Spatiotemporal Characteristics of Urban Land Expansion and Population Growth in Africa from 2001 to 2019: Evidence from Population Density Data

: Africa has been undergoing a rapid urbanization process, which is critical to the achievement of the 11th Sustainable Development Goal (SDG11). Using population density data from LandScan, we proposed a population density-based thresholding method to generate urban land and urban population data in Africa from 2001 to 2019, which were further applied to detect the spatiotemporal characteristics of Africa’s urbanization. The results showed that urban land and urban population have both grown rapidly in Africa, which increased by about 5.92% and 4.91%, respectively. The top three countries with the most intense urbanization process in Africa are Nigeria, the Democratic Republic of the Congo, and Ethiopia. The coupling relationship index of urban land expansion and population growth was 0.76 in Africa during 2001–2019. Meanwhile, the total proportion of uncoordinated development types at the provincial level was getting higher, which indicated an uncoordinated relationship between urban land expansion and population growth in Africa. Cropland, grassland, rural land, and forests were the most land-use types occupied by urban expansion. The proportion of cropland, grassland, and forests occupied was getting higher and higher from 2001 to 2019. The extensive urban land use may have an impact on the environmental and economic beneﬁts brought by urbanization, which needs further research.


Introduction
In 2018, 55% of the world's population lived in urban areas. By 2050, it is estimated that 68% of the world's population will be urban. Africa is the least urbanized region in the world with only about 43% of the population being urban [1], compared with North America (82% of the population in 2018 lives in urban areas), Latin America and the Caribbean (81%), Europe (74%), and Oceania (68%). However, Africa has the highest rate of urban population growth, exceeding 4% per year during , and is expected to remain at 3% or more per year during 2035-2040. Between 1950 and 2018, the urban population of Africa increased more than 16-fold, from 33 million to 548 million. Due to the relatively high growth rate of the urban population in Africa in the future, by 2050, Africa will have 1.5 billion urban residents, second only to Asia's 3.5 billion, although it will still be the geographic region with the lowest degree of urbanization in the world [1]. Whether in the past, present, or future, Africa is undergoing a very rapid process of urbanization. Urbanization is one of the key factors affecting the development of African cities [2,3]. expansion and population growth in Africa coordinated; (3) situation and sustainability of other land types occupied by urban land in Africa during expansion.
To increase the comparability and consistency of urbanization study in Africa across time and space, this study proposes a population density-based thresholding method to generate urban land and urban population. The rest of this study is organized as follows. Section 2 introduces the study area and data source; Section 3 describes details of methodologies adopted in this study; Section 4 presents and analyzes the results, followed by the discussion and conclusion in Sections 5 and 6, respectively.

Study Area
The total area of Africa is about 32 million km 2 , which includes 55 countries or regions, with a total of 751 provinces. In 2018, the urbanization level of Africa was 43%, with 547.6 million people living in cities [1]. Twenty African cities were selected to verify the feasibility of the methods used in our study, and their geographical locations are presented in Figure 1. According to the corresponding urban population in 2018 and the classification standards of [25] and [1], these 20 cities can be divided into six categories, and the detailed information is shown in Table 1. The selection of cities for evaluation is in line with randomness.
expansion and population growth in Africa coordinated; (3) situation and of other land types occupied by urban land in Africa during expansion.
To increase the comparability and consistency of urbanization study in time and space, this study proposes a population density-based thresholdi generate urban land and urban population. The rest of this study is organize Section 2 introduces the study area and data source; Section 3 describes de odologies adopted in this study; Section 4 presents and analyzes the results the discussion and conclusion in Sections 5 and 6, respectively.

Study Area
The total area of Africa is about 32 million km², which includes 55 co gions, with a total of 751 provinces. In 2018, the urbanization level of Africa w 547.6 million people living in cities [1]. Twenty African cities were selected feasibility of the methods used in our study, and their geographical locat sented in Figure 1. According to the corresponding urban population in 2018 sification standards of [25] and [1], these 20 cities can be divided into six ca the detailed information is shown in Table 1. The selection of cities for evalua with randomness.

