University of Ghana http://ugspace.ug.edu.gh EVALUATING THE SUSTAINABIUTY OF LAND AND [:-;"IRONMENTAL RESOURCES l 'T1UZATION IN THE NEW JUABEN DISTRICT AND ITS IMMEDIATE ENVIRONS THIS TI-IESIS IS SUBMITTED TO THE UNIVERSITY OF GHA:\IA IN PARTIAL FULFILLME:-iT OF THE REQUIREMENT FOR THE AWARD OF PII.D IN 'ENVIRON:\1ENTAL SCIENCE' OCTOBE R. 2003 University of GhanaD EhCttLpA:/R/uAgTsIpOaNc e.ug.edu.gh This is to certify that the work presented in this thesis was carried out entirely by myself, and all assistance received has been acknowledged. This thesis has never been presented, either in part or in whole, for a degree of any other university. uu~ . ,1/~/o-r; PROFESSOR K . S. AMANOR University of Ghana http://ugspace.ug.edu.gh DEDICATION To my late parents University of Ghana http://ugspace.ug.edu.gh QUOTATION "people are entitled 10 a healthy and productive life ill harmony with nature" (Rio declaration, /991) University of Ghana http://ugspace.ug.edu.gh ACKNOWLEDGEMENT ~~~~=dh~e ;;,i~a:r:;:~;a~~nt:~l ~~~I~?:~~~~,li~n:a~a~I~~~:a;!~ ~.h~: ~.ecf~~i~~~~ ~l::1::~:~:Z%~];J~::i::~:E:~7:::::~~::~~,:~::t~t~~:?c:ti~~~~ :~~n~~~~~~~~~~oa:,d :::!:.in!!~.I S ~~e~~o~u~~ ~~f-C~~~~m ya;:~~f.~~~~ Director General of Council for Scientific and Industrial Research (CSIR). Ghana played the dual role of supervision and a concerned brother. He commendably balanced these roles. My interactions with Professor S. K. Amanar enhanced the analysis of, particularly, the social dimension of the study Professor P. K. Yankson, the head of Geography and Resource Development Department and Director of the Remote Sensing Applications Laboratory (RSAL), was very cooperative with the provision of resources needed for the project. The ENRECA programme was the primary source of funding for the programme. Thanks to thcrepresentative of the programme to the laboratory, Professor Las serMollerfor making a case for the funding of the project. Encouragement from Professor Chris Gordon and Dr P. K. Ofori·Danson , the first and second coordinators, respectively, of the Environmental Science Programme. Professors 1. Songsoreand L. Nabiia urged me on. Discussions with friends such as Dr E. M. Attus (Lecturer) and Dr Alhassan Osman (Lecturer) who shared the RSAL with me at the beginning of the work were invaluablc in fine-tuning the projectobjeclives. Mr Obeng Kingsley (Ghana Broadcasting Cooperation), Mr Vincent Vordzorgbe (Lecturer), Mr Benard A~u Osei and Ms Esi Duodu Arhimaa read through portions of the drafts. Maxwell Osel-Owusu (nephew) checked to ensure that all cited materials were included in thc list of references. Others wh.o expressed d~p concern and interest in the work include: my siblings: Mr E. 0 .. Pabl, Sa lome~ Osel Owusu and Henry O.Pabi; friends. Kofi Adasi and Gladys Gyasl. I also enjoyed the corporation and goodwill of Mr Appiah of the Environmental Science Secretariat. University of GhanAaB hStTtpR:A//CugTs pace.ug.edu.gh Timely empirical infonnation on the slatus of sus lain ability of land resource usc iS B critical strategic input for achieving sustainable land-use policy obj ectives. ltisalso important for maintaining and improving positive achievements in sustainable resource use. Examining relevant elements of both socio...economic and biophysical factorsasfindingsarereferencedloapproprialeandrecognizedslandard5 iSlhe best approach to sustainability evaluation. Land resources of the New Juaben and the neighbouring districts have served asa medium for various land uses ovcr the years. Yet, the sustainable use of these land resources has been questioned. This study, Iherefore, soughl 10 evaluate the sustainabilityofland resource usc in the area. Four dimensions of sustainable land and environmental resource use were considered: inter-generational equity,protection of water bodies and steep s lopes from unsuitable land uses. efficiency in land allocation for suitable land uses and land-user perspectives of the suitability of cropping systems and land types. The investigation involved vegetation study, mapping of cropping systems with GPS, interviews of farmers. satellite data analysis, spatial data modelingandGeographicallnfonnation Systems(GIS)appJications. The outcome indicated that. land resource utilization has not been sustainable. Per capita land consumption for urban construction has progressively increased from about 0.04 ha to over 0.2 ha. Steep slopes and areas within 100m of the Densu River have been denuded of forest vegetation and rep laced with fallows and annual crop fanns. Cocoa has been cultivated in soil types which are naturally unsuitable. Fanner perception of cropping systems has un dergonesignificant changes, with enhanced rating given to the cultivation of food crops, and diminishing emphasis on tree crops such as cocoa. The trends of resource use could be attributed to various remOle and proximate factors. Continually monitoring and evaluatingtheperfonnanccofland-resourceutilizationcould providefeedback infonnation necessary for red irecling resource use from the current unsustainable course 10 one of sustain ability. University of CGOhNaTnEaN hTtStp ://ugspace.ug.edu.gh DECLARATION ....... . QUOTATION ... DEDiCATION ......•.......... ACKNOWLEDGEMENT..... . ABSTRACT. ........... . CONTENTS ...· ···· .. . LIST Of FIGURES .... . LlSTOFTABLES .. , .. LIST OF PLATES ..... LIST Of ABBREVIATIONS . CHAPTER ONE: INTRODUCTION Al"-1D LlTERATURF. REVIEW .. 1 1.1 Introduction .. .1 1.2 Sustainable land use ..... 1.3 Evaluation of sustainable land use . 1.4 Problemstatemc:nl. ..... 1.5 Goal and specific objectives ... CHAPTER TWO: THE STUDY AREA .• .... 2.1 Location and senlements ..... Demography .... Topography. geology and 5Oils ..... . Climate and drainage ... 2.5 Vegetation 2.6 Land use ..... CHAPTER THREE; MATERJAlS AND METHODS ... 3.1 Dimensions and standards for assessing sustainable rewurce utilization ..... . 35 Data sources. acquisilion and modeling ... 3.2.1 Slope ... 38 ] .2.2 Satel li te data processing and analysis ... 3.3 Vegetation cover and land-use pattems ................................ ... 50 3.3.1 Tree Sampling ... . ...... ............... .... ...................... , .......... 50 3.3.2 Mapping of cropping systems .. . . .... ................ 50 3.4 SUoilnanivdelarnsdistuyit aobfi lGityh ..a..n.a. . http://ugspace.ug.edu.gh 3.5 Spatio-temporal development of the Koforidua Municipality ... .. ........ 53 GIS analyses .... Land-use/coverchangedete(;tion ...... . Intergenerational equity in land allocation for urban land uses ..... . 3.6.3 Steep,lopeproteodonandland-u,econm"... . ......... 56 Protection of the Densu Flood Plain and land-use conflict. .. Land-useefficiency .. ........... .56 Land-user perception of the suitabi lity of land types and cropping systems ... CHAPTER FOUR: RESULTS ... Topographical models ... Land-use/coverdynamics from remote sensing study ........ . .................. 64 Distributionanddescriptivestatisticsofbrighmessvaluesatthetree sampling sites .......................................................................... 64 4.2.2 Trainingstatisticsfortheland-use/coverclasses ..........................6 8 4.2.3 Land-use/coverpanems.. . ................ 75 4.3 Distribution of vegetation and crops ... 4.3. 1 Distribution of trees and fa llows .. ..................... 78 ......... 87 4.4 GIS analyses of sustainable land use Land-use/coverchange .. 4.4.2 Equity in per capita land resource allocated for urban land use inthe KoforiduaMunicipality ... 4.3 Sustainable use of land resources within ecologically sensitive zon es. 4.3.1 Land-use/covercategories of2000 within the nood pJainsofthe Den su 4.3.2 Land-use/covercategoriesof2000 within slope c lasses ... 4.4 Land-use/cover, cropping systems and soil potential or suitability ........... 102 4.5 Perceptions of land use and cover dynamics and sustainable resource use ... 106 4.5.1 Land-usepressure ............................................................... 106 Dynamics of economic suitability valuation of cropping systems ......... 109 4.5.3 Perception of land suitability for different cropping systems. CUHAnPivTeErRs FitIyV Eo:f DGlShCaUnSSaI OhNtt .p..:../ ./.u gspace.ug.edu.gh....•••. 117 Slope surface model... 5.2 Remote sensing srudy ... 5.3 Land-uselcover changes and sustainability of land resource utilization ... Tree cover and sustainability of cropping systems ... Equity in per capita conversion of agriculture lands inlo built-up areas ..... 124 Sustainable utilization of sensitive ecosystems ..... . ......... 127 5.6.1 Sustainable use of sleep landscapes . ....... 127 5.62 Sustainable use of flood plains ... Efficient allocation of suitable land to land-use types for sustainability ... Perceptions of land-use and land suitability and sustainability ... Land-use pressure and sustainable land utilization .... Socio-economic and biophysical dynamics and lhe perception of sustainability of cropping systems and land types ... Views of cropping syslemsand land types and sustainable land use. CHAPTER SIX, SUMMARY, CONCLUSIONS AND RECOMl\1ENDATIONS ....... .. .... . .... .. ............................ 140 Sum mary . . ................... 140 Recommendations UniversLitIyS To fO GF hFaIGnaU RhEttSp ://ugspace.ug.edu.gh ........ 20 2.1: Locationoflhe Studyarea ...... . . .... 22 2.2: Populationgrowthofselected senlements ... . ...... 25 2.3: Topography of the srudy area ...... . 2.4' Soiltypes ofthestudy area ....... . . ... 26 2.5: Mean monthly temperature and rainfall at Koforidua ... Elevationdistributionofthearea ...... . . ........ 60 4.2: Elevation surface of the study area .... . . ......6 1 Experimental variogram orelevation fitted with a power model variogram ..... 62 4.4a: Digital elevation model of the study area... . ... 62 4.4b: Digitalelevationmodeldrapedwithelevation surface..... . ....6 3 4.5: DigitaleJevation model draped with slope surface ... 4.6: Distribution of brightness values ... 4.7: Scaner matnx of spectral valuesofTM 2000 at vegetation sampling sites ...... 67 4.8a: Reflectance distribulion for bands 3. 4 and 5 for the training areas in 1990. 4.8b: Reflectance distributionforbands 3, 4, 5and7forthe trainingareas in2000 ....... 73 4.8c: Scatter diagrams of bands 3, 4and5 for training areas in 1990 ... 4.8d: Scatter diagrams of bands 3, 4, 5 and 7 for training areas in 1990 ... 4.9a: Land·uselcoverpattem of the study area in 1991 ... . 4:9b: Land·uselcoverpattemofthe study area in 2000 . . . 4:10: Land·usc/cover ratios of 1991and2000 . .. 4: 11 : Tree abundance class distribution .... . .. ... .... .. . .... .. ..... . .................... 79 4:12: Relationship between fa llow occurrence and tree density ... 4.13: Group membership of defore slat ion levels by the first two 4: 14: Territorial map of deforestation classes .. . 4.15: Association belWeen tree and cocoa densities . 4.16: Distribution pattern of food crops and trees . .. 4.17: Group membership of con version levels from nonnal cocoa farms 4.18: Territorial map for levels ofc ocoa conversion ... 86 4.19: Land-uselcoverchange pattern from 1991 to 2000 ................ ............ ..8 8 4.20: Amount ofland·uselcover change between 1991 - 2000 ... ::ii~ r=e~~~=~:v:dc~;!~i~ ~~;~bili~~"""'" ..........8 9 ........9 1 University of Ghana http://ugspace.ug.edu.gh ........ 92 4.23 : Historical map of Kofordua Municipality . .... . . .... 93 4.24: Koforidua Municipalityin2000 ....... . 4.25: Physical growth of the Koforidua Municipality from 1920t02000 ........ . 4.26: Population changes in the Koforidua Municipatity from 1960 to 2000 ........ . 4.27 : Per capita Jandconsumption of arable land by urban land use in the KoforiduaMunicipality . .... .... . 94 4.28: Buffers around the Densu River in the 2000 land cover .. . . .... 95 4.29: Segments of Densu River covered with Forest-farms .... . 4.30: Land-cover categories of 2000 within the four buffers around the Densu 4.31 : Slope classes ... 4.32: Land-cover classes within slope classes .. .. 4.33 : Steep slopes under short fallow/annual crop .. .. 4.34: Built-up areas and forest-farms in land-suitabil ity classes .... 4.35: Proportion of land-use/cover categories in each agriculture land su ilability 4.36:Forest-farm per hectare in land suitability classes .. . 4.37: Percentage land suitability for cocoa productiyjty .. . 4.38: Surface of cocoa trees per 90m2 in 2000 .. . 4.39: OccurrenceofthehighestcocoatreedensityinthemostsuitablesoiI and 4.40: Tribal affiliation and landownership .. 4.41: Locationandownershipoffarmland .. . 4.42: Lengthoffallowperiod ....................... .. ... . . .. . ...................... 108 4.43 : Perceptionofdifficullyinfarmlandaccessibililyinfuture .. . ....... .. ... .. ..... 108 4.44a:Pastcommonly cultivated crops . .. . .............. .......... . ..... ... .. ...... 110 4.44b:Currentcommonly cultivated crops .. 4.45 : Economic suitability rating of crops by respondents ....................... . ... 111 4.46: Dendrogramofhierarchica l clusterana l ysisofcrop economlcsuitabilityratings ... .... .. .. ....... .. .. . ... .. .. .... . .. ... ............... 112 4.47: Crop preferences in future ... 4.48: Factors influencing recent crop cultivation by percentage of res pondents 4.49: Viewsofland suitability (crop productivity) based on years offa llow. 4.50: Dendrogram of hierarchical cluster analysis of crop suitability and age of vegetation cover ..... ....... .......... ... ........ .... .......................... 116 UniversLiItSyT o Of FG ThAanBaL EhStt p://ugspace.ug.edu.gh PopuJationgrowth of selected senlements in the Koforidua neighbourhood .... ........ 23 2.2: Characreristicsand crop suitability of soil associations in the study area ...... 27 Land.use pattern in the New Juaben District in 1996... .. .....3 1 Food crops and yields in the New Juaben District ...... . 