30-09-2016, 09:41 AM
Assessment of Urban Expansion in the Sekondi-Takoradi Metropolis of
Ghana Using Remote-Sensing and GIS Approach
1456921723-AssessmentofUrbanExpansionSekondi.pdf (Size: 1.32 MB / Downloads: 13)
ABSTRACT
The conversion of other types of land to uses concerned chiefly with population growth and increase in economic activities is the main
cause of land use land cover changes in human history. Urban expansion has therefore been identified as one of the most evident
examples of human modification of the Earth and has therefore become a very important element in world environmental studies. For
effective monitoring of environmental changes and proper management of natural resources to be carried out, studies on urban growth
patterns needs to be carried out. This makes urban expansion studies extremely important. The Sekondi-Takoradi Metropolis of Ghana
has been experiencing fast urban growth over the past two decades. Forest and agriculture lands are being converted to uses concerned
chiefly with population growth and increased in economic activities. This research sought to assess urban expansion in the metropolis
using an integrated remote sensing and GIS approach. Several remote sensing techniques were used to carry out land-use-land-cover
change detection using two multitemporal Landsat images of the years 1991 and 2008. This assisted in determining the changes that
have taken place over the 17 year period. Urban growth pattern was also analysed using GIS techniques. The results showed that there
has been a significant urban growth in the study area. The annual rate of change of land cover within the 17 year period was determined
to be 1.77%. The results further showed that urban expansion was uneven in different part of the metropolis and that there is a negative
correlation between the density of urban expansion and distance to a major road. The results further showed that the annual rate of
change of urban/built-up land is 4.88%. This urban development has therefore altered the land cover of the metropolis significantly.
INTRODUCTION
Land covers serve as both sources and sinks for most of the
energy and material movements and interaction that occur
between the biosphere and the geosphere. Changes in land-useland-cover
(LULC) therefore have environmental effect both
at local and regional scale. Moreover, changes in land cover
cannot be understood without the knowledge of land use
change that drives them. (Weng, 2001).
The conversion of other types of land to uses concerned with
population growth and increase in economic activities is the
main cause of LULC changes in human history. Urban
expansion has been identified as one of the most evident
examples of human modification of the earth and has therefore
become an important element in world environmental studies
(Trusilova, 2006). Urban growth, both in population and in
areal extent, transforms the landscape from natural cover types
to increasingly impervious urban land. The result of this
change can have significant environmental effects both locally
and regionally (Xian and Mike, 2004).
Sekondi-Takoradi, a coastal city in the Western Region of
Ghana have been experiencing a considerable increase in its
population especially its urban population over the past two
decade (GSS, 2012). This has resulted in changes in the LULC
pattern mainly due to this population increase as well as
accelerated economic development. Urban growth has
increased and pressure to the environment is occurring.
Massive virgin forests and agricultural lands are disappearing,
being converted to urban or associated uses. Land which was
initially covered with vegetation is now being covered with reflective impervious structure such as road and building.
There is therefore the need to assess and evaluate this urban
growth and develop appropriate land use planning policies for
sustainable development.
The integration of remote sensing (RS) and Geographic
Information Systems (GIS) has been widely applied and
recognized as an effective tools in detecting urban LULC
changes (Weng, 2001). Satellite remote sensing has the ability
to collect multitemporal data and turns it into valuable
information for monitoring urban land processes. GIS on the
other hand provides a more flexible environment for entering,
analysing and displaying digital data from various sources
necessary for urban feature identification. These make remote
sensing and GIS more useful tools for urban growth detection
projects (Weng, 2001). Saleh (2011) invested into the
application of the integration of remote sensing and GIS for
detecting urban built up growth for the period 1961-2002, and
evaluated its impact on surface temperature in Baghdad city. A
research carried out by Mohan (2005) also used satellite remote
sensing to evaluate urban LULC change detection in Delhi.
According to his research, he observed that, spatial information
from the remote sensing satellites provides more effective
solution for sustainable environment and urban development.
These show that RS and GIS provide a more effective approach
to detect, analyse and evaluate urban growth patterns.
MATERIALS AND METHOD
Study Area
Sekondi-Takoradi metropolis is located between Latitude 4° 52' 30" N and 5° 04' 00" N and Longitudes 1° 37' 00" W and 1°
52' 30"W. Bounded to the north of the metropolis is the Mpohor
Wassa District, the south by the Gulf of Guinea, the West by
the Ahanta West District and the East by Shama District. The metropolis is strategically located in the south-western part of
the country, about 242 km to the west of Accra and
approximately 280 km from the La Côte d’Ivoire in the west.
