Morphological Characteristics of Informal Settlements and Strategic Suggestions for Urban Sustainable Development in Tanzania: Dar es Salaam, Mwanza, and Kigoma

: Rapid urbanization in developing countries has been accompanied by the spread of informal settlements, which is particularly prominent in sub-Saharan Africa. These settlements have become an important supplement to the inadequate formal housing supply in cities, and their spontaneously formed spatial patterns have important inﬂuences on sustainable development. In this study, qualitative and quantitative approaches were used to examine the morphological characteristics of informal settlements in Tanzania and the associated inﬂuences on urban development. Geographic spatial analyses, landscape pattern indices, and mathematical statistics, along with quick assessments, group discussions, and key informant interviews, were used to obtain detailed information on the spatial forms of informal settlements. The results indicate that the form of the settlements does not conform to the social, economic, or environmental characteristics of sustainable development. The disordered expansion of single-layered buildings with a single function, irregular road networks in poor condition, and a lack of consideration and protection of the ecological environment were found to negatively impact urban function and sustainable development. However, the structure and form of informal settlements could, in addition to formalization projects, be optimized to drive sustainable and socioeconomic development goals as well as environmental conservation.


Introduction
Given rapid urbanization, traditional formal urban sectors in developing countries cannot meet the needs of growing urban populations, and the coexistence of formal and informal sectors has become a major paradigm of urban development [1]. Informal settlements have become an important component of the urban housing sectors of developing countries. 'Informal settlement' refers to a residential area that lacks basic services and infrastructure, where the housing stock may not comply theory-building [42]. However, such studies are challenged by the need to develop far reaching, useful, and accurate descriptions when applied to other contexts [43]. Although enlightening, the results from these studies may not be applicable everywhere; specific local context must be considered.
Tanzania has one of the highest proportions of urban population living in informal settlements in sub-Saharan Africa, at 50-80% [44]. The growth of informal settlements is an important form of urban expansion in most Tanzanian cities. Presently, many studies only focused on the growth and modeling of the informal settlements in Dar es Salaam [45][46][47], with most attention centered on social and economic development problems, regularization approaches, and the impacts of informal settlements [48,49]. Studies on informal settlement morphological characteristics in different cities and their influence on urban sustainable development are lacking, despite Tanzania's secondary and tertiary cities absorbing increasing numbers of urban residents [50,51]. We selected the informal settlements of three cities in Tanzania in different development stages as case studies to analyze the morphological characteristics of informal settlements and their impact on urban sustainable development goals. The results provide useful structural data for the development of informal settlement renovation programs and offer a useful reference for urban planners, managers, and policymakers at regional and national levels.

Methodology
The main steps of the study were as follows: (1) Case area selection, (2) morphological analysis of the informal settlements, (3) functional development analysis of the informal settlements, and (4) evaluation of the influence of the informal settlements on sustainable development. The various input data sets used, methods, results, and the logical dependencies between them are detailed in Figure 1. Mean shape index (MSI), Shannon's diversity index (SDI), Shannon's evenness index (SEI), and the mean Euclidean nearest-neighbor (ENN) distance were used as indicators of building form [52,53]. Graph-theory-based concepts were used by employing α, β, and γ indices to determine the road network's circuitry, complexity, and connectivity [54].
Sustainability 2020, 12, x FOR PEER REVIEW 3 of 23 approach across various disciplines that could be used for testing hypotheses against reality and contribute to an inductive discovery of new hypotheses and theory-building [42]. However, such studies are challenged by the need to develop far reaching, useful, and accurate descriptions when applied to other contexts [43]. Although enlightening, the results from these studies may not be applicable everywhere; specific local context must be considered. Tanzania has one of the highest proportions of urban population living in informal settlements in sub-Saharan Africa, at 50-80% [44]. The growth of informal settlements is an important form of urban expansion in most Tanzanian cities. Presently, many studies only focused on the growth and modeling of the informal settlements in Dar es Salaam [45][46][47], with most attention centered on social and economic development problems, regularization approaches, and the impacts of informal settlements [48,49]. Studies on informal settlement morphological characteristics in different cities and their influence on urban sustainable development are lacking, despite Tanzania's secondary and tertiary cities absorbing increasing numbers of urban residents [50,51]. We selected the informal settlements of three cities in Tanzania in different development stages as case studies to analyze the morphological characteristics of informal settlements and their impact on urban sustainable development goals. The results provide useful structural data for the development of informal settlement renovation programs and offer a useful reference for urban planners, managers, and policymakers at regional and national levels.