3. 1: Slope suitability for maize and cassava ... 4.1: Univariatest3tisticsof brightness values of LandsatTM(2000) of the Study area ... Correlation matrix of the brightness values of Landsat TM (2000) data .......6 8 Statistics on training areas of brightness values of Lands al TM 2000 .. . 4.3b: StatisticsontrainingareasofbrightnessvaluesofLandsalTM 1991 Jeffries-Matusitaseparabilitydistancebetweenlanduse/covertypes inlheyear2000 .. . leffries-Matusitaseparabilitydistancebetween landuse/covertypes in the year 1991.. . . .................................... 72 4.6: Areasoflanduse!coverandland-use!coverchangebetween 1991 and 2000in theareaofstudy ............................................................ 75 Coefficientsofthefirsttwostandardisedcanonicaldiscriminantfunctionof deforestation .................................................. ..... ...............8 0 Structural matrix of the standardised canonical discriminant function of 4.9 Areas of selected soil types used in the analysis .. . Land-use/coverchangematrixforI99 I to2000 ........................... 90 Proportion of slope classes ... xii University of GhaLnISaT hOtFt pP:L/A/uTgEsS pace.ug.edu.gh 2.1: Exampleofforest-fann .. . ..........4 6 2.2: Example of degraded land 4.1: A narrow patch offorest remnant on the bank of the river Densu near 4.2: Sillation of there servo ira I Koforidua ... ...... .. .................................. 97 xiii LIST OF ABUBnRiEvVeIrAsTitIyO NoSf Ghana http://ugspace.ug.edu.gh Cluster Analysis Copenhagen Image Processing System Council For Scientific And Industrial Research (Ghana) Discriminant Function Analysis Digital Elevation Model Proramme for Enhancement of Resean:h Capacity in Developing Environmental Protection Agency Environmental System Resource InstilUie Ghana Environmental Resource Programme Geographic Infonnalion Syslems GPS Geographic Positioning Systems Infra-red Iterative Self-Organizing Data Analysis Institute Of Social, Statistical And Economic Research Landsat Land Satellite Red-Green-Blue Colour System. Remote Sensing Applications Laboratory Triangulated Irregular Network Thematic Mapper xiv University of Ghana http://ugspace.ug.edu.gh CH\I'TF:I{O'\E INTRODLCTIOI\ A'\ll LlTF.RATURE REVIEW L1lnlroduclion Land resources can sustai n the vital life-supporting roles provided the mode of exploitation guarantees sustained utilization, In the past, human preo ccupationhasheen mainly the slrUggie 10 carve out a niche in me biosphere. However, that quest now threatens the very conditions necessary for sustaining human life (Holling and Sanderson, 1996). Sound management of land and environmental resources utili zation should bean important feature in land-use strategies ifsustainabil itycanberealized The importance of sustainable development was endorsed by the Earth Summit in Rio de Janeiro in 1992. which eventually led to the fonnulation and implementation of national strategies for sustainable development. lntemationally, regionally, nationally and locally, this approach to resource use has been acknowledged, and c onstitutesthe subject matter for chapter 10 of Agenda 21(UN, 1993). The need to sustai nably manage land and environmental resources use is a guiding policy for all land-based developments in Ghana (Ministry of Land and Forestry, 1999; EPA, 1994). Managerial activities such as periodic evaluation of land resource qualities and efficiencies of the perfonnances of land-use strategies, in terms of biophysical impacts and productivity,cnsures that implementation leads to achievement of desired results in University of Ghana http://ugspace.ug.edu.gh the long term. Sustainable land-use management must be informed by timely. real and adequate feedback infonnation on land-use perfomlance and the status of land and environmental resources, rather than speculations. In Ghana, however, such a systemalicapproachofregularevaluationasamanagementlOolforsustainablelanduse has not been properly harnessed. Already, this has generated land-use problems with significant implications for long-term use of land and environmental resources in the counuy (Ministr)' of Lands and Forestry, 1999). 1.2 Sustainable land use The concept of sustainable use of land resources is pertinent onlyagai nstabackground of limits to land resource avai lability and use . Ifno such limits exist, ortheyarenol perceived to exist. then it is common that resources are overexploited. However, the coocept of sustain ability becomes increasingly important, as the scarcity of the resource increases due to rising demand and deterioration in resource quality. The dictionary meaningofsustainlsustenance is 'to keep in ex isten ce' or 'to provide for the support of. It is also defined as 'maintaining the same level of activity or pace of plan. method or system without harming its efficiency and the people affected by it' (University of Cambridge, 1996; SBC World Service, 1992). Inferentially, sustainability essentially means keeping life on earth inexistence by maintaining the nat ural levels of the earth's supportive functions intact, or within reasonable limits of 101 erance. Thus, the ultimate goal of sustainable resource use is not to sustain human activity but humans University of Ghana http://ugspace.ug.edu.gh themselves. Whereas measuring sustainability of human well·being is doubtlessly problematic. sustainable production and consumption is easier to measure, an dcouldbe considercd as a proxy for the sustainabilityofhuman well·bcing Many definitions have been proposed to describe aspects of sustaina bledevelopment. The multiplicity and variety of definitions refleci 'the existence of multiple perspectives and their complex inteNeJationships. l-Iowever, this rcview gives the highest attention and priority to sustainabilitydimensions that are relevant to the uti liz ation of land and environmental resources. From an ecological perspective, sustainabilityin land and environmental resource use can be equated to the concept of resilience. Whereas the concept of resilience as an adaptation 10 disturbance appears to be first articulated by Holling (1973), Westman (1985) seems to be the first to formally pose the question of whether the degree, if any, ofasyslem to recover after natural disturbance may assist in the process of restoration after man·made disturbance. Westman (1985) defines ecosystem resilience as the degree. manner, and pace of the restoration of initial st ructure and function in an ecosystem after disturbance Fox and Fox (1986) ill ustrate that disturbance·adapted systems appear to display a degree of resilience to human·induced dist urbance on a scale relative to the magnitude, duration and type of disturbance. University of Ghana http://ugspace.ug.edu.gh For managed ecological systems such asagro-ecosystems, sustainabil itystressestheco- evolution of socioeconomic and natural systems (Reijntjes e/ al., 1992). Sustainable agriculture is a philosophy based on human goals and on understanding the long-term impact of our activities on the environment and on other species. The philosophy seeks to guide the applieation of prior experienee and the latest scienti fic advanees to create integrated. resource-conserving and equitable farming systems which reduce environmental degradation, maintain agricultural productivity, promote economic viability in both the short and long term, and maintain stable rural communities and qualityoflife (Fr.mcis and Youngberg, 1990) Sustainable agricultural land use should successfully manage agricultural resources. satisfy changing human needs while managing or enhancing the quality of the environment and conserving natural resources (TACICGlAR. 1988). The definite reference to and emphasis on changing human needs confers a dynamic connotation to sustainable utilization/management of land and ecological resources in agricultural sustainability (Serageldin, [996). The clear allusion to changes in human needs p'''"pposes a period;, monitoring of human needs to detelm;ne the relevant usc "g;mes~~-.:0> o"~'! i that best respond to prevailing human needs. z.·I · .. . ~ The International Union for the Conservation of Nature (TUCN) emphasizes socio- economic and intergenerational equity as an important issue and dimension i nsustainable resource use. As a value, it prcsupposes thal in the decision making toallocate resources, University of Ghana http://ugspace.ug.edu.gh equal weight should be given to the present and future generations. In other words. sustainable development must be a process of social and economic beHermenl thai satisfies the needsaod valuesofalJ interest groups, while maintaining future opt ions and conservingnatwairesourcesanddiversity(IUCN, 1980). Intergenerational equity requires the maintenance of resource stocks- trees, soil,w8ter, etc, and 10 avoid substantial decline in quality and quantity of resources with time (Markandya and Paree, 1988). In a dynamic system such as human society, sustainability is fundamentally a question of balance maintained over time (Dahl, 1995). Thus, sustainable resouree use may be seen as the lack of forces tending to upset equilibrium over time. Accordingly, most indicators of sustainability are, in fact, measures of unsustainability, that is, the amount or extent of imbalances. lntergenerational equity suggests a judicious use of natural resources in a way that ensures that their capacity to support the development of future generation is not compromised (Ser ageldin.1996). In this case,sustainabledevelopment focuses directly on the populatio n that can be sustained by a given territory, that is, its carrying capacity (FAO, UNFPA and IIASA, 1982; Vitousek el al., 1986). Such capacity depends on the desirable qual ity of life of the population in the territory, which would have to be defined in tenns of essential (minimal) or desirable standards of living Compounding practically all the sustainability dimensions, Smyth and Dumanski (1993) define sustainable land resource management as combining technologies, policies and University of Ghana http://ugspace.ug.edu.gh activities that aim at integrating socioeconomic principles with environmental concerns so as to simultaneously maintain or enhance produclivity; reduce the level of production risk; protect the integrity of natural resources and prevent degrad ation;and besocio- economically viable. Even though this may seem a more comprehensive explanation, it equally complicates assessment of sustainable utilization and management of agro- ecological resources 1.3 Eva luation ofsuslain able land use To move towards the goal of sustainable development, decision-makers at all levels (locally, nationally, regionally and internationally) need information. The primary aim of chapter 40 of Agenda 21 is to help decision-makers at the national level 10 access information. The information is to assist the understanding of how well we are navigating the road to sUSlainable development. The concept of sustainable resource use is dynamic in that what is sustainable in one area, may not be in another, and what was sustainable al onc time may no longer be sustainable today. Although sustainability cannot be measured directly, assessments of sustainabil ity can be made on the performance and direction of the processes thai control the fu nclionsofagivcnsyslem It a specific location (Smyth and Dumanski, 1993). The problems associated with monitoring and evaluating progress towards sustainable development are highlighted in chapter 40 of Agenda 21. The divergent views on University of Ghana http://ugspace.ug.edu.gh indicators of sustainable development has necessitaledthe need to re achconsensusona suitable sel of indicators Utal can adequately reflect Ute wide range of concerns encompassed by sustainable development, as promoted in Agenda 21 . Though several of these indicators have been adopted foruseattheintemational leve l,these are subjected tonationalsovereignryconsidcrations(UnitedNations.1993) Change is not. necessarily. undesirable in relation to the requirements of sustain ability. The effects of change in one unstable factor may be more than offset, ina posiliveor negative sense, by change in some other factor. Herein lies the difference betwecn stability (concerned only wilh the likelihood of change) and sustainabi lity(concemed with the balance between positive and negative change in relation to a particular system). Changes in a land·use system with time reflect instability in one or more individual factors. An understandin.g. o.f the likely direction and n atureofsuchinstability ,-~.~.~~;".