Data Source
LandScan is a time-series community standard for spatially disaggregated global population data produced by the Oak Ridge National Laboratory (ORNL) [26]. The spatial resolution of LandScan is approximately 1 km. LandScan is the most commonly used global population distribution data, as it has the longest time series and the finest resolution. In this study, LandScan data spanning from 2001 to 2019 were generated from the website https://landscan.ornl.gov/ (accessed on 15 May 2021).
The global public land cover dataset MCD12Q1 with a spatial resolution of 500 m from the Moderate Resolution Imaging Spectroradiometer (MODIS) products was used to study the urban land-use change [27]. The time series of this dataset spans from 2001 to 2019, which can be obtained from the EARTHDATA website https://search.earthdata.nasa.gov/ (accessed on 20 May 2021).
The ground truth land-use data used in this study were derived from The Atlas of Urban Expansion (2016 Edition) (http://atlasofurbanexpansion.org/, accessed on 20 May 2021). This dataset selected 200 cities around the world and produced land-use maps circa the years 1990, 2000, and 2014. Landsat imagery, with a spatial resolution of 30 m, was classified into seven classes using the unsupervised classification method, which includes urban built-up [28].
Moreover, the administrative boundary data was collected from the Database of Global Administrative Areas (GADM) (https://gadm.org/, accessed on 8 June 2021). Additionally, the statistical urban population data in Africa were obtained from the UN (https://population.un.org/wup/, accessed on 10 June 2021). In addition, the nighttime light data used in this study was from harmonized global nighttime light dataset produced by Li Xuecao (2020) [29].

Methods
Four main procedures were undertaken to identify spatiotemporal characteristics of urban land expansion and population growth in Africa. Firstly, urban land and urban population were generated using LandScan data with thresholding method and overlay analysis. Secondly, we assessed the accuracy of urban land and urban population obtained from the first step. Thirdly, elasticity coefficient was adopted to detect the coupling development between urban land expansion and population growth. At last, the reclassified MCD12Q1 land-use data were integrated with the acquired urban land to study the urban land-use change ( Figure 2

Identification of Urban Land and Urban Population
Population and population density are common indicators for dividing urban and rural areas [30,31]. As stated above, 1500 people/km² was selected as the threshold to generate spatial layers of urban land and urban population. The steps were as follows: (1) we

Identification of Urban Land and Urban Population
Population and population density are common indicators for dividing urban and rural areas [30,31]. As stated above, 1500 people/km 2 was selected as the threshold to generate spatial layers of urban land and urban population. The steps were as follows: (1) we considered grid cells of LandScan with >1500 people/km 2 as urban land and the images were converted to binary images; (2) the binary images were post-processed using the post-classification sieve function within Envi 5.3.1 (© 2015 Exelis) to eliminate the isolated classified pixels; (3) the population of the corresponding urban area was regarded as urban population. As a result, the spatial layers of urban land and urban population were obtained.

Accuracy Assessment of Urban Population and Urban Land Generated from LandScan
We verified the obtained data to confirm the feasibility of collecting urban population and urban land from LandScan. A linear regression model was applied to the generated urban population and statistical urban population from the UN. As for the generated urban land, we verified it qualitatively and quantitatively. From a qualitative point of view, we visually compared the generated five African cities' urban scope and shape with the existing land-use map and gradient nighttime light level map, in which the calculation of gradient nighttime light types can be referred to Jiang (2021) [32]. From the perspective of quantitative, we selected the generated urban land of 20 African cities to calculate its user's accuracy, producer's accuracy, and kappa coefficient with the ground truth data from The Atlas of Urban Expansion (2016 Edition).