Methodology
The main steps of the study were as follows: (1) Case area selection, (2) morphological analysis of the informal settlements, (3) functional development analysis of the informal settlements, and (4) evaluation of the influence of the informal settlements on sustainable development. The various input data sets used, methods, results, and the logical dependencies between them are detailed in Figure 1. Mean shape index (MSI), Shannon's diversity index (SDI), Shannon's evenness index (SEI), and the mean Euclidean nearest-neighbor (ENN) distance were used as indicators of building form [52,53]. Graph-theory-based concepts were used by employing α, β, and γ indices to determine the road network's circuitry, complexity, and connectivity [54].

Tanzania's Informal Settlements
Tanzania's informal settlements have their own unique characteristics in terms of tenure security, structural quality of housing, and inhabitants [50,55]. First, anyone who erects a structure in the informal settlement has a perceived security of tenure emanating from three generations. Second, the use of modern building materials has increased tremendously since the early 1990s. Third, informal settlements in Tanzania consist of a wide range of socioeconomic groups [56].

Case Area Selection and Description
With a shortage of planned development areas, between 40% and 80% of all built up areas are informal in Tanzania [57,58], and most urban households in all major cities are in informal settlements [59]. Based on the development stage, population growth rate, availability of relevant data, and feasibility of field investigation, we selected Dar es Salaam, Mwanza, and Kigoma as case areas ( Figure 2). Dar es Salaam has long been Tanzania's primary city, with the highest population growth in East African cities in recent years [9], but the city is simultaneously facing serious problems of informal settlement expansion. Mwanza is the second largest city in Tanzania with over 700,000 urban inhabitants in 2012 and a fast-growing population [60]. Mwanza is being challenged by a series of problems related to informal settlements expansion. Kigoma is a major town and transportation hub in Western Tanzania that has been experiencing rapid development in recent years and was selected as the representative of Tanzania's emerging small-and medium-sized cities. The development of informal settlements is causing problems to varying degrees in the three cities. Three case informal settlements were selected in each city based on the criteria that they (1) have shown notable expansion since 2000. Previous studies on urban expansion in Tanzania and Google Earth images (Google Inc., Mountain View, CA, USA) were used for this purpose; (2) should have a rapidly growing population in recent years, thus reflecting rapid urbanization in terms of population growth; and (3) should be representative in terms of urban development and morphological changes. The settlements were recommended by local experts and scholars. The three selected informal settlements were Chamazi (Dar es Salaam), Buhongwa (Mwanza), and Gungu (Kigoma).

Chamazi
Chamazi belongs to the Temeke municipality in Dar es Salaam, with an area of 27 km 2 and an estimated 74,000 inhabitants in 2016. Due to its size, only part of it was selected as the research area. The population data were obtained from the ward executive officer of Chamazi. Two sub-wards,  Chamazi belongs to the Temeke municipality in Dar es Salaam, with an area of 27 km 2 and an estimated 74,000 inhabitants in 2016. Due to its size, only part of it was selected as the research area. The population data were obtained from the ward executive officer of Chamazi. Two sub-wards, Mkondogwa and Vigoa, were selected to compare socioeconomic development differences in areas with differing building densities (Figure 3).

Chamazi
Chamazi belongs to the Temeke municipality in Dar es Salaam, with an area of 27 km 2 and an estimated 74,000 inhabitants in 2016. Due to its size, only part of it was selected as the research area. The population data were obtained from the ward executive officer of Chamazi. Two sub-wards, Mkondogwa and Vigoa, were selected to compare socioeconomic development differences in areas with differing building densities (Figure 3).

Buhongwa
Buhongwa is located in Nyamagana municipality, Mwanza. It has an area of 45 km 2 and an estimated 37,000 inhabitants based on 2016 data. The population data were obtained from the ward executive officer of Buhongwa. Due to its large size, part of the settlement was selected as the research area. With an average annual population growth rate of 8.3%, it is the fastest growing ward in Mwanza. Shibayi and Mitimilefu sub-wards, which have different building densities, were selected for study ( Figure 4). According to the actual situation of the investigation, the Shibayi sub-ward was outside of the area we selected, but this should have no influence on our analysis.