- ",.. ~~ ~.,o! ,.';' . Given the post implementation assessment of sustainable use of land and environmental ~ resources in the presem discussion, it will be appropriate to view 'sustainability' asa measure of Ute extent to which a form of land use is meeting the pillar requirements of productivity, security, protection, viability and acceptability. Thus. sustai nabil ity evaluation refers to a systematic process of post·implementation assessment of real impacts of policy objectives of individuals, institutions and nations in the light of University of Ghana http://ugspace.ug.edu.gh susrainability dimensions. The prime aim is to identifY impacts, such that positive effects maybe enhanced, and action taken to counter negative ones. either through selection of appropriate policy alternatives, or through parallel mitigation and enhancement measures Five years after adopting an agenda for implementing the visions of the 'Earth Summit', it was realized at the Rio +5 forum that there had been little success in translating commitments to realities. The purpose of the summit was to move sustainable development from agenda to action. Before the 2002 Earth Summit, a strong suggestion for a recommitment to the ideals of sustainable development was made by UNEP (2000). But this could be done only after a thorough evaluation of the outcome following the implementation for some time now. To achieve the global, regional and nalionalobjectives. it is obvious that aggregation of local andsi tuationaldevelopments be made first. That will provide insight into inherent patterns of variations in successes, failures and challenges. Knowledge of real experiences will infonn futu re decision- making processes for any needed adjustments. This will ensure that the trajectory of sustainability in the utilization ofagro-ecological resources is mai ntained on the local level and also provide information on the spectrum of opportunities and challenges to the common national and international sustainability in the utilization of natural 1.4 Problem stalemenl University of Ghana http://ugspace.ug.edu.gh Land resources are in limited supply, finite. usable and vulnerable to injudicious use (WeED, 1987). The effects of land-use become critical as it progressively intensifies with growing human populations, and expanding multi-purpose use of land aggravates land-use conflicts. Even with the most technologically advanced modes of harnessing land resowces, inherent deficiencies constrain achievement of optimally sustained productivity,withminimal degradation of the land resource base. Relative viability of land-use types also undergoes constant changes, and should be regularly assessed for utility from the perspectives of land users. This affords the opportuni tyto adopt land-use types that offer the opportunity for optimizing land-use benelits. Among the objectives of Chapter 10 of the United Nations Sustainable Development Agenda, entitled. 'Integrated Approach to the Planning and Management of Land Resources'. improvingandstrengthening ... evaluationsyslemisdeeme dimportant for achieving the ideals of sustainable development. Given the intrinsic limitations of land- use strategies, it becomes a prerequisite that land-use management incorpora te regular and timely evaluation exercises to yield the needed feedback infonnation on the perfonnanceofland-useslrategiesafterimplementations. The deviations of actual efficiency of resource utilization from adopted standards for sustainability furnishes stakeholders with information on current degree of the sustainability of land resources utilization. Havingdonethis.cont rol measures could be applied tore-direct the trajectory of present resource utilization, if found unsustainable. University of Ghana http://ugspace.ug.edu.gh There is also an opportunity to review and adjust management policies, objectives and criteria for resource utilization for achievable options that reflect prevaili ngrealities Such feedback infonnation on challenges and opporrunitiescould guide future decision- making processes by ensuring that land-use performance is optimized to achieve sustainable land-use policy objectives. Land and environmental resources are important media for land-based productive ventures in Ghana. Nationally, high percentageoflhe population directly depends on land and environmental resources for livelihoods. Between 1984 and 2000, the country experienced demographic growth of 53.8% (Ghana Statistical Service, 2000). Demographic growth would likely continue to be a developmental issue. According to statistics, 49.2% of the work force depended on the agricultural and related sectors from the late 1990sto 2000 (Ghana Statistical Service, 1987). During the same period, the agriculture sector contributed approximately 40.0% of the national GOP, a percentage that was far less than the contribution made in the early 1980s, when 61. 1% of the workforce were employed in the sector (Ghana Statistical Service, 1987). In the year 2000, the percentage contribution as reported by the Institute of Statistical and Social Economic Research was 35.5% (ISSER, 2001). Like other developing economies, agriculture and other land-based related sectors are crucial for future economic development University of Ghana http://ugspace.ug.edu.gh Increasing population and over-dependence of me economy may have deleterious consequences for land through land-use presswe where there are deficiencies in land management regimes. It is realized that interaction of rapid population and growth could lead to continuing degradation of the environment and diminished resources productivity (Oti-Boateng et 01., 1989). Other studies have shown that over-population results in accelerated deforestation (Origo, 1999; FAO, 1998; Sera, 2000). Obviously, the overall effect of population growth on land could be mediated by other fa ctors such as existence of employment alternatives and adoption of efficient farming systems. For instance. a study by Simmance (1997) indicated that the degree of imp act of population on land depends on the intensity of land use and the fragility of existing ecology. According to Boserep theory, however, moderate rather than very low population densities encourages efficient land use through intensive and senled agriculture land use (S,hultz,1990). Historically, land resources in Ghana have been managed lhrough various legal instruments and customary practices applied through the coun system (Ministry of Lands and Forest, 1999). The counlry has also ratified several environmenlal and sustainable development-related international and regional agreements. Since 1991 , it has adopted and fonnulated a National Environmental Action Plan to guide sustainable use of land and ecological resources (EPA, 1991). University of Ghana http://ugspace.ug.edu.gh Despite the institution of management strategies and commitments to sustainable policy objectives,thereisconcreteevidencetosuggeslthatthereisasignificanl lag of real achievements behind aspirations raised by national policies, formal commitments and published statements. Investigations on land-use impacts in Ghana (Osman. 2000; Pabi, 1998; Gyasi and Uillo. 1997; Attua, \996) have revealed serious transformations of the pristine vegetation cover into shades of fallow and degraded lands (Pabi and Altua, 2002). The implications for soil fertility decrease and erosion have been strik ing (Attua, 1996; Amanor, 1994),most of the good agricultural lands are)osing their capacity to su pport certain agricultural land uses they previously used to (Amanor, 1994). In 2000. an estimated 220,000 km2 of land in Ghana was affected by sheet and gully erosion (FAO/AGL, 2000). Nutrition depletion is also prevalent; this is due to reduction in fallow cycle, intensive cropping and inadequate application of fertilize r. This is having a longtenn effect in crop yield reduction and annual productivity lossesof2 .9 percem in all crops and livestock (FAOIAGL. 2000). Impacts on Ghana's economy have been On a local scale, the New Juabeng and the neighboring districts express the above scenano in a pronounced manner. In addition to heing the smallest District in Ghana. University of Ghana http://ugspace.ug.edu.gh New Juabeng has many sensitive and fragile, yet, important ecosystems such as river basins and hilly landscapes. The population size in the 2000 was 136.768. which was an increase of 48.0% over the 1984 figure of 92.482. The population density of approximately 1243 persons/km2in2000, which is one of the highesl inlhecountry, is exceeding1yabovethe2000nationalaverageof79.3persons/km2.lntherural communities, the majority of the people directly depend on land through agriculture. Studies indicate that agriculture and urban land uses are affecting the frag ileand sensitiveagro-eco)ogicalsystems (Pabi, 1998; Osman, 2000). Whereas deforestation is a problem in the district, it is pronounced near Koforiduathan the comparatively rural senlements. For instance, in a study of vegetation near Koforidua and other rural settlements, it was found that the area around Koforidua was highly deforested compared with the comparatively rural settings. The pattern was the same for species richness reduction (Pabi, 1998). Over the years, Koforidua has rapidly expanded spatially due to constructional activities; however, per capita ann ual change is unknown. In terms of housing population of the district, the number almost tripled between 1970 and 1995 (Ministry of Local Government, 1996), which obviously, means a corollary consumption of agricultural lands. Whereas,tbe above indications raise fundamental questions about the efficacy of land and environmental resource use, they are not suggeslive ofadefinitive unsustainability or otherwise of land resource use. The reason is that the levels of observed resource utilization efficiency is not measured against any recognized and definite sustainable resource utilization policy objectives or standards of performance-locally, regionally or internationally. Thus, without a substantive evaluation programme, no clear feedback University of Ghana http://ugspace.ug.edu.gh infonnation on the perfonnance of land-use implementation strategies and the degree of achievement of sustainable resource utilization objectives could be identified. This leaves the sustainability of resource utilization question of the di strictsunanswered,and accordingly, no defmite basis for directing future land-use management process. Naturally, without definite infonnation on sustainability of land-use efficiency, management planning decisions will be ineffective. In the five-year (1996-2000) development plans for the administrative districts in Ghana, post implementation evaluation is identified as an important instrument for achieving land-use and environmental managemem policy objectives. The need for building capacity to carry out these managerial activities was also emphasized as an important strategic policy (Ministry of Local Government, 1996). Evaluation of sustainable development at the district level provides a basis fore stablishingaggregated estimates for the nation. Appropriately, the study considered it crucial to investigate relevant sustainability dimensions of land-use in the New Juabeng Di strict in the light of the following issues: inter-generational equity in per capita conversion of land resources into urban effecliveness of allocating suitable land (soil) categories to cro p types that ensures sustained optimum productivity with minimal land degradation; protection of fragile and sensitive but important ecosystems from destructive effects of land use; and University of Ghana http://ugspace.ug.edu.gh viability of cropping systems and suitability of land resources from land-user perspective. Certainly, the need to enhance the capacity fora managi ng land resource has been fully acknowledged, yet, practical achievements have nOI proved encoura ging enough (FAO, 1996; FAD, 1993). Increasingly, integrated or systems approach to finding so lutionsto landscape problems is being recommended (Margerum and Born, 1995). This is due to the complexity and adaptive nature of ecosystems and land-use interactions (Lambi n et aI., 2001). Clay10n and Radcliff (1996) specifically describe a system approach to sustainabi li tystudies as offering a basis for liaison between disciplines on philosophical, methodological and technical levels that will allow their diverse mod el ingtobecoupled with one another. It provides the means of linking the subsystems - hard (HalJ, 1962) and soft (Checkland, 1981),draws on the interactive and control properties of complex and adaptive systems and then integrate biophysical andsocio-econ omicfactorswithina common framework Agenda 21 makes a strong recommendation fo r the use of science and te chnologyasan Important component of the strategies for achieving sustainable development (WCED, 1987). In this respect, the FAD, among other organizations and individuals recommend the integration of remote sensing and geographic information system (G IS) and geographic positioning system (GPS) technologies in land related studies (Yeh and Li. 1997; Hallet et aI, 1996). Particularly, pulling their specific and synergistic advantages University of Ghana http://ugspace.ug.edu.gh to bear upon research makes possible cross-fertilization of ideas and fonnulation of methodologies and analysis that was practically impossible hitherto (Eastman et aJ. 1993; ESlesandStar, 1997; Lange and Gilbert, 1999, Star ef al., 1997) Particularly, as sustainability motivales and guides resourcemanageme ntobjectives,the direction fordevelopmenlandapplicalionofGlS in landscape cons ervation should be towards the linkage and integration of human and physical environmental systems (Flamm and Turner, 1994; Hobbs el ai, 1993). This helps to investigate and reconstruct human-induced Iransfonnation spatially and historically on different scales (Skole and Tuker, 1993; Acevedo er ai, 1996; Brondizio et ai, 1996) an important element of sustainable land resource use evaluation. Evaluation research is pursued in several venues in a GIS environment. For instance, Talen (l998) applied GIS to examine the equity in distribution of public services to various segments of the community, and to search for areas that have not received adequate planning attention and have not been allocated their fair share of public resources. In an earlier study, Talen (1996) used GIS to assess achievement ofopeo space goals by comparing the planned and implemented projects. Knaap el af (1998) have also used GIS in therrevaluation of the relevance and olltcomeso fplanning. University of Ghana http://ugspace.ug.edu.gh However, remote sensing and GIS techniques alone are incapable of capturing the socio· economic dimensions of sustainable resource use. Of course, many of these socio- economic and cultural factors are the underlying driving forces that control thecha nges reflected in remote sensing information. Given this limitation. a case has been made to combine the potentials of relevant disciplines in studies involving land use and land Though,therehave been some attempts to use these technologies in land-lise studies in Ghana. there has been practically no applications to post-implementat ion evaluation in relation to planning purposes. applications are generally restricted to land use/cover mapping that do not adequately yield the needed information for the formulation of land-use management strategies. In this study, though. a comprehensive use was made of these technologies to achieve the objectives of the study. Considering the difficulty imposed by the multi·faceted nature of sustainable land use, the study focused on selected aspects of sustain ability of land use, which were considered important and relevant. I.SGoalandspecificobjectives. The main goal of the study was to provide information for decision making to enhance the achievement of sustainable land and environmental resource management objectives in the study area. To lhisend, it sought to achieve the following specific obj ectives' University of Ghana http://ugspace.ug.edu.gh Find out if land types are allocated to crop systems for which they are potentially suitable. To find out the viability of cropping systems and land suitability from l and~user perspective. To determine whether nat ional land~use directives respecting the protection of sleep slopes and floodplains are being complied with. Provide information on the degree of inter~generalional equity in terms of per capita conversion of potential agriculture lands into urban buil t~up . Determine land~use/cover change between 1991 and 200 I Demonstrate how remote sensing and GIS can improve evaluating procedures of land~useperfonnance in the light of sustain ability criteria University of Ghana http://ugspace.ug.edu.gh CHAPTER TWO T HE STUDY AREA 2. 