Evaluating Land Use Assignment Papers.
A systematic analysis of land use/cover change is so decisive to exactly understand the extent of change and take essential measures to curb down the rate of changes and protect the land cover resources sustainably.Evaluating Land Use Assignment Papers. This land use/land cover change study was conducted in Agarfa district of Bale zone, Oromia Regional State, Southeastern Ethiopia.Evaluating Land Use Assignment Papers.
The objectives of this study were to evaluate the trends, drivers and its socio-economic and environmental implication in study area. A descriptive research method was employed to achieve the intended objectives of the study. In the three years (1976, 1995, and 2014) Landsat Satellite images and socio-economic survey were the main data sources for this study. ERDAS Imagine and Arch-GIS tools were used to classify and generate land use/land cover maps of the study area. Survey questionnaires, key informant interviews, and field observation were employed to obtain information on drivers and its socio-economic and environmental implication in the district.Evaluating Land Use Assignment Papers. The results show that the land use/land cover of the study area had changed dramatically during the period of 38 years. A rapid loss of forest land and shrub land cover in the landscape took place between 1976 and 2014. Conversely, agriculture and grazing lands were increased by 30% and 42% respectively at the expense of the lost land use/land cover types. Forest land is the most converted cover type during the entire study period. In the 38 years, forest lands diminished by over 65% of the original forest cover that was existed at the base year (1976). Local climate change, declining agricultural productivity and livestock quantity and quality and scarcity of fuel wood and constructional materials were some of the socio-economic and livelihood impacts of land use and land cover change of the study area. Thus, this finding affords information to land users and policy makers on extent of the change and social forces leading to this changes and its subsequent implication on local socio-economic and environmental conditions of the study area.
Land Use/Land Cover Change Evaluation, Image Classification, Impacts of Land Use Land Cover Change, Agarfa District, GIS and Remote Sensing
1. IntroductionLand use/cover change is a dynamic, wide-spread and fast-tracking process caused by natural phenomena and exacerbated by human actions, which in turn drives changes that would influence humans. Evaluating Land Use Assignment Papers. Land use/cover conversions to farmland, and settlement and to urban development reduce the extent of lands accessible for food and timber production. Soil erosion, salinization, desertification and other soil degradations allied with intensive agriculture and deforestation decrease the quality of land resources and future agricultural productivity  .Land use alteration trends in many developing countries are both tremendously swift and the direction of changes and rates are influx. Africa is said to have the fastest deforestation in the world as a result of dependence on the primary resources  . Increasing population pressure across Sub-Saharan Africa has been responsible to increase in cultivation and grazing intensity  . Evaluating Land Use Assignment Papers. This has attributed to huge deforestation and conversion of natural habitats to farmlands and settlements with consequences on biodiversity and land degradation. Land use in East Africa has changed rapidly over the last half century: expansion of mixed crop-livestock systems in to former grazing land and other natural areas and intensification of agriculture are the two largest changes that have been noticed  .Ethiopia is a country characterized by rapid environmental changes and modifications accredited to different adverse human actions, like expansion of farm plots at the expense of vegetated lands, substantial fuel wood and charcoal production, overgrazing and encroachment of farmlands into vegetated lands. Like other parts of the country, the Bale Mountains Eco-region has been affected by human activities such as deforestation, overgrazing, illegal logging and hunting  . Residents in the eco-region engage in deforestation to acquire land for crop production, for livestock grazing and to fulfill their demand of timber and fire wood. For example, deforestation rate in the Bale Mountains ranged from 1% to 8%, between 1986 and 2009 with an average rate of 3.7% per annum  . This deforestation rate is about four times the 1% country-wide forest loss  . This implies that forest in the Bale Eco-region is shrinking at an alarming rate.A study in the other parts of the Bale Mountains such as Herenna catchment showed a significant decline in dense forest cover from 26.8% to 6.2% and agricultural