Urbanization is one of the leading drivers of Land Cover Land Use Change (LCLUC) globally, and African countries are at the forefront of urban expansion trends. Predictions indicate that in Africa, urban populations will double by 2050 outnumbering people living in cities around the world (United Nations DESA, 2018; African Union, 2024). By 2030, global urban expansion will increase by 1.2 million km² of new built-up areas (World Bank Group, 2023) exhibiting a 12-fold increase in urban land area (Fourchard, 2011; Angel, 2023). These predictions indicate that a majority of LCLUC globally will transform both natural and managed landscapes to residential neighborhoods, informal settlements, and intensely modified cityscapes. Policy makers and planners need to know more about the quality of different urban areas (e.g., access to services, infrastructure), and LCLUC scientists need to be more explicit in mapping the development of multifunctional land uses of cities over time to better inform planning efforts (McHale et al., 2015).

Africa is leading the way in rapid urbanization in small and medium sized cities. These secondary cities are the new vital economic centers for large regions of the world, often providing safer, less polluted refuge for urban residents seeking a higher quality of life (Roberts, 2014). There is growing evidence that with the development of secondary cities, this urbanization process is divergent from what we have experienced in the Global North historically (McHale et al., 2013; Nagendra et al., 2018). Secondary cities are the fastest growing urban areas, and especially in lower-and middle-income countries (LMIC), are experiencing the negative consequences of unplanned development including managing urbanization, sustainable development, investment for job creation and educational resources, as well as ensuring access to basic services across the range of economic levels. Furthermore, these cities are unique environments that generally have limited data and information on infrastructure, land tenure, and other planning resources (Roberts, 2014; United Nations HABITAT, 2021a, 2021b; Laituri and Sternlieb, 2018), while also largely neglected in remote sensing studies of urban land change (Reba and Seto, 2020). A better understanding of the impacts from LCLUC in rapidly expanding urban centers, such as secondary cities, can enhance efforts to plan for more sustainable development, mitigate long term influences on social and ecosystem services, and assess the health of urban socio-ecological function.

Multi-resolution spatial datasets that are widely available and consistently produced can be used for guiding sustainable urban management to improve the overall quality of urban environments. Such improved datasets and approaches can assist policy makers and planners to assess our progress towards internationally recognized Sustainable Development Goals (SDGs) developed by the United Nations as a universal call to action to end poverty, protect the planet, and ensure that by 2030 all people enjoy peace and prosperity. As the variety, quantity, and quality of Earth Observation platforms increase and computing capabilities and processing techniques rapidly advance, the value of monitoring resources in support of SDG indicators can be enhanced through applied research in the development and testing of improved spatial products and planning-relevant toolsets.

SDG 11 focuses on creating inclusive, safe, and resilient cities and human settlements. This goal tracks a suite of indicators to support more sustainable urban planning using Earth Observation Satellites and other widely available data sources. Specifically, we examine SDG Indicator 11.3.1. This indicator focuses on urban land expansion through a comparison of developed land use areas and population growth rates requiring data at moderate to high resolution capturing where and how cities are expanding to evaluate local urban change (Table 1). Using LCLUC approaches we can identify areas of rapid urban growth—hotspots—that exhibit urban patterns reflecting socio-ecological implications and define indicators of equitable growth (i.e., access to services, availability of green space, encroachment on rural villages). Hotspots may include locations of leap-frog development (new urban areas discontinuous from urban areas but functionally linked), urban corridors (urban development adjacent to major roadways), infill (new development in existing urban areas), extension (development at edge of urban areas), and inclusion (engulfing outlying urban areas, rural urbanization) (United Nations HABITAT, 2021a, 2021b).

