Abstract
Dar es Salaam is one of the most diverse cities in Tanzania in terms of its physical, social, economic, environmental and spatial features. This diversity has contributed to differences in built-up area, population density, as well as the pace of spatial development across different parts of the city. This study aims to examine the relationship between physical built-up area changes in Dar es Salaam, population density change and spatial development using remote sensing images and census data. The study finds that the city population has grown tremendously, with peri-urban wards in particular having experienced positive growth. Dar es Salam’s built-up area change and urban sprawl emerging at the city’s edges distinctly follows the pattern of demographic change. This is accompanied by substantial compact growth in the inner parts of the city. A number of factors such as transport, residential development, migration, high natural growth rates, public policies and land speculation are found to have contributed to these changes. Overall, the study aims to aid planning authorities in effectively responding to the rapid spatial development taking place in the city, for which a holistic approach that combines an understanding of physical and demographic changes is needed. By investigating the changing patterns in land use within this highly urbanizing city, it aims to generate insights into urban development control machineries and identify their underlying dynamics.
Keywords
Introduction
Countries in the Global South are experiencing a high rate of urbanization, including Tanzania. According to the United Nations World Cities Report 2016, 55.3% of the world’s population lived in urban areas (UN-Habitat, 2016). This phenomenon is also reflected in Tanzania, where the rate of urbanization increased from 24.8% in 2005 to 31.6% in 2015 (UN-Habitat, 2016). In 2012, Dar es Salaam city housed 4,364,541 people, comprising approximately 10% of the total Tanzanian population. By 2019, this figure was estimated to have reached approximately 6 million people. It is the second fastest growing city in the world and the fifth largest city in Africa. By 2030, it is expected to join the ranks of the ‘megacities’ of the world. However, the city is also experiencing organic and unplanned growth, with more than 70% of its settlements estimated to be informal as per United Nations findings (Kyessi, 2005). With 5 municipalities, 90 wards and 451 streets, it is the most diverse city in terms of its physical, social, economic, environmental and spatial features. This diversity is assumed to have contributed to differences in built-up area, population density as well as spatial development. In addition, although Dar es Salaam is growing at an alarming rate, the pace of growth is varied across different parts of the city, which poses significant challenges in establishing planning control over its spatial development. Against this backdrop, this article attempts to examine the relationship between changes in physical built-up area, population density and spatial development in Dar es Salaam using remote sensing images and census data. By investigating the changing patterns in land use within this highly urbanizing city, it aims to generate insights into urban development control machineries and identify their underlying dynamics.
Study Area
Dar es Salaam is located on the east coast of Tanzania, with a latitude 6°45″S and 7°25″S and a longitude of 39°E and 39°55″E. The city borders the Indian Ocean to the east and the Coast Region to the north, west and south (see Figure 1). It has a total surface area of 1,628 km2, out of which 235 km2 or 14.4% is covered by water bodies, mainly deriving from the Indian Ocean. The remaining 1393 km2 is covered by land area (Gombe et al., 2017). Immediately after Tanzanian Independence, the city served as the capital until 1973, when this status was shifted to Dodoma City. Nonetheless, the city remains the manufacturing and commercial hub of the country (Todd et al., 2019). According to the 2012 population census, Dar es Salaam had a total of 90 wards and 452 streets (neighbourhoods). It is widely understood that human settlements experience diverse trajectories of development, some more organic in their evolution than others (Msuya & Mosha, 2020). Dar es Salaam is a relatively big city, which has largely developed through informal and erratic construction. This lends it a lot of diversity in terms of types of settlements (formal and informal) and population densities.