Coupling Development of Urban Land Expansion and Population Growth
To depict the human-land relationship in the process of urbanization, the elasticity coefficient was adopted to partition the types of coupling development between urban land expansion and population growth [33]. The elasticity coefficient can be calculated as follows: where EC denotes the elasticity coefficient between urban land expansion and population growth, PR is the average annual change rate of the urban population, whereas LR is the average annual change rate of urban land. According to the increase and decrease changes of PR and LR and the numerical comparison, the coupling relationship between urban population and urban land can be divided into six types. The illustration can be seen in Figure 3 and Table 2. Taking type I as an example, PR and LR are positive at the same time, and the former is greater, indicating that urban population and urban land increase at the same time and the growth of urban population is faster than that of urban land. Based on the numerical comparison between PR and LR, the six coupling types can be divided into two categories: coordinated and uncoordinated development. Detailed information is shown in Table 2. where EC denotes the elasticity coefficient between urban land expansion and population growth, PR is the average annual change rate of the urban population, whereas LR is the average annual change rate of urban land. According to the increase and decrease changes of PR and LR and the numerical comparison, the coupling relationship between urban population and urban land can be divided into six types. The illustration can be seen in Figure 3 and Table 2. Taking type I as an example, PR and LR are positive at the same time, and the former is greater, indicating that urban population and urban land increase at the same time and the growth of urban population is faster than that of urban land. Based on the numerical comparison between PR and LR, the six coupling types can be divided into two categories: coordinated and uncoordinated development. Detailed information is shown in Table 2.

Relationship Types Criteria Explanation
Coordinated development The growth rate of population is greater than that of land, which means the intensive degree of land use is higher, and the relationship between population and land tends to be coordinated.
The population growth rate is lower than that of land, which may lead to extensive use of land resources and an imbalance between population and land.

Land use Change
There are five classification schemes included in the land cover products MCD12Q1, of which the International Geosphere-Biosphere Programme (IGBP) is the most widely used [34]. In this study, the IGBP classification scheme was adopted to extract land-use types, which were reclassified into seven types, namely, forest, grassland, built-up land, cropland, barren land, water, and other. To study the land-use change of urban land, we integrated the urban land generated above with these seven types of land use. Taking Lagos as an example, urban land generated from LandScan mainly overlapped on the builtup land from MCD12Q1 ( Figure 4). Using the simple dichotomous division method [10], the built-up land that did not coincide with urban land was regarded as rural land. As a result, land-use maps comprised eight types, including forest, grassland, urban land, rural land, cropland, barren land, water, and other. At last, through the transfer matrix of urban land-use change, we can get the situation of other land-use types occupied by urban land.

Land use Change
There are five classification schemes included in the land cover produc of which the International Geosphere-Biosphere Programme (IGBP) is the used [34]. In this study, the IGBP classification scheme was adopted to ex types, which were reclassified into seven types, namely, forest, grassland, cropland, barren land, water, and other. To study the land-use change of u integrated the urban land generated above with these seven types of lan Lagos as an example, urban land generated from LandScan mainly overl built-up land from MCD12Q1 ( Figure 4). Using the simple dichotomous div [10], the built-up land that did not coincide with urban land was regarded As a result, land-use maps comprised eight types, including forest, grasslan rural land, cropland, barren land, water, and other. At last, through the tran urban land-use change, we can get the situation of other land-use types occu land.

Verifying the Urban Population and Urban Land Generated from LandScan
To verify the accuracy of the urban population generated from LandScan in this study, statistical urban population from the UN was used for comparison. Linear regression analysis was applied to the two datasets and the comparison result is shown in Figure 5. The results show that the urban population obtained from LandScan has a linear correlation with the statistical urban population of UN, with the R 2 reaching 0.9871. At the same time, the linear correlation coefficient is 0.8903, which is close to 1. As a result, the urban population generated from LandScan has a strong relationship with that from the UN.