The study of urbanization and its consequences has traditionally occurred at quite different scales and resolutions: (1) regional scales (including national and global), coarse (> 30 m pixels) and moderate (30 m pixels) resolution urbanization patterns, and (2) fine scale (city, neighborhood, parcel), very high resolution (< 3 m pixels) studies of urban heterogeneity. The regional scale and coarse/moderate resolution approaches are good for understanding the dynamic boundaries of urban growth, and how urban land expands into other surrounding land uses over time (Meentemeyer et al., 2013), while the strengths of the fine scale method enable scientists to analyze urban heterogeneity and accurately view all land covers within each land use with high spatial precision (Cadenasso et al., 2007). Studies that implement the first type of analysis are termed “red blob urbanization” analyses, where a series of maps over time presents a single urban class expanding across the landscape (Feng et al., 2018; Xu et al., 2016). Studies implementing fine scale methods attempt to understand landscape quality, human-environmental interactions, and tradeoffs in ecosystem services (Wang et al., 2022; Guan et al., 2023) but are often restricted in space and time due to limited access to Earth Observation data that support analyses at this resolution consistently across broad extents.

Urban mappable units derived from image processing and analysis of pixelated surfaces are based upon classification methods and urban mapping concepts (e.g., urban gradient, boundary, and form) (Figure 1). The configuration and composition of urban mappable units enable calculations of landscape metrics such as pattern, cluster, fragmentation, and urban extent (Pontius and Cheuk, 2006). Higher spatial resolution of urban pixels increases accuracy of land use capturing urban heterogeneity (Murtaza and Romshoo, 2014). The classification of urban mappable units yields added information to identify urban expansion and related hotspots to inform land planning strategies.

Our overarching goal was to interrogate traditional LCLUC approaches and show where and how available remotely sensed data and geospatial applications can advance a cross-scale “pixels to people” approach to define and clarify urban change patterns and hot spots of urban growth in secondary cities. Using a two-tiered multi-resolution case study analysis of urban change hotspots in three Africa counties (Ethiopia, Nigeria, and South Africa) and example secondary cities (Mekelle, Benin City, Polokwane, respectively), we demonstrate how LCLUC classification methods drive the process of identifying and defining mappable urban units. We explore the conceptual basis of urban change mapping, applying SDG 11.3.1 to demonstrate how elements such as urban extent, dynamic urban boundaries, urban clusters, and land consumption rates characterize urban change. We observe that monitoring and management efforts utilizing remote sensing products and geospatial approaches require a nuanced depiction of urbanization, one that dissects the homogenous, amorphous “red blob” representing the urban landscape to better capture urbanization driven land cover changes (Cadenasso et al., 2007; Reba and Seto, 2020; Zhu et al., 2019). Finally, we discuss how these results can help scientists and policy makers reflect on equitable processes for using integrated datasets and information in the Global South to inform policy regarding high-impact urban change hotspots.

To highlight the potential value of integrated multi-resolution urban mapping product suites for assessing impacts of urbanization-driven LCLUC hotspots, we present a two-tiered land imaging approach in Ethiopia, Nigeria, and South Africa. Our study countries and cities exhibit a range of ecosystems and societal dynamics that demonstrate the flexibility of this method. This multi-tiered approach provides: 1) insights from SDG 11.3.1 calculation about urban expansion (land consumption) and urban land use change patterns (where and type); and 2) the application of landscape metrics to observe urban heterogeneity by assessing the relationship between building density to green space. We selected three representative secondary cities (Mekelle, Ethiopia; Benin City, Nigeria; Polokwane, South Africa) of urban expansion hotspots according to SDG 11.3.1 to present examples of the indicator outputs and to examine urban land use change patterns (Figure 2). As a synthesis paper, we summarize this methodology and reference Cardenas-Ritzert et al. (2024a) and Shah Heydari et al. (2024), which provide detailed methods and procedures, summarized as follows (Figure 3):

We found that our multi-tiered LCLUC approach supported the identification of hotspots of urban expansion and identify urban agglomerations across the 5-year study period (2016–2020). Shah Heydari et al. (2024) examined the land use patterns in the three selected countries (Tier 1) and observed (1) the urban area was quite small in comparison to country size (~1.5%) and (2) urban change in the three countries were generally the result of the transformation of agricultural and rangelands to urban land use. While land use classification provides an overview of where urban areas are located, the small areas of cities do not adequately reflect the impact of urbanization (i.e., the urban footprint) and the supply chains that support them. The Tier 2 urban land cover analysis using high resolution remotely sensed data (and Tier 1 products of urban delineation) enables the assessment of land cover changes as well as the application of landscape metrics (i.e., building density and distance between green spaces) to begin to address urban heterogeneity. Landscape metrics that describe spatial distribution patterns, intermixing of land cover classes, aggregation of buildings, and distance between vegetation land cover characterize urban form and configuration enabling comparisons between the three countries. Landscape pattern metrics revealed that Nigerian and South African cities generally have higher building densities which may indicate greater intermixing of land cover, while Ethiopia tends towards less dense and more interspersed land cover configurations and lower building densities.

