Akademische Arbeit, 2021
32 Seiten, Note: 100%
INTRODUCTION
1.1 Background
1.2 Description of the study area
1.3 Statement of research problem
1.4 Objectives
1.4.1 Main objective
1.4.2 Specific objectives
1.5 Research question
1.6 Significance of the research
1.7 Beneficiaries
1.8 Tools, equipment and facilities
LITERATURE REVIEW
2.1 Introduction
2.2 Remote Sensing and Urban Growth
2.3 Spatial Metrics
2.3.1 Densification
2.4 Image Classification
2.4.1 Supervised Classification
2.4.1.1 Support Vector Machine (SVM)
2.4.2 Unsupervised Classification
2.5 Change Detection
2.5.1 Post Classification Comparison
2.6 Resampling
2.6.1 Nearest Neighbour
CHAPTER THREE
METHODOLOGY
3.1 Introduction
3.2 Data Acquisition
3.3 Image pre-processing
3.3.1 Layer Stacking
3.3.2 Image re-projection
3.3.3 Resampling
3.3.4 Image Sub-setting
3.3.5 Training Samples
3.4 Image Processing
3.4.1 Image Classification
3.4.2 Accuracy Assessment
3.4.3 Change Detection
3.5 Densification
3.5.1 Area stratification
3.5.2 Building Density Extraction
CHAPTER FOUR
RESULTS AND ANALYSIS
4.1 Introduction
4.2 Results of classification
4.3 Result of accuracy assessment
4.4 Statistics of land cover changes between 1995 and 2009
4.5 Land covers change
4.6 Urban Sprawl
4.7 Result of densification
4.7.1 Building density in 2005
4.7.2 Building density in 2017
CONCLUSION AND RECOMMENDATION
5.1 Conclusion
5.2 Recommendation
Urbanization is a phenomenon that is of major concern in both developing and developed countries. According to the census record between 1988 and 2002 there was a near-doubling of the urbanized population in Urban-West of region Zanzibar Island and continued high growth since 2002 are unprecedented in intensity, scope and impacts on the lands, particular in Urban-West. Such rapid urbanization of West district has meant that the city has doubled in areal extent. Currently there is lack of timely and quantitative information of the urban growth and its long-term impacts, thus planners haven’t been able to assess and analyse consistently growth of Urban-West in Zanzibar. This research aimed at analysing and quantifying urban growth pattern of Urban-West in Zanzibar using Remote Sensing, GIS and Spatial Metrics. To achieve this goal various data were used such as Landsat images, Sentinel 2A image and Building data have been utilized. Urban sprawl has been analysed from Landsat Thematic Mapper images of 1995, 2009 and Sentinel 2A of 2020.Support Vector Machine (SVM) supervised learning has been applied for land cover mapping followed by change detection analysis. The overall classification accuracies for the land cover were 88.3%, 91.4% and 96.6% and a kappa statistic of 0.8, 0.8, and 0.9 for 1995, 2009 and 2020 respectively. Result obtained in the three-land cover change was that the built-up areas has increased while non-built up areas have decreased. This was complemented by Spatial Metrics specifically building density as a composition metrics that was obtained by area stratification. Building data of 2005 and 2017 were used; the study area was stratified into 55 grids of 3km by 3km followed by extraction of the building density from each of the 55grids.Result obtained was from 2005 to 2017 there was an increase in building density of about 9180.6 per Kilometre squared. This result has concluded that the eastern part of Urban-West has experienced low density extending to the western part while the northern part and southern part have experienced medium density. There is a radial expansion following the main roads networks, from sub-centres or from main centre. The sprawling was due to an increase in urban development and improved services that have resulted from population growth. From this the planners should be able to asses and analyse such consistently growth of Urban-West for better urban growth and development plan. Also, they should establish monitoring mechanism on the eruption of the informal settlement for better development control.
Keywords: GIS, Land cover, Remote sensing, spatial, Urban growth.
Rapid urbanization and urban area expansion of sub-Saharan Africa are megatrends of 21st century. Addressing environment and social problems related to these megatrends requires faster and more efficient urban planning that is based on measured information of the expansion patterns. Urbanization has been a universal phenomenon and important social and economic aspect taking place all over the world (Magidi, J. & Ahmed, F., 2018). The process is going rapidly leading to fundamental changes in urban growth in terms of expansion and densification.