Buhongwa
Buhongwa is located in Nyamagana municipality, Mwanza. It has an area of 45 km 2 and an estimated 37,000 inhabitants based on 2016 data. The population data were obtained from the ward executive officer of Buhongwa. Due to its large size, part of the settlement was selected as the research area. With an average annual population growth rate of 8.3%, it is the fastest growing ward in Mwanza. Shibayi and Mitimilefu sub-wards, which have different building densities, were selected for study ( Figure 4). According to the actual situation of the investigation, the Shibayi sub-ward was outside of the area we selected, but this should have no influence on our analysis.

Gungu
Gungu belongs to Kigoma-Ujiji municipality, Kigoma. It has an area of 5.6 km 2 and an estimated 29,355 inhabitants according to data from 2016, with an average annual population growth rate of 6.8%. The population data were obtained from the Ward Executive Officer of Gungu.

Gungu
Gungu belongs to Kigoma-Ujiji municipality, Kigoma. It has an area of 5.6 km 2 and an estimated 29,355 inhabitants according to data from 2016, with an average annual population growth rate of 6.8%. The population data were obtained from the Ward Executive Officer of Gungu. Two sub-wards with different building densities, Mwenge and Masanga, were selected for this study ( Figure 5).

Gungu
Gungu belongs to Kigoma-Ujiji municipality, Kigoma. It has an area of 5.6 km 2 and an estimated 29,355 inhabitants according to data from 2016, with an average annual population growth rate of 6.8%. The population data were obtained from the Ward Executive Officer of Gungu. Two sub-wards with different building densities, Mwenge and Masanga, were selected for this study ( Figure 5).

Field Survey Data
In August 2017, the research team conducted field surveys involving quick assessments, group discussions, and key informant interviews. Quick assessments recorded the layout of buildings, number of building floors, road conditions, socioeconomic activities, and environmental conditions in each settlement. A scoring criterion that considered the courtyard situation, road conditions, number of building floors, degree of mixed land use, and roofing and wall materials was used to evaluate the quality of buildings ( Table 1). Most of the previous studies considered wall materials, roof materials, and floor materials; other factors were considered here to judge the building quality more comprehensively. From the field survey, we found that the three settlements basically include one-storey buildings, with almost no access to paved roads. These two indicators were not included in the comprehensive score. Each settlement was divided into a grid. Chamazi was divided into 82,500 × 500 m grids, Buhongwa into 159,500 × 500 m grids, and Gungu into 100,250 × 250 m grids. Grids with different building densities were selected in each settlement, and 1-3 points (total 182 points) were assigned in each grid to produce a score of building quality.
A group discussion was organized in each of the six sub-wards, involving the ward executive officer, sub-ward leader, and four randomly selected local residents. In total, 33 local people participated in group discussions. Through pre-prepared questions, information about the infrastructure and basic service supply, land right status, development driving force, and main challenges and current countermeasures of each ward and sub-ward were obtained.
For each city, we interviewed one city planner, one ward executive officer, and two sub-ward leaders. In total, 12 key informants were selected for in-depth interviews to obtain detailed information about the development history, current situation, problems, and latest progress of urban planning efforts regarding the respective informal settlements. Limited by research time and resources, we were unable to interview the other sustainable-development-related professionals such as economists or social and environmental scientists.

Evaluating Indicator Observation Score
Courtyard 30% and below have a courtyard 1 30-70% have a courtyard 3 70% and above have a courtyard 5 Roofing materials 30% and below are concrete, ceramic tile, or high quality iron sheet 1 30-70% are concrete, ceramic tile, or high quality iron sheet 3 70% and above are concrete, ceramic tile, or high quality iron sheet 5 Wall materials 30% and below are permanent materials such as bricks or cement 1 30-70% are permanent materials such as bricks or cement 3 70% and above are permanent materials such as bricks or cement 5 Degree of mixed land use Mainly used for living 1 Mix of residence and agriculture 3 Mix of residence, industry, and commerce 5

GIS and Remote Sensing Data
GIS data, including administrative boundaries and road networks, were obtained from the Tanzania National Bureau of Statistics and OpenstreetMap (Geofabrik GmbH, Karlsruhe, Germany), respectively. Digital elevation model data were obtained from the United States Geological Survey and used to calculate the gradients of each settlement. High-resolution remote sensing images were used to extract the profiles of buildings using eCogniton 8.7 software (Trimble Germany GmbH, Munich, Germany). The influence of the cloud amount in the remote sensing images can seriously impact interpretation results. According to the quality of the images in each year for each city, we finally chose remote sensing (RS) images in different years in the three informal settlements. Road layers were obtained by manual vectorization using road data in combination with RS images.

