and Gemed, (25.87 Mg ha -1) sites (Figure 10).
4.2.2. Belowground Root Biomass Carbon Stock
The mean belowground root carbon stock in the Adi Goshu site estimated to be 4.18 Mg ha -1 for the untapped Boswellia papyrifera woodlands and 4.82 Mg ha -1 for the tapped
Boswellia papyrifera woodland (Figure 5), and obviously no significant differences
(P=0.6209) were observed between them. Similarly, in the farmland of Adi Goshu site, the mean belowground root biomass carbon stock of untapped and tapped Boswellia papyrifera woodlands was 2.23 and 4.19 Mg ha -1, respectively (Figure 6). And no significant difference was observed within these stratum (P=0.3733).
40
In the Lemlem Terara site, the estimated mean belowground root carbon stock was 6.98
Mg ha -1 for untapped Boswellia papyrifera woodlands and 6.65 Mg ha -1 for the tapped
Boswellia papyrifera woodland (Figure 7) and similarly no significant differences
(P=0.7697) were observed between the root carbon stock of untapped and tapped stratum.
With regard to the Gemed site, the mean belowground root biomass carbon stock of the tapped Boswellia papyrifera woodland was estimated 6.47 Mg ha -1 (Figure 8).
The variation of the mean belowground carbon stocks of the untapped Boswellia papyrifera woodlands was evaluated between sites. As Figure 9 showed the mean belowground root biomass carbon stock for Lemlem Terara was 6.98 Mg ha -1, which was significantly larger (P=0.0074) than Adi Goshu (4.18 Mg ha -1) site.
On the other hand, for the tapped Boswellia papyrifera woodlands, no significant differences (P=0.4748) were seen in the amount of belowground root biomass carbon stock among the different study sites, Adi Goshu (4.82 Mg ha -1), Lemelem Terara (6.65 Mg ha -1) and Gemed (6.47 Mg ha -1) (Figure 10).
4.2.3. Dead Wood Biomass Carbon Stock
In Adi Goshu site, the mean dead wood biomass carbon stock in the stratum of untapped and tapped Boswellia papyrifera woodlands estimated to be 0.48 Mg ha -1 and 2.89 Mg ha -1
(Figure 5), respectively. And no significant differences (P=0.1461) were observed in the mean dead wood carbon between the two stratum within this site. Furthermore, the result also indicated a significant amount of dead wood biomass carbon which was accounted
2.32 Mg ha -1 (Figure 6) in the farmland of untapped Boswellia papyrifera woodlands.
Similarly, in Lemlem Terara site, no significant differences (P=0.0684) were found in the mean dead wood carbon stock between the untapped (2.79 Mg ha -1) and tapped (0.40 Mg
41
ha -1) (Figure 7) Boswellia papyrifera woodlands. In Gemed site, the dead wood carbon estimated to be 0.80 Mg ha -1 for the tapped Boswellia papyrifera woodland (Figure 8).
When we evaluate whether there was statistical differences of dead wood carbon stock amount between Adi Goshu and Lemlem Terara, it was found a significant variation
(P=0.0238) within the untapped Boswellia papyrifera woodlands (Figure 9). On the other hand, when we evaluate the mean dead wood carbon stock in the tapped Boswellia papyrifera woodlands, no significant differences (P=0.1502) among Adi Goshu (2.89 Mg ha -1), Lemlem Terara (0.40 Mg ha -1) and Gemed (0.80 Mg ha -1) sites (Figure 10).
4.2.4. Herb, Annual Crop and Litter Carbon Stock
In Adi Goshu site, the estimated mean carbon stock of the herb biomass for the untapped
Boswellia papyrifera woodlands was 0.28 Mg ha -1 while in the tapped stratum the herb carbon stock was estimated 0.45 Mg ha -1 (Figure 5). Obviously, no significant differences
(P=0.4496) were observed between the two stratum of the herb carbon stocks within the same site. In contrast, Figure 6 indicated that the mean herb biomass carbon stock in the farm land of the untapped Boswellia papyrifera woodlands was 0.52 Mg ha -1 and 0.12 Mg ha -1 for farm land of the tapped stratum and significant differences (P=0.0230) were found between them.
In the Adi Goshu site, the estimated mean annual crop biomass carbon stocks in the farm land of the untapped Boswellia papyrifera woodlands was, 2.27 Mg ha -1 while 0.70 Mg ha -
1 for the tapped stratum (Figure 6). On the other hand, the estimated annual crop root biomass carbon stocks were 0.18 for the untapped, while 0.07 Mg ha -1 in the tapped
Boswellia papyrifera woodlands. As a result, in both crop biomass and root carbon pools, no significant differences (P=0.3599 and P=0.2777) were observed within the stratum.
42
In the case of Lemlem Terara site, the herb biomass carbon stocks for the untapped and tapped Boswellia papyrifera woodlands were 0.26 Mg ha -1 and 0.42 Mg ha -1, respectively
(Figure 7). Whereas, the mean litter biomass carbon stocks of this site were 0.61 Mg ha -1 for untapped and 0.46 Mg ha -1 for the tapped Boswellia papyrifera woodlands. Similarly, there were no significant differences (P=0.2386 and P=0.7131) found in both carbon pools between similar stratum. In the Gemed site, the mean herb biomass carbon stock in the tapped Boswellia papyrifera woodland was 1.19 Mg ha -1 (Figure 8).
The herb biomass carbon stock in the untapped Boswellia papyrifera woodlands, no significant differences (P=0.9191) were observed between Adi Goshu (0.28 Mg ha -1) and
Lemlem Terara (0.26 Mg ha -1) sites (Figure 9). However, in Gemed site, the tapped
Boswellia papyrifera woodland of the mean herb biomass carbon stock was 1.19 Mg ha -1, which was significantly different (P=0.0207) from the other two study sites (Figure 10).
4.2.5. Soil Carbon Stock
In Adi Goshu site, the estimated mean soil carbon stocks were 33.61 Mg ha -1 for the untapped Boswellia papyrifera woodlands and 38.48 Mg ha -1 for the tapped woodlands
(Figure 5). The mean soil carbon stocks in the farmlands were 49.00 Mg ha -1 for the untapped and 23.33 Mg ha -1 for the tapped Boswellia papyrifera woodlands (Figure 6).
Accordingly, no significant differences were found between the woodlands (P=0.5501) and farmlands (P=0.1737) of the mean soil carbon stocks.
In the case of the Lemlem Terara site, the mean soil carbon stock for the untapped
Boswellia papyrifera woodland was 58.19 Mg ha -1 while 34.25 Mg ha -1 (Figure 7) was for the tapped woodland. Correspondingly, a significant differences (P=0.0316) were observed between the two strata of soil carbon stocks. Whilst comparing the mean soil carbon stocks across sites in the untapped woodland stratum, the Lemlem Terara site was significantly
43
(P=0.0019) richer than that of Adi Goshu (33.61 Mg ha -1) (Figure 9). However, no variations in soil carbon were observed in the tapped stratum (P=0.8884) (Figure 10).
50 a 45 -1 40 UW TW a 35 30 a 25 a 20 15 10 a Carbon Stock in Mg ha Mg in Stock Carbon a a 5 a a a 0 AGBC BGRBC DWBC HBC SOC Carbon Pools
Figure 5: Mean Carbon Stock of Different Carbon Pools of Woodland in Adi-Goshu
70 a
-1 60 UW TW 50 40 30 a a 20 a
Carbon Stock in Mg ha Mg in Stock Carbon 10 a a a b a b a a a a 0 AGBC BGRBC DWBC HBC CBC CRBC SOC Carbon Pools
Figure 6: Mean Carbon Stock of Different Carbon Pools of Farmland in Adi-Goshu
44
70 a 60 -1 50 UW TW b 40 a a 30
20 Carbon Stock in ha Mg Stock Carbon 10 a a a a a a a a 0 AGBC BGRBC DWBC HBC LBC SOC Carbon Pools
Figure 7: Mean Carbon Stock of Different Carbon Pools in Lemlem-Terara
45 40 -1 35 TW 30 25 20 15 10 Carbon Stock in Mg ha Mg Stock in Carbon 5 0 AGBC BGRBC DWBC HBC SOC Carbon Pools
Figure 8: Mean Carbon Stock of Different Carbon Pools in Gemed
45
70 a 60 -1 Adi-Goshu Lemlem-Terara 50
40 b a 30
20 b Carbon Stock in Mg ha Mg in Stock Carbon 10 a b a b a a 0 AGBC BGRBC DWBC HBC LBC SOC Carbon Pools
Figure 9: Mean Carbon Stock of Different Carbon Pools in Untapped Stratum
50 a 45 a -1 40 Adi-Goshu Lemlem-Terara Gemed a 35 a 30 a 25 a 20 15
Carbon Stock in Mg ha Mg in Stock Carbon 10 a a a 5 a a a b b a 0 AGBC BGRBC DWBC HBC LBC SOC Carbon Pools
Figure 10: Mean Carbon Stock of Different Carbon Pools in Tapped Stratum
AGBC=Above Ground Biomass Carbon; BGRBC=Below Ground Root Biomass Carbon; DWBC=Dead Wood Biomass Carbon; HBC=Herb Biomass Carbon; LBC=Litter Biomass Carbon; CBC=Crop Biomass Carbon; CRBC=Crop Root Biomass Carbon; SOC=Soil Organic Carbon; UW= Untapped Boswellia papyrifera Woodland; TW=Tapped Boswellia papyrifera Woodland. Mean with same later are not significant difference at P < 0.05
46
4.2.6. Total Carbon Stock Density
In the Adi Goshu site, the estimated total carbon stock density for the untapped and tapped
Boswellia papyrifera woodlands were 55.26 Mg ha -1 and 65.93 Mg ha -1, respectively
(Figure 11). While in the farm land of the same study site the carbon stocks for the untapped were 65.44 Mg ha -1 and for the tapped, 45.18 Mg ha -1 (Figure 11). In the Lemlem
Terara site, the total carbon stock for the untapped woodland was 96.74 Mg ha -1, which was found the biggest carbon amount in the total study area and for the tapped Boswellia papyrifera woodlands it was 68.77 Mg ha -1 (Figure 11). In the Gemed site, the total carbon stock density in the tapped Boswellia papyrifera woodland was 71.01 Mg ha -1 (Figure 11).
120 UW TW
-1 96.74 100
80 71.01 65.93 65.44 68.77 60 55.26 45.18 40
20 Carbon Stock in Mg ha Mg in Stock Carbon
0 Adi-Goshu-W Adi-Goshu-F Lemlem-Terara Gemed
Figure 11: Total Carbon Stock Density of Different stratum
(W=Woodland; F=Farmland)
4.3. Yield of Gum and Resin
Long-term data (since 1999) on production of gum olibanum and gum arabic was obtained from Natural Gum Processing and Marketing Enterprise (NGPME) (Table 2).
Accordingly, since 1999, in Kafta Humera the maximum yield in gum production was recorded in 2001, i.e. 878.30 tons (4.57 million Birr) whereas the minimum yield obtained was 28.35 tons (0.25 million Birr), in 2006. On the other hand, the highest gum yield in
47
Gondar (Metema, Merab Armachoho, Tsegede and Aelfa Takusa districts) was attained
835.53 tons (5.35 million Birr) in 2004 and the lowest was 151.97 tons (0.35 million Birr) in 1999. In Assosa site (Sherkole, Sirba Abay and Mao Komo districts) the collection was began in 2005. The maximum yield recorded was 327.99 tons (12.02 million Birr) in 2011 while the minimum was 1.68 ton (0.01 million Birr) in 2006. Therefore, the total production of gum olibanum and gum arabic in the three sites was recorded approximately
12,112.98 tons with the net income of 129.34 million Birr since 1999.
According Tilahun et al . (2007) showed that the Boswellia papyrifera woodland yield 127 kg ha -1 for the exclosure land and 84.5 kg ha -1 year -1 for the free grazing land. In addition the annual frankincense yield at tree level yielded on the average of 471.5 g tree -1 in Tigray region (Tilahun et al ., 2011) and according Eshete et al . (2012) yield per tree per year largely varied and ranged from 41 to 1829 g. This resin yield increased with tapping intensity.
Table 2: Yield and Income of Gum Olibanum and Gum Arabic
Kafta-Humera Gondar Assosa Year Yield Income Yield Income Yield Income (Tons) (Million Birr) (Tons) (Million Birr) (Tons) (Million Birr) 1999 772.72 1.77 151.97 0.35 ------2000 856.18 3.73 267.32 1.16 ------2001 878.30 4.56 439.18 2.28 ------2002 529.29 2.85 575.62 3.10 ------2003 229.74 1.81 654.60 5.15 ------2004 149.73 0.96 835.53 5.35 ------2005 111.44 0.83 686.02 5.14 3.04 0.02 2006 28.35 0.25 360.61 2.29 1.68 0.01 2007 39.91 0.39 714.25 3.90 79.01 0.43 2008 147.46 1.37 613.60 4.31 110.22 0.86 2009 201.61 3.51 557.01 10.41 75.93 1.38 2010 205.64 5.70 764.28 21.89 290.87 7.13 2011 210.16 7.93 243.74 6.50 327.99 12.02 Source: Natural Gum Processing and Marketing Enterprise, 2012
48
4.4. Land Use and Land Cover Changes
The interpretation and analysis of remote sensing imagery involves the identification and measurement of various targets in images in order to extract information. In this study unsupervised classifications was carried out first to identify the overall land use and land cover clusters without training data. Then supervised classification was carried out for the three images of (1985, 1995 and 2010) based on the training areas and the different false color composites of 4, 3, 2. Then the change detection analysis was carried out by visual comparison of features and detailed quantitative approaches.
4.4.1. Land Use and Land Cover Mapping
Using the application of supervised image classification methods, six major land use and land cover types were identified. These include Wood land, shrub/bush land, grass land, agricultural land, bare land and water body, based on the characteristics of Landsat images of the year 1985, 1995 and 2010.