Using overlay analysis of dynamic urban agglomerations (2016 and 2020) and administrative boundaries we demonstrate mapping urban concepts (Figure 4). Frames A, B, C includes the boundaries for Mekelle, Ethiopia that includes dynamic urban boundaries (A—2016; B—2020) and the urban woreda or district administrative boundary. Similarly, Frames D, E, and F show the Polokwane, South Africa boundaries and Frames G, I, J show the Benin City, Nigeria boundaries. Frames C, F, and J overlay the 2016 and 2020 boundary with the administrative boundary for all study site cities. All cities have clearly defined urban clusters (core and non-core) for 2016 and 2020 with an increase in the size and number of urban clusters in 2020 indicating leapfrog development. The urban woreda for Mekelle is the unit of governance, and this overlay analysis reveals how the governance unit is outpaced by unplanned urban growth. In contrast the administrative boundaries for both Polokwane and Benin City are much larger (blue area in Frames F and J, respectively), include multiple urban agglomerations, and can obscure the type of urban change that is occurring and can be overlooked by local planners who do not have access to these types of analyses.

The SDG 11.3.1 ratio varied across the study site countries; our study sites are representative of secondary cities with high SDG 11.3.1 ratios. All three cities exhibited increases in developed land use per capita over the study period with urban expansion occurring primarily in the form of extension (contiguous urban growth) and leapfrog development with limited infill development. Spatial analysis revealed four observations: 1) the dynamic nature of urban expansion exhibited by the increasing number of urban clusters; 2) the identification of urban hotspots can direct efforts for future planning and assessment of access to services when urban growth surpasses administrative boundaries; 3) the type of expansion (extension, leapfrog, infill) yields insights about patterns of connectivity, proximity, and contiguity and 4) landscape metrics can be calculated to assess urban heterogeneity. For example, both green space and building density changed between 2016 and 2020 exposing patterns of urban expansion with more green space (tall and short vegetation) on the periphery of each city, fewer patches of green space within the urban area, and increased building density (Table 3; Figure 5).

Mekelle, Ethiopia is an example of how mapping the urban agglomeration can contribute to urban planning. Figure 6 overlays the 2016 (blue) and 2020 (red) dynamic boundaries on OpenStreetMap. Platforms such as OpenStreetMap provide a means to fill in data gaps that include attributes of landscape features such as street names, whether roads are paved or not, building footprints, building materials, number of floors, and type of business or household. These results provide critical information for urban planners about where urban change is occurring but is limited in terms of understanding urban function and equitable growth—why, where, and what is located on the urban landscape.

We learned that how an urban pixel is defined informs decisions about how the spatial extent of the “red blob” is determined. Through our multi-resolution datasets and analyses we observe patterns of urban change (Cardenas-Ritzert et al., 2024b). The process of deriving SDG metrics requires decisions about land cover, land use, and urban functions where decisions about what type of land cover or use can have real impacts on those people living in the city. For example, the urban core has a different spatial structure than an informal settlement. However, the terms imply both land use and land cover. Green space lumps all kinds of socio-ecological functions together and obscures both function and form of the classified pixel. Our land cover classification products dissect urban agglomerations using landscape metrics that include identification of fine resolution urban objects, such as green spaces, and building density aiding research and planning efforts to better understand trends of urban change and configuration.

Applying SDG 11.3.1 is an example of data integration where population data and satellite imagery reveal land consumption patterns. Land consumption alludes to tradeoffs and synergies of ecosystem service provisioning. The type of land consumption (why LULUC is occurring) is the result of drivers beyond the scope of the SDG 11.3.1 ratio and reflect changing urban functions such as access to basic services (i.e., installation of new water infrastructure), types of development (i.e., leapfrog or infilling), and disruption of ecosystem services (i.e., changes to urban green space and parks).