1 Locafion and settlements The study area is sandwiched between latitude 0°, 10' W and 0°.24 ' W, and longitude 5°.55 ' N and 6° . 15' N. It is located in the upper region afthe Densu River Basin . Though the New Juaben District was the primary focus of the study, portions of neighbouring districts· Vi la Kroho, Akuapem North. Suhum Kraboa Coaltar, and East Akirn Districts. that interface directly with the New Juaben were considered for a couple of the remote sensing studies and the GIS analyses. The reasons for inc1uding parts of the neighbouringdiSlriclS in the study were due to the irregular shape of New Juaben Districl,spatial configuration of the settlements in the area and expansi on ofseuJements in New Juaben into the neighbouring districts, the need for investigating specific environmental phenomenon over a wider spatial scale andconvenienc eo By size, New Juabeng District is the smallest district in Ghana. Of a total land area of IIOkm square, about 27.0% was built-up, making it one of the most built-up districts in Ghana in 1996 (Ministry of Local Government, 1996). In space, the settlements are made ofa centrally positioned conurbation and other minor settlements at diffe rent stages of socio-economic development. Koforidua, the most urbanized settlement, fonns the central core of the conurbation. Surrounding Koforidua are relatively less urbanized towns including Koforidua Ada., Effiduase, Asokore and Adjeso (Fig.2. 1). Structurally, University of Ghana http://ugspace.ug.edu.gh it is impracticable to differentiate Koforidua per se from the other towns; it has grown and physically merged with them to fonn a single massive built~up conurbation (Fig 2.2). Practically. the conurbation has expanded into the Akuapem North District; a siruation that has created administrative and planning challenges. Acoording to the Medium Term Development Plan publication, there were 52 settlements in 1996 - most of which were villages. The area is heavily urbanized and there is an acute land shortage for other important land uses. Land scarcity is acknowledged as a major developmental constraint in the district (Ministry of Local Government, 1996).ln I 996,about 90.00/0 of the population lived in urban centers. Figure2.1:Locationoftbestudyarea University of Ghana http://ugspace.ug.edu.gh 2.2 Demography Demographically. New Juaben District is multi·ethnic, with the dominant tribe being the Ashanti. Other tribes such as the Akwapems, Krobos. Akims, Ewes and other Northern Ghana tribes from northern Ghana are fairly represented. The population of the New Juabeng District was 52,796 in 1960, increasing to 68,607 in 1970. By 1984, the population was 92.482, of which 77,608 lived in the conurbation. The total population for the district in 2000 was 136,768 (Statistical Service,2002).)n 1984.82.8%(82.566) lived in urban senlements, whereas the remaining lived in rural areas, The proportion of urban population was 83.4%, which isa slight increase above that of 1984 (Ghana Statistical Service, 1987,2002). Population density of the district was far higher than lheregiona1 average in 1984. II must be stated that the bu lk of the population lived in Koforidua. Comparatively, Koforidua has been undergoing significant demographic growth. In 1960, the population of Koforidua was 34,856; rose to 46,234 in 1970, and reached 58.298 by 1984. In 2000, the population stood at 88,093. The historical population pattems of important senlements in the conurbation are presented in Fi gure2.2. University of Ghana http://ugspace.ug.edu.gh Figure 2.2: Popuilltion growtb of selected Hnleme. . " University of Ghana http://ugspace.ug.edu.gh TABLE 2.10 POPULATION GROwn! OF SELECTED SETTLEMENTS IN TIlE KOFORJDUA NEIGHBOURHOOD. Settlement 1960 1970 1984 2000 Koforidua 556 773 1609 2543 Ada Asokore 2965 4206 5892 10068 Effiduase 3907 6207 7681 10063 Koforidua 34256 46235 58731 87315 Oyako 4113 5061 6612 Akwadum 298 2114 Karle 251 1392 Nkwanta Source: Ghana StatiSlical Service, Accra University of Ghana http://ugspace.ug.edu.gh 2,3 Topograpby,geoloe,yandsoil!i The Voltruan scarp is an unportant geological feature of the northern section of the region. This consists of chalk and coarse sandstone wilh some clay shale inlerstratified (Junner and Hirst, 1946 The Landscape (Figure 2.3) does not show much attitudinal variarions; it ranges from about 152 to 200 m above sea level The plateau overlooking Koforidua andlrendmg a north-west direction forms an imposing feature in the region. It peaks at about 200mintheAboaboarea. The presence of exposed outcrops is a major characteristic of these highlands. To the southeast lies another prominent plateau range with a relatively low height. Lying between these two is a valley that serves as a passage for Nsukwao (a stream) and the Koforidua-Nkurakan road. From Koforidua towards the western section of the area. the terrain is Iow-l)1ng and flat Wlth gentle undulations and occasionaUy low hills rising abruptly ahove the generally low elevation On a broad scale of soil classification. Forest Ochrosol is the dominant soil type in the atea, with the Adawso-SaWJI3Se Association being the oomrronest (Figure 2.4). Other soil associations such as Koforidua and the Nankesc Feric Luvisols (FAD, 1988) that are dissec1cd by streams and the Densu River arc restricted to a few locaJities (Walker, 1962). These are developed from parental rock material of granite. Characteristically, these soils are dark brown, sandy with clay and gravels of quartz. loam and humus COtq)Onents (Adu and Asiamah, 1992)_ The characteristics and suitability of soil types in the area are summarised in Table 2. 2 (Adu and Asiamah, 1992) University of Ghana http://ugspace.ug.edu.gh Figure 2.3: Topograpby oftbestudy area University of Ghana http://ugspace.ug.edu.gh Fi~ure2A : Soil Iypesofthe study area University of Ghana http://ugspace.ug.edu.gh TABLE 2. 2, CHARACTERISTICS AND CRO P SUITABILITY OF SOIL ASSOCIA nONS IN THE STUDY AREA. CHARACTERISTICS CROP SUITABILITY ASSOCIATION Koforidua-Nankesel a Brown and reddish clay Cocoa, oil palm, coffee. citrus. Nta-Ofin Compound loarns developed on plantain, maize and cocoyam summits and upper slopes over weathered biotite? and granodiorite b Strong brown loarns developed in colluviwn on middle slopes Fete-Bediesi c. Brown soils developed Plantain, Complex over hard quartz. vegetables and rice. d. Red and brown sandy c1aysandloamsdeveloped in deep piedmond drifts; redconcrecretionaryclays on upper sJopes. e. Yellow -brown and grey- brown al luvial sandy clays on lower slopes and on valley bottoms University of Ghana http://ugspace.ug.edu.gh Adawso-Ba~iasel a. Pale yellow brown and Cocoa. coffee. oil palm. cilrUS. Nta-OffinCompound redish brown gravely and maize, cassava. plantain and concrecretionary clay cocoyam. loams and clays developed di rectly little weathered biotite granites on swrunits and upper slopes with areas of red and brown loarns derived from peneplain drifts. b Pale brown and mettle grey alluvial sands and clays on slopes and in valley bottoms Soun::e:AduandAsiamah,1992. 2.4 ClimateaDddr-ainage Being tropical, thetemperarure is consistently high the whole year round (Figure 2.5). The minimwn and maximum temperatures occur in the months of July to August and February to April respectively The mean temperature is approximately 27 °C, occurring from February to March and 26 °c from July to August (Dickson and Benneh, 1988) Rainfall regime is hi-modal (Figure 2.5). The major rainy season occurs between March/April to July. while the minor season starts from around September to November University of Ghana http://ugspace.ug.edu.gh (Fig 2.5). Rainfall is normally heavy with associated thunderslOrms. with a mean annual rainfall range between 1200-1700mm. Specifically. the mean annual rainfall of Koforidua and Oyoko are 1425 and 1504mm respcctively (Adu and Asiama. 1992). Results ofanaJysis of rainfall records from 1966 to 1996 revealed a steady decline in the amountofannuaJ rains in the area, though rainfaJl days remained practicaJlyunchanged (pabi,I998). From December to February, the area comes under the influence of the HannaHan winds that precipitate the onset of the dry season. Humidity is generally high, with the values reducing slightly during the dry season. It is usual for the humidity to approach IOO.O'/.in the night Figurt' 2.S: Mean monthly temperature and raiofall University of Ghana http://ugspace.ug.edu.gh The Densu River is the biggest in the area, with Asuoyaa, Kwniabena, Okumi and Nsukwao being its tributaries. The plateau ranges in the northeastern and the southeastern parts of the area serve as recharging sources for the waler bodies. The scanty recharging capacity ofthesehlghlands is incapable of sustaining these rivers the year round. Human activities such as fanning and timber felling and bushfires have destroyed mosl of the vegetation along the banks of the river. Consequently. they are rendered seasonal,reducing to shal low pools in the dry season while swelling in the rainy season 2.5 Vegetation Consistent with the above climatic regime, the area has a semi-deciduous forest of the Celtis-Triplochiton Association (Taylor, 1952). However, with progressive intense land- use pressure such as logging and farming, the originally dense vegetation has been completely modified,losingits pristine status. Only patches of the original vegetation exist in the most inaccessible areas such as relatively high elevations (Adu and Asiamah,I992). Species richness has reduced in many areas, particularly, around the urban senlements (pabi, 1998). Most of the tree species remaining of the natural vegetation are Triplochifon sc1eroxylon. Anfioris ofricono. Ceiba pentandro, Cola giganlea. etc .• most of which are characteristic of relic forests. Large trnctsofland are covered with food University of Ghana http://ugspace.ug.edu.gh crop fallows at various stages of maturity, Pan;cum maximum and Chromolaena odorata (Adu and Asiamah. 1992). 