5.
The results show that the urban population obtained from LandScan ha lation with the statistical urban population of UN, with the R 2 reaching 0.98 time, the linear correlation coefficient is 0.8903, which is close to 1. As a re population generated from LandScan has a strong relationship with that f  (Figure 6c). It is obvious that the spatial distribution generated from LandScan is similar to the urban built-up from The Atlas o sion (2016 Edition) and built-up land from MCD12Q1. As for the distributi light types, the urban land generated from LandScan was not completely high nighttime light, except for Cairo. However, the shape of high nightti bution is very similar to that of urban land. This also supports demograph urban transformation in Sub-Saharan Africa can be better interpreted as phenomenon rather than a strictly economic one [35,36]. For urban land generated from LandScan, we verified it from both qualitative and quantitative perspectives.  (Figure 6c). It is obvious that the spatial distribution of urban land generated from LandScan is similar to the urban built-up from The Atlas of Urban Expansion (2016 Edition) and built-up land from MCD12Q1. As for the distribution of nighttime light types, the urban land generated from LandScan was not completely covered by the high nighttime light, except for Cairo. However, the shape of high nighttime light distribution is very similar to that of urban land. This also supports demographers' views that urban transformation in Sub-Saharan Africa can be better interpreted as a demographic phenomenon rather than a strictly economic one [35,36].
To better verify the urban land generated from LandScan, 20 cities were selected to conduct the accuracy assessment. The accuracy was evaluated by the ground true land-use map collected from The Atlas of Urban Expansion (2016 Edition). The kappa coefficient, producer's accuracy, and user's accuracy of 20 cities are shown in Figure 7. The kappa coefficient ranges from 0.66 to 0.99. The user's accuracy in all cities is higher than 60%, and some even reach 96%, except for Gombe. Meanwhile, the producer's accuracy of the 19 cities is higher than 50%, and the highest is 86%. The derived urban center in LandScan is smaller than the ground truth data, which includes most artificial impervious areas and associated infrastructures, parks, lakes, and so on within the urban boundary [37]. Additionally, the spatial resolution of LandScan is 1 km, which is much coarser than that of the ground truth data (30 m). As a result, we think the method used in this study is reasonable and can be effectively implemented. Int. J. Geo-Inf. 2021, 10, x FOR PEER REVIEW 9 of 21 To better verify the urban land generated from LandScan, 20 cities were selected to conduct the accuracy assessment. The accuracy was evaluated by the ground true landuse map collected from The Atlas of Urban Expansion (2016 Edition). The kappa coefficient, producer's accuracy, and user's accuracy of 20 cities are shown in Figure 7. The kappa coefficient ranges from 0.66 to 0.99. The user's accuracy in all cities is higher than 60%, and some even reach 96%, except for Gombe. Meanwhile, the producer's accuracy of the 19 cities is higher than 50%, and the highest is 86%. The derived urban center in LandScan is smaller than the ground truth data, which includes most artificial impervious areas and associated infrastructures, parks, lakes, and so on within the urban boundary [37]. Additionally, the spatial resolution of LandScan is 1 km, which is much coarser than that of the ground truth data (30 m). As a result, we think the method used in this study is reasonable and can be effectively implemented. 10, x FOR PEER REVIEW 10 of Figure 7. The producer's and user's accuracy of 20 cities in ascending order of the kappa coefficient.

Time-Series Changes of Urban Population and Urban Land in Africa
Urban population and urban land showed a synchronous growth trend in Africa from 2001 to 2019 (Figure 8). The urban land increased from 37,160 to 104,647 km 2 , with an average annual growth rate of 5.92%. At the same time, the urban population increased from 233.73 million to 554.33 million, with an average annual growth rate of 4.91%. The growth rate of urban land was higher than that of the urban population on the whole. Taking 2001 as the benchmark, we estimated the year-on-year growth rate of the urban population and urban land in Africa, and world urban population, which can be seen in Figure 8. The growth rate of the urban population in Africa has increased steadily from 2001 to 2019, which was significantly higher than the average growth rate of the world's urban population. At the same time, the expansion of urban land fluctuated greatly. The growth rate of urban land area exceeded that of the urban population in 2003.