Our approach developed an integrated mapping framework supporting multi-resolution datasets and openly accessible tools to conduct LCLUC assessments of urbanization. We developed an automated approach to define urban agglomerations within different countries in Africa. We highlighted three African countries as examples and identified rapidly growing secondary cities to assess hotspots of urban expansion to provide information for urban monitoring objectives. Our analysis synchronized two different tiers of multi-resolution datasets. Tier 1 used moderate resolution imagery to identify urban boundaries and land use drivers of change that informed the Tier 2 analysis using higher resolution to examine the “red blob” and extract metrics of urban heterogeneity. This approach can provide the basis for ongoing monitoring, can be applied to different years of imagery, and be applied to new locations.

Coupled with socio-economic metrics, our suite of analytical products assists in identifying drivers of land use change that fuel urban expansion. For example, economic indicators regarding public investment in Ethiopia in form of roads, infrastructure, and housing (World Bank Group, 2023) and service sector and industrial development employment opportunities in urban South Africa (World Bank Group, 2023) contribute to urban growth. Alternatively, Nigeria has the highest fertility rate (5.3) of the three countries where the natural increase outpaces migration fueling urban increase (Menashe-Oren and Bocquier, 2021). Using LCLUC approaches we can identify areas of urban expansion that reflect socio-ecological conditions and begin to define spatial indicators of equitable growth (i.e., access to services, availability of green space, encroachment on rural villages).

We are challenged by identifying urban functions that require both social and remote sensing in situ. The availability and structure of existing data (i.e., resolution of images; access to images) is incompatible with socio-ecological data (i.e., different scales of analysis, data formats). Indeed, we are dependent upon deriving surrogates for socio-ecological data where characteristics of the spectral bands of satellite images influence decisions in classification compounded by disciplinary and cultural backgrounds of the researchers. Approaches that are holistic and comprehensive are needed as we build more integrated datasets that become the basis for long term planning, scenario-building, and forecasting (Pickett et al., 2020). Tracking the potential for equitable distribution of social and ecological services within secondary cities is needed to support policy and sustainable planning goals informed by the SDGs.

Liverman et al. (1998) discussed this conundrum and identified the need for transdisciplinary action to facilitate complex analyses of urban change. The integration of human geography data (i.e., data about people) and satellite sensors create an information pixel useful to people for planning urban landscapes. For example, the Normalized Difference Vegetation Index (NDVI), the Modified Normalized Difference Water Index (MNDWI), and the bare soil index (BI) are all calculations derived from Landsat satellite imagery to identify built-up, or developed, area extents (Dolean et al., 2020). The Modified Socio-environmental Vulnerability Index (M-SEVI) and the Soil and Water Assessment Tool can be used to measure ecosystem services (Norman et al., 2012). These methods can identify tradeoffs in provisioning of ecosystem services with urban growth and LCLUC around, and within, urbanization hotspots.

Coupling remote sensing products with local information (e.g., private/public land) assists in defining urban function by creating more robust data. While remotely sensed nighttime-light and spectral information are helpful in delineating urban areas, cities are far more complicated; efforts to assess, monitor, and manage urban development are dependent upon understanding the urban socio-ecological function that creates healthy cities. And the use of very high-resolution imagery is challenging due to image availability, calibrations between sensors and cities, and high computational needs necessary for consistent products for planning and monitoring across space and time. Hot spots of urban change reveal varying urban function and the impacts on socio-ecological services. “Pixels to people” refers to how people experience and use the land on which they live and work—an aspect of urban form and function to derive from robust classification strategies (Figure 7). Metrics such as SDG 11.3.1 and SDG 11.7.1 require higher resolution information at the local scale to understand the changing nature of urban function. This requires very high-resolution imagery (≤3 m) for object-based analysis linked to social sensing techniques (i.e., participatory mapping) and local knowledge (i.e., mobile mapping on site, surveys) to capture urban function (Lui et al., 2015). Satellite imagery using different sensors (e.g., Planet Dove, Maxar WorldView) can enrich the analysis to measure development of impacts on ecosystems services, access to urban green space (SDG 11.7.1), and the distribution of social and ecological services—the nature and combination of urban functions across the cityscape.