Methodology
The present study used remote sensing to capture the built-up area changes in the city over time. For this, satellite images of different time periods, obtained through free source images from the United States Geological Survey (USGS), were compared to estimate the urban built-up area at different times. Due to some challenges in acquiring images with low cloud cover, captured in the last 20 years with 5-year intervals, images from 1995, 2000, 2015 and 2018 were acquired instead. Since the images were raw satellite images, a number of pre-processing tasks had to be undertaken before using them to classify land cover. With the aid of ENVI software, images extracted from the USGS website 1 were pre-processed. The rationale for pre-processing the images was to get the area of interest (Dar es Salaam city) and undertake the removal of atmospheric effects. Alongside this, radiometric calibration, quick atmospheric correction, dark subtraction, image masking and pan sharpening were also undertaken.
The pre-processed images were then used as inputs in image classification, in which both unsupervised and supervised classifications were applied (Figure 2). The reason for using both was to improve accuracy check, whereby unsupervised classifications were first applied with many classes to see the distribution of pixels depicting different land covers, followed by supervised classifications. Within the supervised classifications, four classes were created as training samples, namely, built-up area, vegetation, water and bare land. Under this method of classification, the maximum likelihood method was applied. The output results (land cover classes of each input feature and their statistics) were then used to calculate land cover change over time and to create a time series of urban sprawl. The results from land cover changes were then analysed with integration of socio-economic data from census tracts and literature in order to establish a pattern of built-up area changes in relation to population densities and the spatial development of the city.

Analysis and Discussion
The study took an explorative approach to understand the existing relationship between the physical and socio-economic features of a city and how they affect city spatial growth. Population growth rate and other demographic patterns and trends were analysed along with land use and built-up area change through mapping the city land use cover over different periods. Built-up area remains a key focus and has been analysed by assessing its growth, density, direction and the socio-economic factors influencing its development. The analysis ultimately aims to decode the compactness or sprawling nature of Dar es Salaam city and locate the driving forces behind its development.
Population Growth
The population of Dar es Salaam city has grown tremendously in the span of 10 years (two census rounds). From 2002 to 2012, it grew from 2,487,288 to 4,364,541, which is equivalent to a 43% change. Between the periods 1988 and 2002 and 2002 and 2012 (inter-censual periods), population growth in Dar es Salaam increased from 4.3% to 5.6% (see Figure 3), which is above the national annual average growth rate of 2.7%. A number of factors have attributed to this growth in population. Urban to rural migration remains one of the biggest factors, associated with poverty and lack of livelihood opportunities in other regions (Wang et al., 2005). However, other factors such as the expansion of the city boundaries, which saw a number of villages become engulfed by the urban area, have also been significant. As a result, the number of wards increased from 73 to 90 from 2002 to 2012. The new wards were formed through the sub-division of existing wards and inclusion of some peri-urban wards (URT, 2013). Peri-urban wards continued to densify, with residential and commercial developments taking place in the outskirts due to the influx of new rural-urban migrants.

Social Group Distributions
The population of Dar es Salaam is comprised of various social groups. Their categorization is based on a number of demographic characteristics: age, sex, marital status, relationship to the household member and education (literacy). By identifying these social groups and their distribution across the city, the social and economic stratification of the city over time and space can be located.
Age and Sex Profile
According to 2012 census reports, the overall sex ratio for Dar es Salaam city was dominated by females, with 95 males for every 100 females. However, the sex ratio for the population that was below 5 years and within the age group of 30–74 years was above 100, thus indicating an excess of males over females in these particular age groups.
High fertility and mortality rates were revealed in 2012 census, whereby 32% of the total Dar es Salaam population was below the age of 15 and only 2% was above 65 years of age (see Table 1). In understanding rural to urban migration flows, age and sex structure play a greater role in developing countries since data for migration is not well recorded. The migration pattern for youth in sub-Saharan Africa is estimated to feature age groups between 10 and 14 for females and 15 and 19 for males (Menashe-Oren & Stecklov, 2017). However, since the youth (15–24 years) and young adult population (15–35 years) in Dar es Salaam was recorded as making up 23.8% and 46.8% of the population, respectively, this indicates that the city is experiencing high rural to urban migration.
Population by Key Age Groups; Dar es Salaam City, 2012 Census
Rural to Urban Migration
Rural to urban migration is the main factor contributing to population growth in Tanzanian cities. This type of migration is also reflected in the social stratifications within households, where the prevalence of nuclear families is sparse as compared to extended families. The relationship between the head of the household and other household members in Dar es Salaam was found to be relatively high, with the percentage increasing towards peri-urban areas. The large presence of extended families and the nature of their relationship indicate a high rate of rural to urban migration in Dar es Salaam. This population is well represented in terms of population structure, with many African cities having a bulge in the age group 15–34. This indicates a high level of youth in-migration from rural to urban areas (see Figure 4), which tends to create pressures in housing and inversely affects land use and spatial development in the city.