Table 3: Description of Land Use and Land Cover Types Identified
Land Use and Land LULC Description Cover Classes Woodland Land under stands of trees with a canopy cover of at least 20%, including integral open space, and including felled areas that are awaiting restocking, predominant species like Boswellia papyrifera Shrub/Bush land Land covered by small trees, bushes, and shrubs, in some cases mixed with grasses; less dense than woodland Grass land Lands where small grasses are the predominant natural vegetations. It also includes land with scattered or patches of trees and it is used for grazing and browsing Agricultural Land Areas allotted to rain fed crop production, mostly oil seed, cereals and pulses Bare land and Land surface which is mainly covered by bare soil and rock out crops Settlement and land covered by structures which included urban towns and rural villages Water Body Lakes, rivers and streams
49
4.4.1.1. Land Use and Land Cover of Kafta Humera District
Based on Figure 12 and Table 4, the land use and land cover classes of wood land
(42.17%) and agricultural land (23.51%) show the largest share in 1985 for the Kafta
Humera site. Whereas, the aerial coverage of the bare land and water body were the least coverage of the district occupied about 27,832.68 ha (4.40%) and 694.53 ha (0.11%) respectively. Moreover, much of the district coverage of the year 1985 was the natural vegetation consist of wood land, shrub land and grass land that occupied about 71.98% of the total area while the remaining 28.02% was arable land, bare land and water body.
Similarly, in the case of the year 1995, the land use and land cover classes of the wood land (35.64%) and Agricultural land (28.23%) show the largest coverage of the district.
The least area coverage was still bare land and water body, which occupied only 28,433.61 ha (4.49%) and 491.22 ha (0.08%) respectively.
In 2010, as showed in Figure 12 and Table 4 the highest share of all classes were agricultural land, which was occupied 247,509 ha (39.11%), followed by wood land cover
(25.75%). The least area coverage was bare land and water body, which covered 33,796.62 ha (5.34%) and 351.81 ha (0.06%).
Table 4: Areas of LULC of Kafta Humera District for the Years 1985, 1995 and 2010
1985 1995 2010 Land Use Type Area (ha) (%) Area (ha) (%) Area (ha) (%) Agricultural land 148772.34 23.51 178640.37 28.23 247509 39.11 Bare land 27832.68 4.40 28433.61 4.49 33796.62 5.34 Wood land 266879.88 42.17 225538.11 35.64 162973.26 25.75 Shrub land 96540.03 15.25 108364.5 17.12 107579.07 17.00 Grass land 92158.29 14.56 91409.94 14.44 80667.99 12.75 Water body 694.53 0.11 491.22 0.08 351.81 0.06
50
Figure 12: LULC Map of Kafta Humera District for the Years 1985, 1995 and 2010
51
4.4.1.1.1. Land Use and Land Cover Change Matrices of Kafta Humera District The land use and land cover change matrix depicts the direction of changes from 1985 to
1995 (Table 5).
Table 5: LULCC Matrices of the Kafta Humera District (1985-1995) Change to LULC 1995 (ha) LULC Agricultural Bare Wood Shrub Grass Water Total land land land land land body Agricultural 132624 874 14230.1 866.31 178 0 land 148772.34 Bare land 606.15 19469.2 2483.63 2137.2 3136.47 0 27832.68 Wood land 42365.7 4603.29 179023 21141.9 19745.9 0 266879.88 Shrub land 1593.61 1629.52 13862.5 64653 14801.4 0 96540.03 Grass land 1450.98 1822.38 15938.8 19459.8 53486.4 0 92158.29 Water body 0 35.19 0 106.29 61.83 491.22 694.53 Change from LULC 1985 from Change (ha) Total 178640.37 28433.61 225538.11 108364.5 91409.94 491.22 632877.75
As shown in (Figure 13) between 1985 and 1995, there was a remarkable increase of
agricultural land from 148,772.34 ha in 1985 to 178,640.37 ha in 1995; although small
portion of it was converted to Bare land (874 ha), wood land (14,230.1 ha), shrub land
(866.31 ha) and grass land (178 ha) in 1995. Whereas, the wood lands were decreasing by
41,341.77 ha between the years 1985 to 1995.
40 30 20 10 0 -10 -20 Area in Thousands ha in Area -30 -40 -50 Agricultural land Bare land Wood land Shrub land Grass land Water body
Figure 13: LULCC (1985-1995) of Kafta Humera District
52
The negative values indicate a decline while the positive values indicate the expansion of
that particular land use and land cover.
Table 6: LULCC Matrices of Kafta Humera District (1995-2010)
Change to LULC 2010 (ha) LULC Agricultural Bare Wood Shrub Grass Water Total land land land land land body Agricultural 167613.93 845 9704 367.09 110.35 0 178640.37 land Bare land 1080 11034.5 6366.6 5411.61 4540.86 0 28433.61 Wood land 60062.96 9248.59 124986.66 21106.78 10133.1 0 225538.11 Shrub land 15551.72 6178.31 11188 61306.78 14139.7 0 108364.5 Grass land 3200.39 6427 10728 19375.29 51679.3 0 91409.94 Water body 0 63.18 0 11.52 64.71 351.81 491.22 Change from LULC 1995 from Change (ha) Total 247509 33796.62 162973.26 107579.07 80667.99 351.81 632877.75
In similar fashion , between 1995 and 2010, there was a remarkable increase of agricultural
land from 178,640.37 ha in 1995 to 247,509 ha in 2010; although small portion of it was
converted to Bare land (845 ha), wood land (9,704 ha), shrub land (367.09 ha) and grass
land (110.35 ha) in 2010. Whereas, the wood lands were decreasing by 62,564.85 ha
between the years 1995 to 2010 (Figure 14).
80
60 40
20 0 -20 -40 Thousands ha in Area -60
-80 Agricultural land Bare land Wood land Shrub land Grass land Water body
Figure 14: LULCC (1995-2010) of Kafta Humera District
53
4.4.1.1.2. Rate of Land Use and Land Cover Change in Kafta Humera District
Based on Table 7, between 1985 and 1995, remarkably, agricultural land has increased with a rate of 2,986.80 ha/year and further increased to the second period with an accelerated rate of change by 4,591.24 ha/year. The expansion of agricultural land was preceded turned over by the outflow of the bush/shrub land, wood land and grass land areas in the study site.
Table 7: Rate of Changes in LULC Classes (1985-2010)
1985 to 1995 1995 to 2010 Land Use and Land Cover ha/year % per year ha/year % per year
Agricultural land 2986.80 2.01 4591.24 2.57 Bare land 60.09 0.22 357.53 1.26 Wood land -4134.18 -1.55 -4170.99 -1.85 Shrub land 1182.45 1.22 -52.36 -0.05 Grass land -74.84 -0.08 -716.13 -0.78 Water body -20.33 -2.93 -9.29 -1.89
On the contrary, grass land and wood land had decreased from 1985 to 1995 at the rate of
74.84 ha/year and 4,134.18 ha/year respectively and further decreased in 2010 with the rate of 716.13 and 4170.99 ha/year.
4.4.1.1.3. Accuracy Assessment of Kafta Humera District
The accuracy assessment was conducted via a standard method. For all maps, producer accuracy, user accuracy and Kappa statistics were computed. Overall, all the three maps met the minimum 85% accuracy (Appendix 11).
54
4.4.1.2. Land Use and Land Cover of Metema District
Based on Figure 15 and Table 8, in 1985, the biggest share of the land use and land cover classes in Metema district was the shrub land which covers an area of 114,570.9 ha
(30.10%). And followed by wood land (28.46%) and agricultural land (24.14%).
The highest share of the land use and land cover classes for the year 1995 (Figure 15 and
Table 8) acquired shrub land, which was occupied 30.94% of the district total area and agricultural land accounted, about 30.76%. Water body was identified the least cover than any others land use and land cover classes from the total size of the district.
In 2010, the dominant land use and land cover classes were agricultural land (42.21%) & shrub land (24.45%) from the total area of the district and followed by wood land coverage, 63,423.09 ha (16.66%) while the least area coverage occupied by the water body was only 0.05%.
Table 8: Areas of LULC of Metema District for the Years 1985, 1995 and 2010
1985 1995 2010 Land Use Type Area (ha) (%) Area (ha) (%) Area (ha) (%) Agricultural land 91889.73 24.14 117110.16 30.76 160699.41 42.21 Bare land 26137.71 6.87 31481.91 8.27 36627.93 9.62 Wood land 108333.54 28.46 73851.48 19.40 63423.09 16.66 Shrub land 114570.9 30.10 117772.29 30.94 93076.65 24.45 Grass land 39300.93 10.32 40250.34 10.57 26651.25 7.00 Water body 444.51 0.12 211.14 0.06 198.99 0.05
55
Figure 15: LULC Map of Metema District for the Years 1985, 1995 and 2010
56
4.4.1.2.1. Land Use and Land Cover Change Matrices of Metema District
Table 9: LULCC Matrices of Metema District (1985-1995)
Change to LULC 1995 (ha) LULC Agricultural Bare Wood Shrub Grass Water Total land land land land land body Agricultural 83961.95 713.36 3545.34 1865.47 1803.61 0 91889.73 land Bare land 2254 12288.61 5000.49 5296.05 1298.56 0 26137.71 Wood land 15315.3 7163.82 40896.99 34688.43 10269 0 108333.54 Shrub land 10033 8279.07 19287.99 64670.45 12300.39 0 114570.9 Grass land 5545.91 2889.45 5120.67 11167.83 14577.07 0 39300.93 Water body 0 147.6 0 84.06 1.71 211.14 444.51 Change from LULC 1985 from Change (ha) Total 117110.16 31481.91 73851.48 117772.29 40250.34 211.14 380677.32
In the district, the land use and land cover change between 1985 and 1995 was showed an
increase in the area coverage of agricultural land, bare land and shrub land (Table 9).
Whereas, the reduction in the wood land area was dramatic and was about 34,482.06 ha
(Figure 16).
30
20
10 0 -10
Area in Thousandsha in Area -20
-30
-40 Agricultural land Bare land Wood land Shrub land Grass land Water body
Figure 16: LULCC (1985-1995) of Metema District
57
Table 10: LULCC Matrices of Metema District (1995-2010)
Change to LULC 2010 (ha) LULC Agricultural Bare Wood Shrub Grass Water Total land land land land land body Agricultural 111568.87 431.98 2591.22 625.6 1892.49 0 117110.16 land Bare land 3133.84 16245.77 2889.18 6779.79 2433.33 0 31481.91 Wood land 27115.11 5172.22 31017.43 5581.69 4965.03 0 73851.48 Shrub land 12968.4 9897.84 20834.6 65732.59 8338.86 0 117772.29 Grass land 5913.19 4877.55 6090.66 14352.53 9016.41 0 40250.34 Water body 0 2.57 0 4.45 5.13 198.99 211.14 Change from LULC 1995 from Change (ha) Total 160699.41 36627.93 63423.09 93076.65 26651.25 198.99 380677.32
Figure 17 and Table 10 show the land use and land cover matrices of the district between
1995 and 2010. In this study period, like other districts, agricultural land (43,589.25 ha)
has increased in the expense of other classes from the original extent and occupies the
biggest area in Metema district and followed by the bare land (5,146.02 ha) class. Whereas,
shrub land, grass land and wood land classes were lost their original extents and
transformed to other classes.
50 40
30 20 10
0 -10 Thousandsha in Area -20
-30 Agricultural land Bare land Wood land Shrub land Grass land Water body
Figure 17: LULCC (1995-2010) of Metema District
58
4.4.1.2.2. Rate of Land Use and Land Cover Change in Metema District
Table 11 demonstrates the rate of changes in land use and land cover classes for the whole period of the study. The total wood land area between the years 1985 and 1995 had been shrunken due to its clearance and mainly transformed to agricultural land. As a result, in the first study period (1985-1995), the annual rate of wood land was estimated about
3,448.21 ha/year. In a similar manner, the area coverage of wood land in the second study period (1995-2010) was continuing to decline at a rate of 695.23 ha/year. Remarkably, the class shrub land also was transformed into other LULC classes by the rate of 1,646.38 ha/year.
Table 11: Rate of Changes in LULC Classes (1985-2010)
1985 to 1995 1995 to 2010 Land Use and Land Cover ha/year % per year ha/year % per year
Agricultural land 2522.04 2.74 2905.95 2.48 Bare land 534.42 2.04 343.07 1.09 Wood land -3448.21 -3.18 -695.23 -0.94 Shrub land 320.14 0.28 -1646.4 -1.40 Grass land 94.94 0.24 -906.61 -2.25 Water body -23.34 -5.25 -0.81 -0.38
On the other hand, during the first period of the study (1985-1995), the total area dominated by the class agricultural land was increasing by an extent of 2,522.04 ha/year.
In addition, it was also observed that this increase of agricultural land was accelerated by
2,905.95 ha/year between the year 1995 and 2010. The grass land class extent has been fluctuated in terms of changes. The expansion of grass land during the period of 1985 to
1995 has been recognized and the estimated size of the land added to be in this class was
59
94.94 ha/year, but during the second period (1995 to 2010) there has been a reduction in the rate of 906.61 ha/year and was transformed to other LULC type.
4.4.1.2.3. Accuracy Assessment of Metema District
Like the previous district, the accuracy assessment of all the images for Metema district was conducted. For all maps, the producer accuracy, user accuracy and Kappa statistics were computed. Overall, the maps met the minimum 85% accuracy (Appendix 12).
4.4.1.3. Land Use and Land Cover of Sherkole District
In the district, in 1985, wood land is the biggest share of the land use and land cover classes, which occupied an area of 148,055.67 ha (42.08 %) followed by shrub land
(32.90%) and grass land (13.92%) (Figure 18 and Table 12). The least area coverage are bare land, agricultural land and water body with an area coverage of 25,394.4 ha (7.22%),
12,431.34 ha (3.53%) and 1,245.15 ha (0.35%), respectively.
The dominant land cover in the year 1995 includes wood land (38.64%) and shrub land
(33.84%). Whereas, the coverage of agricultural land and water body were 4.76% and
0.36%, respectively are the least area coverage’s.
In 2010, the largest part of land use and land cover classes were shrub land (30.86%) and wood land (24.39%). On the other hand, agricultural land (6.89%) and the water body
(0.52%) were the least area coverage from the total extent of the study area.
60
Figure 18: LULC Map of Sherkole District for the Years 1985, 1995 and 2010
61
Table 12: Areas of LULC of Sherkole District for the Years 1985, 1995 and 2010
1985 1995 2010 Land Use Type Area (ha) (%) Area (ha) (%) Area (ha) (%) Agricultural land 12431.34 3.53 16749.09 4.76 24246.27 6.89 Bare land 25394.4 7.22 36622.35 10.41 65745.63 18.69 Wood land 148055.67 42.08 135946.89 38.64 85800.87 24.39 Shrub land 115767.9 32.90 119077.74 33.84 108567 30.86 Grass land 48962.61 13.92 42207.3 12.00 65670.3 18.66 Water body 1245.15 0.35 1253.7 0.36 1827 0.52
4.4.1.3.1. Land Use and Land Cover Change Matrices of Sherkole District
Table 13 and 14 demonstrated the results of change detection matrices that depict what
land use and land cover class is changed to which land use and land cover. In this regard
image differencing was applied. The diagonal values of the table show the unchanged area
coverage of that particular land use and land cover during the two periods.