Building Form
A set of spatial pattern indices (MSI, SDI, SEI, and ENN) were used to analyze the morphological characteristics of buildings: shape, size, density, and distribution patterns. The density of buildings is key to identifying informal settlements. A settlement can be characterized by both the building unit density (number of buildings per unit area) and building coverage (the ratio of the total area of buildings to the area of the corresponding administrative region).

Road Connectivity
The α index measures the circuitry of a network using the ratio of the number of circuits in the network to the maximum number of circuits possible (Equation (1)). The α index ranges between 0 to 1, with greater values indicative of a higher degree of network circuitry. The β index reflects the complexity and completeness of the road network and is expressed as the ratio of links to nodes (Equation (2)). The β index ranges from 0 to 3; β < 1 indicates a disconnected network, β = 1 indicates that the network is one single loop, and β > 1 means the network connectivity is rather complex. The γ index is used to measure the degree of node connection in a network, expressed as the ratio of links in a network to the maximum number of links between nodes (Equation (3)) [54]. where L is the number of links and N is the number of nodes.

Change in Building Area
Changes in the spatial distributions of buildings over time in Chamazi, Buhongwa, and Gungu are shown in Figures 6-8 Table 2).

Change in Vegetation Coverage
Vegetation coverage in each settlement was found to decline over time. Chamazi had the largest decrease in vegetation cover from 6.5 km 2 in 2005 to 2.5 km 2 in 2013, indicating that the conversion of green space (mainly woodland and cultivated land) was extensive during settlement expansion. The vegetated area in Buhongwa declined from 26% in 2004 to 19% in 2016, suggesting an increase in farmland encroachment. Vegetation coverage (mainly shrubs) in Gungu declined comparatively less than in the other two settlements, but nonetheless declined from 33% in 2000 to 25% in 2013 (Table  3). . The average building area in Chamazi and Gungu was basically unchanged and tended to be stable; in Buhongwa, it increased gradually, suggesting an improvement in living conditions. The standard deviation (SD) of building size in each settlement was large, reflecting the lack of unified planning guidance for housing construction ( Table 2).  Table 2).

Change in Vegetation Coverage
Vegetation coverage in each settlement was found to decline over time. Chamazi had the largest decrease in vegetation cover from 6.5 km 2 in 2005 to 2.5 km 2 in 2013, indicating that the conversion of green space (mainly woodland and cultivated land) was extensive during settlement expansion. The vegetated area in Buhongwa declined from 26% in 2004 to 19% in 2016, suggesting an increase in farmland encroachment. Vegetation coverage (mainly shrubs) in Gungu declined comparatively less than in the other two settlements, but nonetheless declined from 33% in 2000 to 25% in 2013 (Table 3). The average shape of buildings was simple and almost rectangular. The length and width ratio of buildings in Gungu was the largest, followed by Buhongwa, then Chamazi. Standard deviation values indicated the degree of variation in building shape, with the largest values found for Gungu, followed by Buhongwa and Chamazi.
Building diversity, as indicated by SDI, was found to increase gradually, with Chamazi having the largest values, followed by Buhongwa and Gungu. This increase was indicative of the coexistence of broad social groups in the informal settlements of Tanzania. SEI values indicate that the building distribution in Chamazi remained relatively even, whereas that of Buhongwa was relatively heterogeneous, perhaps due to its rocky-hilly terrain and large mining areas.
The ENN values suggested that the greatest changes in building landscape occurred in Chamazi.  (Table 4).

Block Texture
Building density in Chamazi was mainly medium and high, composed mostly of single-storey detached buildings with few high-rise buildings. We found no clear textural features, the building sizes and building shapes varied widely, and the orientation was changeable; the blocks showed a twig-like distribution that followed the unplanned road network.
Building density in Buhongwa was mainly medium and low, containing mostly single-story detached houses except for some houses built on rocky-hilly terrain. With the exception of a large school in Southern Buhongwa, few other high-rise buildings were present. With the exception of the relatively high building density along Kenyatta Road, most areas had low building density. The housing distribution was very complex, with varied sizes, orientations, and densities; the blocks showed an irregular twig-like distribution.
Building density in Gungu was mainly medium and high, with no multi-story buildings except for some churches and training schools. Buildings were dense and exhibited a certain texture overall. The orientations, sizes, and distributions of the houses were relatively regular compared to the other two settlements. The blocks showed a mixture of checkerboard and twig-like distributions.