2.6 Land Use As already stated, land scarcity is a critical developmental issue in the New Juabcn District. Table 2.3 shows the land-use pattern as it existed in 1996 (NOpe, 1996) TABLE 2.3, LAND-USE PATTERN IN THE NEW JUABENG DISTRICT IN 1996 TOWNSIVILLAGE AREA (km ) % OF AREA Land available for agriculture and otherde"e!opment 50 45.0 Settlements 27 24.0 Rivers and Forest Belts 17 15.0 Steep slopes Roads, Railways, and High tension lines Total Land Area Sow-ce. National Development Plaruung Commission, 1996 Fanning remains an important occupation, especia11y, in the rural settlement s.ln 1984, though, the percentage of the labour force engaged in farmi ng was 24.6 %, which was below the regionaJ level of more than fifty percenl (Ghana Statistical Service , 1981). A ba.selinesurvey conducted in 1996 on occupations in the area indicated that 60.0 %were engaged in farming (Ministry of Loca1 Govemment, 1996). University of Ghana http://ugspace.ug.edu.gh Conditions of the soil and climate make the area suitable for the cultivation of cocoa, cola nut, oil palm and other such tree crops (Table 2.2). These crops are sti ll important. especially, around rura1 towns like Oyoko, Akwadum and Suhyen where forest vegetation persists. Though cocoa used to be very imponant around Koforidua, replanting after swollen shoot and capsid attacks in the 1950s has seen minimal success. This is principally due to the prevalence of drought, bush fires, deforestation and unsustainabJe food cropping systems. As usual,cocoais planted under the canopy offoresltrees. TABLE 2.4, FOOD CROPS AND Y1ELDS IN THE NEW JUABEN DlSTRlCT Major Crops Number Percentages Actual Yield of farmers of fanners (hectares) (tonnes) 12.8 900 1600 Cassava 300 17600 Cocoyam 200 17.0 100 Plantain 100 Tomatoes 150 100 Garden eggs 10.3 100 Pepper 60 1000 Oil Palm 100 Citrus 50 1000 Total 1170 100 3150 Source, National Development Planning Commission, 1996 University of Ghana http://ugspace.ug.edu.gh Fanners cultivate food crops bom for consumption and sale. Essentially. me traditional slash-and-bum memod is used in clearing land for fanning. Of the major food crops cultivated in the district, cassava and maize are the most popular among the farmers (Table2.4). In the year 2000, statistics by the Research and Information Directorate of the Ministry of food and Agriculrure indicated mat cassava and maize were the most common food crops in tenns of size of area cultivated. Vegetables, cocoyam and plantain are also cultivated but on a relatively minor scale (Ministry of Local Goverrunent, 1996). A survey showed that majority of the fanners engage in small-scale fanning; 61 .1 % of the farms were less than 3 hectares (Ministry of local Government, 1996). The problem of land accessibility for large farming is compounded by the existence of poor soils in some areas, oUlcropsand the uneven nature oft he landscape Wood-fuel fills the domestic energy needs of most inhabitants of the area. Logging is another important economic activity in the area. The Densu River is the main source of water for all activities. A reservoir constructed on it serves Koforiduaand surrounding townships with potable water. Unfortunately, seasonal water shortage has become an annual experience in the towns University of Ghana http://ugspace.ug.edu.gh CHAPTER THREE MATERIALS AND METHODS A typical evaluation study requires a series of standard management procedures. including: (I) adoption of standards of performance for sustainable land resource utilization, (2) datagalhering on actual performance of land resource uti lization and (3) comparison oflhe standards and Iheactual perfonnance (Cole. 1996; Hellriegel and Slocum, 1978). These broad procedures were expanded by incorporating relevant elements of the BelJagio Principles· Guidelines for Practical Assessment of Progress Toward Sustainable Development (IISD, 1996). Among others. the principles make the fo llowing recommendations for the assessment procedure' guidance bya clear vision of sustainable development and goals that define the vision; consideration for the well-being of social, ecological. and economic sub- systems. their stale as well as ihe direction and rate of change of the state of their compllnent parts, and the interactions bet\veen parts; consideration for cquity and disparity wilhin the current population and between presenland future generations, dealing with such concerns as resource use, over-consumption, etc., as appropriate; an explicit set of categories or an organ izing framework that links vision and goals to indicators and assessment criteria; comparing indicator values to targets: and reference values, ranges, thresholds, ordirecti on of trends, as appropriate. 34 University of Ghana http://ugspace.ug.edu.gh 3.1 DIMENSIONS AND STANDARDS FOR ASSESSING SUSTAINABLE RESOURCE UTILIZATION Four dimensions were adopted for the conceptofsus[ainability ofresour ceutilization. Corresponding indicators and standards/criteria for sus[ainability of land and environmental resource Ulilization for the assessment procedures were adopted by making references to national sus[ainability policy objectives and recommendations byinstitutionsandindividualsonsustainablemanagementforresourceutilization. The four dimensions were as follows • Equilable per capita allocation ofp otential agricultllre lal/d resource for urban This assessment was carried oul on the Koforidua Municipality, mainly against the matrix of pronounced scarcity of land and widespread inefficienlconvers ion of prime agricullure lands into urban land uses in the New Juabeng District (Minislry of local Government, 1996). Such a situation necessitates efficient use of land for urban development to ensure inter-generational equity and accessibility of land for other competing land uses such as agriculture. The hypothesis for the evaluation is that differences in per capita land allocated for urban land use (built-up areas) in the KoforiduaMunicipality should not be significantly different betwee ngenerations. • Allocalion ofl and types 10 land use.'ifor which they are inheremly most sui/able for optimum and sustained prodllctivity. The logic is that different crops perform ditTerentlyon different soil types under varied agriculture inpul regimes. Moreover, sustainable agriculture land use requires University of Ghana http://ugspace.ug.edu.gh that the natural productive potential of land resource base be maintained or en hanced. and to ensure that the ecological principle of resilience is achieve d through the strict observance of complementarity (SeA, 1991). To attain the objectives of Ihis dimension of sustainable land use, suitable landsmusl be identified for appropriate Ecologicalsuslainability This ensures thaIland users protect or cause minimal damage to sensitive but vital ecosystems. Protection of ecosystems on steep slopes and the Densu River were the main foci in this context. In assessing the latter, a national environmental policy injunction that forbids farming within 100 m of water bodies (Ministry of Lands and Forestry. 1999) was adopted as a standard. For steep slopes, the same guidelines prohibit the use of slopes >30°. However. this investigation adopted the FAO (1985) slope classification scheme that was used for the national agricult ure land suitability scheme. since it provides appropriate details for intended analyses in the study. Besides. it has the same unit of measurement (percentage) as the one used in the study. Slope suitabilities for annual crops (cassava and maize) whi ch are extensively cultivated in the area. and whose cultivation may cause greater damage to fragile ecosystems was considered. The slope suitability classes are shown in Table 3.1' University of Ghana http://ugspace.ug.edu.gh Table 3. 1: Slope suitability for maize and cassava 0-8 %: suitable 0-8 %: suitable 8-16%: suitable or marginally suitable 8-16 %: marginally 16-30%: marginally suitable or not suitable suitable >30%: nol suitable > 16%; not suitable Source:SRJ(CSIR),Accra • Land-user perception ofs ustainable land uses and land types The economic viability of agriculture land-use types from land-user pers pectiveisan important detenninant of the sustainability of cropping systems. Essentially, this is infonned by experience from years of cultivating crop types on different lands. This significantly innuences filture decisions on choices of crops and land typeslhatwould optimize productivity and utility (SeA, 1991). Accordingly. fanners' opinions of sustainability(suitability)ofcroppingsYSlemsandlandtypeswereevaluated 3.2 DATA SOURCES, ACQUISITION AND MODELING A good knowJedge of the conditions of land and environmental resources and their uses are imponant for evaluating sustainable utilization (LiJlesand and Kiefer, 19 94}. Collins and Rhind (1997) have argued for a consistent and readily available framework data for scientific monitoring and modeling of resources. Estes ef al. (1995) have recommended ten relevant core data sets: topography, hydrology, 31 University of Ghana http://ugspace.ug.edu.gh infrastructure.cJimatology, demography. land use,landcover. soils. air quality and economyforsustainabledevelopmentassessmenLv,:ithduecognizanceofconstraints oftimeanddataavailabiJity,weightierandthe most sensitive indicato rswereselected for the dimensions of sustainable resource utilization considered. Other factors considered were the need for a reasonable balance between accuracy and simplification of data modeling, scale of analysis and usability ofo utputinformation 3.3SlopeJ\.'lodeling Slope is a critical function of erosion hazard (Desmet and Govers, 1996), and therefore a good indicatoroferosivepOlential for landscapes. In the stud y,slopcis expressedintermsofgradient,themaximumrateofchangeofaltilUde (Peukerel ai" 1978; Markselal., 1984), Tooperationalize the national environmental policies on the protection of land resources such as soil and vegetation for susta ined use, slope has been adopted as an indicator to guide identification of areas that re quire conservation measures (Ministry of Lands and Forestry, 1999). This provided a justifiable basis for examining conflicts between land resource management policy objectives and existing or real land resource utilization panerns. Elevatiansampling Slope is a derivative ofa digital elevation model (OEM). Hence, the construction ofa digital elevation model (OEM), which is a 3-dimensional raster representation of terrains, is an important requirement for slope modeling. In the study. e levation was used as the voxel(the volume elements from which OEMs are generated). A 'progressive sampling· method by Makarovic (1984) was adopted and manually University of Ghana http://ugspace.ug.edu.gh adapted for lack of software for carrying out the. normally, automated tasks. The techniquea[[ows for objective sampJing of terrain with varyingcompJexity. as is the case in the srudy area. In this case, highly varied topographic terrains were comprehensivelysampJedlhanlessvariedones. First, the boundary of the area of interest was marked OUI on a topographic mapwi tha scale of 1:50.000. and a contour interval of 50 ft. Lines of varying intervals (depending on thc:: closeness of contours) were drawn paralic I to the Ia t i(Udes across a lopographic sheet. Points of intersection between the horizontal an dthecontourlines were then marked OUI. Due to practical difficul ty of estimating elevation measurementsbt:tweenadjacentcontours,samplingofclevationpointswererestricted to the points of intersections only. A total of 1605 poims were marked for the elevation modeling. The sampled points were digitized on an eleclronic digitizing board configured 10 Arc:lnfo3.5.I,and the coverage projcctedand rcgislered 10 the Ghana lopographical map. In ArcView 3.2 environment, the coverage was converted into a point shape file and elevation data entered into an attribute table. The 811ribute table was converted into an Excel format for subsequent use in the Surfer 7.02 Pa ckage. Spuria~ interpolation by Kriging and variogram rechniques for DEM and slope Modelmg Naturally, it is impossible to make exhaustive sampling for environmental phenomena under investigations at every desired location. This necessitated the use of spatial interpolation techniques to estimate values for areas not sampled {Burrough, 1986; University of Ghana http://ugspace.ug.edu.gh Franke and Nielson. 1991; Watson. 1992). Two broad categories ofgeo·statistical interpolation techniques can be identified global and the local (Petrie and Kennie. 1990; Oakes and Thrift. 1975). Tn global interpolation, a function incorporates information from all the samples, thereby suppressing local variations. Therefore. this was considered inappropriate for the study since the scale of analyses required the computation of local variations. The local techniques are based onthe assumption that each point innuences the resulting surface only up toa certain defini tedistance. Therefore. two local methods • Kriging (Houlding. 1994) and Triangular tessellation or Triangular Irregular Network-based technique (TIN) (Nielson. 1993) were used in the study. For one thing, these are quantitative interpolating methods, which have been found to be suitable for computer contouring algorithms (La m, 1990; Burrough, 1986). Also, they can be justifiably used for comparatively small da tao Kriging is a local geo·statisticai estimation technique synonymous with "optimal prediction" (Joumel and Huijbregts, 1978). The stochastic nature of Kriging interpolation procedure allows for the computation of statistical significance of the surface and unccrtaintyofthe predicted values to be calculated. This is expressed ina variogram that shows how the averageditTerence between values at points change with distance between the points. Semi·variancemay be defined as half the expected squared difference between the random variables Z (x) and Z (x + h) at a particu lar lag h. The variogram defined as a parameter of the random variable, is then the function that relates semi-variance to lagh(1). ",~. n"'"O".., T. . y(h)= ~E[Z(X) . Z(x+h)f (I) ~.:~ ._- University of Ghana http://ugspace.ug.edu.gh The sample or experimental variogram - y(h), can be estimated for p(Jr) pairs of observations.z(x,+h),I= 1.2, 3 . . .. p(h) by equalion (2) y(h)= ~P(h)LZ(X.)- Z(L-h)' (2) The attribute data of elevation and coordinate (feet) were used to generate experimental variogram with f0l1y(40) lags and alotal separation distance of 2200 meters in Surfer. Experimental variogramsare, invariably. discontinuous and do not meet the conditionality of non-negative variance (Cressie, 1993). To achieve a continuous modd of the experimental variogram, the power functional model (~.508:risanexponcnt)(Panna[ier.1996).wasfittedtothcexperimentalvariogram since Ihey strongly matched. The interpolated elevation surface wasgridded to 192.80 x 192.80m and used to construct the OEM. Subsequently, a slope surface was constructed as a derivative of the DEM and measured in percentage. It was convel1ed into a line shape file and exponed into ArcView \\hcre it was interpolated using TIN- base technique into slope surface. 3.2.2Satellitedataproctssing:mdanalysis Maps are limiting for ecological analysis since they usually provide d ataasdiscrete, sharply oounded and intcmally homogenous polygons (Aspinall and Pearson, 1995; Goodchild et 01. . 1992). This type of spatial organizational model imposes a fixed scaJeon analyses (Stoms, 1992). Satellite imagery, however, provides a varied surface that can be usc:d to invesligate generalization and scale effecis (Simmons et al. . 