Built-up Area Change Mapping
One of the indicators of Goal 11 of the SDGs (Sustainable Cities and Communities under target 11.3) is a ratio of land consumption rate to population growth rate. This indicator can be defined by examining the percentage of land used, preserved or developed in relation to population growth as indicated/shown by satellite images and census data. In line with this, the primary aim of this study was to examine how city land cover changes over time in relation to different socio-demographic factors, with a focus on changes in built-up area. This involves identifying whether the city population is becoming increasingly concentrated and dense or deconcentrated/ dispersed as per consumed land. The findings on land cover changes from 1995 to 2018 across various intervals are presented below.
1995–2000
From 1995 to 2000 the population of Dar es Salaam grew from 1,830,067 to 2,272,483 at a growth rate of 4.4%. Built-up area grew from 144.89 km2 (8.90%) to 253.48 km2 (15.57%). A total of 6.67% increase is equivalent to a 1.33% change per year (see Figure 5). A number of factors have contributed to the observed changes. This includes improved infrastructure in the outskirts of the city, rural to urban migration and the expansion of city boundaries through planning. An example is the 2,000 plot projects, which majorly contributed to the opening up of new areas in the city, but also land speculation in the peri-urban areas.

2000–2015
From 2000 to 2015, the population of Dar es Salaam grew from 2,272,483 to 5,115,698 at growth rate of 5.74. Built-up area grew from 15.57% to 38.73% (23.16% increase) 253.48 and 630.52 km2, respectively, which is equivalent to 1.54% change per year and 7.72% per each 5-year interval (see Figure 6). These figures almost quadrupled from the previous years, yet this can be attributed to the fact that, due to some difficulties in acquiring data for 5 years intervals, a 15-year interval has been used instead. A number of factors have contributed to the observed changes including improved infrastructures in the outskirts; this has driven city expansion and rural to urban migration. The opening up of city outskirts through roads has also been significant as many peri-urban areas have become more accessible through the development of municipal infrastructures. Other factors include various housing projects in peri-urban areas and resettlement schemes, which have moved a large share of the city population away from hazardous areas located near the city centre to the city outskirts.

2015–2018
From 2015 to 2018, the population of Dar es Salaam grew from 5,115,698 to 6,368,272 at growth rate of 5.69%. Built-up area grew from 38.73% to 51.80% (13.07%) 630.52 and 843.30 km2. respectively, which is equivalent to 4.35% change per year (see Figure 7). The percentage change between these years was found to be high compared to other years, though it was only a 4-year interval. A number of the aforementioned factors continued to contribute to the observed changes, namely, improved infrastructure in the outskirts of the city, rural to urban migration and the expansion of city boundaries. In addition, a reduction of government local expenditure has led to an increasing rent burden in the central city wards, which has, in turn, driven more people to purchase land in the outskirts of the city and begin constructing their own houses.

From the above findings on land cover changes in the city of Dar es Salaam, from 1995 to 2018, built-up area seems to have grown year by year, largely through sprawl. The direction of sprawl extends towards the edge of the city and is influenced by a number of factors (see Table 2). While new areas seem to be emerging in the outskirts, some places also depict an increasing population density through concentration of built area. Overall, although the city is sprawling, compact development can still be observed in some parts of the city with a high population density; hence, measures for sustainability need to be taken to ensure land consumption and growth rate are well aligned. The findings present empirical evidence of changes that, in transforming the built-up area, are also affecting the spatial development of the city. The rate of change can help inform policymakers in addressing issues of spatial development in highly urbanizing cities, as well as urban planning authorities, which are responsible for ensuring land is used productively and sustainably. The pattern of change is discussed below, with a focus on how internal structures are affected by demographic changes.
Summary Table, Land Cover Changes
Pattern of Built-up Area Change and Population Dynamics
The population of Dar es Salaam is growing very fast. As aforementioned, over the span of 10 years (two census periods), it almost doubled with a 43% increase. While this continues to rise, the internal structure of the city is also undergoing changes. According to the 2002 Census, the lowest ward population was recorded as being 2,944 people and the highest was 83,392. However, by the 2012 Census this rose to the lowest ward population being 6,411, while the highest had a population of 202,582. This is equivalent to a 54.07% increase in people in the lowest ward population and a 58.83% increase in the highest ward population, as recorded over two census rounds (see Figures 8 and 9). The changes which were observed to have taken place within the internal structure of the city population have also had a direct effect on changes in the built-up area, as evidenced through the increase in population within wards located on the city’s outskirts. Generally, the city is growing from the centre outwards towards its major transport corridors (arterial roads) through a finger-like structure model. There is a much higher concentration in the northern, southern and western sides, as the eastern side is constrained by the Indian ocean (Hill & Lindner, 2009).