land increased from 11.1% to 17.7%  . It is also estimated that the montane forest which comprises more than 40% of the total area of BMNP has been lost at an average annual rate of 3.74 km2 between 1973 and 2005  . According to  between 1973 and 2008 a total of 26.65% of the original area of Afromontane grasslands, a total of 15.41% of upper montane forest, and a total of 14.58% of Afromontane dwarf shrubs and herbaceous formations were converted to agricultural lands and hence, agricultural fields increased from 1.71% to 9.34%.Most of these studies focused on the assessment of biophysical (trends and extent) condition of LULCC over time using remote sensing data and expert opinion. However, these studies missed to integrate socio-economic survey to better understand the causes and impacts of the change.Evaluating Land Use Assignment Papers. Analysis of land use/land cover change at a local scale is vital to comprehend about the complex relationships between environmental, economic and social drivers that induce changes and the likely impacts at local level. Moreover, a systematic analysis of land use/cover change is so decisive to exactly understand the extent of change and take essential measures to curb down the rate of changes and protect the land cover resources sustainably  . In the study area, there are observable land cover change problems which include widespread deforestation for several different purposes such as, encroachment of agriculture and settlements into forest areas, charcoal and construction materials production and severe soil erosion along hilly slopes. However, there are scarce local scale land use/cover change studies in southeastern Ethiopia in general and Agarfa district in particular. As the study area borders BMNP which is one of the world’s hotspot areas that harbors variety of endemic animal and plant species, evaluation of the impacts of land use/cover change on these local ecosystem and livelihoods in this area is relevant and timely.Evaluating Land Use Assignment Papers. Moreover, information on land use/cover change and its possibilities for ideal use is essential for the selection, planning and implementation of land use schemes to meet the increasing demands of basic human needs and wellbeing. Thus, the study evaluated the spatio-temporal pattern changes of land use/land cover with associated causes of alteration in the study area. This information also assists in monitoring the changes in land use resulting from changing demands of increasing human population  .The main objective of this study is to evaluate the spatio-temporal changes of land use/land cover over the four decades (1976-2014) in Agarfa District.Specific objectives of this study were:1) To classify and quantify the land use/land cover changes in the study area.2) To identify the major driving forces of the land use/land cover changes.3) To assess socio-economic implications of the observed land use/land cover changes.2. Materials and MethodsAgarfa district is found in the Bale Administrative Zone of Oromia Regional State, in Southeastern part of Ethiopia. It lies between 7˚8’N to 7˚28’N latitude and 39˚31’E to 40˚5’E longitude (Figure 1). Its surface area is about 1216.34 km2 (Figure 1).Figure 1. Location map of the study area.The elevation of Agarfadistrict ranges from 1400 m to 3800 m above mean sea level (a.m.s.l). About 61% of the district is plain with slope ranging from 0 to 8 degrees and the majority of this area lies in the southeastern and western parts of the study area. Wabeshabelle river gorges and related rugged terrains make about 31% of the district. Mountain ranges comprise about 8% of the district which lies along southwestern corridor of the district that borders Bale Mountains National Park  . The district is drained by Wabeshebelle tributaries, Weybriverand other perennial rivers that drain the district are Wuchima, Makkalla, Fawwa, Tugumma. Agarfa district falls within three traditional agro-climatic zones, vernacularly termed as Gamoji (hot), Bada-dare (temperate), and Bada (cold). Mean maximum and mean minimum temperatures are 25˚C and 10˚C respectively. Evaluating Land Use Assignment Papers. The amount of maximum and minimum rainfall received in the area ranges between 1200 mm and 400 mm, respectively. The dominant soil types in the district include: vertisols, cambisols, luvisols and lithosols that are derived from tertiary volcanic rocks  . The common natural vegetation types in the district comprise of juniper and podocarpus trees, bamboos, scattered woods, shrubs and bushes along relatively lower altitudes. The district is also endowed with varieties of wild lives which include Bush back, monkey, Columbus monkey, hyena, jackal, apes, different species of birds, and varieties of reptiles and amphibians.About 87.36% of the population lives in rural areas whose livelihood is