Road Network Structure and Accessibility
The circuitry, complexity, connectivity, and density of the road network in Gungu were the highest, followed by that in Chamazi and Buhongwa (Table 5), which we suggest is related to proximity to the city center. Gungu is close to the city center and one of its two main roads runs through central Gungu. Chamazi and Buhongwa are located slightly farther from the city center. Except for its northwest corner, most of Gungu is rather flat and contains no river crossings. There is a seasonal river in the southwestern part of Chamazi, a large sand mining area and a small lake in the west, and large areas of woodland in the south and east. In Buhongwa, large areas of farmland are found in the south and north, and a mining area in the central part. In addition, the railway line has a certain impact on road connectivity. Notes: The α index, β index, and γ index measures the circuitry, complexity, and connectivity of a network, respectively.

Distribution Characteristics of Building Density
(1) With the increase in the road network density, building density tended to increase (Figure 9). Convenient transportation can promote economic development within settlements since the main economic activities are conducted along the main roads. Buildings were also found to be concentrated around the roads. In informal settlements, buildings tend to be established first, with roads formed around them. The road network was complex and irregular with poor road conditions, which reduce residents' quality of life and limit the economic development potential of the settlements. (2) The results showed that building density decreased with increases in slope ( Figure 10). Many of the buildings in informal settlements were constructed ignoring slope conditions. The total area and average number of buildings with a slope of more than 15 degrees were largest in Chamazi, followed by Buhongwa and Gungu. Floods and landslides are significant environmental problems in these settlements. The demolition and reconstruction of residential areas have also caused difficulties. (3) With the increase in the ENN distance, building density tended to decrease ( Figure 11). We found that the average distance between buildings was about 10 m, which gradually decreased over time. Many buildings are very close together and have a disordered orientation. were both separated and intermixed. There was an obvious social gap between the residents of these two building types.
(1) With the increase in the road network density, building density tended to increase (Figure 9). Convenient transportation can promote economic development within settlements since the main economic activities are conducted along the main roads. Buildings were also found to be concentrated around the roads. In informal settlements, buildings tend to be established first, with roads formed around them. The road network was complex and irregular with poor road conditions, which reduce residents' quality of life and limit the economic development potential of the settlements. (2) The results showed that building density decreased with increases in slope ( Figure 10). Many of the buildings in informal settlements were constructed ignoring slope conditions. The total area and average number of buildings with a slope of more than 15 degrees were largest in Chamazi, followed by Buhongwa and Gungu. Floods and landslides are significant environmental  (3) With the increase in the ENN distance, building density tended to decrease ( Figure 11). We found that the average distance between buildings was about 10 m, which gradually decreased over time. Many buildings are very close together and have a disordered orientation.   (3) With the increase in the ENN distance, building density tended to decrease ( Figure 11). We found that the average distance between buildings was about 10 m, which gradually decreased over time. Many buildings are very close together and have a disordered orientation.  separated and intermixed. There was an obvious social gap between the residents of these two building types. (5) With increasing building density, building quality tended to worsen ( Figure 13). In 2013, the building densities of Chamazi and Gungu were 6.712% and 23.237%, respectively; in 2016, the building density of Buhongwa was only 1.615%. New, good quality housing in Chamazi and Gungu was mostly distributed in newly developed areas with a lower building density, low population, and poor infrastructure conditions. Settlements mainly expanded horizontally, with building quality and density showing a negative correlation. In most areas of Buhongwa, building density remained low over time, with newly added houses interspersed with old houses. Building construction was mainly in the process of densification. Therefore, the relationship between building quality and building density was not obvious.  (5) With increasing building density, building quality tended to worsen ( Figure 13). In 2013, the building densities of Chamazi and Gungu were 6.712% and 23.237%, respectively; in 2016, the building density of Buhongwa was only 1.615%. New, good quality housing in Chamazi and Gungu was mostly distributed in newly developed areas with a lower building density, low population, and poor infrastructure conditions. Settlements mainly expanded horizontally, with building quality and density showing a negative correlation. In most areas of Buhongwa, building density remained low over time, with newly added houses interspersed with old houses. Building construction was mainly in the process of densification. Therefore, the relationship between building quality and building density was not obvious.
Sustainability 2020, 12, x FOR PEER REVIEW 14 of 23 separated and intermixed. There was an obvious social gap between the residents of these two building types. (5) With increasing building density, building quality tended to worsen ( Figure 13). In 2013, the building densities of Chamazi and Gungu were 6.712% and 23.237%, respectively; in 2016, the building density of Buhongwa was only 1.615%. New, good quality housing in Chamazi and Gungu was mostly distributed in newly developed areas with a lower building density, low population, and poor infrastructure conditions. Settlements mainly expanded horizontally, with building quality and density showing a negative correlation. In most areas of Buhongwa, building density remained low over time, with newly added houses interspersed with old houses. Building construction was mainly in the process of densification. Therefore, the relationship between building quality and building density was not obvious.