1992). Additionally. the consistent regularity of covering the same area provides data that University of Ghana http://ugspace.ug.edu.gh make for regularly temporal inventorying and monitoring for management ofland and ecological resources that undergo varioustransformarions The Landsat Thematic Mapper(TM) satellite data used for this study has relatively high spatial, temporal, spectral and radiometric resolutions (SBRC, 1994). The spatial resolution of30x30 m for the multi-spectral bands, 1-5 and 7 were particularly suitable for the study since they are capable ofdetecling the typically, small-sized fanns prevalent in the area of study. However, this high spatial resolution meant a corresponding high data volume. The thermal-infrared - band 6, has a coarse resolulionofl20 x 120m. It measures the amount of infrared radiam flux emitted from the earth's surface. It was used for mapping oul broad temperature patterns of the area using the equation (Anuta et 01., 1984): T:-12.5809D-O.000233DxD, where T is the temperature in degrees Celsius and D is the relative digital counts for the TM thermal infra·red(lR) band The high radiometric and spectral resolutions characteristic of Landsat TM data enable them 10 discriminate the heterogeneous cover of the tropical landscape effectively. Landsat TM has been used for change detection studies with results comparable to those by aerial photos (Boresjo-Brongeand Thulin, 1985). Remole sensing data are not received ina perfect state due to errors introduced during the imaging processes (Duggin and Robinove, 1990; Lunnetta et aI., 1991). This may atreclthe accuracy of subsequent image analysis ifnotcorrecled (Maye reta!., 1993), University of Ghana http://ugspace.ug.edu.gh especially for change delection analysis. Therefore. the image analysis was preceded by correclion of these dislortions (Teillct, 1986). Radiometric and geometric errors - the most common I)'pesofrandom errors, are nonnallycorrected at the use r level of processing. All tbe data for Ihe two periods were secured in a dry season with little cloud cover and haziness. so radiometric correction was nOl conducted. Geometric rectification and registration ofs atellite dala For accurate area. direclion and distance measurements, geometric rectification must beperfonned.lmageregistrationisalsoanecessaryrcquirementfor the detection of changes in land-use!cover from multi-temporal satellite data sets. Sin ceallthesewere relevant in the study, a hybrid process of simultaneous image-Io- image geometric rectificalionandregisuation (Jensen et al. • 1993) was conducted. A subscene of 1991 Landsat TM image, which was previously rectified and registered to the nationaJ topographic map, was used asa base year image. The registration of the 2000 image to the 1991 image involved the seleclion of closely matching pairs of ground control points (aep) in the IWO images. It is often difficult to locate good ground control points in rural areas. so much use was made of features associated with urban environment such as roads, junclions of and betweenroad s,rivers,etc. In al~ 16 acps were collected for the registration process. The second power defonnation model was used for the registration, as a compromised choice. The basis is that, besides producing a more accurate result, it does not distort the resultant image. as does the third power model. Moreover, il is bener to use lower power models for a sub-scene less than half scene. The first power model could have been University of Ghana http://ugspace.ug.edu.gh moreappropriatc,bulfortheprofoundvariedelevationalnatureofthelerrain(Jenson, 1996) Bond selection and Image enhancement Univariate and mullivariate slatistical analyses were conducted 0 nthesalellitedatalo idc:ntify bands with high infonnalioncontent for the enhancemenl process. Suitable band combination is important for optimizing infonnation extraction. Appropriate band combination is delennined by terrain, climate and the nature of project interpretalion(Sabins. 1987). Nonnally, it is inappropriate to use high correlated bands in a singl e analysis to avoid infonnationredundancy. Theusefulnessofagivenband-combination may be judged by the correlation between the bands (Epem&, 1986). Generally, bands with low correlalioncoefficientsarecapableofdiscriminatingbelween land-use/coverclasses (Mulders et al, 1992}. Therefore, a correlation matrix of the bands was computed in SPSS 10.0 from me TM digital numbers of the point shape extracted in CHIPS for the vegetation sampling sites. Apal1 frorn their comparatively lowcorre lation coefficients between mem, bands 3,4 and 5 thai were selected for the enhancement procedure have these vital information values critical for the study: J. Band 3(red) represents one of the most imponant bands for vegetation discrimination. It is also useful for soil-and geological-bounda rydelinc:ation. 2 Band 4 (rejlective injrared) is responsive to the amount ofvegetslion biomass present in a scene. emphasizes soil-crop and land-watercontrasl. 3. Band 5(mid-infrared) is sensitive to the amount of moisture in plants. University of Ghana http://ugspace.ug.edu.gh Bands 3, 4 and 5 were loaded onto the red-green-blue (RGB) color system in the order of 5.4 and J respectively, A conlTaSt-stretchenhancemcni procedure was applied in the image analysis 10 optimize information content of the false color image. This was appropriately used since the spectral distribution was Gaussian (Jensen. 1996 ).Inlhe enhanced image, thicker vegetation appeared dark green, with light er vegetation cover appearing light green. Young or short fallowslannual crops appeared greenish yellow 0( yellow, with built·up/exposed or bare ground appearing reddish and purple respectively Unsupervised and supervised classificolions TheclassificalionprocessstartedwithinitialexplOfalionandidentification of possible distinct cover classcs using an unsupervised classification algorithm called Iterative Self-Organizing Dala Analysis Technique (ISODATA) (Jain, 1989). Coupled wilh previously acquired in situ information, :lcombined functional and structural (land· uselcover)c1assirlcation scheme thai fits local condilionsand the objectives of the study was designed to guide characterization of the land-use/covertype s. The adopted land-use/cover classification scheme was made up of following five (5) broad categories: Buih·uplexposed (exp) areas: constructcd features such as buildings and road 5; Forest-fanns(for-fann): patches of dense forest or tree crop fannso ramixture ofthelwo(Plate2.1); Long fallow (Iongfall): dense and high thicket with a number oflrees of differcntsizes; ShoUrt nfailvloewr s(sithyfa loJ)f/ aGnnhual ncaro ph tftaprm:/s/(uangnscrpoap)c: ey.ouugng.e fdalulo.wgsh or annual food- crop fanns or a mixture of them, usually with widely spaced isolated trees; Degraded lands (deg): comparatively short vegetation of grass and other herbs, invariably associated with highlands with a history of intensive cultivation and bushbuming Plate2.2: An example ofa forest-farm. 46 University of Ghana http://ugspace.ug.edu.gh Plate 2.2: Pictorial illustration of degraded land Despite the measures taken to reduce errors in the classification process, the problem of mixed pixel (pixels of different cover types)(Smith et al., 1985) would have introduced some errors. Practical constraints in separating some land~use/cover types necessitated the combinations of some categories of land~use/cover types, which also simplified the scheme. This generalization may have further reduced the accuracy of the output image. Supervisedclassijicalion Prior to designing the classification scheme, the satellite data was examined for the level of nonnality in the distribution pattern oftbe reflectance values. This was to ensure that the distribution pattern meets the assumed normality criterion that justifies University of Ghana http://ugspace.ug.edu.gh the use of the Maximum Likelihood supervised classification algorithm used in the study (Griffith and Amrhain, 1991;Schalkoff, 1992). Training areas were selected for the various land-use/cover classes. The selection procedure carefully considered homogeneous representative areas for the various land_use/covercategoriesbasedoncolour,size,shape,andcontextproperties (Richards, 1993; Lam, 1990; Bamsley and Barr, 1997). Previous geo-referenced field information facilitated the selection of the training areas. These were on-screen digitized as polygon shapes. Training statistics: mean, standard deviation, covariance and correlation matrices of the spectral values were generated for the five land-use/coverclas ses. The separability matrix of Jeffries-Matusita, which has a continuous scale of 0 - 1.44, was also generated for the training areas to determine the degree of separation or similarity between them. Here, cover classes with separability distance of zero (0) are considered identical classes, with those with a separation distance of 1.44 being totally separated orabsolulely different or dissimilar classes. Histogram and scatter plots also facililated the classification process by progressively in dicating the degree of between-class separation prior 10 the final classification of the entire image. Polygons with multimodal data were nOl used in the classifier-training process to ensure that each class has only one [and-use/covertype. Maximum Likelihood classifier algorithm was used 10 classify the image. Though highly reputed for its high accuracy of classification, and hence one of the mostly University of Ghana http://ugspace.ug.edu.gh used, it relies on the assumption of a multivariate-normality of data distribution However, this normality condition is not fully met bysalelliledata, since the values are discrete, with non-negative values; me values range from 0-225 (Malher,1999). Fieldverificationandclassificationaccuracy Verification procedures were designed to balance scientific rigor and defensibi lity with practical limitations of time, general limitations/challenges associated with the use of remote sensing techniques in the tropics and availability of appropriate historical test reference information. The actual procedure involved the use of a combination offield and limited number of suitable aerial photo graphic information. Samples collected from the vegetation ground truthing were a valuable source of test reference information since these were randomly collected, photog raphed,adequately characterized floristically, extensively distributed throughout a II the cover types and geo-referencedwith GPS. As earlier mentioned,a field verification s urvey carried out after the unsupervised classification yielded important information that improved the subsequent supervised classification. The pixels used to train the Maximum Likelihood algorithm were not used as test samples. This was to ensure that the reference information was as independent or unbiased as possible. Statislically.classificationsschemeswithfewcategories,such asthatadopted for this study do not demand the use of many test samples for error calculation. However, the high imponanceaccorded the vegetated classes, and the apprec iable leve I of within cover variability necessitated the useofa relatively high number 0 fsamples( 150) for the analysis. The test sample data were converted into a poinl shape, rasterisedand University of Ghana http://ugspace.ug.edu.gh used for a pixel-by-pixel comparison with the classified image, and accuracy compuled using Kappa slalistical technique (Carstensen, 1987). which measures the level of agreement between two images captured in differenl periods. 3.3. VEGETATION COVER AND LAND-USE PAlTERNS 3.3.1 Tree Sampling Trees with a diameter of iO.Ocm or more, al breasl height (1.301) were counted in 150 quadrats of 30m x 30m. The quadrars were located at variable distances from footpaths at difTerent intervals apart. Theinter-quadrat interval was wider where tree distribution was unifonn. The quadrats were geo-referenced using a 12XL Gamin GPS receiver. whose accuracy was previously detennined (±7.6 m). This was computed by using 30 coordinate measurements ata triangulation point that were recorded. at five-minutes intervals. Ateachoflhe 150 sampling site, the general state of the vegelation was observed. This was 10 help with interpretation of images for classification and vaiidation of the classified image. The choice of trees as indicator is based on the assumption that they tend to be more persistent features on Ian dscapes than herbs overtime, barring significant natural or anthropologica l Iy induced impacts Though,this is not discounting seasonal climatic variations that a ffectplant phenology and canopy structure. 3.3.1 Mapping ofc ropping systems Land use is appropriately described in terms of function rather than phys ieo-chemical properties; it is human use of the land. In this context, cultivated and recently harvestedfannlandsprovidcdindicationsforidentifyingeroppingsysterns. 50 University of Ghana http://ugspace.ug.edu.gh Cocoa, maize and cassava were identified in the same quadrats in which the trees were counted. Cocoa trees in the quadrats were counted but only the presence or absence of maize and cassava were recorded Apart from the graphical and tabular representations, univariate and multivariate analyses were conducted to explore and confirm some relationships. The satellite bands and the number of trees were used as predictorvariahles to assign lhesampling sitcsofthe vegetation ground truthingto pre·determined groups of three (3) levels of cocoa tree densities using discriminant function analysis (OA) (Manly, 1992). The groups were: 1) cocoa farms with normal tree densities. 