Table 13: LCLCC Matrices of Sherkole District (1985-1995)
Change to LULC 1995 (ha) LULC Agricultural Bare Wood Shrub Grass Water Total land land land land land body Agricultural 9486.89 364.41 943.23 487.19 1149.62 0 12431.34 land Bare land 306.92 4080.22 9387.72 8539.92 3079.62 0 25394.4 Wood land 4422.85 13719.96 89763.66 23063.42 17085.78 0 148055.67 Shrub land 1136.1 12557.43 19359.42 68284.8 14421.6 8.55 115767.9 Grass land 1396.33 5900.33 16492.86 18702.41 6470.68 0 48962.61 Water body 0 0 0 0 0 1245.15 1245.15 Change from LULC 1985 from Change (ha) Total 16749.09 36622.35 135946.89 119077.74 42207.3 1253.7 351857.07
In the first period of the study (1985-1995), 2,944.45 ha of the agriculture land area was
transformed its original extent and at the same time gained about 7,262.2 ha from other
classes (more from wood land, 4,422.85 ha) and as a result the class was increased by
4,317.75 ha. In contrast, wood land transformed its 58,292.01 ha (more to shrub land
62
23,063.42 ha) of its original coverage and simultaneously gained 46,183.23 ha from other
classes (more from shrub land, 19,359.42 ha) but not totally compensated at least its lost
area, 12,108.78 ha.
15 10 5 0 -5
Area in Thousandsha in Area -10
-15 Agricultural land Bare land Wood land
Shrub land Grass land Water body
Figure 19: LULCC (1985-1995) of Sherkole District
Table 14: LULCC Matrices of Sherkole District (1995-2010)
Change to LULC 2010 (ha) LULC Agricultural Bare Wood Shrub Grass Water Total land land land land land body Agricultural 13360.4 697.44 979.1 571.17 1140.99 0 16749.09 land Bare land 760 9173.48 6002.1 12468.6 8218.17 0 36622.35 Wood land 5060.19 26305.3 46139.9 37397.16 21044.3 0 135946.89 Shrub land 1955.58 21831.6 24024.3 46280.69 24737.5 248.13 119077.74 Grass land 3110.11 7737.85 8655.48 11849.38 10529.3 325.17 42207.3 Water body 0 0 0 0 0 1253.7 1253.7 Change from LULC 1995 from Change (ha) Total 24246.27 65745.63 85800.87 108567 65670.3 1827 351857.07
Similar to the first period, in the second period of the study (1995-2010), agricultural land
gained a remarkable land area (10,885.88 ha) from other classes and at the same time lost
3,388.7 ha (Table 14) to other classes and the balance was an increment of the land area by
63
7,497.18 ha. Whereas, wood land transformed more of its coverage (89,806.98 ha) to other classes than the area gained (39,660.96 ha).
40
20
0
-20
Area in Thousands ha in Area -40
-60
Agricultural land Bare land Wood land
Shrub land Grass land Water body
Figure 20: LULCC (1995-2010) of Sherkole District
4.4.1.3.2. Rate of Land Use and Land Cover Change in Sherkole District
Table 15 shows the rate of changes in land use and land cover classes. The wood land area between the years 1985 and 1995 had been reduced by an annual rate of change, about
1,210.88 ha/year. And in the second study period (1995-2010) the wood land was continued to decline rapidly at a rate 3,343.07 ha/year. Amazingly, bare land area has been increase from other LULC classes at a rate of 1,122.80 ha/year and the class continued at a rate of 1,941.55 ha/year, in the second period.
On the other hand, shrub land area was fluctuated, in the first period, it was increasing at a rate of 330.98 ha/year but in the second unfortunately it was reduced at a rate of 700.72 ha/year. The extent of grass land has been also fluctuated in terms of rate of changes, in the first period, the reduction of grass land has been recognized at a rate of 675.53 ha/year but
64
during the second period (1995 to 2010) there has been increasing at the rate of 1,564.2 ha/year.
Table 15: Rate of Changes in LULC Classes (1985-2010)
1985 to 1995 1995 to 2010 Land Use and Land Cover ha/year % per year ha/year % per year
Agricultural land 431.78 3.47 499.81 2.98
Bare land 1122.80 4.42 1941.55 5.30
Wood land -1210.90 -0.82 -3343.10 -2.46
Shrub land 330.98 0.29 -700.72 -0.59
Grass land -675.53 -1.38 1564.20 3.71
Water body 0.86 0.07 38.22 3.05
4.4.1.3.3. Accuracy Assessment of Sherkole District
To assess the classification accuracy, confusion matrix was used. For all maps, the producer accuracy, user accuracy and Kappa statistics were computed. As usual, all the maps met the minimum 85% accuracy (Appendix 13).
65
5. DISCUSSIONS
5.1. Vegetation Characterstics
The structure of wood land forest is consistent with other studies on dry forests. For example, the observed range of basal area (BA) is comparable with the ranges (3.84-10.36 m2 ha -1) reported by Singh and Singh (1991) and 6.58-23.21 m2 ha -1 reported by Jha and
Singh (1990) for several dry tropical communities in India. Similarly, the overall population structure of woodlands can help understand the status of the forest stand
(Tesfaye et al ., 2010; Worku et al ., 2012). Reverse J-shaped distributions such as those shared by Group I species in the Gemed site indicate more or less a healthy or stable regeneration. In contrast, bell-shaped distributions (Group II species of Adi-Goshu and
Lemlem-Terara sites) suggest hampered regeneration (Belayneh and Demissew, 2011).
Despite indications of hampered regeneration, most of the species in Group II had a considerable number of individuals in the middle diameter classes that could be managed sustainably to improve their regeneration and produce gums and resins, in addition, the potential of carbon sequestration would enhance.
Comparison of the ranges of tree diameters with respect to the above ground biomass accumulation revealed that tree species with lower range of diameter possess more density but accumulated less biomass. On the other hand, trees having bigger diameters were few in number but accumulated more biomass. Therefore, an inverse relationship was seen between tree density and DBH whereas a direct relationship was observed between the above ground biomass and DBH. In this regard, the findings from Terakunpisut et al .
(2007) and Juwarkar et al . (2011), indicate similar results with the current study. Trees during their initial stages of growth i.e. when their DBH is smaller could sequester less carbon but gradually increases in DBH and would accumulate more carbon. Moreover, it
66
has been observed the younger trees grow much faster as compared to older ones. Thomas
(1996) suggested that fast growing tree species are expected to have higher growth rates, and may accumulate large amounts of carbon in the first stage of their lifespan while the high specific gravity of slower-growing species allows them to accumulate more carbon in the long-term.
5.2. Carbon Stocks in Different Carbon Pools
5.2.1. Aboveground Carbon Stock
The estimation mean aboveground biomass carbon stock for Adi Goshu (16.71-19.29 Mg ha -1), Lemlem Terara (26.59-27.91 Mg ha -1) and Gemed (25.87 Mg ha -1) sites was in agreement with the reported result (14.66-43.22 Mg ha -1) in Chaiyo et al . (2011) in the tropical deciduous forests. In addition, the overall range of the aboveground carbon stock was within the range of 14-123 Mg ha -1 (Murphy and Lugo, 1986). The aboveground biomass of the current study sites was estimated by the allometry correlation with DBH.
Consequently, the most of the finding are in agreement with other studies. Moreover, the variation of carbon stock in aboveground dependent on many factors such as the stand structure and composition, topography, altitude, disturbance, forest fire and fuel wood collection, micro climate.
The above ground carbon stock in Lemlem Terara (27.91 Mg ha -1) was higher than in Adi
Goshu (16.71 Mg ha -1) site in the untapped Boswellia papyrifera woodland. Firstly, the variation is perhaps due to the considerable differences in the presence and densities of the field trees. Hence, the tree density and basal areas was 379.81 tree/ha and 11.59 m 2/ha respectively in Lemlem Terara site and 259.50 tree/ha and 7.17 m 2/ha in Adi Goshu site.
And it is possible to suggest that the more protected woodlands could enhance the ecosystem carbon stock. Secondly, a difference in carbon stock amount was also observed
67
because of the accumulation of litter fall in Lemlem Terara site than the other, as high litter accumulation may maintain soil moisture reserve that could promote tree biomass productivity. This is also explained by the decrease in soil erosion rate and the increase in the overall species diversity and aboveground biomass within Lemlem Terara site.
5.2.2. Belowground Carbon Stock
In the findings the estimated mean belowground carbon stock of untapped Boswellia papyrifera woodlands for Lemlem Terara site (6.98 Mg ha -1) was significantly higher than that of the Adi Goshu site (4.18 Mg ha -1). This was because, trees have much more potential to produce larger quantities of belowground biomass compared to shrubs or herbs. As revealed by Lemma et al . (2007), more biomass production increased the aboveground litter and the belowground root activity and these make trees are an important factor for SOC.
Therefore, roots are a key component of the belowground system, and they are a main source of SOM, which are influencing the soil microbial activity and decomposition processes (Cheng, 1999; Janssens et al ., 2002). On a mass basis, coarse roots contribute more to the total ecosystem biomass than the fine roots. And when establishing carbon budgets and carbon allocation at an ecosystem level, coarse root biomass and production data should be collected (Vogt et al ., 1998). However, coarse roots account for only a small portion of annual root production. In contrast, fine roots have a high turnover rate despite the fact that they account for only a small fraction of total root biomass (Chen et al ., 2004). Due to their high metabolic activity, fine root production and turnover plays a critical role in forest carbon dynamics. Further, fine roots are an important structural and functional component of forested ecosystems, representing an important carbon input to the soil and can equal or exceed aboveground detritus production in some systems (Vogt,
68
1991). A large proportion of forest production is allocated to fine roots, resulting in a large carbon flux into the belowground system (Kurz et al ., 1996).
5.2.3. Dead Wood Carbon Stock
Dead woods are an important component of the carbon pool in many forests. The mean dead wood carbon stock of untapped Boswellia papyrifera woodland stratum in the
Lemlem Terara (2.89 Mg ha -1) was significantly greater than Adi Goshu (0.48 Mg ha -1) sites because of the extent of protection and there was removal of the dead wood for fire wood in Adi Goshu site. The overall dead wood carbon stock fall within the range of the reported result (1.2-3.3 Mg ha -1) for the drier life zones in tropical forests (Delaney et al .,
1998).
In addition, deadwood dynamics are closely related to forest management, such as harvesting operations. In the current study, the findings indicated that the carbon stock for the deadwood is generally smaller. Sakai et al . (2008) suggested that the warm and humid climate induces quick decomposition of dead-wood. This may result in low accumulation of deadwood carbon in the lowland woodland of Ethiopia.
5.2.4. Herbaceous and Litter Carbon Stock
The overall result of the herbaceous biomass carbon stocks of the present study were in line with the study by Roshetko et al . (2002), which ranges between 0.1 and 1.4 Mg ha -1.
Herbaceous plants contribute to the soil carbon sequestration. However, they are shorter- lived than tree which can survive as high as 100 years or more. Whereas, in the Lemlem
Terara site, the untapped and tapped Boswellia papyrifera woodlands had the mean litter biomass carbon stock of 0.61 and 0.46 Mg ha -1, respectively. Hence, the results indicated that low quantity carbon stocks when compared to the reported carbon by Hirai et al .
(2006) that was undertaken in the widely planted coniferous species litter biomass carbon
69
stocks (ranges 3.11 to 4.35 Mg ha -1). The other study conducted by Roshetko et al . (2002) in the homegarden system of Indonesia reported the mean litter biomass carbon stock was
2.0 Mg ha -1, which is greater than the findings of the current study because the litter fall of the homegarden system was utilized for soil fertility management. In most parts of
Ethiopia tree litter layers are cleared for fuel wood; and this may explain the relatively lower carbon stock observed in the site.
This litter decomposition is one of the major sources of soil organic carbon and the decomposition process was dependent on the quality of the litter fall (Mafongoya et al .,
1998) and the plant species (Lemma et al ., 2007). Litter decomposition rates are also frequently considered to be regulated by soil organisms, environmental conditions and chemical nature of the litter (Gallardo and Merino, 1993). The physical environment, especially soil moisture, temperature and relative humidity are important in litter decomposing as these regulate the biological activity in soil (Sayer, 2006). The climatic condition in the study areas is very hot and humid during the rainy season; and this might have been causing high rate of litter decomposition leading to relatively low litter accumulation at the ground surface beneath the tree canopy.
5.2.5. Soil Carbon Stock
The availability of the mean soil carbon stock in Lemlem Terara site is high in the untapped (58.19 Mg ha -1) than the tapped (34.25 Mg ha -1) Boswellia papyrifera woodland.
This variation was due to the disturbance in the tapped Boswellia papyrifera woodlands such as a free grazing and this results poor soil organic matter formation due to removal of the litter fall (Tschakert, 2004).
In conformity with carbon stock distribution, the amount of soil carbon stored in tropical dry forest was indicated those areas with more aboveground carbon density could also hold
70
high carbon density in the soil. With this regard, the Lemlem Terara site was one of the study sites that possessed more aboveground biomass carbon stocks of 27.91 Mg ha -1 in the untapped Boswellia papyrifera woodland while, 58.19 Mg ha -1 was estimated the mean soil carbon stock. These results are more or less similar with the study reported by Mekuria et al . (2009). And the above argument was supported by the significant positive correlation between the soil carbon and the aboveground biomass and cover of carbon stock. On the other hand, other related studies were reported the increment of soil carbon stocks in the ecosystem as the number of plant species and aboveground biomass increases (Solomon et al ., 2002; Lemenih and Itanna, 2004; Lemenih et al ., 2005). Therefore, trees having more aboveground and belowground biomass contribute more to the soil carbon sequestration compared to shrubs or herbs. The accumulation of soil organic carbon also depends on the quantity of litter (Lemma et al ., 2007) and root activity such as rhizo-deposition and decomposition (Rees et al ., 2005).
In addition, Solomon et al . (2002) was revealed that land use and land cover change could affect the amount and composition of soil organic matter through their influence on decomposition and humification processes. Moreover, the present findings agree with Lal
(2001) who was reported the increasing soil organic carbon pool with the addition of biomass to soils when the pool has been depleted as a consequence of land use and land cover change. Proper forest management could increase the aboveground vegetation carbon density and this possibility to enhance the carbon storage in the soil. Degraded soils on the restoration practices could have the potential to provide terrestrial carbon sinks and reduce the rate of enrichment of atmospheric CO 2 (Marks et al ., 2009).