Zanzibar is a semi-autonomous polity within the United Republic of Tanzania consisting of two main islands, Pemba and Unguja and many smaller islets. Unguja is home to the polity’s capital city, known as Zanzibar. Although smaller settlements were found into vicinity long ago, Zanzibar city can be traced essentially to the establishment of the Omani seat of power there in the 1690’s.The settlement expanded dramatically when the sultan of Oman moved to Zanzibar early in the nineteenth century and incited a period of ‘clove mania’ Zanzibar sultanate collected customs from the nineteenth century caravan trade in eastern Africa ,and these tariff’s, combined with profits from plantation agriculture , and international trade in slaves,ivory,cloves and other commodities gave rise to a building boom and urban expansion. The city’s population only climbed above 30,000 towards the nineteenth century’s end and it stagnated through most of the British colonial era (1890-1963). By 1931, the city had 45,276 people, while the 1948 count increased by only a few dozen. The 1950s brought a new wave of urbanization, with the population of the city and its suburbs growing to nearly 70,000 by 1958.As in much of Africa, independence in Zanzibar created freedom of movement that allowed for greater urban population expansion, such that the urban and peri-urban settings combined held 142,041 people by 1978, and 208,137 by 1988. The near-doubling of the urbanized population between 1988 and 2002 when the census (NBS, 2002) recorded 390,074 people in Urban-West region and continued high growth since 2002 are unprecedented in intensity, scope and impacts on the lands, particular in West districts. According to the (NBS,2012)Tanzania National census, the population of the Zanzibar Urban-West region was 593,678. The rapid urbanization of West district has meant that the city has doubled in areal extent in the last twenty five years (Myers, G.A., 2010)
Although urban growth is unescapably process, efforts can be made to protect natural resource, reduce natural hazards such as flooding and improve the livelihoods of urban dwellers through proper ways of urban planning and management (Abebe, G.A., 2013). To do so city planners, policy makers and resource managers need more advanced and quick techniques to acquire quantitative information on urban growth processes and patterns (Abebe, G.A., 2013). To understand the pattern, ones need quantification information that will help to know the impacts and the measure to undertake. A study by (Jr, C.A., & Davidson, F, 1996) indicate density as one of the most important design parameters in housing and human settlements quantification.
Densification refers to number of units per area occupied. There are various types of density but in this study the focus is on building density which is referred to as a number of buildings per unit area. Densification has a significant impact on health, urban environment, and productivity of cities and on human development as a whole. Urban densities affect urban development processes at the city and neighbourhood levels. Measuring density is important in estimating the nature and scale of activities in total population which helps in knowing the trend of building, environmental impacts and other phenomena associated with cities. Urban expansion is the increase in built up area from time to time involving a change from non-built up to built-up areas. Studies indicate that built up area can expand radially following main road networks, from sub centres or from main centres (Sall, O., 2008).The density can be cooperated by the use of Spatial Metrics in order to understand the spatial pattern of urban growth and sprawl.
Spatial Metrics are numerical indices that describe the structure and patterns of a landscape (Ramachandra, T.V, Aithal, B.H., & Sreekantha, S.,2012 2012). Other researchers (Herold, M., Liu, X., & Clarke, K. C.,2003) also defined spatial metrics as “quantitative and aggregate measurements derived from digital analysis of thematic categorical maps showing spatial heterogeneity at a specific scale and resolution”. Thus, both definitions emphasize on the quantitative nature of the metrics. Spatial Metrics are one of the important aspects to measure and analyse the spatial patterns of urban growth and sprawl. The complex patterns of land use/cover can be represented in any landscape using Spatial Metrics (Eric J. Gustafson, G. R. P., 1998). There are three categories of spatial metrics which are; Class Metrics, Patch Metrics, and Landscape metrics. In Landscape Metrics there are two categories which are Composition Metrics and Configuration Metrics. Composition metrics are indices that are related to the presence, proposition, variety, and the richness of patch type without considering the spatial character of patches (Eric J. Gustafson, G.R.P.,1998). Configuration Metrics refers to the spatial arrangement, character, orientation and position of patches in the landscape (Eric J. Gustafson, G.R.P., 1998). Composition Metrics was used throughout the research.