Our study leveraging the two-tiered urban monitoring approach lays the groundwork for further urban analysis to inform decision making and assess links to socio-ecological functions in different cities using widely accessible data sources.

The development of a 2- or even 3-tiered methodological framework fusing remote sensing data sources at varying spatial and temporal resolutions and extents supports quantifying temporal trends in urban development and analyzing the societal and ecological impacts of rapid population growth within urban centers. Our data products and analyses directly inform SDG monitoring for policy development, ecological impact assessments, and provide insights into the patterns of access to ecological and social services. There is an urgent need for baseline information about urban expansion by world health organizations and national and local planning entities. As these methods and products related to LCLUC and fine resolution depictions of urban features are made publicly available, they serve as important steppingstones for future research efforts including forecast modeling under various urban planning and natural disaster scenarios, as well as informing future NASA Earth Observation missions.

We are poised to conduct the next step in our methodology which is to add a third focal spatial tier to our multi-tiered framework and to apply our frameworks to SDG 11.7.1, which focuses on the equitable distribution and access to urban green spaces, promoting access to urban ecosystem services through sustainable urban planning. Tier-1 and Tier-2 LCLUC data products enable future analyses coupled with very high resolution imagery (≤ 3 m resolution imagery) to extract green spaces and evaluate riparian/non-riparian area, surrounding building density, distance to edge of the urban area, connectivity between spaces, and other potential surrogates for characterizing the quality, multi-use potential, and access of the public areas (Cardenas-Ritzert et al., 2024b). We can quantify changes in these green spaces and other public areas within the 5-year time series in relation to expansions of other urban features and population growth. Developing multi-tiered frameworks may fill data gaps in support of policy and planning applications to use multi-resolution LCLUC classifications within other secondary cities to determine the equitable distribution of social and ecological services.

However, identifying where land consumption is taking place requires engagement with local representatives to ascertain why particular locations are hotspots. Such observations can capture functions which require local knowledge, ancillary data, and field verification to map the multifunctional land uses within a city. Urban extent, expansion and hotspots need to be harmonized between multiple perspectives to calibrate urban mapping concepts. For example, juxtaposing the administrative boundary of the city—the urban unit of management for city planners—with the dynamic urban extent captured from time series analysis using remotely sensed data, can identify rapidly changing areas for monitoring and management and characterize types of urban change. The drivers of urbanization are different than they were in the past, they are highly variable in different geographic locations, with correlations and relationships uncovering more questions about how to achieve sustainability than answers (Robinson, 2002; Henderson and Turner, 2020).

African urbanization, inclusive of new forms of urbanism due to economic influences (Koti, 2022) and the ongoing rural–urban transformation (Sakketa, 2023) coupled with a young growing labor pool (i.e., the median age is 19.5 years) (Worldometer, 2024) mean that urban transformation in Africa will be at the forefront of the urban century (Organization for Economic Co-operation and Development, 2022). Many socio-ecological services are not only spatial in nature, such as access to services, green spaces, and supply chains, but also demonstrate the (in)equitable process of urban expansion. Cross-scale and multi-resolution data about African cities is essential to understand ecological, economic, and social outcomes (Zhou et al., 2022). African cities are the economic centers of improved outcomes and higher standards of living that often outpace national averages where secondary cities are central to promoting economic and political integration (Institut Europeen de Cooperation et de Development, 2022). The challenges of African urban growth require planning, innovative policies, and sound data-driven decisions to ensure equitable and inclusive outcomes.

Multi-resolution data sets such as those presented within our study can be enhanced for additional analyses, such as the distribution of services, through the fusion with locally collected citizen scientist data sets. Workshops and training can aid in how to use these types of data enhancements through instruction of the data to stakeholders with a range of spatial data expertise. Receiving feedback as to how stakeholders perceive value in LCLUC maps and other NASA data products for incorporation into local and national level research and planning efforts provides insight into the best way to support and promote those efforts. By engaging local partners and conducting outreach we can help put NASA-driven data products into the hands of those who are working to promote more sustainable urban land planning and policy development. Next steps aim to enrich the definition and usefulness of hotspots of urban change to enable cohesive planning through improved and accessible remote sensing products and participatory approaches to include local communities to empower local governments and people (Espey et al., 2024).