Besides, in Lemlem Terara site, high litter fall accumulation and sufficient organic matter inputs (Tschakert, 2004) were investigated. However, in Adi Goshu site, the soil organic
71
matter was one of the poorest pools. According to some findings, soil carbon is highly affected by soil chemistry and physical soil characteristics through disturbances (Nave et al ., 2010; Powers et al ., 2011). Sagar et al . (2008) also argued that disturbances control the soil quality mainly due to the biomass removal that is limiting the amount of organic matter inputs into the soil. Evidently, the degree of soil fertility was reflected in the aboveground carbon density.
5.2.6. Total Carbon Stock Density
Plants capture CO 2 from the atmosphere through the process of photosynthesis. And the standing vegetation, most of which is eventually added to the soil as plant organic litter and then to the soil as SOC by microbial activity. Therefore, the estimation of carbon stock both in the aboveground and in soil becomes imperative to assess the carbon sequestration potential (Ramachandran et al ., 2007).
In the Adi Goshu site, the estimated ratios between the mean SOC and biomass carbon were 2.01 and 1.99 for untapped and tapped Boswellia papyrifera woodlands respectively.
Similarly, in the Lemlem site, the ratio between the SOC and biomass carbon stock were
2.09 for the untapped and 1.29 for the tapped Boswellia papyrifera woodlands. On the other hand, in the Gemed site where only exist tapped Boswellia papyrifera woodland, the estimated mean ratios between SOC and biomass carbon stock were 1.42.
In the case of the present study, the carbon content in the soil was higher than the above- ground biomass carbon due to extensive expansion of agricultural land by shrinking the area coverage of the woodlands. Higher content of SOC than the above ground biomass carbon suggested that the sequestered SOC might have been originated from the decomposition roots of the original vegetation that had been there before the new land use and land cover; indicating more recent changes in land use and land covers. The present
72
study also indicates the SOC is higher than that of the biomass carbon but not 2.5 to 3 times of biomass carbon as recorded in well-managed terrestrial ecosystems Post et al .
(1990). However, in the tropical forest, the carbon in the soil is roughly equivalent or lesser than the above ground biomass due to soil and vegetation degradation (Walkley and
Black, 1934). In turn degradation is occurred because of the land use and land cover changes over time with the need-based forestry practices such as fuel wood, shifting cultivation, overgrazing and overall anthropogenic disturbances. Therefore, these mining practices cause the reduction of tree densities and carbon-storing capacity (Vishwanathan et al ., 1999).
5.3. Value and Functions of Gum and Resin
Ethiopia is one of the major producers and exporters of the natural gum and resins. Oleo- gum resins, such as frankincense, have been items of the great historical commerce in the
Horn of Africa, in general, and in Ethiopia, in particular, (Fitwi, 2000; Lemenih and
Teketay, 2003). Boswellia papyrifera is known to occur in various parts of Ethiopia including Tigray, Amhara and Benshangul Gumuz regions (Demissew, 1993). In addition to this there are different acacia species ( Acacia seyal, Acacia tortilis, Acacia abyssinica , etc) which are the main source of the different gum and resin products. Tapping of frankincense provides not only considerable cash income and employment opportunity
(Gebremedhin, 1997; Lemenih and Teketay, 2003), but also fetches valuable foreign currency.
Boswellia papyrifera grows on dry and rocky sites where other tree species often fail
(Gebrehiwot et al., 2005). In northern Ethiopia, where the majority of the soils (60–80%) are about 20 cm deep (Hurni, 1988), Boswellia papyrifera trees are found on steep slopes with an average gradient of 30-40%. The species makes economic use of the marginal
73
areas on which other species cannot grow. In these sites it provides plant cover and produces high biomass and hence protects the soil and provides shade. When Planted this species can economically and socially improve the overall value of a degraded area.
Incense collection offers off-farm employment for many farmers. In western Tigray alone, annually about 7000 seasonal laborers are employed; among which 31% are women. Men are mainly involved in tapping and collecting incense from the forest while women undertake sorting and grading of the product. A tapper can collect about 10-15 quintals of incense per annum and receives a net income of US $ 100 to 150 (Berhe, 1997). Women accrue an average income of US$ 16 per month (Gebremedhin, 1997).
Incense is being burnt in many churches in worldwide and used as oil extract for a number of applications, such as modern perfumery, traditional medicine, pharmaceuticals, fumigation powders, fabrication of varnishes, adhesives, painting, and chewing gum industries. It also gives a flavor in food industry, e.g. bakery, milk products, different alcoholic and soft drinks (Tucker, 1986; Coppen, 1995; Deffar, 1998). Therefore, Ethiopia will be more benefited from the export of these products, provided efforts are made to develop and manage these resources in a sustainable way.
5.4. Land Use and Land Cover Change
Change in land use and land cover may result in land degradation that manifests itself in many ways depending on the magnitude of changes. For example, woody vegetation, which may provide fuel and fodder, becomes increasingly scarce, water courses are drying up, and soils become thin and stony. All of these manifestations have potentially severe impacts on land users and people who rely for their living on the products from a healthy landscape.
74
The change detected in Kafta Humera, Metema and Sherkole districts in accordance with the classified classes between 1985 and 2010 revealed that wood land, shrub land and grazing land covers were forced to be transformed into agricultural land and degraded land.
The latter definitely showed how changes in land use and land cover aggravate land degradation. The land use and land cover change observed in the study area has a negative impact on both the environment and socio-economic settings. For instance, the expansion of cultivated land in the study area was at the expense of grass land, shrub land and wood land. Due to the expansion of farmland on the steep slopes, land has greatly deteriorated and degraded.
As a result, there was a dramatic expansion of agricultural land within the specified time period because of population pressure and poor land administration. The expansion of agricultural land between 1985 and 2010 in the districts in general, could be directly related to rapid population growth and resettlement programs. In kafta humera district the agricultural land covers the largest area in 2010 which shows conversion of other land cover classes to agricultural land. As shown in (Table 4), 39.11% of the area was covered by agricultural land in 2010 due to the expansion of agricultural activities and settlement programs areas in 2003 in to woods, shrubs and grass lands. There was also increase agricultural land from time to time in the Metema district due to the new settlers from
Wollo areas, since the Wollo areas was affected in drought in 1985.
Similarly, the shrub land was reduced in size in between 1995 to 2010 with rate of 52.36 ha/year (Kafta Humera), 1,646.4 ha/year (Metema) and 700.72 ha/year (Sherkole). This is because of the degraded wood lands were changed to shrub lands. Whereas, bare land was continuously increased in between 1985 to 2010 in the three districts, because of the new settlers lose of much vegetation for infrastructure and fire wood purpose. Susceptibility to
75
forest degradation is understood that the forest resources can be influenced or degraded by human activities. In reality, forest resources are degraded not only by human activities but also due to other natural factors too. However, in this research human activities were taken in to consideration, because the unplanned actions such as illegal logging, exploitation of forest resources for fuel wood and charcoal production as well as expansion of agricultural lands are the main factors that cause forest degradation. Generally, the average rate of the agricultural expansion from 1985 to 2010 was between 1,980.21 ha/year and 2,665.67 ha/year.
5.4.1. The Causes of Wood Land Cover Change
Forest cover change is triggered by various factors that undermine the forest use potential and its productivity which leads to irreversible deterioration. Besides, forest cover change is the direct reflection of the dynamics of socio-economic development. Likewise, several factors stimulated by the activity of man are responsible for massive conversion of forest land cover into other land forms and land use units in the three study sites (Kafta Humera,
Metema and Sherkole) districts. All the factors of deforestation such as the prevalence of various types of agricultural activities, fire wood and charcoal production, cutting trees to fulfill the demand of constructional materials, settlement expansion and income generation are directly or indirectly related to population growth and new settlements.
The agro ecological condition of the districts is convenient for agriculture. Due to this, crop production and livestock rearing is the basic economic activity in the districts especially in Metema and Kafta Humera. Most of the farmers rear livestock and want to maintain large number with little care of their quality. According to the informants, the larger number of cattle population in a given family is both a source of wealth and status.
76
Indeed, this mental attitude is not limited to the study area and has been prevalent throughout Ethiopia.
The land use and land cover change results in the study areas indicated that there is an increasing trend of agricultural land from 1985 to 2010. The implication of increased agricultural land in terms of aerial coverage means other land use and land cover classes have been converted into agricultural lands. For instance, between 1985 and 1995, about
2,986.80 ha/year in Kafta Humera, 2,522.04 ha/year in Metema and 431.78 ha/year of wood land coverage drastically changed into agricultural lands. In addition, according to the views of respondents, the expansion of various types of agricultural activities is the major sources of forest cover change in the study area. Therefore, the presence of farmers with their various types of agricultural activities (both crop production and livestock rearing) inside and along the margin of the districts wood land’s coverage and considered to be the major factor for forest cover change in the study area.
The expansion of agricultural land together with the bare land in the districts was increasing from time to time. The agricultural land expansion was in expense of other
LULC types to satisfy food demand for the ever increasing human population and new settlers. This alteration of LULC type coupled with poor land management practice in the region resulted in exposing of land for erosion hazard, which was later turned to accelerated land degradation. Most of these studies identified that deforestation and expansion of cultivation land in to marginal areas were the principal cause of land degradation (Zeleke and Hurni, 2000).
In the rural areas fire wood (collected from the nearby forest areas) and cow dung are the two most important sources of energy. According to the informants, over the recent years fire wood is commercialized as its demand has increased particularly in those areas which
77
are devoid of trees and in the urban areas of the district. Moreover, as the agricultural officers identified fire wood and charcoal productions are the major causes of forest cover change. Hence, the increasing demand of forest products, in the form of fire wood and charcoal within and outside the district has been causing of deforestation in all study area.
The demand of forest products for the construction of house and fence has been aggravated the destruction of forest in the study areas. From the respondent’s point of view, it was evident that cutting trees to fulfill the demand of constructional material is considered to be the causes of deforestation in the districts. Field observation data also indicated that woody biomass was found to be the most important house construction material in the districts. In addition, selling of wood and wood products are means of income generation for the resource poor people, especially jobless youths and women in the districts. Referring this, agricultural officers and forestry experts argued that these groups of people illegally cut down trees from the forest area so as to supply large quantity of forest products for urban and rural dwellers through the nearby markets.
78
6. CONCLUSIONS AND RECOMMENDATIONS
6.1. Conclusions
Carbon stock in different carbon pools (aboveground and belowground biomass, dead wood, litter, herbaceous biomass and soil) has a potential to decrease the rate of enrichment of atmospheric concentration of CO 2. Increase in carbon stock in woodlands can be achieved through sustainable woodland management including site preparation, species selection and afforestation. Woodlands appear to offer a relatively low-cost approach to sequestering carbon. Furthermore, the total carbon stock density in our study sites was high and ranges from 45.18 to 96.74 Mg ha-1, and hence it has a high potential for absorbing carbon dioxide from the atmosphere. And the country could contribute to the effort of global climate change mitigation. On top of this the farmers around the woodland area would be beneficial for both non-timber products (gum and resin) and the income from the carbon trading; and this increases the farmer’s adaptive capacity through diversified livelihoods.
The most serious problems with woodlands to sequester carbon would be the occurrence of large scale expansion of commercial agricultural land for sesame production. In this case, a substantial leakage problem would also be expected to increase. This potentially could reduce the feasibility of forest carbon project to attract voluntary carbon market in such areas. The result shows that agricultural and bare lands have been increased, while wood land has decreased in the area coverage in all study areas. The reason behind is that the aggravated population pressure from settlement program in the previous periods. Hence, the new settlers need shelter and food consumption, therefore agricultural expansion is inevitable and this leads to drastic deforestation in the areas.
79
Remote sensing data are very popular in land use and land cover change classification and are the most fundamental tools to know the causes of environmental risks in many cases and they are the main inputs for the land use planning. GIS has been used in establishing technical support to planning and decision making through the use of comparable, integrated maps and related data type’s. Therefore, the use of RS and GIS in decision support system for the area of resource allocation and policy decisions are a paramount importance for natural resource management.
6.2. Recommendations
The potential role of woodland in sequestrating carbon to reduce the buildup of greenhouse gases in the atmosphere is now well recognized. A number of alternative approaches to utilize woodland management for carbon sequestration can be applied. These include woodland protection from deforestation, the management of these trees for carbon trading and commercializing gum and resin products as well as establishing plantations for buffering the natural woodlands. These interventions may build a system of sustainable wood land resource management and utilization for enhancing the potential of the carbon trading and livelihood diversification in the woodlands of Ethiopia. The following points should be seriously considered for managing the woodland resources.
Adequate understanding on climate change issues and more powerful methods to
implement cost/benefit analyses of woodland-based GHG mitigation and livelihood
improvement to define incentives for wide scale adoption of the protection of the
woodlands should be implemented.
Incorporation of available knowledge on the production and management of gums
and resins. Hence, appropriate policies should exist and be implemented to use
these resources on a sustainable basis, both locally and nationally.
80
Land use and land cover change analysis is a good indicator of the trend of their dynamics. However, to use the result of the analysis for decision making; intensive socio-economic data needs to increase its efficiency and accuracy.
Creating awareness at the grass root level how to utilize the forest resource properly with respect to environmental, socio-economic benefit.
The government has to engage efficiently towards the reduction of the rate of population growth and its consequent effect on the environment and food security.
The community should be engaged in continues training and awareness on health, reproduction and the general effect of high population growth on the limited natural resources.
NGPME and the regional government should made fundamental thinking in the policy of woodland management in such a way of promoting carbon trading for additional financial incentive to the local community who are depending on the woodland resource.
81
REFERENCES
Ahadnejad, M., 1999. Environmental Land Use Change Detection and Assessment Using Multi-temporal Satellite Imagery. Zanjan University.
Albrecht, A., and Kandji, S.T., 2003. Carbon sequestration in tropical agroforestry systems. Agriculture, Ecosystems and Environment 99: 15-27.
Atlas of Woody Biomass Inventory and Strategic Planning Project (AWBISPP), 2003. The Federal Democratic Republic of Ethiopia Ministry of Agriculture. Benshangul Gumuz National Regional State. Addis Ababa, Ethiopia.
Baker, D.F., 2007. Reassessing carbon sinks. Science 316: 1708-1709.