Number and Composition of Residents
The three settlements were previously mainly composed of woodland, agricultural land, and open space. However, with the development of the national economy and market economy reforms, more land began to be sold to outsiders from around the year 2000 and onward. This led to the coexistence of different social groups within settlements, which are currently mainly of medium-and low-income levels. Of the 24 residents interviewed, 15 relocated to the study area because the settlement was close to the main road, land was available at a low price, and the costs of living were low.
The increase in the number of migrants has driven rapid population growth in the settlements. The population density in the planned residential areas was much higher than that in the informal case settlements, indicating that settlements have much room for improvement in terms of population carrying capacity. The three settlements had an average of four to six people per household and an average of 5 to 10 people per building due to the existence of tenants. Chamazi has a large floating population (50%), mainly consisting of mobile vendors and workers from Illala and Kigoboni. The floating population of Buhongwa (22%) and Gungu (4%) are mainly students and mobile vendors from nearby wards. The residents of the three settlements are mainly engaged in small business activities (mostly informal), fisheries, animal husbandry, agriculture, and mining. The proportion of residents with a fixed occupation and income was 50% in Chamazi, 20% in Gungu, and 15% in Buhongwa. In all three settlements, the illiteracy rate was below 5%, with most residents having some level of primary or secondary education. These data were obtained through group discussions and key informants interviews.

Infrastructure Development
Despite the relatively adequate supply of electricity, the three settlements are far from meeting the needs of local residents in terms of the provision of public water, drainage, solid waste collection, education, and medical care. Chamazi has not been connected to the urban public water supply network, and only 20% of the residents of Buhongwa and Gungu have access to the public water supply. Most of the residents use pit latrines. There are no public drainage facilities. Groundwater and springs are easily contaminated by toilet waste, so diarrhea is a common problem. Solid waste is mainly collected by private contractors, but their services are very limited. Other wastes are mainly burned or thrown away. The prevalence of informal economic activities substitutes the shortage of commercial service facilities in the settlements to a large extent. Due to the lack of educational facilities, some primary school classrooms in Chamazi contain up to 600 students. Medical facilities in Buhongwa are especially inadequate.

Living Environment and Health
There are virtually no measures to protect the ecological environment in the informal settlements. With the wide use of dry toilets, and the lack of public drainage facilities and solid waste collection services, the living environment is often highly deteriorated, posing threats to residents' health. The three rivers of Chamazi often flood during the rainy season, and houses built along the river typically collapse during the rainy season. Much open-air waste near the Mzinga River ( Figure 14) contributes to water pollution problems. Numerous sand mining activities occur around the seasonal river in the southwestern part of Chamazi. To prevent flooding during the rainy season, local residents have constructed a flood embankment using living garbage (Figure 15), which may be polluting the water.    There are some industrial activities, a pharmaceutical factory, and a large garbage dump along the Nyashishi River in Southern Buhongwa ( Figure 16). Nearly 85% of the garbage in Mwanza is processed at the Buhongwa dump, and this has negatively affected the surrounding environment. Gungu encounters light or heavy mountain torrents almost every year, causing serious waterlogging. The increased building activities have led to fewer trees and consequent soil erosion.  There are some industrial activities, a pharmaceutical factory, and a large garbage dump along the Nyashishi River in Southern Buhongwa ( Figure 16). Nearly 85% of the garbage in Mwanza is processed at the Buhongwa dump, and this has negatively affected the surrounding environment. Gungu encounters light or heavy mountain torrents almost every year, causing serious waterlogging. The increased building activities have led to fewer trees and consequent soil erosion.  There are some industrial activities, a pharmaceutical factory, and a large garbage dump along the Nyashishi River in Southern Buhongwa ( Figure 16). Nearly 85% of the garbage in Mwanza is processed at the Buhongwa dump, and this has negatively affected the surrounding environment. Gungu encounters light or heavy mountain torrents almost every year, causing serious waterlogging. The increased building activities have led to fewer trees and consequent soil erosion.