Materials
The study was based on the use of a time series of satellite
Landsat images –Thematic Mapper and Enhanced
Thematic Mapper Plus (ETM+) as remote sensing data
acquired in the years 1991 and 2008. The satellite images were
obtained from the U.S. Geological Survey (USGS). Among the
reference data used are topographical maps, aerial photographs
and land cover map of the study area. Also geographic data
(GPS points) were collected for all the various land cover
types.
Methods
Assessment of urban is a complex process involving several
activities. Such research involves the processing of multitemporal
images to obtain essential, precise and accurate
information on the changes that the earth’s environment is
undergoing. The methodology adopted in the research is
divided into three main categories namely: Image Preprocessing,
Image Classification and Urban Expansion
Detection and Analysis
Image Pre-processing
To achieve accurate change detection, multi-temporal images
must be pre-processed both geometrically and radiometrically
to correct errors arising from imaging sensors, atmospheric
effect and earth’s curvature. Three separate bands (bands 4, 3,
2) were combined into a single layer using the layer stack tool
in Erdas Imagine software. The original unrectified images
were distorted; they were therefore rectified and re-project unto the Ghana Datum War Office coordinate system using a total
of 50 ground control points (GCP). The rectified image
produced a root mean square (RMS) value of 0.17 which was
accepted because it was within half a pixel (Jianya et al., 2008).
After the rectification, the images were then resampled to a 30
m by 30 m pixel resolution using the nearest neighbour
resampling technique. Using a four corner coordinates, a
subset which covers the study area was created.
Image Classification
Supervised classification was used to classify the individual
images into three land cover classes namely non-urban, urban
and water. Training samples for all the various land cover types
were obtained from an aerial photograph, a land cover map and
GPS coordinates that were picked during the field navigation.
The Maximum Likelihood Algorithm which classifies images
according to the covariance and variance of the spectral
response patterns of a pixel was the parametric rule used during
the classification. The accuracy of the classified images was
then assessed. This allowed a degree of confidence to be
attached to the results. During the accuracy assessment of the
images, the overall accuracy as well as the kappa statistic was
also computed for the classified images.
Post Classification Change Detection
Change detection studies have various meanings to diverse
users (Singh, 1989) nonetheless; the commonest understanding
of change detection applications is the fact that it has the ability
to provide information on changes in terms of the trend, spatial distribution and extent of change. Transition contingency
matrix was generated to test the independence that existed
between the land cover classes in the different years. A crosstabulation
(matrix) was generated from the 1991 and 2008
thematic maps.
Urban Expansion Analysis
In order to analyse the rate, nature and location of the urban
land change, an image of urban/built-up was extracted from
both classified images. The 1991 and 2008 urban/built-up
images were overlaid and recoded to obtain an urban expansion
image. For further analysis of urban expansion, the urban
growth image was overlaid with some geographic features
such as major roads and some urban centres. The metropolis
was also divided into four zones namely, Takoradi, Sekondi,
Effia- Kwesimintsim and Essikado-Ketan zone. The amount of
urban growth in each of these four zones where also computed
accordingly.
Urban expansion process usually shows a relationship with
distance from certain geographic features such as road (Weng,
2001). Using the buffer function in GIS, ten buffers with a
width of 400 m cumulatively were created around a major road
in the metropolis. This buffer distance was chosen based on
several city planning factors. The major road chosen was the
main Takoradi-Accra road. Each of the buffers created were
then overlaid with the urban expansion feature to compute the
amount of urban expansion in each buffer zone. To calculate
the amount of land in each buffer, each of the created buffers
was overlaid with the land-use-land-cover change map. In order to construct a distance decay function of urban
expansion, the density of urban expansion in each buffer zone
was computed using the formula below.
Total amt. of land in each buffer
Amt. of urban expansion in each buffer
Density of urban expansion
To determine whether a relationship exist between the amount
of urban expansion and the distance to a major road as well as
the density of urban expansion and distance to a major road, a
distance decay function was constructed. This distance decay
assisted in analysing the pattern of urban expansion in the
metropolis
To investigate the relationship between the density of urban
expansion and distance to a major road, Pearson Correlation
Analysis was employed. This was done to determine the
pattern of the urban growth in relation to the available social
amenities such as schools, hospitals and recreation facilities.
3. RESULTS AND DISCUSSION
Results
Image Classification and Accuracy Assessment
The classification scheme used resulted in two land cover maps
from the two multi-temporal images. The land cover maps that
were obtained from the 1991TM and the 2008EMT+ images
are shown in Figure 2 and 3 respectively