Bale Eco-Region Sustainable Management Programme (BERSMP), 2010. REDD in Ethiopia: Opportunities, Challenges and the PFM Approach. FARM-Africa/SOS Sahel Ethiopia Participatory Natural Resources Management Unit (PNRMU), Addis Ababa, Ethiopia.
Bekele, A., 2007. Useful trees and shrubs of Ethiopia: Identification, Propagation and Management for 17 Agroclimatic Zones. Technical Manual No 6. RELMA in ICRAF Project, Nairobi, Kenya. 552 pp.
Bekele, A., Birnie, A. and Tengas, B., 1993. Useful trees and shrubs of Ethiopia: identification, propagation, and management for agricultural and pastoral communities. Regional Soil Conservation Unit, Technical Handbook No. 5. Swedish International Development Cooperation Agency, Nairobi, Kenya.
Belayneh, A, and Demissew, S., 2011. Diversity and Population Structure of Woody
Species Browsed by Elephants in Babile Elephant Sanctuary, eastern Ethiopia: an
implication for conservation. Agriculture and Forestry 3(1): 20-32.
Berhe, A., 1997. Preliminary Survey on Forest Products Utilization and Marketing in Tigray. Bureau of Agricultural Development and Natural Resources. Mekelle, Ethiopia. 67 pp.
82
Billah, M., and Rahman, G.A., 2004. Land cover mapping of Khulna city applying remote sensing technique. Geoinformatics 707-714.
Bohannon, J., 2007. From greener production to carbon trading: Sustainable energy careers. Science 315: 868–869.
Bolin, B., Doos, B. R., Jager, J., and Warrick, R., 1986. The Greenhouse Effect, Climate Change and Ecosystem, Scope 29 John Wiley, Chichester.
Brown, S., 1997. Estimating biomass and biomass change of tropical forests: a primer. FAO Forestry Paper 134. Food and Agriculture Organization of the United Nations, Rome, Italy.
Brown, S., 2002. Measuring carbon in forests: current status and future challenges. Environmental Pollution 116: 363-372.
Brown, S., Sathaye, J., Cannel, M. and Kauppi, P., 1996. Management of forests for mitigation of greenhouse gas emissions. In Watson, R.T., Zinyowera, M.C., and Moss, R.H., eds. Climate change 1995, impacts, adaptations and mitigation of climate change: scientific-technical analyses. Report of Working Group II, Assessment Report, IPCC. Cambridge, UK, Cambridge University Press. pp 773- 797.
Brown, S., Swingl, J.R., Tenison, R.H., Prance, G.T., and Myers, N., 2002. Changes in the use and management of forests for abating carbon emissions: issues and challenges under the Kyoto Protocol. Philos. Trans. R. Soc. Lond. A 360: 1593-1605.
Cairns, M.A, Brown, S., Helmer, E.H., and Baumgardner, G.A., 1997. Root biomass allocation in the world's upland forests. Oeclogia 111(1): 1-11.
Canadell, J.G., and Raupach, M.R., 2008. Managing Forests for Climate Change Mitigation. Science 320: 1456-1457.
Central Statistical Agency (CSA), 2007. Agricultural Sample Survey. Report on Area and Production - Tigray Region. Version 1.1.
83
Central Statistics Agency (CSA), 1995. The 1994 Population and Housing Census of Ethiopia. Federal Democratic Republic of Ethiopia Office of Population and Housing Census Commission central Statistical Authority, Addis Ababa, Ethiopia.
Central Statistics Agency (CSA), 2008. Summary and Statistical Report of the 2007 Population and Housing Census. Population Size by Age and Sex. Federal Democratic Republic of Ethiopia Population Census Commission, Addis Ababa, Ethiopia.
Chaiyo, U., Garivait, S., and Wanthongchai, K., 2011. Carbon Storage in Above-Ground Biomass of Tropical Deciduous Forest in Ratchaburi Province, Thailand. World Academy of Science, Engineering and Technology 58: 636-641.
Chaturvedi, R.K., Raghubanshi, A.S., and Singh, J.S., 2011. Carbon density and accumulation in woody species of tropical dry forest in India. Forest Ecology and Management 262: 1576-1588.
Chave, J., Andalo, C., Brown, S., Cairns, M.A., Chambers, J.Q., Eamus, D., Folster, H., Fromard, F., Higuchi, N., Kira, T., Lescure, J.P., Nelson, B.W., Ogawa, H., Puig, H., Riera, B., Yamakura, T., 2005. Tree allometry and improved estimation of carbon stocks and balance in tropical forests. Oecologia 145: 87-99.
Chen, X., Eamus, D., and Hutley, L.B., 2004. Seasonal patterns of fine-root productivity and turnover in a tropical savanna of northern Australia. Journal of Tropical Ecology 20: 221-224.
Cheng, W., 1999. Rhizosphere feedbacks in elevated CO 2. Tree Physiology 19: 313-320.
Congalton, R., 1991. A Review of Assessing the Accuracy of Classifications of Remotely Sensed Data. Remote Sensing of Environment 37: 35-46.
Coppen, J., 1995. Flavors and Fragrances of plant origin. Food and Agriculture Organization of the United Nations (FAO). Rome, Italy.
Deffar, G., 1998. Non-wood forest products in Ethiopia. EC-FAO Partnership program (1998-2000), Addis Ababa, Ethiopia. 15 pp.
84
Delaney, M., Brown, S., Lugo, A.E., Torres-Lezama, A., Bello Quintero, N., 1998. The quantity and turnover of dead wood in permanent forest plots in six life zones of Venezuela. Biotropica 30: 2-11.
Demissew, S., 1993. The Ethiopian herbs: their aromatic and medicinal uses. Selamta , 13: 16-22.
Demissew, S., 1996. A description of some essential oil bearing plants in Ethiopia and their indigenous uses. Journal of Essential Oil Research 5: 465-478.
Diakoulaki, D., Georgiou, P., Tourkolias, C., 2007. A multi-criteria approach to identify investment opportunities for the exploitation of the clean development mechanism. Energy Policy 35: 1088-1099.
Dixon, R.K., 1995. Agroforestry systems: Sources or sink of greenhouse gases? Agroforestry Systems 31: 99-116.
Dixon, R.K., Brown, S., Houghton, R.A., Solomon, A.M., Trexler, M.C., Wisniewski, J., 1994. Carbon pools and flux of global forest ecosystems. Science 263: 185-190.
Dossa, E.L., Fernandes, E.C.M., Reid, W.S., and Ezui, K., 2008. Above and belowground biomass, nutrient and carbon stocks contrasting an open-grown and a shaded coffee plantation. Agroforest Syst 72:103-115.
EMA (Ethiopian Mapping Authority), 1988. National Atlas of Ethiopia . EMA, Addis Ababa, Ethiopia. 76 pp.
Eshete, A. Sterck, F. and Bongers, F., 2011. Diversity and production of Ethiopian dry woodlands explained by climate and soil stress gradients. Forest Ecology and Management 261: 1499-1509.
Eshete, A., Sterck, F.J., and Bongers, F., 2012. Frankincense production is determined by tree size and tapping frequency and intensity. Forest Ecology and Management 274: 136-142.
Eshete, A., Teketay, D., and Hulten, H., 2005. The socio-economic importance and status of populations of Boswellia papyrifera (Del.) Hochst in north Gondar zone. Forests, Trees and Livelihoods 15: 55-74.
85
Eshete, A., Teketay, D., Lemenih, M., and Bongers, F., 2012. Effects of resin tapping and tree size on the purity, germination and storage behavior of Boswellia papyrifera (Del.) Hochst. Seeds from Metema District, northwestern Ethiopia. Forest Ecology and Management 269: 31-36.
FAO, 2000. The digital soil and terrain database of east Africa (sea) notes on the Arc/Info files Version 1.
FAO, 2005. Global Forest Resource Assessment. FAO Forestry Paper 147. Food and Agriculture Organization of the United Nations, Rome, Italy.
Fitiwi, G., 2000. The Status of Gum Arabic and Resins in Ethiopia. Report of the Meeting of the Network for Natural Gum and Resin in Africa (NGARA) 29 th - 31 th May, Nairobi, Kenya. pp 14-22.
Franklin, J.F., Shugart, H.H., and Harmon, M.E., 1987. Tree death as an ecological process. BioScience 37:550-556 .
Gallardo, A., and Merino, J., 1993. Leaf decomposition of two Mediterranean ecosystems of South West Spain: Influence of substrate quality. Ecology 74: 152-161.
Gautam, A.P., Webb, E.L., Shivakoti, G.P., and Zoebisch, M.A., 2003. Land use dynamics and landscape change pattern in a mountain watershed in Nepal. Agriculture, Ecosystems and Environment 99: 83-96.
Gebrehiwot, K., Muys, B., Haile, M., and Mitloehner R., 2002. Boswellia papyrifera (Del). Hochst: a tropical key species in northern Ethiopia. Conference on International Agricultural Research for Development. October 9-11, Deutscher tropentag. Germany.
Gebrehiwot, K., Muys, B., Haile, M., and Mitloehner, R., 2005. The use of plant water potential to characterize tree species and sites in dry lands of northern Ethiopia. Journal of Arid Environments 60: 581-592.
Gebremedihn, T., 1997. Boswellia papyrifera (Del.) Hochst. from western Tigray: opportunities, constraints and seed germination responses. Ethiopian MSc in Forestry Program thesis works. Report No. 1996: 12. Swedish University of Agricultural Sciences. Faculty of Forestry. Sweden. 58 pp.
86
German Bundestag, 1990. Protecting the Tropical Forests: A High Priority International Task. Bonn University-Printing Office, German.
Hairiah, K., Sitompul, S.M., van Noordwijk, M., and Palm, C., 2001. Carbon Stocks of Tropical Land Use Systems as Part of the Global Carbon Balance: Effects of Forest Conversion and Options for ‘Clean Development' Activities. ICRAF , Indonesia. 49 pp.
Haripriya, G.S., 2002. Biomass carbon of truncated diameter classes in Indian forests. Forest Ecology and Management 168: 1-13.
Harmon, M.E., Franklin, J.F., Swanson, E., Sollins, P., Gregory, S.V., Lattin, J.D., Anderson, N.H., Cline, S.P., Aumen, N.G., Sedell, J.R., Lienkaemper, G.W., Cromack, K., and Cummins, K.W., 1986. Ecology of Coarse Woody Debris in Temperate Ecosystems. Advance in Ecological Research 15: 133-302.
Harmon, M.E., Sexton, J., 1996. Guidelines for Measurements of Woody Detritus in Forest Ecosystems (US LTER Publication No. 20). US LTER Network Office, University of Washington, Seattle, WA, USA.
Henry, H., Tittonell, P., Manlay, R.J., Bernoux, M., Albrecht, A., and Vanlauwe, B., 2009. Biodiversity, carbon stocks and sequestration potential in aboveground biomass in smallholder farming systems of western Kenya. Agriculture, Ecosystems and Environment 129: 238-252.
Hirai, K., Sakata, T., and Morishita, T., 2006. Characteristics of nitrogen mineralization in the soil of Japanese cedar (Cryptomeria japonica) and their responses to environmental changes and forest management. J. Jpn. For. Soc. 88: 302-311.
Houghton, R. A., 2005. Aboveground Forest Biomass and the Global Carbon Balance. Global Change Biology , 11: 945-958.
Houghton, R.A., Lawrence, K.T., Hackler, J.L., Brown, S., 2001. The spatial distribution of forest biomass in the Brazilian Amazon: a comparison of estimates. Global Change Biology 7: 731-746.
Howes, F.N., 1950. Age-old resin of the Mediterranean region and their uses. Economic Botany 4: 307-316.
87
Hurni, H., 1988. Degradation and conservation of the resources in the Ethiopian highlands. Mountain Research and Development 8 (2/3): 123-130.
IPCC, 2000. The intergovernmental panel on climate change: A special report on land-use, land-use change and forestry. Intergovernmental Panel on Climate Change, Cambridge University Press, Cambridge, UK.
IPCC, 2005. Special report: Carbon dioxide capture and storage: Summary for policymakers: A report of Working Group III of the IPCC and Technical Summary. Intergovernmental Panel on Climate Change on the invitation of the United Nations Framework Convention on Climate Change. ISBN 92-9169-119-4.
Issac, S.R., and Nair, M.A., 2006. Litter dynamics of six multipurpose trees in a homegarden in southern Kerala India. Agroforestry Systems 67: 203-213.
Janisch, J.E., and Harmon, M.E., 2002. Successional changes in live and dead wood carbon stores: implications for net ecosystem productivity. Tree Physiology 22: 77- 89.
Janssens, I.A., Sampson, D.A., Curiel-Yuste, J., Carrara, A., and Ceulemans, R., 2002. The carbon cost of fine root turnover in a Scots pine forest. Forest Ecology and Management 168: 231-240.
Jensen, J.R., 1996. Introductory Digital Image Processing: A Remote Sensing Perspective, Second Edition. Prentice Hall, Upper Saddle River, N.J. 316 pp.
Jha, C.S., and Singh, J.S., 1990. Composition and dynamics of dry tropical forest in relation to soil texture. Journal of Vegetation Science 1: 609-614.
Jina, B.S., Sah, P., Bhatt, M.D., and Rawat, Y.S., 2008. Estimating Carbon Sequestration Rates and Total Carbon Stockpile in Degraded and Non-Degraded Sites of Oak and Pine Forest of Kumaun Central Himalaya. ECOPRINT 15: 75-81.
Johnson, E., and Heinen, R., 2004. Carbon trading: Time for industry involvement. Environment International 30: 279-288.
88
Juwarkar, A.A., Varghese, A.O., Singh, S.K., Aher, V.V., Thawale, P.R., 2011. Carbon Sequestration Potential in Aboveground Biomass of Natural Reserve Forest of Central India. International Journal of Agriculture 1(2): 80-86.
Kafta Humera District Livelihood Report (KHDLR), 2007. Disaster Prevention and Preparedness Agency, Ethiopia.
Kandlikar, M., 1996. Indices for comparing greenhouse gas emissions: Integrating science and economics. Energy Economics 18: 265-281.
Kurz, W.A., Beukema, S.J., and Apps, M.J., 1996. Estimation of root biomass and dynamics for the carbon budget model of the Canadian forest sector. Can. J. For. Res. 26: 1973-1979.
Lal, R., 2001. World cropland soils as a source or sink for atmospheric Carbon. Advances in Agronomy 71: 145-191.
Lal, R., 2004. Soil carbon sequestration impacts on global change and food security. Science 304: 1623-1627.
Lal, R., 2005. Forest soils and carbon sequestration. Forest Ecology and Management 220: 242-258.