Surveying techniques such as conventional surveying and mapping techniques tend to be more costly to cater for estimation of urban sprawl in developing countries (Dadhich, A., & Goyal,R., 2017). Increasingly, research are being directed to do mapping and monitoring urban sprawl using Remote Sensing and GIS techniques (Abebe, G,A., 2013). Remote Sensing is cost effective and technologically reliable for analysis of urban sprawl. It is used to provide temporal land-cover maps that explicitly exhibit the dynamic of urban growth. Some underlying patterns cannot be well visualized since it cannot provide full description of underlying processing responsible for the changing of urban landscape. Thus, currently urban growth is effectively measured using Spatial Metrics due to its capability of describing the underlying process on urban landscape.
Zanzibar is a semi-autonomous part of Tanzania and it consists of two islands which are Unguja and Pemba. Unguja is the larger of its two main islands. This island is located in East Africa, 40 km east of Tanzanian mainland and slightly south from the Equator 6˚00ʹ50.18ʺto 6˚18ʹ50.18ʺ S and from 39°18ʹ52.55ʺ to 39°24ʹ52.55ʺ E see (Figure 1.1). Unguja covers a total area of 1,660 kilometres square. The total population of Unguja Island is 896,721 and is most concentrated in part of Zanzibar Urban-West region(Bank, 2018). Unguja Island is divided into 3 administrative regions which are Zanzibar urban region, North region and the South region.
Zanzibar Urban-West is of the 31 regions of Tanzania. Located on the island of Unguja, Zanzibar city serves as the region’s capital. In 2012, according to Tanzania national Census the population of the Zanzibar Urban-West region was 593,678. The Zanzibar Urban-West covers a total area of 230 kilometres square with a density of 2600 kilometres square. The climate is typically insular, tropical and humid, with an average annual rainfall of 1500 to 2000 mm. The rainfall regime is split into two main seasons, a primary maximum in March, April and May in association with the southwest monsoon and a secondary maximum in November and December. The months in between receive less rain, with a minimum in July. The economy of Zanzibar Urban-West depends on agriculture and fishing.
Abbildung in dieser Leseprobe nicht enthalten
Figure 1.1: Location map of the study area.
Urbanization is universal phenomena taking place all over the world and has led to significant land use change/cover changes in both developing and developed countries. Decisions made on aspects of urbanization have significant impacts on health, urban environment for the productivity of cities, and on human development at large. Zanzibar is highly urbanized compared to the Tanzanian mainland with 66% of the population living in urban areas on its main island Unguja and the main city is expanding rapidly. Zanzibar, like other places in Tanzania, struggles with overpopulation which is a big problem for a sustainable city. As the population is fast increasing in Zanzibar, peri-urban areas and small towns often grow in an unplanned way that makes service delivery challenging, degrading the environments, leading to insufficient infrastructure, evolution of informal settlement, insufficient sanitation and water supply, increase in number of dwellers who lead to increase in carbon emission thereby causing climate change, environmental hazards such as flooding and reducing quality of life. However, there is lack of timely and quantitative information of the urban growth and its long-term impacts, thus planners have not been able to assess and analyse consistently growth of Urban-west in Zanzibar. Thus, municipal authorities and decision makers need to know the urban sprawl and growth phenomena of Urban-west in Zanzibar for better infrastructure planning and future urban development. Hence, there is a need to expand the application of spatial metrics in order to develop a quantitative approach for effective and informed decision making for the development control.
This study is set to analyse and characterize urban growth in Urban-West in Zanzibar area using GIS, Remote Sensing and Spatial Metrics. Thus, the lack of quantitative information with Spatial Metrics will be addressed by the study output.
To analyse and quantify urban growth pattern of Urban-West in Zanzibar using Remote Sensing, GIS and Spatial Metrics.
The specific objectives of this research are:
1) To produce Land cover and change detection map of the study area for the years 1995, 2009 and 2020.
2) To compute building density of the study area for the years 2005 and 2017.
3) To analyse and describe changes of urban growth of the study area for the years 1995, 2005, 2009, 2017 and 2020.