Although the city population seems to have increased over time, the rate of change is different from ward to ward. According to the 2002 Census, many wards in the Central Business District (CBD) and its vicinity were recorded as having a high population. However, in the 2012 Census, this trend changed as higher population growth was found further away from the city centre. Most of the wards in the CBD and its vicinity were recorded as having a negative change in population while the wards in the outskirts and peri-urban areas were recorded as having a positive change, with the highest being a 700% change from 2002 to 2012 (see Figure 10). While many factors have contributed to this trend, the most notable is rising housing rents in the city centre, as a result of which many people moved to the outskirts where rent prices are considerably lower. This has also led to an increase in private housing construction on the outskirts where unplanned land is cheap and easily available. Consequently, over a period of 23 years, the city’s built-up area has increased by 36.23% which is equivalent to a 1.5 yearly increase. These changes are mainly driven by population growth and the affordability of housing, which have directly affected land use and built-up area. Notably, the improvement of physical infrastructure such as roads has significantly contributed to increasing changes in built-up area in the city, regardless of the presence of planned or unplanned land/settlements. This is depicted in the Figure 10, which shows the population change and tentative changes of built-up area as extracted from satellite images between 1995 and 2018 (over unequal intervals, owing to availability of data).

While the population change in the city centre was found to be negative from 2002 to 2012, population density was still found to be high in the central wards. However, there have been slight changes in the distribution of population density as it has been growing outwards from the central parts of the city to the wards on the outskirts (see Figure 11).

While observing the distribution, population density was found to be high in wards located near the centre of the city, yet the percentage changes in population density from 2002 to 2012 were found to be high in the peripheral wards with negative changes observed in the centre (see Figure 12). Further investigation reveals that the wards with a high change in population density were those that were particularly involved in planning projects. This increased people’s access to land, through which many urban residents were able to purchase land and construct their own houses, leaving rental housing in the city centre. On the other hand, findings in terms of population density are fittingly aligned with findings in overall population change as per the ward analysis made across the two census rounds (2002–2012).
This pattern of growth not only indicates the need for proper planning in the outskirts, but also suggests that a change in the conventional approaches to urban planning is much required. Since there is a high influx of people migrating to the peripheral wards, this raises concerns regarding the nature of spatial development taking place in these areas.
Pattern of Urban Sprawl
Urban sprawl, as identified by the UN-Habitat (2008), refers to a visible misalignment of population growth and the physical expansion of the city. The city of Dar es Salaam has presented a remarkable pattern of development in this regard, channelled through both compact and sprawling growth. Similar to the outskirts of most African cities which are unbalanced in relation to their settlement patterns, population density and land use (Eckert, 2011), the results of the study indicate that urban sprawl in Dar es Salaam radiates from the city centre to the peripheries, which are highly dominated by informal low-density settlements (Bhanjee & Zhang, 2018). This pattern is also evident through the difference which was found between the 2002 and 2012 census periods, in which the percentage change of discontinuously built-up area and demographic changes were significant, at 192% to 75%, respectively. Continuously built-up area on the other hand, almost followed the demographic change with a 76% change (Congedo & Macchi, 2015). Although the city is sprawling, compact development can still be observed in some parts of the city with a high population density (see Figure 12). This shows that urban planning authorities have been incapable of keeping up with the pace of urban growth in the city, with informal settlements expanding rapidly and creating a sprawling pattern of growth along the major transport corridors and the peri-urban areas of the city (Congedo & Munafò, 2012).