Lal, R., 2008. Soil carbon stocks under present and future climate with specific reference to European ecoregions. Nutr. Cycl. Agroecosyst 81: 113-127.
Lal, R., 2009. Sequestering Carbon in Soils of Arid Ecosystems. Land Degrad. Develop . 20: 441-454.
Lal, R., Bruce, J.P., 1999. The potential of world cropland soils to sequester carbon and mitigate the greenhouse effect. Environ. Sci. Policy 2: 177-185.
Lemenih, M., Abebe, T. and Olsson, M., 2003. Gum and resin resources from some Acacia, Boswellia and Commiphora species and their economic contributions in Liban, Southeast Ethiopia. Journal of Arid Environments , 55: 465-482.
Lemenih, M., and Itanna, F., 2004. Soil carbon stocks and turnovers in various vegetation types and arable lands along an elevation gradient in southern Ethiopia. Geoderma 123: 177-188.
89
Lemenih, M., and Teketay, D., 2003. Frankincense and Myrrh Resources of Ethiopia I. Distribution, Production, Opportunities for Dry Land Development and Research Needs’, Senet: Ethiopian Journal of Sciences 26(1): 63-73.
Lemenih, M., and Teketay, D., 2004. Restoration of native forest flora in the degraded highlands of Ethiopia: constraints and opportunities. Ethiopian Journal of Science 27: 75-90.
Lemenih, M., Karltun, E. and Olsson M., 2005. Soil organic matter dynamics after deforestation along a farm field chronosequence in southern highlands of Ethiopia. Agriculture, Ecosystems and Environment 109: 9-19.
Lemma, B., Kleja, D., B., Olsson, M., and Nilsson, I., 2007. Factors controlling soil organic carbon sequestration under exotic tree plantations: A case study using the
CO 2 Fix model in southwestern Ethiopia. Forest Ecology and Management 252(1- 3): 124-131.
Lillesand, T.M., 2000. Remote Sensing and Image Interpretation, 4 th edition, John Wiley and Sons, Inc. New York.
Lunetta, R.S., 1999. Remote Sensing Change Detection: Environmental Monitoring Methods and Applications, Ann Arbor Press.
Lynch, J.M., and Whipps, J.M., 1990. Substrate flow in the rhizosphere. Plant and Soil 129: 1-10.
Lyngbaek, A.E., Muschler, R.G., Sinclair, F.L., 2001. Productivity and profitability of multistrata organic versus conventional coffee farms in Costa Rica. Agroforestry Systems , 53: 205-213.
MacDicken, K.G., 1997. A Guide to Monitoring Carbon Storage in Forestry and Agroforestry Projects. Winrock International, Arlington, Virginia, USA.
Mafongoya, P.L., Nair, P.K.R., and Dzowela, B.H., 1998. Mineralization of nitrogen from decomposing leaves of multipurpose trees as affected by their chemical composition. Biol. Fertil. Soils 27: 143-148.
90
Malhi, Y., Baldochi, D.D., and Jarvis, P.G., 1999. The carbon balance of tropical temperate and boreal forests. Plant Cell Environment 22: 715-740.
Malhi, Y., Meir, P., and Brown, S., 2002. Forests, carbon and global climate. Philos. Trans. R. Soc. Lond. A 360: 1567-1591.
Marks, E., Aflakpui, G.K.S., Nkem, J., Poch, R.M., Khouma, M., Kokou, K., Sagoe, R., and Sebastià, M.-T., 2009. Conservation of soil organic carbon, biodiversity and the provision of other ecosystem services along climatic gradients in West Africa. Biogeosciences Discussions 6: 1825-1838.
MAS, J.F., 1999. Monitoring land cover changes: a comparison change detection techniques. International journal of remote sensing 20(1): 139-152.
Matamala, R., Gonzàlez- Meler, M.A., Jastrow, J.D., Norby, R.J., and Schlesinger, W.H., 2003. Impacts of fine root turnover on forest NPP and soil carbon sequestration potential. Science 302: 1385-1387.
Mekuria, W., Veldkamp, E., and Haile, M., 2009. Carbon stock changes with relation to land use conversion in the lowlands of Tigray, Ethiopia. Tropentag, University of Hamburg, October 6-8.
Moges, Y., Eshetu, Z., and Nune, S., 2010. Ethiopian Forest Resources: Current Status and Future Management Options In View of Access to Carbon Finances. Ethiopian Climate Research and Networking and the United Nations Development Programme (UNDP) Addis Ababa, Ethiopia. 56 pp.
Montagnini, F., and Nair, P.K.R., 2004. Carbon sequestration: An underexploited environmental benefit of agroforestry systems. Agroforestry Systems 61: 281-295.
Murphy, P.G., and Lugo, A.E., 1986. Ecology of tropical dry forest. Annual Review of Ecology and Systematics 17: 67-88.
Nair, P.K.R., Kumar, B.M., and Nair, V.D., 2009. Agroforestry as a strategy for carbon sequestration. J. Plant Nutr. Soil Sci. 172: 10-23.
91
National Conservation Strategy Secretariat (NCSS), 1993. National Conservation Strategy, National Policy on the Resources Base, its Utilization and Planning for Sustainability, Addis Ababa, Ethiopia, volume 1.
Nave, L.E., Vance, E.D., Swanston, C.W., and Curtis, P.S., 2010. Harvest impacts on soil carbon storage in temperate forests. Forest Ecology and Management 259: 857- 866.
Nguyen, C., 2003. Rhizodeposition of organic carbon by plants: Mechanisms and controls. Agronomie 23: 375-396.
Olsson, L., Ardo, J., 2002. Soil carbon sequestration in degraded semiarid agro-ecosystems Perils and potentials. AMBIO: A Journal of the Human Environment 31: 471-477.
Paul, K.I., Polglase, P.J., Nyakuengama, J.G., and Khanna, P.K., 2002. Change in soil carbon following afforestation. Forest Ecology and Management , 168: 241-257.
Pearson, T., Walker, S., and Brown, S., 2005. Sourcebook for land-use, land-use change and forestry projects. Winrock International and the Biocarbon fund of the Worldbank. 57 pp.
Persson, T.A., Azar, C., Lindghen, K., 2006. Allocation of CO 2 emission permits: Economic incentives for emission reductions in developing countries. Energy Policy 34: 1889-1899.
Post, W.M., Izaurralde, R.C., Mann, L.K., and Bliss, N., 2001. Monitoring and verifying changes of organic carbon in soil. Climate Change 51: 73-99.
Post, W.M., Pengh, T.H., Emanuel, W.R., King, A.W., Dale, V.H. and Delnglis, D.L., 1990. The global carbon cycle. Am. Sci. 78: 310-326.
Powers, M., Kolka, R., Palik, B., McDonald, R., and Jurgensen, M., 2011. Long-term management impacts on carbon storage in Lake States forests. Forest Ecology and Management 262: 424-431.
92
Ramachandran, A., Jayakumar, S., Haroon, R. M., Bhaskaran, A., and Arockiasamy, D. I., 2007. Carbon sequestration: estimation of carbon stock in natural forests using geospatial technology in the Eastern Ghats of Tamil Nadu, India. Current Science 92 (3): 323-331.
REDD Methodological Module, 2009. “Estimation of carbon stocks and changes in carbon stocks in the dead wood carbon pool”. Avoided Deforestation Partners Version 1.0
Rees, R.M., Bingham, I., Baddeley, J., and Watson, C.A., 2005. The role of plants and land management in sequestering soil carbon in temperate arable and grassland ecosystems. Geoderma 128: 130-154.
Reyes-Reyes, G., Baron-Ocampo, L., Cuali-Alvarez, I., Frias-Hernandez, J.T., Olalde- Portugal, V., Fregoso, L.V., and Dendooven, L., 2002. Carbon and nitrogen dynamics in soil from the central highlands of Mexico as affected by mesquite (Prosopis spp. ) and huizache ( Acacia tortuoso ): A laboratory investigation. Applied Soil Ecology 19: 27-34.
Roshetko, J.M., Delaney, M., Hairiah, K., and Purnomosidhi, P., 2002. Carbon stocks in Indonesian homegarden systems. American Journal of Alternative Agriculture 17(2): 1-11.
Ros-Tonen, M., Dijkman, W., and van Bueren, E. L., 1995. Commercial and Sustainable Extraction of Non-timber Forest Products. Towards a Policy and Management Oriented Research Strategy. The Tropenbos Foundation, Wageningen.
Sagar, R., Raghubanshi, A.S., and Singh, J.S., 2008. Comparison of community composition and species diversity of understorey and over storey tree species in a dry tropical forest of northern India. Journal of Environmental Management 88: 1037-1046.
Sakai, Y., Takahashi, M., and Ishizuka S., 2008. Estimating decay rates of dead wood by changes in wood density in coniferous plantations in Japan. Jpn. J. For. Environ. 50: 153-165.
Sayer, E.J., 2006. Using experimental manipulation to assess the roles of leaf litter in the functioning of forest ecosystems. Biological Reviews 81: 1-31.
93
Schlamadinger, B., Marland, G., 1998. The Kyoto Protocol: Provisions and unresolved issues relevant to landuse change and forestry. Environmental Science and Policy 1: 313-327.
Sedjo, R.A., 2001. Forest Carbon Sequestration: Some Issues for Forest Investments. Resources for the Future. Washington, D.C.
Sheikh M.A., Kumar M., and Bussmann R.W., 2009. Altitudinal variation in soil organic carbon stock in coniferous subtropical and broadleaf temperate forests in Garhwal Himalaya. Carbon Balance and Management 4:6.
Singh, A., 1989. Digital Change Detection Techniques Using Remotely Sensed Data. International Journal of Remote Sensing 10(6): 989-1003.
Singh, L., and Singh, J.S., 1991. Species structure, dry matter dynamics and carbon flux of a dry tropical forest in India. Annals of Botany 68: 263-273.
Six J., and Jastrow, J.D., 2002. Organic matter turnover. Encycl Soil Sci 936-942.
Smith, T.M., Cramer, W.P., Dixon, R.K., Leemans, R., Neilson, R.P., and Solomon, A.M., 1993. The global terrestrial carbon cycle. Water, Air and Soil Pollution 70: 19-38.
Soil Science Society of America, 2001. Carbon Sequestration: Position of the Soil Science Society of America (SSSA). SSSA available at www.soils.org/files/about- society/carbon-sequestration-paper.pdf (accessed: Dec. 15, 2011).
Solomon, D., Fritzszhe, F., Lehmann, J., Tekalign, M., and Zech, W., 2002. Soil Organic Matter Dynamics in the Subhumid Agroecosystems of the Ethiopian Highlands: Evidence From Natural 13 C Abundance and Particle Size Fractionation. Soil Science Society of America Journal 66: 969-978.
Strand, A.E., Pritchard, S.G., McCormack, M.L., Davis, M.A., and Oren, R., 2008. Irreconcilable differences: Fine-root life spans and soil carbon persistence. Science 319: 456-458.
Streck, C., and Scholz, S.M., 2006. The role of forests in global climate change: Whence we come and where we go. International Affairs 82: 861-879.
94
Taddese, W., Teketay, D., Lemenih, M., Fitiwi, G., 2002. Boswellia country report for Ethiopia: Chikamai B., ed. Review and synthesis on the state of knowledge of Boswellia species and commercialization of frankincense in the dry land of Eastern Africa. FAO/EU/FORNESSA Publication. pp 11-33.
Takahashi, M., Ishizuka, S., Ugawa, S., Sakai, Y., Sakai, H., Ono, K., Hashimoto, S., Matsuura, Y., and Morisada, K., 2010. Carbon stock in litter, deadwood and soil in Japan’s forest sector and its comparison with carbon stock in agricultural soils. Soil Science and Plant Nutrition 56:19-30.
Takimoto, A., Ramachandran Nair, P.K., and D. Nair, V., 2008. Carbon stock and sequestration potential of traditional and improved agroforestry systems in the West African Sahel. Agriculture, Ecosystems and Environment , 125: 159-166.
Tegene, B., 2002. Land-Cover/Land-Use Changes in the Derekolli Catchment of the South Welo Zone of Amhara Region, Ethiopia. Eastern Africa Social Science Research Review 18 (1): 1-20.
Tekle, K., Hedlund, L., 2000. Land cover changes between 1958 and 1986 in Kalu District, Southern Wello, Ethiopia. Mountain Res. Dev. 20 (1): 42-51.
Terakunpisut, J., Gajaseni, N., and Ruankawe, N., 2007. Carbon sequestration potential in aboveground biomass of thong pha phum national forest, Thailand. Applied Ecology and Environmental Research 5(2): 93-102.
Tesfaye, G., Teketay, D., Fetene, M., and Beck, E., 2010. Regeneration of seven
indigenous tree species in a dry Afromontane forest, southern Ethiopia. FLORA
205: 135-143.
Thomas, S.C., 1996. Asymptotic height as a predictor of growth and allometric characteristics in Malaysian rainforest trees. American Journal of Botany 83(5): 556-566.
95
Tilahun, M., Muys, B., Mathijs, E., Kleinn, C., Olschewski, R., and Gebrehiwot, K., 2012. Frankincense yield assessment and modeling in closed and grazed Boswellia papyrifera woodlands of Tigray, Northern Ethiopia. Journal of Arid Environments 75: 695-702.
Tilahun, M., Olschewski, R., Kleinn, C., and Gebrehiwot, K., 2007. Economic analysis of closing degraded Boswellia papyrifera dry forest from human interventions. A study from Tigray, Northern Ethiopia. Forest Policy and Economics 9: 996-1005.
Tschakert, P., 2004. Carbon for Farmers: Assessing the Potential for Soil Carbon Sequestration in the Old Peanut Basin of Senegal. Climatic Change 67: 273–290.
Tucker, A.O., 1986. Frankincense and Myrrh. Economic botany 40(4): 425-433.
Tucker, M., 2001. Trading carbon tradable offsets under Kyoto’s Clean Development Mechanism: The economic advantages to buyers and sellers of suing call options. Ecological Economics 37: 173-182.
Tupek, B., Zanchi, G., Verkerk, P.J., Churkina, G., Viovy, N., Hughes, J.K., Lindner, M., 2010. A comparison of alternative modeling approaches to evaluate the European forest carbon fluxes. Forest Ecology and Management 260: 241-251.
Vishwanathan, M.K., Samra, J.S., Sharma, A.R., 1999. Biomass production of trees and grasses in a silvipasture system on marginal lands of Down Valley of north-west India. Agroforestry Systems 46: 181-196.