predominantly dependent on rain-fed, subsistence agriculture along with rearing of livestock. The major crops produced are barley, wheat and teff, maize, beans, and peas. The crops which are produced in the drier, low land areas are mainly maize, sorghum, and teff  .2.1. Data SourcesThe three years satellite images were used to scrutinize the spatio-temporal changes in land use/land cover in the study area (Table 1). Landsat satellite images of Agarfa district, the Landsat MSS for the time period 1976 was obtained from Global Land Cover Facility (GLCF), while Landsat TM7 and Landsat ETM+8 for the periods 1995 and 2014, respectively were acquired from USGS Global Visualization Viewer an Earth Resources Observation and Science Center (EROS). Apart from satellite images, field observation and questionnaire were conducted. Other secondary sources of data such as topo sheet of the study area, district annual reports, and other published and unpublished materials were consulted.2.2. Data Processing ProceduresBase map of the study area was prepared from the topographic map on 1:50,000 scale. Various permanent features like roads, rivers or any other land based features were transferred to the base map. The administrative district boundary map was brought to Universal Transverse Mercator (UTM) project in zone 37 and later the satellite imageries were clipped with the administrative boundary of Agarfa.Evaluating Land Use Assignment Papers. Thereafter, preliminary interpretation of satellite data was carried out and a preliminary interpretation key was prepared. The preliminary interpreted maps thus prepared were taken to field for ground checking. Initially three dates (1976, 1995, and 2014) satellite imageries were downloaded from their respective sources.Evaluating Land Use Assignment Papers. To conform the pixel grids and remove any geometric distortions in the imagery, the first Landsat MSS of 1976 image was registered and geo-referenced to the WGS_1984_UTM_Zone 37 coordinate system based on 1:50,000 scale topographic maps. Then, each of the Landsat TM of 1995 and ETM+ of 2014 images were registered to the 1976 image using image to image rectification technique. Then, image enhancement was undertaken in order to enhance the quality of the image and readability of the features.The preparation of thematic maps from the digital satellite data was carried out by using ERDAS Imagine ver. 2014 and ArcGIS ver. 10.3 software and platforms. Then field visit to site was carried out to obtain ground control points using GPS for ground truth data collection.2.2.1. Image ClassificationClassification is the hypothetical representation of a real life situation using easily defined and well delineated criteria.Evaluating Land Use Assignment Papers. Classification schemes are usually hierarchically organized into numerous levels with various degrees of details and have certain criteria to differentiate land cover categorization from one another  .According to  , more common land use/land cover classification schemes are: Anderson, national land cover data and FAO land cover classification systems. Consequently, for this study, FAO land cover classification system in Table 2 was employed for classifying and setting land use/land cover types of the study area.supervised classification method was used with maximum likelihood classifier decision rule assisted by ground control points (GCPs) collected during field surveys for the latter two periods of images while the researcher’s prior knowledge and different physical features patterns recognition system was used for the former one(2014). Figure 2 shows spatial distribution of GCPs draped over Google Earth. In this study, a total of 60 ground truth points collected during the field survey were used for the classification (30) and validation (30) of 2014 image. For the classification and validation of image in 1976 and 1995 periods, aerial photographs of 1957 and 1982 were used to generate random points of 200 and 232, respectively, and half these were used for training classification while the remaining half were used for the accuracy assessment. Evaluating Land Use Assignment Papers. A representation of the regions of interest known as the training sites were digitized giving them different IDs and unique colours.Training areas for all spectral classes were developed by composing each information class to be identified by the classifier. Since there was more than oneTable 1. Satellite data.Table 2. Description of the land use/land cover categories.Figure 2. GCPs collected by the researcher during the field survey.spectrally different signature found for each information class. For example, an information class such as cultivated land contains several crop types, active agriculture and inactive agriculture areas, and each of them must be represented by several spectral classes. The