Based on the above research objectives, the following research questions will help to assist the analysis:
1. What are changes in the land cover in Urban-West (Zanzibar)?
2. Which zone of Urban-West (Zanzibar) has low building density compared to others?
3. What are the changes in urban growth in Urban-West (Zanzibar)?
This research will help to provide information on the urban expansion extent and densification in Unguja Island specifically in the region of Urban-West that will quantitatively help in making informed decision on the urban growth and development of Unguja Island. Also, the information will help the resource manager in allocation of the services, the planners in the planning process, city managers in the process of controlling and managing at a city level and decision makers in their process of making decision over urban growth and sprawl.
1. Planners.
2. Decision makers.
3. City managers.
4. Urban designers.
In this research computer hardware, software and internet were used. The computer software’s used were ArcGIS 10.4, ENVI 5.3, QGIS 3.10, RStudio 3.4 and Microsoft excel.
ArcGIS 10.4 was used for the preparation of the layout of all of the thematic maps. ENVI 5.3 was used for image pre-processing, image classification and change detection. QGIS 3.10 was used in preparation of building data and extraction of building density. Microsoft excel was used in data entry that was to be used in plotting. RStudio 3.4 was used in plotting of the building density graphs.
This chapter reviews important literatures regarding remote sensing and urban growth, spatial metrics, image classification, change detection method, resampling and densification.
Remote sensing is helpful tool to better understand the spatiotemporal trends of urbanization and monitor the spatial pattern of urban landscape compared to traditional socioeconomic indicators such as population growth (Wang, J, Ju, W., & Li, M., 2009). However, the availability of multi-temporal data is important to analyse the dynamics of land cover change over time and space and in order to detect changes and pattern of different spatial phenomenon the images should be of the same season. This will help to reduce data inaccuracy generated due to seasonal variations. Medium resolution Landsat images play the key role in the analysis of urban change at different spatial scale (Abebe, G.A., 2013).Different studies have been conducted on urban changes using medium resolution Landsat images. For example, (Yuan, F .,et al., 2005)used multi-temporal Landsat images to analyse urban growth pattern and to monitor land cover changes of two twin cities in Minnesota metropolitan area and demonstrated the potential of multi-temporal Landsat data to provide an accurate and economical means to map and analyse changes in land cover over time.
Spatial metrics can be defined as measurements derived from the digital analysis of thematic-categorical maps exhibiting spatial heterogeneity at a specific scale and resolution (Herold, M., Couclelis, H., & Clarke, K.C., 2005). This definition emphasizes the quantitative and aggregate nature of the metrics, since they provide global summary descriptors of individual measured or mapped features of the landscape (patches, patch classes or the whole map). There are three categories of spatial metrics which are; Class metrics are computed for all patch of a particular type or class in the landscape which in the context to LULC classes (Jagadish,K., 2004). Examples of class metrics is class area. Patch metrics are computed for individual patches which represent discrete areas of similar characteristics (Jagadish, K.,2004). Landscape metrics are computed for the entire patch mosaic of the whole landscape (Jagadish, K., 2004). They result as the combination of patch and class types.
In Landscape metrics there are two categories which are composition metrics and configuration metrics. Composition metrics are indices that are related to the presence, proposition, variety, and the richness of patch type without considering the spatial character of patches. Configuration metrics refers to the spatial arrangement, character, orientation and position of patches in the landscape. Composition metrics will be utilized throughout the research, and density been it measure parameter.
Densification refers to number of units per area it occupies. There are various types of density like building density, population density, land use density, gross density, net density is used in different perspective in this research the focus is on building density. Density has been categorized as low, medium and high density with the given standards. Measuring density is important in estimating the nature and scale of activities in total population, knowing the trend of building, environmental impacts and in modelling other phenomena associated with cities, rural and natural habitats. Unguided densification results into eruption of informal settlement with all kind of environmental pollution.
Image classification is the process of extracting information from multi-band raster images such as satellite images (Lillesand, T.M, Kiefer, R.W., & Chipman, J., 2008). The majority of image classification is based on the detection of the spectral response pattern of land cover classes(Brito, P.L; and Quintanilha, J., 2012).The choice of image classification method mostly depends on the objectives of the research , the nature of the image and the level of detail or accuracy requires for specific application(Lillesand, T.M, Kiefer, R.W., and Chipman, J., 2008).There are two broadly image classification methods which are; Supervised Classification and Unsupervised Classification.