Vogt, K.A., 1991. Carbon budgets of temperate forest ecosystems. Tree Physiology 9: 69- 86.
Vogt, K.A., Vogt, D.J., and Bloomfield, J., 1998. Analysis of some direct and indirect methods for estimating root biomass and production of forests at an ecosystem level. Plant Soil 200: 71-89.
Walkley, A., and Black, I.A., 1934. An examination of the Degtjareff method for determining soil organic matter and a proposed modification of the chronic acid titration method. Soil Sci. 37: 29-38.
96
Watson, R.T., Noble, I.R., Bolin, B., Ravindranath, N.H., Verardo, D.J., and Dokken, D.J. (eds), 2000. Land Use, Land-Use Change, and Forestry. Intergovernmental Panel on Climate Change (IPCC), Special report. Cambridge University Press, UK. 375 pp.
WBISPP (Woody Biomass Inventory and Strategic Planning Project), 2000. Manual for woody biomass inventory. Woody Biomass Inventory and Strategic Planning Project, Ministry of Agriculture, Addiss Ababa, Ethiopia.
Weintraub, M.N., Scott-Denton, L.E., Schmidt, S.K., and Monson, R.K., 2007. The effects of tree rhizodeposition on soil exoenzyme activity dissolved organic carbon and nutrient availability in a subalpine forest ecosystem. Oecologia 154: 327-338.
White, M., Gower, S., and Ahl, D., 2005. Life cycle inventories of round wood production in northern Wisconsin: inputs into an industrial forest carbon budget. Forest Ecology and Management 219: 13-28.
Worku, A., Teketay, D., Lemenih, M., and Fetene, M., 2012. Diversity, regeneration status, and population structures of gum and resin producing woody species in Borana, Southern Ethiopia. Forests, Trees and Livelihoods , DOI:10.1080/14728028.2012.716993.
Yelenik, S.G., Stock, W.D., and Richardson, D.M., 2004. Ecosystem level impacts of invasive Acacia saligna in the South African Fynbos. Restoration Ecology 12: 44- 51.
Zeleke, G., and Hurni, H., 2000. Implications of Land-use and Land-cover Dynamics for Mountain Resource Degradation in the Northwestern Ethiopian Highlands. Mountain Research and Development . 22: 184-191.
97
APPENDICES
Appendix 1: Height and DBH Ranges of the Vegetation Characterstics
No Scientific Name Strata Kafta Humera Metema Sherkole Height DBH Height DBH Height DBH 1 Abubuka (Bertegna) UW ------TW ------6-8 7-19 2 Acacia mellifera UW 5-6 7.5-10.4 ------TW 7 12 ------3 Anogeissus UW 3-6 13.4-20.1 ------leiocarpus TW 4-7 9.1-20.3 6-8 12.5-28.5 6.5-9 14.5-34 4 Balanites aegyptiaca UW ------TW 6.5-8.5 26-40.2 ------5-8.5 8.1-33 5 Boswellia papyrifera UW 2.7-8.5 11.5-31.2 5-8.5 9-33.5 ------TW 4.5-8.5 10-33.6 4.5-8 9.5-27 3-9 4.5-36 6 Combretum UW ------adinogonium TW 3-7.5 5.4-32.5 6 16 3.5-9 6.5-27 7 Combretum allinum UW ------TW 3 11.5 ------8 Combretum molle UW 4-4.5 7.7-16 5-7 7-16.5 ------TW 3-5 13-16.5 ------9 Dalbergia UW 4.5-5.5 8.5-15.4 4.5-5 5.5-21 ------melanoxylm TW 2.5-5 8.5-16 ------6-7 14.5-21 10 Dicrostachy cineria UW ------TW 6 13.5 ------11 Entada abyssinica UW ------7.5 25.5 ------TW ------5-7.5 9-20.5 12 Ficus sur UW ------TW ------16 90 13 Gardenia ternifolia UW ------5-6.5 8.5-23 ------TW ------5 6.8 14 Kudkuda (Amharic) UW ------5-6 6-15.5 ------TW ------15 Lannea froictosa UW ------5.5-6.5 9-25 ------TW 2-5.5 8.5-18.8 4.6-5.5 13.5-14 5-8 5.5-21.5 16 Pteospurmum UW ------kuntianum TW ------5-6.5 9.6-12 17 Pterocarpus lucense UW ------5.5-8.5 12.5-29.5 ------TW ------6-10 14.8-33.5 18 Sterculia setigera UW 5.5-6 23.5-32.2 6-7.5 22-39.5 ------TW 6.5-8 21.9-49 4.5-8.5 8.5-40 3.8-10 4.5-57 19 Tamarindus indica UW 3-7 13-29 ------TW ------
98
No Scientific Name Strata Kafta Humera Metema Sherkole Height DBH Height DBH Height DBH 20 Terminalia UW ------macroptera TW ------7-9 19-32.5 21 Ximenia americana UW ------4.5 7.5 ------TW ------22 Yedebene fus UW ------6-7 17.5-21 ------(Amharic) TW ------23 Zengorifa (Amharic) UW ------5.5-7 11-19.5 ------TW ------5-5.5 15-17 ------24 Ziziphus mucronata UW -- --- 5 12.5 ------TW 3 12.4 ------
Height range (m), DBH range (cm) of tree species on three sites of western lowland woodlands of Ethiopia
Appendix 2: Mean Carbon Stock of the Three Sites (Mean ± SE)
Carbon pool Strata Adi Goshu Lemlem-Terara Gemed P-value AGWB carbon stock UW 16.71±2.15 a (b) 27.91±2.57 a (a) ----- 0.0074 -1 Mg ha TW 19.29±4.57 a 26.59±3.30 a 25.87±3.70 a 0.4747 BGRB carbon stock UW 4.18±0.54 a (b) 6.98±0.64 a (a) ----- 0.0074 Mg ha -1 TW 4.82±1.14 a 6.65±0.83 a 6.47±0.93 a 0.4748 DWB carbon stock UW 0.48±0.43 a (b) 2.89±0.75 a (a) ----- 0.0238 Mg ha -1 TW 2.89±1.47 a 0.40±0.22 a 0.80±0.33 a 0.1502 Herb Biomass carbon UW 0.28±0.11 a 0.26±0.06 a ----- 0.9191 stock Mg ha -1 TW 0.45±0.19 a (b) 0.42±0.13 a (b) 1.19±0.19 a 0.0207 Litter Biomass UW ----- 0.61±0.27 a ------carbon stock Mg ha -1 TW ----- 0.46±0.09 a ------Soil carbon stock UW 33.61±2.24 a (b) 58.19±5.46 a (a) ----- 0.0019 (0-30 cm depth) Mg TW 38.48±7.55 a (a) 34.25±6.13 b (a) 36.68±2.39 (a) 0.8884 ha -1 Total carbon stock UW 55.26 96.74 ----- Mg ha -1 TW 65.93 68.77 71.01 AGWB=aboveground woodland biomass; BGRB=belowground root biomass; DWB=dead wood biomass; UW=untapped Boswellia papyrifera woodland; TW=tapped Boswellia papyrifera woodland. Mean with same later between raw and column are not significant at P < 0.05.
99
Appendix 3: Mean Carbon Stock of Adi Goshu Farm Land (Mean ± SE)
Adi Goshu Carbon pool UF TF P-value AGWB carbon stock Mg ha -1 8.92±6.56 a 16.77±4.43 a 0.3732 BGRB carbon stock Mg ha -1 2.23±1.64 a 4.19±1.11 a 0.3733 DWB carbon stock Mg ha -1 2.32±0.17 a 0.00±0.00 b 0.0004 Herb Biomass carbon stock Mg ha -1 0.52±0.02 a 0.12±0.07 b 0.0230 Crop Biomass carbon stock Mg ha -1 2.27±1.92 a 0.70±0.20 a 0.3599 Crop Root Biomass carbon stock Mg ha -1 0.18±0.12 a 0.07±0.01 a 0.2777 Soil carbon stock (0-30 cm depth) Mg ha -1 49.00±17.07 a 23.33±5.31 a 0.1737 Total carbon stock Mg ha -1 65.44 45.18 AGWB=aboveground woodland biomass; BGRB=belowground root biomass; DWB=dead wood biomass; UF=untapped farmland; TF=tapped farmland. Mean with same later between raw are not significant at P < 0.05.
Appendix 4: Adi Goshu Woodland ANOVA Results
a) Dependent Variable: AGBC Mg ha -1
Source DF Sum of Squares Mean Square F Value Pr > F
Model 1 19.9602565 19.9602565 0.26 0.6207
Error 10 765.6935420 76.5693542
Corrected Total 11 785.6537985
R-Square Coeff Var Root MSE AGBC Mg ha -1 Mean 0.025406 48.61416 8.750392 17.99968
b) Dependent Variable: BGRBC Mg ha -1
Source DF Sum of Squares Mean Square F Value Pr > F
Model 1 1.24646302 1.24646302 0.26 0.6209
Error 10 47.86333021 4.78633302
Corrected Total 11 49.10979323 R-Square Coeff Var Root MSE BGRBC Mg ha -1 Mean 0.025381 48.61484 2.187769 4.500208
100
c) Dependent Variable: TBDW C Mg ha -1 Source DF Sum of Squares Mean Square F Value Pr > F
Model 1 17.44011915 17.44011915 2.48 0.1461
Error 10 70.21384285 7.02138428
Corrected Total 11 87.65396200 R-Square Coeff Var Root MSE TBDW C Mg ha -1 Mean 0.198966 157.0659 2.649789 1.687056 d) Dependent Variable: HBC Mg ha -1 Source DF Sum of Squares Mean Square F Value Pr > F
Model 1 0.08975096 0.08975096 0.62 0.4496
Error 10 1.44942781 0.14494278
Corrected Total 11 1.53917877 R-Square Coeff Var Root MSE HBC Mg ha -1 Mean 0.058311 105.1451 0.380714 0.362084 e) Dependent Variable: SOC Mg ha -1 Source DF Sum of Squares Mean Square F Value Pr > F
Model 1 71.114696 71.114696 0.38 0.5501
Error 10 1859.578505 185.957850
Corrected Total 11 1930.693201 R-Square Coeff Var Root MSE SOC Mg ha -1 Mean 0.036834 37.83545 13.63664 36.04195
Appendix 5: Adi Goshu Farm Land ANOVA Results
a) Dependent Variable: AGBC Mg ha -1
Source DF Sum of Squares Mean Square F Value Pr > F
Model 1 73.9609361 73.9609361 1.09 0.3732
Error 3 203.6025616 67.8675205
Corrected Total 4 277.5634977
R-Square Coeff Var Root MSE AGBC Mg ha -1 Mean 0.266465 60.43240 8.238175 13.63205
101
b) Dependent Variable: BGRBC Mg ha -1 Source DF Sum of Squares Mean Square F Value Pr > F Model 1 4.62168750 4.62168750 1.09 0.3733
Error 3 12.72906250 4.24302083
Corrected Total 4 17.35075000
R-Square Coeff Var Root MSE BGRBC Mg ha -1 Mean 0.266368 60.45075 2.059859 3.407500 c) Dependent Variable: TBDW C Mg ha -1 Source DF Sum of Squares Mean Square F Value Pr > F
Model 1 6.48069788 6.48069788 331.68 0.0004
Error 3 0.05861718 0.01953906
Corrected Total 4 6.53931506
R-Square Coeff Var Root MSE TBDW C Mg ha -1 Mean 0.991036 15.03666 0.139782 0.929610 d) Dependent Variable: HBC Mg ha -1 Source DF Sum of Squares Mean Square F Value Pr > F
Model 1 0.18737321 0.18737321 18.55 0.0230
Error 3 0.03029623 0.01009874
Corrected Total 4 0.21766945
R-Square Coeff Var Root MSE HBC Mg ha-1 Mean 0.860815 35.81583 0.100493 0.280581 e) Dependent Variable: CBC Mg ha -1 Source DF Sum of Squares Mean Square F Value Pr > F
Model 1 2.94925130 2.94925130 1.16 0.3599
Error 3 7.61161382 2.53720461
Corrected Total 4 10.56086512
R-Square Coeff Var Root MSE CBC Mg ha-1 Mean 0.279262 119.8766 1.592861 1.328750
102
f) Dependent Variable: CRBC Mg ha -1 Source DF Sum of Squares Mean Square F Value Pr > F
Model 1 0.01654401 0.01654401 1.75 0.2777
Error 3 0.02836829 0.00945610
Corrected Total 4 0.04491230
R-Square Coeff Var Root MSE CRBC Mg ha-1 Mean 0.368363 86.11243 0.097242 0.112925 g) Dependent Variable: SOC Mg ha -1 Source DF Sum of Squares Mean Square F Value Pr > F
Model 1 790.713886 790.713886 3.16 0.1737
Error 3 751.653816 250.551272
Corrected Total 4 1542.367701
R-Square Coeff Var Root MSE SOC Mg ha-1 Mean 0.512662 47.11332 15.82881 33.59731
Appendix 6: Lemlem Terara Woodland ANOVA Results
a) Dependent Variable: AGBC Mg ha -1 Source DF Sum of Squares Mean Square F Value Pr > F
Model 1 3.4859171 3.4859171 0.09 0.7695
Error 7 262.8745161 37.5535023
Corrected Total 8 266.3604332
R-Square Coeff Var Root MSE AGBC Mg ha -1 Mean 0.013087 22.30801 6.128091 27.47036 b) Dependent Variable: BGRBC Mg ha -1 Source DF Sum of Squares Mean Square F Value Pr > F
Model 1 0.21725035 0.21725035 0.09 0.7697
Error 7 16.42323021 2.34617574
Corrected Total 8 16.64048056
R-Square Coeff Var Root MSE BGRBC Mg ha -1 Mean 0.013056 22.30484 1.531723 6.867222
103
c) Dependent Variable: TBDW C Mg ha -1 Source DF Sum of Squares Mean Square F Value Pr > F
Model 1 11.40712111 11.40712111 4.63 0.0684
Error 7 17.23847468 2.46263924
Corrected Total 8 28.64559579
R-Square Coeff Var Root MSE TBDW C Mg ha -1 Mean 0.398216 78.64173 1.569280 1.995480 d) Dependent Variable: HBC Mg ha -1
Source DF Sum of Squares Mean Square F Value Pr > F
Model 1 0.05145324 0.05145324 1.66 0.2386 Error 7 0.21705142 0.03100735 Corrected Total 8 0.26850467
R-Square Coeff Var Root MSE HBC Mg ha -1 Mean 0.191629 55.70491 0.176089 0.316110 e) Dependent Variable: LBC Mg ha -1
Source DF Sum of Squares Mean Square F Value Pr > F
Model 1 0.04656153 0.04656153 0.15 0.7131
Error 7 2.22155596 0.31736514