Spatial Distribution of Urban Land Expansion and Population Growth in Africa in Different Periods
In order to explore the spatial development dynamics

Spatial Distribution of Urban Land Expansion and Population Growth in Africa in Different Periods
In order to explore the spatial development dynamics and distribution pattern of As for the spatial distribution of urban land growth in Africa during the different periods, there were similar spatial distribution characteristics, which are demonstrated in Figure 9. The growth of urban land was mainly concentrated in countries such as Nigeria, Egypt, the Democratic Republic of the Congo, and Ethiopia. Urban land has been growing rapidly in all provinces of some countries like Nigeria, the Democratic Republic of the Congo, and Kenya. Meanwhile, urban land growth in some countries is concentrated in some areas. For example, urban land growth in South Africa is mainly concentrated in coastal areas, Gauteng, and North West provinces. Nigeria has always been the leading country of urban land change during the study period. In addition, the proportion of urban land change in Nigeria to the total urban land change in Africa was getting higher and higher, with 17.73%, 20.85%, and 26.43% from 2001 to 2007, 2007-2013, 2013-2019, which was much higher than that of other African countries (Table 3). At the same time, the urban land expansion of Ethiopia and the Democratic Republic of the Congo was getting faster and faster. Together with Nigeria, they constituted the top three countries in Africa with the largest urban land expansion from 2001 to 2019. The top 10 countries with the largest urban land expansion account for more than 60% of the total urban land expansion in Africa during the study periods. As for the spatial distribution of urban land growth in Africa during the different periods, there were similar spatial distribution characteristics, which are demonstrated in Figure  9. The growth of urban land was mainly concentrated in countries such as Nigeria, Egypt, the Democratic Republic of the Congo, and Ethiopia. Urban land has been growing rapidly in all provinces of some countries like Nigeria, the Democratic Republic of the Congo, and Kenya. Meanwhile, urban land growth in some countries is concentrated in some areas. For example, urban land growth in South Africa is mainly concentrated in coastal areas, Gauteng, and North West provinces. Nigeria has always been the leading country of urban land change during the study period. In addition, the proportion of urban land  Table 4). The urban population growth in Ethiopia has also been relatively rapid, except in 2007-2013. Ethiopia's urban population growth was mainly concentrated in Addis Ababa, Amhara, and Southern Nations, Nationalities and People ( Figure 10). During the study period, the top 10 countries with the largest urban population growth accounted for more than 60% of the total urban population growth in Africa. nt. J. Geo-Inf. 2021, 10, x FOR PEER REVIEW 14 of 21

Spatiotemporal Characteristics of Coupling Development between Urban Land and Urban Population
From 2001 to 2019, the coupling relationship index of urban population and urban land was 0.76. In general, the growth rate of urban land was greater than that of the urban population in Africa, and the type of coupling relationship was in type VI. Figure 11 shows the spatial change of the coordination relationship between urban land and urban population in different provinces of Africa in different periods. As shown in Figure 11, the urban development in African provinces was mainly in type I and type VI. Nigeria has always been dominated by type VI. The urban development of the Democratic Republic of the Congo was mainly in type VI except from 2013 to 2019. Meanwhile, the urban development of Ethiopia was more complex. As the most urbanized area in Ethiopia, Addis Ababa was mainly dominated by type VI, except 2013-2019. From 2001 to 2019, the EC values of Nigeria, the Democratic Republic of the Congo, and Ethiopia were 0.90, 0.49, and 0.82, respectively, which means that the growth rate of the urban population in these countries was lower than that of urban land, and urban land showed an extensive growth.
We counted the proportion of urban development types in different periods, and the results are displayed in Figure 12. According to the change of the proportion of different types of cities over time, type VI was increasing, while type I was decreasing. The total proportion of uncoordinated development types (IV, V, and VI) was getting higher and higher. Therefore, the balance of urban land expansion and population growth in Africa is not optimistic. Among them, type VI represents the synchronous growth of urban population and urban land, and the growth of urban land is faster than that of population, which indicates that most provinces are still in the stage of rapid urbanization, and it is speculated that the imbalance of urban land expansion and population growth will continue or even intensify in the future.
Ethiopia, Addis Ababa was mainly dominated by type Ⅵ, except 2013-2019. From 2001 to 2019, the EC values of Nigeria, the Democratic Republic of the Congo, and Ethiopia were 0.90, 0.49, and 0.82, respectively, which means that the growth rate of the urban population in these countries was lower than that of urban land, and urban land showed an extensive growth. We counted the proportion of urban development types in different periods, and the results are displayed in Figure 12. According to the change of the proportion of different types of cities over time, type Ⅵ was increasing, while type Ⅰ was decreasing. The total proportion of uncoordinated development types (Ⅳ, Ⅴ, and Ⅵ) was getting higher and higher. Therefore, the balance of urban land expansion and population growth in Africa is not optimistic. Among them, type VI represents the synchronous growth of urban population and urban land, and the growth of urban land is faster than that of population, which indicates that most provinces are still in the stage of rapid urbanization, and it is speculated that the imbalance of urban land expansion and population growth will continue or even intensify in the future.