Supervised classification are based on external/ground knowledge of the area shown in the image (Mather, P.M., 2004).This method require input from the user before chosen algorithm is applied. This input may be derived from the fieldwork, air photo analysis or from the study of appropriate maps. There are common algorithm/classifier used in this method which are box classifier, centroid classifier and maximum likelihood methods. Also, there is more advanced approach of image classification such as support vector machine and decision trees.
Support vector machine are effective because they produce classification accuracy values as high , if not higher than other classification methods and they are efficient because they need small amounts of training data that is located in those area of feature space that lie near to interclass boundaries (Mather, P.M., 2004). The mapping function that is used by Support Vector Machine (SVM) is kernel-based method. The reason as why using Support Vector Machine (SVM) ,according to (Candade, N., & Dixon, D.B., 2004), claims that Support Vector Machine (SVM) behaviour is independently of dimensionality and thus this method does not requires as much training data as say a Maximum Likelihood (ML) classifier which has a need for exponentially increasing amount of training data as dimensionality increases. These authors also conclude that Artificial Neural Network (ANN) and Support Vector Machine (SVM) give more accurate results than Maximum Likelihood (ML) classifier in a study of land cover classification.
In supervised classification, clustering algorithms are used to partition the feature space into a number of clusters. The algorithms examine the unknown pixels in an image and aggregate them into a number of classes based on the natural grouping or clusters present in an image. In performing classifications, image values within a given class are close together while image values in different classes are comparatively well separated (Mather, P.M., 2004).
Change detection is the process of identifying differences in the state of an object or phenomenon by observing it at different times (Singh, A., 1989) . Change detection involves the application of multi-temporal datasets to quantitatively analyse the temporal effects of phenomenon. There are different change detection methods have been developed and documented with their own advantage and disadvantage some of them are; Image differencing, principal component analysis, Post classification comparison, image rationing, and Change vector analysis. This research utilizes post classification comparison as the method of change detection.
Post classification comparison is the most intuitive method of change detection, is GIS overlay of two independently produced classified images. This method can produce from-to information (Lu, D. et al., 2004). This method is done on the classified image. Thus, the accuracy of this method depends on the accuracy of the classification result.
Resampling is the technique of manipulating a digital image and transforming it into another form. This manipulation can be done for various reasons including change of orientation, change of resolution and change of sampling points etc. (Gurjar, S. B., & Padmanabhan, N., 2005).Usually resampling method involves determination of digital values (DN) to fill in the output matrix of the rectified or registered image (Baboo, D. S. S., & Devi, M. R., 2010). Basically, there are three resampling methods which are nearest neighbour, bilinear interpolation and Cubic convolution.
This is the resampling method which is mostly used in remote sensing. It uses the digital value from the pixel in the original image, which is nearest to the new pixel location in the corrected image (Baboo, D. S. S & Devi, M. R., 2010). The main advantages of nearest neighbour include simplicity and the ability to preserve original values in the unaltered scene. Its disadvantages include noticeable position errors, especially along linear features where realignment of pixels is obvious (Baboo, D. S. S., & Devi, M. R., 2010).
[...]
Der GRIN Verlag hat sich seit 1998 auf die Veröffentlichung akademischer eBooks und Bücher spezialisiert. Der GRIN Verlag steht damit als erstes Unternehmen für User Generated Quality Content. Die Verlagsseiten GRIN.com, Hausarbeiten.de und Diplomarbeiten24 bieten für Hochschullehrer, Absolventen und Studenten die ideale Plattform, wissenschaftliche Texte wie Hausarbeiten, Referate, Bachelorarbeiten, Masterarbeiten, Diplomarbeiten, Dissertationen und wissenschaftliche Aufsätze einem breiten Publikum zu präsentieren.
Kostenfreie Veröffentlichung: Hausarbeit, Bachelorarbeit, Diplomarbeit, Dissertation, Masterarbeit, Interpretation oder Referat jetzt veröffentlichen!
Kommentare