Corrected Total 8 2.26811749
R-Square Coeff Var Root MSE LBC Mg ha -1 Mean 0.020529 100.2973 0.563352 0.561682 f) Dependent Variable: SOC Mg ha -1
Source DF Sum of Squares Mean Square F Value Pr > F
Model 1 1146.742626 1146.742626 7.17 0.0316
Error 7 1119.623716 159.946245
Corrected Total 8 2266.366342
R-Square Coeff Var Root MSE SOC Mg ha -1 Mean 0.505983 25.18769 12.64699 50.21098
104
Appendix 7: Overall Untapped Boswellia papyrifera Woodlands ANOVA Results
a) Dependent Variable: AGBC Mg ha -1 Source DF Sum of Squares Mean Square F Value Pr > F
Model 1 376.3510083 376.3510083 11.20 0.0074
Error 10 336.1239494 33.6123949
Corrected Total 11 712.4749577
R-Square Coeff Var Root MSE AGBC Mg ha -1 Mean 0.528231 25.98641 5.797620 22.31020 b) Dependent Variable: BGRBC Mg ha -1 Source DF Sum of Squares Mean Square F Value Pr > F
Model 1 23.50600208 23.50600208 11.19 0.0074
Error 10 21.00408542 2.10040854
Corrected Total 11 44.51008750
R-Square Coeff Var Root MSE BGRBC Mg ha -1 Mean 0.528105 25.98438 1.449279 5.577500 c) Dependent Variable: TBDW C Mg ha -1 Source DF Sum of Squares Mean Square F Value Pr > F
Model 1 16.00888101 16.00888101 7.08 0.0238
Error 10 22.59666582 2.25966658
Corrected Total 11 38.60554683
R-Square Coeff Var Root MSE TBDW C Mg ha -1 Mean 0.414678 91.85403 1.503219 1.636530 d) Dependent Variable: HBC Mg ha -1 Source DF Sum of Squares Mean Square F Value Pr > F Model 1 0.00050356 0.00050356 0.01 0.9191
Error 10 0.46408979 0.04640898
Corrected Total 11 0.46459335
R-Square Coeff Var Root MSE HBC Mg ha -1 Mean 0.001084 80.04787 0.215427 0.269123
105
e) Dependent Variable: SOC Mg ha -1 Source DF Sum of Squares Mean Square F Value Pr > F
Model 1 1813.287256 1813.287256 17.36 0.0019
Error 10 1044.668670 104.466867
Corrected Total 11 2857.955926
R-Square Coeff Var Root MSE SOC Mg ha -1 Mean 0.634470 22.26770 10.22090 45.90014
Appendix 8: Overall Tapped Boswellia papyrifera Woodlands ANOVA Results
a) Dependent Variable: AGBC Mg ha -1 Source DF Sum of Squares Mean Square F Value Pr > F
Model 2 187.933853 93.966926 0.78 0.4747
Error 16 1925.599999 120.350000
Corrected Total 18 2113.533852
R-Square Coeff Var Root MSE AGBC Mg ha -1 Mean 0.088919 45.89392 10.97041 23.90385 b) Dependent Variable: BGRBC Mg ha -1 Source DF Sum of Squares Mean Square F Value Pr > F
Model 2 11.7444658 5.8722329 0.78 0.4748
Error 16 120.3443131 7.5215196
Corrected Total 18 132.0887789
R-Square Coeff Var Root MSE BGRBC Mg ha -1 Mean 0.088913 45.89215 2.742539 5.976053 c) Dependent Variable: TBDW C Mg ha -1 Source DF Sum of Squares Mean Square F Value Pr > F
Model 2 19.92277310 9.96138655 2.14 0.1502
Error 16 74.50299336 4.65643709
Corrected Total 18 94.42576647
R-Square Coeff Var Root MSE TBDW C Mg ha -1 Mean 0.210989 154.2153 2.157878 1.399263
106
d) Dependent Variable: HBC Mg ha -1 Source DF Sum of Squares Mean Square F Value Pr > F
Model 2 2.68635125 1.34317563 4.99 0.0207
Error 16 4.30706091 0.26919131
Corrected Total 18 6.99341216 R-Square Coeff Var Root MSE HBC Mg ha -1 Mean 0.384126 62.03835 0.518836 0.836316 e) Dependent Variable: SOC Mg ha -1 Source DF Sum of Squares Mean Square F Value Pr > F
Model 2 36.471250 18.235625 0.12 0.8884
Error 16 2447.343717 152.958982
Corrected Total 18 2483.814967 R-Square Coeff Var Root MSE SOC Mg ha -1 Mean 0.014684 33.54980 12.36766 36.86358
Appendix 9: The Overall Comparison of Models
a) Dependent Variable: AGBC Mg ha -1
Source DF Sum of Squares Mean Square F Value Pr > F
Model 1 4978.076433 4978.076433 94.40 <.0001
Error 70 3691.197929 52.731399
Corrected Total 71 8669.274362 R-Square Coeff Var Root MSE AGBC Mg ha -1 Mean
0.574221 53.27317 7.261639 13.63095
b) Means with the same letter are not significantly different
Tukey Grouping Mean N Model
A 21.946 36 Brown (1997)
B 5.316 36 WBISPP (2000)
107
Appendix 10: Biomass and Carbon Stock Estimation Using WBISPP (2000)
a) Carbon Stock in Three Different Sites and Two Strata of Woodlands (Mean ± SE) Carbon Pool Strata Kafta Humera Metema Sherkole P-value AGWB Carbon UW 5.42±0.59 a (b) 7.78±0.66 a (a) ----- 0.0235 -1 Stock Mg ha TW 5.55±1.08 a 6.82±0.87 a 4.31±0.35 a 0.1072 BGRB Carbon UW 1.36±0.15 a (b) 1.95±0.16 a (a) ----- 0.0236 -1 Stock Mg ha TW 1.39±0.27 a 1.70±0.22 a 1.08±0.09 a 0.1094
AGWB=Above Ground Woodland Biomass; BGRB=Below Ground Root Biomass; UW=Untapped Boswellia papyrifera Woodland; TW=Tapped Boswellia papyrifera Woodland. Mean with same later between raw and column are not significant at P < 0.05
b) Carbon Stock in Kafta Humera Farm Land of Two Strata of Woodlands (Mean ± SE)
Kafta Humera Carbon Pool UF UF P-value AGWB carbon stock Mg ha -1 2.41±1.97 a 3.47±0.88 a 0.6057 BGRB carbon stock Mg ha -1 0.60±0.49 a 0.87±0.22 a 0.6033
AGWB=Above Ground Woodland Biomass; BGRB=Below Ground Root Biomass; UF=Untapped Farmland; TF=Tapped Farmland. Mean with same later between raw are not significant at P < 0.05
Appendix 11: Accuracy Assessment of Kafta Humera
a) Classification accuracy assessment of 1985
Agricultural Bare Wood Shrub Grass Water Land Use land land land land land body Agricultural land 80.39 2.27 0.00 2.22 0.00 0.00 Bare land 0.00 86.36 2.27 4.44 4.88 0.00 Wood land 0.00 0.00 88.64 6.67 2.44 0.00 Shrub land 1.96 4.55 4.55 80.00 4.88 0.00 Grass land 1.96 6.82 2.27 4.44 87.80 0.00 Water body 15.69 0.00 2.27 2.22 0.00 100 Total 100 100 100 100 100 100
Overall Accuracy = 86.43% and Kappa Coefficient = 0.8372
108
b) Classification accuracy assessment of 1995
Agricultural Bare Wood Shrub Grass Water Land Use land land land land land body Agricultural land 90.91 2.27 2.56 2.17 0.00 0.00 Bare land 2.27 84.09 2.56 4.35 4.65 0.00 Wood land 2.27 0.00 87.18 8.70 9.30 0.00 Shrub land 2.27 4.55 5.13 78.26 4.65 0.00 Grass land 2.27 6.82 2.56 6.52 81.40 0.00 Water body 0.00 2.27 0.00 0.00 0.00 100 Total 100 100 100 100 100 100
Overall Accuracy = 86.82% and Kappa Coefficient = 0.8419
c) Classification accuracy assessment of 2010
Agricultural Bare Wood Shrub Grass Water Land use land land land land land body Agricultural land 88.64 2.38 2.38 2.22 2.38 0.00 Bare land 2.27 85.71 2.38 4.44 7.14 0.00 Wood land 4.55 2.38 88.10 4.44 2.38 0.00 Shrub land 2.27 2.38 4.76 82.22 4.76 0.00 Grass land 2.27 7.14 2.38 6.67 83.33 0.00 Water body 0.00 0.00 0.00 0.00 0.00 100 Total 100 100 100 100 100 100
Overall Accuracy = 87.98% and Kappa Coefficient = 0.8558
Appendix 12: Accuracy Assessment of Metema
a) Classification accuracy assessment of 1985
Agricultural Bare Wood Shrub Grass Water Land Use land land land land land body Agricultural land 94.87 7.69 0.00 2.22 4.17 0.00 Bare land 2.56 87.18 4.26 6.67 6.25 0.00 Wood land 0.00 0.00 82.98 4.44 4.17 0.00 Shrub land 0.00 2.56 4.26 80.00 8.33 0.00 Grass land 2.56 2.56 4.26 4.44 77.08 0.00 Water body 0.00 0.00 4.26 2.22 0.00 100 Total 100 100 100 100 100 100
Overall Accuracy = 86.43% and Kappa Coefficient = 0.8372
109
b) Classification accuracy assessment of 1995
Agricultural Bare Wood Shrub Grass Water Land Use land land land land land body Agricultural land 88.89 0.00 2.63 0.00 4.44 0.00 Bare land 4.44 86.36 0.00 4.65 2.22 0.00 Wood land 0.00 2.27 89.47 13.95 4.44 0.00 Shrub land 2.22 6.82 5.26 76.74 8.89 0.00 Grass land 4.44 4.55 2.63 4.65 80.00 0.00 Water body 0.00 0.00 0.00 0.00 0.00 100 Total 100 100 100 100 100 100
Overall Accuracy = 86.82% and Kappa Coefficient = 0.8419
c) Classification accuracy assessment of 2010
Agricultural Bare Wood Shrub Grass Water Land Use land land land land land body Agricultural land 93.02 4.65 0.00 0.00 2.13 0.00 Bare land 2.33 86.05 0.00 4.65 6.38 0.00 Wood land 0.00 4.65 92.31 9.30 2.13 0.00 Shrub land 2.33 2.33 5.13 81.40 8.51 0.00 Grass land 2.33 2.33 2.56 4.65 80.85 0.00 Water body 0.00 0.00 0.00 0.00 0.00 100 Total 100 100 100 100 100 100
Overall Accuracy = 88.76% and Kappa Coefficient = 0.8651
Appendix 13: Accuracy Assessment of Sherkole
a) Classification accuracy assessment of 1985
Agricultural Bare Wood Shrub Grass Water Land Use land land land land land body Agricultural land 100.00 0.00 5.66 6.38 6.98 0.00 Bare land 0.00 92.31 5.66 4.26 4.65 0.00 Wood land 0.00 0.00 75.47 4.26 2.33 0.00 Shrub land 0.00 0.00 7.55 76.60 6.98 0.00 Grass land 0.00 5.13 5.66 8.51 79.07 0.00 Water body 0.00 2.56 0.00 0.00 0.00 100 Total 100 100 100 100 100 100
Overall Accuracy = 86.05% and Kappa Coefficient = 0.8326
110
b) Classification accuracy assessment of 1995
Agricultural Bare Wood Shrub Grass Water Land Use land land land land land body Agricultural land 100.00 2.50 3.92 2.33 2.33 0.00 Bare land 0.00 87.50 5.88 4.65 6.98 0.00 Wood land 0.00 2.50 74.51 4.65 4.65 0.00 Shrub land 0.00 5.00 5.88 83.72 4.65 0.00 Grass land 0.00 2.50 9.80 4.65 81.40 0.00 Water body 0.00 0.00 0.00 0.00 0.00 100 Total 100 100 100 100 100 100
Overall Accuracy = 87.21% and Kappa Coefficient = 0.8465
c) Classification accuracy assessment of 2010
Agricultural Bare Wood Shrub Grass Water Land Use land land land land land body Agricultural land 97.56 0.00 2.04 2.17 2.27 0.00 Bare land 2.44 91.89 6.12 4.35 6.82 0.00 Wood land 0.00 0.00 79.59 4.35 4.55 0.00 Shrub land 0.00 5.41 4.08 80.43 4.55 0.00 Grass land 0.00 0.00 8.16 8.70 79.55 0.00 Water body 0.00 2.70 0.00 0.00 2.27 100 Total 100 100 100 100 100 100
Overall Accuracy = 87.60% and Kappa Coefficient = 0.8512
111
Appendix 14: Tree Data Collection Form
Recorder: ______Date: ______Stratum: ______Plot no: ______Location: N ______E ______Altitude (m a.s.l.):______Soil type: ______
Scientific Name Local Name DSH (cm) DBH Total height (m) Remark (cm)
112
Appendix 15: Herb, Litter and Soil Sample Form
Stratum: ______Plot no: ______Location: N ______E ______Altitude (m a.s.l.):______Stratum and plot no Sample type Fresh weight Sample fresh (g/4m 2) weight (g) Herb Litter Soil Herb Litter Soil Herb Litter Soil Herb Litter Soil Herb Litter Soil Herb Litter Soil
Appendix 16: Dead Wood Data Collection Form
Tree no Standing Logged tree Felled tree Remark DBH Height Mid Length Mid Length (cm) (m) diameter (m) diameter (m) (cm) (cm)
113
BIOGRAPHICAL SKETCH
Binyam Alemu Yosef was born in Shire Enda-Sillasie, Ethiopia in 1987. He completed his elementary and secondary education at Shire Enda-Sillasie in July 2005. He attended a
B.Sc. degree in Natural Resource Management at Hawassa University, Wondo Genet
College of Forestry and Natural Resources from October, 2005 - June, 2008. After graduation, he was employed as an Expert of Land Use at the Delanta District of
Agriculture and Rural Development (December 2008 to August 2009). After that, He was employed as Graduate Assistance at Wollo University in September 2009. In 2010, he joined again Hawassa University, Wondo Genet College of Forestry and Natural
Resources for his study of M.Sc. degree in Integrated Watershed Management.
114