2) those thai have been partially convened into fa llow and annual crops (moderate conversion) and 3) those with only remnant cocoa trees or totally converted into fallowand annual rood crops (total/extreme conversion). In the case of the last group, farmers were asked of the history of cropping systems at the sites. These were displayed ina canonical plot of the first two canonical functions. The same analysis was carried out usi ngtheLandsat TM (2000) but substituting the density of cocoa trees for the density of other tree species, and used 10 assign the sites to categories or levels of de for estationbasedon 3.4 SOIL AND LAND SU ITABILITY A digital copy of soil association and agriculture [and suitabili ty maps prepared under the Ghana Environmental Resources Management Programme (GERMP) by Soil Research Institute,al the scale of I: 250, 000, were secured and rasterised. The land suitability classification was based on the FAG (1985) scheme. These included the University of Ghana http://ugspace.ug.edu.gh general agriculture land suitability of crops and land suitability for co coaunder medium input regime. The most extensively occurring soil associations and suitability categories were used. Those included in the investigation were: Nankese·KoforiduaINta-Ofin Compound Association (area): I Medium or coarse-textured soils with good structure il. Moderately gravely, stony or concretionary. iii. Deep non-gravely soils on lower slopes. iv Occurs on topography that is undulating with (5-12 %)slopes. v. Has moderate to severe susceptibility 10 erosion and maintains fairly high moisture retention capacity although surface dryness is common in the dry vi. Roolpenetrationiseasy. vii. It is marginal for mechanized farming with hand tillage recommended Suitable fora wide range of tree and arable crops Adawso-Bawjiasi Assciaction(Area) i This upland association consists of Adawso and Bawjiasi series. ii. It is confined to the summilSofiow hills and lower slopes. iii. The soils are gravely and stony. iv. Maybewell,moderatelyorimperfectlydrained,susceptibletodrought,and v They support maize, cassava, yams, pineapples, yet permanent crops suchas citrus and mangoes are suitable. University of Ghana http://ugspace.ug.edu.gh Fete-Bediesi Complex i. A mixture of good and bad lands, with thin reddish and brown soils overlying ii Some lands have clayey pan soils with tightly packed stones. iii. Forestry is recommended for these areas iv. The good lands occur on slopes, with soi ls that are moderately. well or imperfectly drained. v. This area is considered suitable for tree and arable crops Map data compilation by Town and Country Planning of the New Juabeng District from the 1920s through 1990 were redrawn to a common scale, digitised and geo- referenced to the national topographic map. It was then exported from the Arclnfo environment into ArcView. Area was measured in hectares University of Ghana http://ugspace.ug.edu.gh 3.6 GIS ANALYSES Apart from the topographic modeling that was constructed with a Surferpacka ge,all other GIS analyses were carried out in ArcView 3.2 (ESRJ) and IORISI 32 (Clark University) in a grid format. It is computationally easier and faster for a computer to handle overlay analyses in raster than the vector model (Martin, 1996). The analysis carriedoulincludedBoolean,mapquery,areasummaries,elc. 3.6.1 Laod-use/covercbange detectioD In any definition of sustain ability, the key element is change. This is particularly important in management of natural resource utilization. In determining the land- use/cover change, a method referred to as delta by Nelson (1982), and post classification by Riordan (1981) was adopted. It involved the overla yingandpixel-by- pixel comparison of the classified images of the two periods to determine locations, directions and areas ofland-use/coverchange. The post-classification technique has advantages of producing from-to image and change detection matrix that makes it easier to detect indefinable changes. This approach has been successfully used by Olhers (Royer and Charboneau, 1988; Pabi and Anua, 2002). A major drawback of the post-classification method is that errors introduced by classityingthe two images separately would be present in the finalcha nge-detection map (Rutchey and VeJcheck, 1994). Such random and independent errors could mask, especially, areas with limited changes. Hence, this necessitates the need for high classification accuracy (Gordon, 1980; Howarth and Boasson. 1983). This should be ofparticularconcem when images captured at dose time interval are being used. University of Ghana http://ugspace.ug.edu.gh However. the use of images captured in the same season would have reduced this error (Collins and Woodcock. 1994). Rectification of the images and the detailed ground survey may also have further reduced these errors. The degraded land class was further reclassified as short fallow!annua!cropto reduce the number of from-to land-use/cover categories to ease comprehension. The two were assigned to one class since they were the most similar categories . Additionally, the degraded land is assumed to develop from short-fallow!annual crop fields . 3.6.2 l otergeoeralional equityin land allocation ror urball 1311d uses Intergenerationalland-useefficiencyexaminedthedegreeofequity in per capita land allocation among generations. Based on the theory of environmental economics, Tietenberg (1992) has proposed a dynamic efficient allocation model, which can be used to realize efficient land allocation and maintain equity between generations in different time periods. An abridged and simplified form of the model was used in this sludy. The model allows depletable and non-recyc1able resource to be allocated using dynamic efficient criteria. For equity in the use of arable land for urban land use, it is expected that per capita land resource used (Q) be the same for all generations at different times (t - 1.2,3 ... n). That is, QI = Q2 = Q3" .Qn. If total land resource used by a generation of population (Pr) is QT, then per capita land use Q! =QTIPT. An important assumption for this simplification is thai current generation cannot be certain of how future generations may use land resources. It will, therefore, be simplistic loralionalize that future generations will be satisfied with a heritage of capital converted from land resources through inefficient land-use practices University of Ghana http://ugspace.ug.edu.gh Population census data of 1970, 1984, 2000 and a few projected ones were obtained from the Ghana statistical Service. Some of the population sizes were estimated by interpolation. Estimates of built-up sections of the settlements were computed in ArcView from historical maps and classified images. 3.6.3 Steepsiopeprotectionandiand-usecontlici The slope map. was overlaid with a land-use/cover image. This involved initial appJication of Boolean analysis of the layers. Subsequently, map query anaJysis was conducted to identify zones where short fallows fell within s]opes g reaterthan 15°. 3.6.4 Proteclionoflhc Densu Flood Plain and land-use conflict A layer of series ofbufTers was constructed around the segment of the Densu River in the study area. Three buffers were constructed at distances of 100 m, 200 m and 300 m from the river. This was followed by Boolean analysis of the buffer and the land- use/cover maps. Map query was conducted on the two maps to identify land-use/cover types that fall within the various buffer corridors. Another map query was also carried out to identify segmems within the 100m buffer that have forest-fann s. 3.6.5 Laod-use efficiency This dimension examined whether agricultural land-uses were being carried out on land types for which they are most suitable. To this end, a series of map queries between the soil types and the land suitability map on the one hand, and the land- use/cover map and the cocoa land suitability map on the other were c arriedout. University of Ghana http://ugspace.ug.edu.gh 3.7 Land.userperceptionsorthesuitabilil)'orJand typesandcro ppingsystems Human inlentions, decisions and activilies conslilUle the core of management failures or successes ofagro.ecological resources. Humans are an important nexus between land use and sustainable utilization of agro-ecological resources. Hence. of all relevant stakeholders in the evaluation of land use for sustainabi lity, landusersare verylmportanl. On balance, it is local knowledge infonned by experiments that should gel highest priority in sustainable evaluation. Local fanners live more closely to real problems, constraints, and opportunities afforded by the environment and land use. Their experience will pick up com~ interactions of variables which may be very specific to the local circumstances but which would have been rejected in scientific experimentation because of their local nature and the impossibility of researching every possible pennutalion offaclor values (e.g. rocky, steep slop es; high rainfall; restricted growing season; cerealsinlercropped with legumes; no fertili zer;handtools; family labour). Perhaps the most pressing reason to give local knowledge the first consideration is that, in the end, it is the local fanners who suffer from poor advice, failedlechnologies, and expensive solulions. Accordingly. a limited discussion with representatives of the district planning committee was carried out. The purpose was !O exam ine their activit ies on land use and environmental management, the degree of corroboration among them, their current priorities on land use and environmental issues, challenges to policy and planning implementation programmes and how any of these are being confronted. University of Ghana http://ugspace.ug.edu.gh Since individual landowners and tenants are the ultimate decision-make rsofland-use adoption,lheywereinterviewedfortheirperceptionoflandandcropsuitability. With this, one gains insight into underlying socio-economic and bio physical forces thatdefinecurrentandfuturetrendsthroughspatialinter-connectivitybetweenland resources and activity or fanners. To this end, a questionnaire (Appendix I) was designed as an instrument to source for infonnation on lhe past. presen tandpossible future land-use/cover trend. The interviews also examined land-use intensities by changes in the length offallowcycles. environmenlally harmful agri cu ltural land-use practicesandeaseofacquisitionofproductivelandsforfarming.Perceivedeconomic (profitability) and productive suitability of different crops and land types were also A pre-testing of the draft questionnaires using ten respondents and limited focus group discussions helped to refine some of the questions. Individual farmers were interviewed random ly. The interviews were carried out in areas where the selected crop types were cultivated. Interviews were conducted both at homes and on farms. A total of 126 individuals were interviewed The data analyses carried out included both univariate and multivatate statistical techniques. Confirmatory statistical test to compare the rating of the suitability of cropping systems and land types were carried out by the Mann-Whitne ytechnique (Conover, 1980). This was because the variables were independent and the comparisons were made between two or more samples at a time. Cluster analysis {CAl was carried out in SPSS (version 10.0) to explore simi larities between crop types based on their land and economic suitability ratings This was displayed in dendrograms (Hull and Nie, 1981 ; Manly, 1992) University of Ghana http://ugspace.ug.edu.gh CHAPTER FOUR RESl'LTS 4.1. TOPOGRAPHIC MODELS The spatial distribution of the elevation samples renects the nature of the topography in the area. The elevation distribution (Figure 4.1) and model (Figure 4.4a), indicate a broad spectrum of elevation. with a wide range of 527.304m (1730.00ft). and a minimum and maximum elevations of I07.94m (350.00ft) and 318.991m (2080.DOft) respectively. There was a standard deviation of 126.248m (416.00 ft), an indication of extremes in topographic characteristics of the area (Figures 4.1 and 4.2). The distribution of the elevation data appears 10 be nonnally distribution (Figure 4.1}. The experimental aUlo-variogram indicates a strong relationship between change in location and altitudinal variation (Figw-e 4.3). The variogram showed little variability, for which reason no correction by smoothing was applied. A nugget semi variance of about 1000 was reported. Theoretically, a sample separation or lag of zero is supposed to have a semi variance of zero. There was no definite range (separation distance beyond which semivariance remains the same) and sill (no levelling off of the curve). The fit of the power model to the experimental one was a near perfect one (Figure 4.3). Steep slopes were common in the areas of the mountain range overlooking Koforidua (Figures 4.4a and4.4b). Figure 4.5 sbows the distribution of slope across the landscape. 59 University of Ghana http://ugspace.ug.edu.gh Altitudefft Figure 4.1: Elevation distribution 60 University of Ghana http://ugspace.ug.edu.gh 0800.816 Km A Figure 4.2: Contours and c1evation surface University of Ghana http://ugspace.ug.edu.gh ------U-H1-ICtl-on,-U.u- r-Or-f"'ln-ce-:VO-.U- Figure 4.3: Experimental variogram of elevation fitted with a power model variogram Figure 4.4a: Digital elevation model of the study area 62 University of Ghana http://ugspace.ug.edu.gh Figure4.4a: Digital elevation model draped with elevation surface Figure 4.5: Digital elevation Model draped with slope surface University of Ghana http://ugspace.ug.edu.gh 4.2 LAND-USE/COVER DYNAMICS FROM REMOTE SENSING STUDY ..& .2. 