Characteristics of Other Land Types Occupied by Urban Expansion
According to the above analysis, the land expansion rate of African cities is higher than that of the urban population, which may lead to extensive use of land resources. Therefore, it is necessary to monitor the situation of other land types occupied by African

Characteristics of Other Land Types Occupied by Urban Expansion
According to the above analysis, the land expansion rate of African cities is higher than that of the urban population, which may lead to extensive use of land resources. Therefore, it is necessary to monitor the situation of other land types occupied by African cities. Figure 13 is a Pareto diagram of other land types occupied by urban expansion in different periods. Except for urban land, land use in Africa is divided into forest, grassland, rural land, cropland, barren land, water, and other. As can be seen from Figure 13, forest, grassland, rural land, and cropland are the four types that have been occupied mostly by urban expansion, accounting for more than 95% in different periods. From 2001 to 2019, among the types of land occupied by urban expansion, cropland was occupied up to 29%, followed by grassland, rural land, and forest (Figure 13d). During the study period, the proportion of cropland occupied by urban expansion has gradually increased and has become the most occupied land-use type (Figure 13a-c). At the same time, the proportion of grassland and forest was gradually increasing too, but the proportion of rural land occupied by urban land has dropped sharply. To sum up, with the rapid expansion of urban land in Africa, more and more cropland, grassland and forest have been invaded.

Discussion
In this study, we propose a population density-based thresholding method to iden tify urban land and urban population. This is a simple, fast, and effective method to iden tify urban area, which can provide continuous and comparable urban land and urba population data for different regions. Using this method, we can get the data of urba population and urban land in African countries and conduct further comparative researc Findings show that Africa has experienced considerable growth in urban land an urban population from 2001 to 2019. Urban land averagely increased by 5.92% per annum while urban population increased averagely by about 4.91% per annum during 2001-2019 The speed of urban land expansion is relatively faster than the growth of urban popula tion, which is consistent with previous studies [17,38]. Nigeria, the Democratic Republi