Discussion
Like many other developing countries, informal settlements are playing an increasingly important role in providing living and working spaces for the burgeoning urban population of Tanzania. The property and business formalization program launched in 2004 and implemented in 2008 is now the main method to cope with informal development in Tanzania, which proved to be helpful for regulating the real estate market, reducing land conflicts, and promoting economic development of informal settlements [61,62]. Many researchers reported that the morphological characteristics of informal settlements may have some important influence on sustainable development [15,37], but how to respond is still open for discussion, and little has been done to specifically tackle this problem. According to our findings, the spatial forms of informal settlements in Tanzania do not conform well to the principles of sustainable development in terms of economic, societal, and environmental protection. Through the optimization of the structure and form of the elements of informal settlements, their spatial development can better align with the sustainable development goals, which may further enhance the effectiveness of property and business regularization projects.

Building Density Control
Buildings in the informal settlements tend to have simple shapes and small areas; most are single-story detached buildings mainly used for living. The land use function is singular and there is a notable lack of markets, hospitals, schools, and other buildings for social service provision, resulting in high commuting costs and low life quality. The average distance between buildings is short, and many houses have no distinct boundary. They are intermingled, leading to poor privacy and poor living conditions. Building quality tended to worsen with increasing building density, with most of the low-quality buildings spreading horizontally. Building quality showed considerable disparity, the orientation of the buildings was highly variable, and high-and low-density building areas were mixed in a disorderly manner, failing to conform to the principles of compact communities. Areas of excessively high building density suffer from ventilation and lighting issues, and congested living conditions can easily cause land disputes and neighborhood conflicts. Areas of low building density reflect inefficient land use.
Building density is an important index that can be used to identify informal settlements since it is quite different from the planned area [16,63]. Building density needs to be controlled to improve the living conditions and land use efficiency. Areas with excessively high building densities (typically the areas with poor building quality) should be controlled to prevent further densification. Measures should be taken to transfer residents to low building density areas if necessary, thus creating space for the provision of roads and sanitation facilities. Areas with excessively low building densities should be merged and consolidated with each other to increase the population density, thus leading to improvements in land use efficiency and the provision of more economical infrastructure and social services. Two-story or multi-story buildings should be encouraged. The government, non-governmental organizations (NGOs), and community-based organizations should adopt a two-story or multi-story building scheme for the transformation of existing houses and the planning of new plots. Although the middle classes are increasingly entering informal settlements and are keen to invest in the building of single-layered luxurious houses, informal settlements continue to be mainly used to meet the housing needs of the urban low-income class. The development of multi-story housing is a possible method to deal with this demand and will help control the disordered low density horizontal expansion.

Environmental Protection
The layout of the buildings reflects the low amount of consideration of local environmental conditions, with many buildings constructed in or near environmentally risky areas, which jeopardizes the lives and properties of the residents. Safe drinking water is lacking, as rivers and groundwater sources are polluted by garbage and feces. Buildings have replaced much of the farmland and vegetated areas, the service functions of ecological space in informal settlements are declining continuously, and the landscape pattern is generally fragmented. The buildings scattered in the environmental risk area are difficult to centrally manage and providing infrastructure is inconvenient; thus, the living conditions are harsh and residents' health and property are at risk.
Measures should be implemented to improve residents' awareness of environmental risks and the need for environment protection to limit further expansion into environmentally risky areas and enhance the local capacity for mitigating environmental risks. Departments should gradually relocate residents from these areas when necessary. The characteristics and landscape value of the natural environment should be factored into local development plans, with protection given to ecological spaces. These places are not only important for entertainment activities, but also for economic opportunities.

Road Planning
Roads in informal settlements are short and narrow and have few regular sections. Accessibility is poor, with a high proportions of dead ends. Road networks are often irregular with changeable lengths, widths, and pavement types. Since unplanned roads are mixed amongst disorderly distributed buildings, blocks have no clear textural features, instead having a semi-natural, loosely distributed, and branch-shaped pattern. This has not been conducive to the accumulation of economic activities and the effective provisioning of infrastructure and social services. Poor road conditions also restrict the development of public transport.
Road planning should be scheduled first. Tanzania Development Vision 2025 and National Growth and Reduction of Poverty II both point out that local road investment should be a high priority for inclusive growth and poverty eradication [64,65]. Since 2000, most of Tanzania's transportation budget has been devoted to the construction, maintenance, and upgrading of trunk roads, but local road conditions have rarely improved, especially in informal settlements [66]. Planned road networks can be used to regulate and control the layout of settlements, and create compact and orderly spatial forms, thus enhancing the accessibility and livability of settlements. This could also help to create activity centers, improve work and service conditions, and facilitate the transformation and upgrading of settlements in the future.