Driving Forces Behind Built-up Area Changes and Urbanization
As identified by Estes et al. (2012), ‘Analysis of land-cover change can act as a good proxy to track population increases while also yielding clues about the drivers of these trends’. The high rate of change in built-up area happening in the city is largely associated with residential development and urbanization. In addition to this, factors such as the development of transport and communication infrastructures between the city’s settlements have proved to have high influence on spatial temporal expansion and land cover change in the city (Mkalawa & Haixiao, 2014). This is evident through the fact that areas which are well connected to infrastructure tend to be more developed, such as the southern part of the city in Kigamboni Municipal (see Figure 9). Other driving factors behind built-up area change include migration, high natural growth rates of urban populations, public policies such as the housing policy in 1972 which legalized the existence of squatters (Armstrong, 1986), and the development of agglomeration economies (Mkalawa, 2016). The city is currently undergoing multiple processes of urban change, aggravated by population growth and poor control over urban development. New spatial development changes have been driving urban planners and managers, in their operation with most of their intervention plans being reactive of the processes such as land regularization (Hill & Lindner, 2010).
The need to accommodate the increased urban population with adequate housing has also sped up land use changes in the city, with most of the agricultural land in the city’s outskirts being converted for residential and other purposes. This phenomenon is taking place in an unprecedented manner, with most of the development taking a sprawling pattern, characterized by informal spatial development. Supporting this, some studies have found that the need for affordable land by the urban poor has been a major driving force towards the informalization of land use changes in the peri-urban areas of the city (Mkalawa, 2016). This is further influenced by a land market and tenure system which inversely affects residential location preferences (Andreasen et al., 2017). It is clear that in Dar es Salaam, planning authorities have failed to meet the demands for land and housing, with informal urban spatial development outpacing their ability to plan and survey land in the city (Hill et al., 2014). In this context, understanding the change in built-up areas, population density and its relation to spatial development can help inform policy decisions and enhance the preparedness of urban planning authorities in taking measures towards sustainable urban development.
Conclusion
For cities located in developing countries, land cover and use change can be used as a monitoring tool in spatial planning to keep up with urban growth while ensuring sustainable land development (Congedo & Munafò, 2014). This assumes more importance in light of SDG 11. Among other factors, land speculation, which dominates most of the land market in Dar es Salaam, is found to be a major driving force behind land use and cover changes in the city and its peripheries (Kombe, 2005). For planning authorities to effectively respond to rapid urban growth taking place in the city, a holistic approach is needed for understanding physical and demographic changes, which can be attained through a critical analysis of urban land patterns. With sprawl dominating the urban land pattern of the city, housing settlements are becoming exposed to many other urban challenges such as poor accessibility, vulnerability to floods, lack of tenure security and inadequate access to services such as education and health. Although planned areas of sprawl are said to be less socially vulnerable than informally developed sprawl areas (Bhanjee, 2019), both patterns entail unsustainable changes in land use and land cover. Through the findings of this study, it is evident that the city will continue to grow in a sprawling manner while also continuing to densify in the centre. This will occur regardless of the negative population change observed in smaller urban wards that are more dominated by mixed-use, high-rise buildings (as compared to wards in the outskirts). Hence, appropriate measures should be taken by planning authorities to ensure the city development is sustainable with increased urbanization.
Footnotes
Acknowledgements
A special thanks to Dr Richard Mitchell (University of Glasgow), Professor David Everett and Professor Caryn Abrahams (University of Witwatersrand), Dr Ivan Turok and Andreas Scheba (Human Sciences Research Council, Cape Town) and SHLC partners for their comments.
Declaration of Conflicting Interests
The authors declared no potential conflicts of interest with respect to the research, authorship and/or publication of this article.
Funding
This articles was written as part of the work of the Centre for Sustainable, Healthy and Learning Cities and Neighbourhoods (SHLC), which is funded via UK Research and Innovation, and administered through the Economic and Social Research Council, as part of the UK Government’s Global Challenges Research Fund. Project Reference: ES/P011020/1.