Discussion
In this study, we propose a population density-based thresholding method to identify urban land and urban population. This is a simple, fast, and effective method to identify urban area, which can provide continuous and comparable urban land and urban population data for different regions. Using this method, we can get the data of urban population and urban land in African countries and conduct further comparative research.
Findings show that Africa has experienced considerable growth in urban land and urban population from 2001 to 2019. Urban land averagely increased by 5.92% per annum, while urban population increased averagely by about 4.91% per annum during 2001-2019. The speed of urban land expansion is relatively faster than the growth of urban population, which is consistent with previous studies [17,38]. Nigeria, the Democratic Republic of the Congo, and Ethiopia are the three countries with the fastest urban land expansion and urban population growth in Africa. This is mainly due to the reason that Africa's urbanization is commonly driven by demography [39][40][41]. In 2019, Nigeria was the most populous country in Africa with a population of 201 million, followed by Ethiopia (70 million), Egypt (68 million), and the Democratic Republic of the Congo (51 million) [42]. As for the average annual growth rate of population from 2001 to 2019, the Democratic Republic of the Congo was the highest with a growth rate of 3.8%, followed by Nigeria (3.7%), Ethiopia (3.3%), and Egypt (2.6%) [42,43].
Whether from the perspective of Africa as a whole or most provinces of African countries, the growth rate of urban land is greater than that of urban population, which is in line with Angel's prediction that urban land cover in sub-Saharan Africa will expand at the fastest speed during 2000-2050 [17]. From a global perspective, urban growth brought about by urbanization is reflected in the outward expansion, inward compression, or upward high-rise buildings [44,45]. However, the spatial expansion of African cities is characterized by large-scale and uncontrollable, which is technically known as urban sprawl. Studies on African urbanization show that urban sprawl leads to unsustainable land development and often engulfs the surrounding urban areas, transforming nonurban land, especially agricultural land, into urban land [46,47]. It is consistent with the results of this study that cropland was the most land-use type occupied by urban land. Food insecurity has been a long-standing problem in Africa. About 257 million people in Africa are in malnutrition, accounting for nearly one-third of the world's malnutrition population [48]. Additionally, a quarter of Africa's population is still starving. Food insecurity in Africa has been a great challenge to the achievement of the Second Sustainable Development Goal (SDG2) that living in a world with no hunger proposed by the 2030 Agenda. There is no doubt that the disordered urban expansion in Africa will hinder the realization of SDG2 even more.
The future development of cities and the subsequent occupation of land and natural resources will determine whether we can move towards an environmentally sustainable future. The unplanned or improperly managed urban expansion will lead to rapid sprawl [49], pollution [50], and environmental degradation [51], as well as unsustainable production and consumption patterns [52]. Unfortunately, many urban areas in Africa have been undergoing dense, unplanned, and unsustainable urbanization, which is considered to be eroding the socio-economic and environmental benefits associated with urbanization and sustainable development [41]. Africa's forests are mainly distributed in the equatorial region, stretching from Sierra Leone to Victoria Great Lakes region [53], including the Democratic Republic of the Congo, parts of Nigeria, and Ethiopia, which are undergoing rapid urbanization. As shown in the results of this article, encroachment of forests by the extensive urban land use in Africa is also obvious. Such urban expansion has the potential to destroy the habitats of key biodiversity hotspots and lead to carbon emissions associated with tropical deforestation and land use change.
Sustainable development is still one of the most advocated development concepts in the world. However, there are still limited signs of progress in achieving sustainable development in Africa. Rapid and unplanned urbanization is a major threat to the sustainable development of Africa. This study provides a simple and easy way to monitor the urban population growth and urban land expansion in African cities. This will help to fully understand the complex urbanization process in Africa and formulate better urban development policies. With the rapid increase in the population of African cities, more and more related issues will emerge. Therefore, more research is needed to focus on the process of urbanization in Africa.

Conclusions
Urbanization is a complex geographical process with the simultaneous growth of population and land, closely related to social economy and environment, etc. Africa is experiencing rapid urbanization, and some problems related to rapid urbanization are emerging. In this study, we propose a population density-based thresholding method to generate urban land and urban population data in Africa, which can be applied to further study the characteristics of Africa's urbanization process. Our method is proved to be feasible. The results show that from 2001 to 2019, the average annual growth rate of urban land in Africa was 5.92%, which was greater than that of the urban population (4.91%). Moreover, the growth rate of the urban population in African is greater than that of the world, which is consistent with previous research conclusions. The regional analysis of urban population and urban land change in Africa shows that the top three countries with the fastest urbanization process in Africa are Nigeria, the Democratic Republic of the Congo, and Ethiopia. As for the coordinated relationship between urban population growth and urban land expansion, results show that the total proportion of uncoordinated development types (IV, V, and VI) was getting higher, which means that the urban land use in Africa tends to be extensive, and land-use efficiency is reduced. With the rapid urbanization in Africa and extensive urban land use, more and more cropland, grassland, and forest are occupied. The encroachment of cropland may exacerbate food security in Africa. Additionally, with the expansion of cities, the deforestation of forests will cause more carbon emissions and will affect biodiversity, which will have an impact on the environment.
According to the results of this paper, the extensive land use induced by rapid urbanization process in Africa is likely to damage the environmental benefits of Africa. As the second-largest continent in the world and home to a population of 1.29 billion, sustainable urbanization is very important for Africa and even the world. As a result, it is very important to monitor urban dynamics in Africa and formulate sustainable urban development policies. More efforts are needed to achieve the goal of sustainable urban environment in Africa.