Local Autonomy
Overall, previous studies from Tanzania suggested that although local government units may not be able to deal with all land planning and administration matters, they stand a better chance of being able to do so than the central government due to their ability to respond quickly and affordably [67]. Therefore, planning rights could be moderately decentralized and encourage local autonomy. In an interview with the city planner of Kigoma, we learned that despite the efforts by the city planning department to formulate a detailed land use plan, the planning was difficult to implement due to the lack of funds for land surveys and related working personnel. Some sub-ward leaders in Gungu divided land boundaries by themselves to deal with the deficiency of government planning and to avoid land conflicts. This can promote orderly housing construction and layout and is also conducive to speeding up the process of land rights formalization. Sub-ward leaders have advantages in terms of the local residents' identity and land ownership authentication. Therefore, under the current constraints of the limited financial, management, and executive capacity of the governments at higher levels, planning rights could be decentralized to the ward level, making full use of local knowledge, and promoting the process of local autonomy.

Conclusions
The findings here provide three possible contributions. First, we obtained detailed information on the social and economic development of informal settlements in different cities of Tanzania using a field survey, group discussion, and interviews, which could provide useful basic information for formalization and upgrading for each informal settlement. Secondly, the temporal and morphological characteristics of informal settlements were analyzed, and the results showed that measures must be taken as soon as possible to prevent the further disordered spread of informal settlements. Most importantly, we discussed the possible relationship between the spatial form of informal settlements and their social economic development. We found that the spatial forms of informal settlements in Tanzania, basically the chaotic layout of single-layer dominated buildings and irregular road networks with poor accessibility, negatively impact the function and sustainable development of cities, and also weakened the development potential of informal settlements. The implementation of the Government of Tanzania's Property and Business Formalization program and other settlement upgrading strategies have generally emphasized policy and management issues related to land, housing, and social services. According to the results of this study, the government can promote the sustainable development of its society by optimizing the spatial forms of the informal settlements. First, building density should be controlled to improve living conditions and land use efficiency. Second, land for construction should be selected to avoid areas with high environmental risks, and ecological spaces should be protected. Third, land space can be maximized through the development of two-or multi-story buildings. In addition, by planning roads first, the layout of settlements could be regulated and controlled. Decentralization is another possible method to encourage local residential planning. In general, land use planning should be better integrated for the formalization of property rights [13].
An additional contribution the extraction of informal settlements from remote sensing images. Building density, area, and height are the most commonly used indicators for delineating an informal settlement and describing its morphology [23][24][25]. However, due to size differences, locations, development stages, populations, economic capacities, and the variability in the research scale, defining and dividing informal settlements with the standard general form are difficult [68]. Concordant with our findings, Sirueri [39] found different spatial patterns for building density, average building area, average distance between buildings, and road connectivity in the informal settlements in Nairobi and Dar es Salaam, as did Kuffer [40] in the slums of three cities in Asia and Africa. However, commonalities exist in the average shape of the buildings, which tend to be simple and rectangular; the buildings are mostly one story, areas ≤90 m 2 , an average distance between buildings of 10 m, and vegetation cover <25%. These indicators can be used for the identification and collection of basic knowledge of the current informal settlements in Tanzania, but data will need to be amended as informal settlements develop.
Urban sustainable development emphasizes people-orientation and capacity-building, as the three-dimensional coordinated development of society, economy, and ecology [69]. All three dimensions can be affected by the spatial organization of the city on different scales [70]. Determining how to guide a city to form a reasonable spatial form and promote the coordinated and healthy development of other elements is a hot issue in the study of urban sustainable development [71,72]. Compact city and smart growth theory embody this research. This study was also based on these theories and concepts, further indicating the important connection between spatial form and social economic development, attaching the importance of morphology in the theory of sustainable development. However, we mainly focused on the collection and processing of relevant data and the corresponding empirical analysis of the informal settlements; we did not conduct a theoretical analysis of local urban development and spatial form. More in-depth research is still needed to analyze the mechanism behind the relationship between the morphological elements of informal settlements and their functional development to propose more innovative suggestions to promote the sustainable development of the informal settlements.