1 Statistics of brightness val lies for Landsat TM (2000) d ata allhe samplin g lable 4.1 shows the statistics of brightness values forme Landsat TM bands. Though the visible bands (1 and 3) had the highest minimwn and maximum brightness "aJues. they recorded very low standard deviations. The infrared bands (5 and 7) recorded the highest range, lowest minimum brightness values and the highest standard deviations. Practically. the bands expressed appreciable degree of nonnality in brightness distribution (Figure 4.6). Bands 7 and, especially. 3 were positively skewed whereas Band 4 was aJso negatively skewed There was a significant correlation between the bands (P~<.~ I I/' I, ~;T"', 1/ /I j L~~ -"---_ 1 ~ - - -- Figure -1 .'7 "r;,IIl"r 1ll;lIri\~~ UI"plTlraJ \allJ~~ HfT'! 2utlU at \~J!dali 67 University of Ghana http://ugspace.ug.edu.gh TABLE 4.2 ; CORRELATION MATRIX OF THE BRIGHTNESS VALUES OF LANDSATTM (2000) DATA 4.2.2 Training area slatislics for tbe laod-use/coverclasses In the enhanced LandsatTM composite, high vegetation cover zones appeared dark green, with areas of sparse vegetation cover being reddish (Appendices 2 and 3). Land-use/cover types with high and uniform amount of vegetation had low within variations in brightness values. All the tand--use/cover types had relatively low standard deviations of brightness values for band 3 in LandsatTM 2000, whereas 5 and 7 had the highest. The forest-farms generally recorded very low standard deviations for all the bands (0.959-4.t82), whereas built-up/exposed areas recorded the highest (1.977-13.862) in 2000 (Table 4.3a and Figures 4.8a and 4.8b). The pattemwassimilartotheI991trainingstatistics(Table4.3b). University of Ghana http://ugspace.ug.edu.gh Individual bands for both the 2000 and 1991 TMdata separated the land·use/cover classes to different degrees (Figures4.8c and 4.8d). Generally, similarity followed a definite order, with land-use/cover classes displaying high degree of similarity depending on the amount of vegetation cover. Band 3, differentiated the cover categories into two broad groups: the comparatively more vegetated land-use/cover classes (forest-farms and thetwofallowcategories),occupying region s of low digital numbers and less vegetated lands (degraded and the built·up/exposed areas), occupying a region of high digital numbers. The order of arrangement of land- usclcoverclassesalong the visible red and the infra-red bands was opposite that of near infra-red band. This manner of arrangement was common to both 199 1 and 2000 Bands 3 and 5 were very poor at differentiating between the degraded land and the senlementsJcxposcdgrounds. The Jefries Matisutaseparability distance matrix (Table 4.4) indica ted a definite trend of land-usc/cover transformation from a high-vegetated land-use/coverthrough low vegetatcd cover types. This trend showed appreciabJe level of correspondence to the panemdescribedinthescanerdiagram(Figure4.8cand4.8d). ln 2000 and J991, lhe forest-farm on one the hand and the degraded land and the senlemenllex posed land- use/cover classes on thc other were totally different (separabil ity dist ancewas 1.4 1) University of Ghana http://ugspace.ug.edu.gh TABLE 4.3A: STATISTICS ON TRAINING AREAS OF BRIGHTNESS VALUES OF LANDSAT TM 2000 Band 4 Degraded land 140.000 58.000 92.000 53.000 75.000 149.000 135 .000 64.777 111.644 74.088 1.815 3.227 7.067 Forest-farms 134.000 60.000 58.000 31.000 141.000 83.000 103.000 74.000 Moan 136.380 76.775 42. 115 Sid. De .... 0.959 2.510 3.168 Long fallow Min. 134 .000 68.000 69.000 36.000 88.000 104.000 69.000 136.517 76.992 85.052 SId. lk. .. 1.427 2.851 4.911 Short fallow-annua l crops 68.000 69.000 39.000 146.000 95.000 161.000 164.000 138.715 77.669 96.859 56.891 Sid. De ... 1.340 7.539 8.591 Buiit-up/exposedarea 138 .000 60.000 &3.000 50.000 Max. 152.000 83.000 198.000 234.000 Mean 144.8 11 69.505 110.432 83.965 Std. Oev. 1.977 3.616 9.342 13.862 University of Ghana http://ugspace.ug.edu.gh TABLE 4.3B: STATISTICS ON TRAIN ING AREAS OF BR IGHTNESS VALUES OF LANDSATTM 1991 Band 3 Oegradedland 33.000 46.000 66.000 92.000 56.743 73.027 SId. Dey 2.953 4.931 Forest-farms 38.000 27.000 66.000 54.269 47.373 Sid. Dey. 0.839 3.808 Long fallow 54.000 43 .000 71.000 69.000 61.009 54.320 3.828 Short fa llow-annual crops Min. 54.000 55.000 81.000 125.000 63 .348 69.103 Std. Dey. 3.345 5.18) Built-uplexpo.sedareas 34.000 41.000 53.000 65.000 116.000 50.4 16 72 .016 Std. Dey. 2.510 4. 129 6.969 71 University of Ghana http://ugspace.ug.edu.gh TABLE 4.4: lEFFRlES-MAllJSITA SEPARABILITY DISTANCE BETWEEN LAND USE'lCOVER TYPES IN TI-IE YEAR 2000 Region Degraded Long Short fallow- Built-upl land fallow annua[crop exposed . Degraded land 0.00 Forest-farms long fallows 1.39 Short fallow-ann 1.37 1.30 cron Built-up/exposed TABLE 4.5 :JEFFRlES-MAllJSITA SEPARABILITY DISTANCE BETWEEN LAND USE'lCOVER TYPES IN THE YEAR 1991 Region Degraded Forest- Long Short fallow- Built-up! Land farms fa llows ann.croo exposed. Degraded land 0.00 Forest-fanns Long fallows 1.32 Short fallow and 0.90 tnn.croo Built-up! 1.11 exposed University of Ghana http://ugspace.ug.edu.gh I Figun: 4.80: Reflectance distribution fur bands of bands 3, 4 and 5 for the _ ueas m 1990 Fogun:4.!Ib: Rd!ectancedistributionforbaodsofbands3, 4, 5 and 7 for the IIliDinBtImIS iu2000. University of Ghana http://ugspace.ug.edu.gh Ir~:C5§:~ c~e 63-· .. - I ~bo~.~.. _ :.. -.~ 0 ~I © -~ ..- ":.~ --.rc ~ _. " _-"_ - Figure 4.8c: Seaner diagrams of bands of bands 3, 4 and 5 for the training areas in 1990 ~cY=- .~g . ~g ..---.. - I. tOO ,.- k:: ::: : ': =0G D:=-: 0 t ~ ~:--= 'II... "'Il ' - ~ - t~ '. - Flgure4.8d: Scatter diagrams of bands 3. 4. 5 and 7 for the training areas in 2000. University of Ghana http://ugspace.ug.edu.gh 991 Figure 4.9.: Land-use/cover pattern in 1991 75 University of Ghana http://ugspace.ug.edu.gh Figure 4. 9b: Land-use/conr patt~m in 2000 76 University of Ghana http://ugspace.ug.edu.gh 4.1.3 Land-use/conrpatterns The five land-uselcoverclasses for 1991 and 2000 are showed in figures 4.9a and 4.9b. Theamounland proportion of the cover types are also shown in table 4.6 and figure 4.10. lnthe periods investigaled.the forest-fannswere restricted to Ihe north- western sections of the area, with the fallow dominating the southern sections. However, some appreciable amount of dense vegetation was found in the south- eastern and western sections. During the period considered, the Koforidua Municipality significantly expanded in the direction of the Akwapem North District. The municipality has practically merged with Okorase in the Akwapem North District. The distribution of vegetation cover panern showed a negative spatial correspondence with temperature intensity (Appendix 4). ABLE 4.6: AREAS OF LAND USE/COVER IN 1991 AND 2000 IN HE AREA OF STUDY f\fe.(Ha) 2000 IArea(Ha) - 1991 peg,-adJExpgrd. orest-farms 129333.500 20764.392 ngfall-fanns 72901.500 19501.573 htfalllAnncrps 52294.000 112323.988 enlements 13572.120 University of Ghana http://ugspace.ug.edu.gh Figure 4.10: Laod-use/£overralio50fI991 and 1000 4. 3 DISTRJBUTION OF VEGETATION AND CROPS 4.3.1 DistributioD of trees and fallows There were six hundred and eighteen (618) trees in all the 150 quadrats (13.5 hal, whkhconstituled an average of48 trees perha. Abundance-classdi stribulion(Figure 4.11) was positively skewed, presupposing a high degree of rarity. There was a mean of 4 trees per quadrat. A significant inverse relationship (ETA = 0.294, P.:: ~::' ;":, . 8] University of Ghana http://ugspace.ug.edu.gh '1 -'1 -'I ' j '1 '1 t t n t ::':, t . t ~:2 U " t,;' f t " t:: t t " -.J -.! -,J r , j .1' .J Figure 4.14: Territorial map of deforestation classes SY!!I!;zQ1 I!:velsof~foreslatio!l Extremeltotaldeforeslation Moderate deforestation Nonna] vegetation cover Oroupcentroid 82 University of Ghana http://ugspace.ug.edu.gh Fi =O.9J6C - 0.460b5 - 0.452bJ - 0.420b2 - 0.J82b7 + O.130b4.... .. .. .. ... (1) Fj; = - 12OC + 0.262b5 + O.096b3 +0.69Ib2 + 0430b7 + 0.190b4 ....... .. ...... (2) The annual food crops - cassava and maize, and cocoa (perennial cash crop) were distributed with high spatial variability. There were 2,527 cocoa trees in the 150 quadrats, with a mean of 19 cocoa trees per quadrat . This was equiva lent to 133 cocoa trecs per ha. Cocoa tree abundance distribution was heavily skewed loward sthelcast abundanccclasses, with 58 .0% of the quadrats recording zero(O) cocoa trees and 65.3% of the quadrats containing from 0-5 cocoa trees. This may indicate a convcrsionofcocoa farms into either fallow or annual crop farms. Thcre was alsoa strong positive relation between me distribution 0 fnaturalvegelalion cover and crop Iypes. Generally, cocoa tree density was high where the number of trees was also high and vice vcrsa (Figure 4.15). The correlation was a highly significantone(p.L---- I ~m-----~-~--~-~--~-~--~-=----------------- Figure 4.27: Percapila land consump1i6o of arable land by urbaoland uuio tbe Koforidua Muoicipalitv University of Ghana http://ugspace.ug.edu.gh 4.4.3 Sustalaable me of Iud resounes within. ecologically leasitinzoDes. Land-use/cover categories of 2000 wilhin the flood plains ofI he Densu RiW!r There was no appreciable difference in the relative compositions of land-use/cover in the series of buffers created at the various distances from the Densu River in the year 2000 (Flp'es 4.28 aod 4.29). Within the buffers, long fallow constituted the highest proportion of the land-use/cover types, with forest-farms and short fallow/annual crops following the order. Some amount of built-up/exposed areas was located within aU the buffers (Figure 4.30). Figure ".28: Buff~rs around tb~ D~nsu River in tht 2000 land C'ovrr 95 University of Ghana http://ugspace.ug.edu.gh F;gu .. 4.29: Segm .... of Dons. 1Uv. . <• .,red w;,h Fo",H.rms ~. Plat~ 4.2: A Darrow patch orromt remnant on the bank oftbe river Den5u near h:oforidua. For most part, land within 100 m ortbe river is totally deforested. 96 University of Ghana http://ugspace.ug.edu.gh oriaialted from tbe ellpO!!led !!Iurf.as in Denlu Basin j:: : 0_004 I~ :: 0.001 '00-200 8lnerdlst. .e e(m) Figun 4.30: .... nd-use/cover C.tegOM of 2000 witbin the four burren aroand tbe~nsuRiver 97 Land-use!cover cOlUegnoriivese 0r/s2i0ty00 o wfi tGhinh salonpea c lhaststeps: //ugspace.ug.edu.gh The soil associations have a strong spatial relationship with eleva tion(Figure4 :3Ia) Koforidua-NankeselNta-Ofm, Adawso-BawjiaseINta-Ofin and Chichiwere-Ayensu- Kakum occupy lower elevations. Fele-Bediesi, Bediesi-Sikaben and Wenchi- Kumayili are located on slopes and mountain summits. Of the total land surface for which slope was generated, about 80.0% of the area were within the slope categories of 0-2% and 2-5% (Figures 4.30 and 4.31). Only 4.59% of lheareawasfound in the reJatively Sleep range of 16 and 4So/• . The steep slopes were located in the mountainous areas, with the less steep slopes found in the flat and undulating zone Different proportions of the 2000 land-use!covertypes were located wi thin the various slope classes (Figure 4.32). Forest-farms, short fallow- and long fallow farm s dominated the less steep slope (nat and undulating) classes. However, at the steeper smpes.theproportionoflong fallow and the short fallow-annual crop farms largely prevailed. with a decrease in the amount offorest-fanns. Shon fallow-annual crops dominated the Sicepest slopes (Figure 4.33). Some built-up/exposed zones were also found in the steep slopes (>16%) 98 University of Ghana http://ugspace.ug.edu.gh Adawso-Bawdii~la-Ofin Figure 4.3la: Soil Associations draped over elev81ion model 99 University of Ghana http://ugspace.ug.edu.gh • Setll~nt Iopetlngrtd D~~ -1-- PROPORTION OF SLOPE LASSES !Slope class (%) Percentage of class 12-5 158208 765 000 University of Ghana http://ugspace.ug.edu.gh • BuilHlplup S shor1lall 9 l.onJiaU l1li For../arm • Dog""" Slope L---__ ~ ____ Figure 4.32: Land-co\'crclasses witbin slope classes , ..II + -~. Figure 4.33: St«p slopes under short faliow/anDual crop 101 University of Ghana http://ugspace.ug.edu.gh 4.4.4 Land-use/cover, cropping systems and soil potential or suitability SettJements have been established in almost all the soil associationslsuitability cat egories (Figures 4.34 and 4.35). However, two soil associations have been widely used for constructional activities, particularly. in the Koforidua Municipality. These were the Adawso-Bawjiase and Nankese-Koforidua Associations. The latter is rated good for agriculture, and was the most extensively cultivated with cocoa and also the most densely forested (Figures 4.34 and 4.36). There was a remarkable relationship between the soil types/land suitability for cocoa cultivation (Figures 4.37). A more striking parallel relationship was apparent between the zone designated as good for cocoa cultivation and the prevailing spatial distribution of cocoa farms (Figures 4.38). The areas with the most extensive cocoa coverage were those designated as suitable by the land suit ability scheme for cocoa production (Figure 4.39). Again. forest cover has a strong positive relationship with the distribution of cocoa. Obviously, both forest cover and land suitability seemed to have influenced cocoa distribution in the area -iiii..uJI Ue8S Foresf·lann lkllisultib ~::. 11&1:: I '" ~84.." 102 University of Ghana http://ugspace.ug.edu.gh :::=. I [IJ Shortfllll-annCfop Il Buill·upllt~P I I Figure 4.35: Proportion oflaod-useicovcrcalegorics in each 3ericulturelalldsuitabilitvclassio2000 Figure 4.36: Forest-farm per bectare in land suitability classc~ 103 University of Ghana http://ugspace.ug.edu.gh ... a.ttI.meaL ii~~:.b:.' "' :.:. _101 Figure 4.37: Pertentagr land suitability for cocoa production ',p.poIR_ .•• , i::·::"'··~'· .~ Figure 4.31: Sulfa« 0' cocoa lrea per 90m2 in 1000 104 University of Ghana http://ugspace.ug.edu.gh -Kofuridua • I- Figure 4.39: The highest cocoa tree density occun-ed in the most suitable soil and rOl'Ht~rarms 105 4.5 PrUrcenpitivoeDsrosfiltlyin do-fu sGeahnda lnanad -h