researcher used more than thirty training samples for each classification. A recode function was used to merge spectrally different classes to generate final information classes. Landsat multispectral bands, except the thermal band, were used to identify the classes.The post-classification approach was used for mapping detailed LULC determination.Evaluating Land Use Assignment Papers. This approach is generally considered the most obvious approach to change detection  . It requires the comparison of independently classified images of the same study area acquired over two different time periods  . By properly coding the classification results for times 1976, 1995 and 2014, the analysis were produced a change map showing a complete matrix of changes (e.g. change from wetland to cultivated land and grassland to forest). Three land use/cover maps from 1976, 1995 and 2014 were produced using the ENVI 4.3 software.2.2.2. Accuracy Assessment of Supervised ClassificationAccuracy assessment is a procedure that compares a classified map against reference data or facts from the field to evaluate how well the classification represents the real world phenomena. An interpretation is then made of how closely the newly produced map from the remotely-sensed data matches the reference (base) map. For this purpose, the ground control points (GCPs) collected from field survey were superimposed to the classified maps and used to compare the value of facts from the field with the value of the classified map. This comparison produces error matrix which is the basis of accuracy assessment process (Table 3).2.2.3. Results of Accuracy AssessmentClassified LULC maps from remotely sensed images may contain various types of errors. The confusion matrix presented in Table 3 shows the overall classification accuracies and accuracies of the single land-use/land cover classes.The standard method of Confusion Matrix was used to assess classification accuracy for each image date by comparing classification results with ground truth region of interest (ROIs). The methods of accuracy assessment used included the Kappa statistic and Google Earth. The Kappa statistic is a statistical method of assessing the accuracy that took into account the chance of random agreement. Accordingly, the Kappa statistic average accuracy of classification was 78%, 80% and 93% for 1976, 1995 and 2014 images respectively (Table 3). The result falls within the range of very good to excellent  .2.3. Socio-Economic Data Collection MethodQualitative and quantitative data were collected by employing survey questionnaires, key informant interviews, and field observation to obtain reliable results.Table 3. Error matrix of land use/land cover maps derived from Landsat images.Accordingly, socio-economic survey was conducted to collect data on causes and impacts of land use and land cover change. Evaluating Land Use Assignment Papers. For this purpose five representative peasant administrations (PAs) were selected from the three agro-ecologies for household heads sample selection. From selected PAs, a total of 120 household heads were selected using random sampling techniques. Survey questionnaires were prepared in local language (Afan Oromo) for the interview and then translated into English for analysis. In addition to this, key informant interview was conducted along with field observation to obtain additional information on the long year experience of land use/land cover change deriving practices in the district. The collected household survey data was concerned with the current socio-economic activities of the study area whereas data collected from the elders concerned with trends, causes and effects of land use/land cover change that traced back to 1976 and 1995 and linked to the current existing situation.2.4. Data AnalysisTo analyze quantitative data of satellite imagery, ERDAS Imagine version 2014 and ArchGIS version 10.3 softwares were employed. Accordingly, the major land use/land cover classes for the three study periods were classified using ERDAS Imagine version 2014 and separately mapped using ArchGIS version 10.3 softwares and then, areal coverage of the major land use/land cover types for each mapping period were calculated both in hectares and percentages.Evaluating Land Use Assignment Papers. The comparison between values of the LULC types was carried out toidentify the percentage change, trend and rate of change between the entire study periods, i.e. 1976 and 2014. The collected socio-economic data were analyzed both qualitatively and quantitatively. Quantitative socio-economic data were analyzed by SPSS version 16 software. Field observation and qualitative socio-economic data were analyzed qualitatively. And the results were compared with the major land use/land cover change maps generated for the three study periods.3. Results and Discussions