Pathways to Industrialisation and Urban Spatial Outcomes: The Case of Addis Ababa - Bishoftu Economic Growth Corridor, Ethiopia

ABSTRACT Economic developments affect urban land-use changes. Using Lefebvre’s production of (urban) space as a conceptual framework, this study investigated the industrialisation process and its urban spatial outcomes in three cities (Galan, Dukem, and Bishoftu) along the southern outlet of Addis Ababa, Ethiopia’s capital. Landsat images were used to quantify the changing patterns of the three cities from the period 2006 to 2021. Interviews with key informants and document analysis were used to supplement these findings. The analysis reveals that in Addis Ababa-Bishoftu economic growth corridor, the unruly urban land-use changes have become a serious issue, challenging the sustainable development of cities. The study confirms that the proximity to Addis Ababa has contributed to the urbanisation process and urban corridor formation through investment expansion. This calls for a high level of expertise as well as integrated regional development for sustainable urban and industrial development.


Introduction
There is no all-encompassing formula for development (Li 2017).Nonetheless, industrialisation is a transf ormative process shaping countries' economic structure and impacting their social and institutional fabric (UNIDO 2020).Industrial development is considered to be the key to fostering economic growth and improving living conditions in urban areas (Nguyen et al. 2019).On the other hand, the industrialisation process has been a tale of two stories -the success of the Asian region (Akkemik 2009) while the sluggish in the African region (Shapiro 2007;Altenburg 2010;Mulu 2013;Morris and Fessehaie 2014;Oqubay 2015;Mbate 2017;UNECA 2017;Abreha et al. 2021).
The process of industrialisation must be the starting point for presenting and explaining the 'urban problematic.'Forthe last century, this process has been the driving force behind societal transformations (Lefebvre 1996).The development of cities is a continuous process.If this process is not appropriately managed, industrialisation coupled with burgeoning urbanisation will affect the liveability of cities.As cities become more industrialised, urban environments become more commodified (Zieleniec 2018;Serin et al. 2020), and a similar process has been shown in Ethiopia (Dadi et al. 2016;Giannecchini and Taylor 2018;Beyer et al. 2021;Koroso et al. 2021).As a result, land uses such as open spaces, agricultural lands, water bodies, and green areas have come under increasing human pressure in cities.The growing commodification of urban spaces and the subsequent changes or decline of urban land uses give rise to the issues of equity, social justice, and sustainability.The notion of the 'right to the city' came to address this issue (Lefebvre 1996;Harvey 2008).
The industrialisation process has received considerable theoretical and empirical attention.As some scholars argue (Shapiro 2007;Deng et al. 2008;Altenburg 2010;Mulu 2013;Oqubay 2015;Mbate 2017), industrialisation is indispensable if Africa is to transform its steady growth rates into socioeconomic development.The state of the industrialisation process and its subsequent rapid urban growth result in an opportunity to promote a dynamic self-sustaining urbanisation process (Schmidt et al. 2018).Industri alisation may result in urban anomalies, including social, physical, and economic stress (Koroso et al. 2021).Urbanisation that is not supported by economic growth but simply by population explosion may result in poverty, unemployment, hopelessness, criminality, and social evils (Tadesse and Imana 2006;Kassahun and Tiwari 2012).Planned urbanisation and indus trialisation could make cities conducive to the growth of jobs in the manufacturing and service sectors (UNECA 2017; Cheru and Aklilu 2020), through agglomeration of economies and could be drivers of long-term structural transformation in Africa (Dorosh and Thurlow 2013;Schmidt et al. 2018).Li (2017) points out that no one will be left behind if industrial development is inclusive and sustainable.
Industrial development and investment decisions could worsen a community's living conditions.Studies (Tian 2015;Nguyen et al. 2019;Dadi et al. 2020) indicate that industrial expansion negatively impacts local people's livelihoods by transforming agricultural land uses into industries.Urbanisation is a process that requires objective management and an institutional role (Özden and Enwere 2012).Thus, framing and guiding the growth and development of urban centres requires apt policy, strategies, and functional integration of urban centres.Studies by Buckley and Annaz (2009) and Duranton (2009) state that neglecting cities can impose high costs, and managing urbanisation is a crucial aspect of supporting growth.
In Ethiopia, the urban development agenda was given little attention until the mid-2000s because the Ethiopian People's Revolutionary Democratic Front (EPRDF) led government that came into power in 1991 had focused on Agricultural Development-Led Industrialization (ADLI) policy (Befekadu and Nega 2000;Abreha et al. 2021).However, the government ratified a national urban development policy in 2005 that aims to make urban centres become industrial hubs.The GoE has also ratified the 10 years perspective plan (2020)(2021)(2022)(2023)(2024)(2025)(2026)(2027)(2028)(2029)(2030) in 2020 that includes industry growth as one of its cornerstones (UN-Habitat 2014; NPC 2020).Such deliberate policies and strategies are needed to ensure the long-term viability of both cities and industries (UNECA 2017).Despite multiple interventions, the industry sector remains underdeveloped in Ethiopia, unable to contribute towards the desired structural transformation of the economy, and its development process has produced more challenges (Mulu 2013;Newman et al. 2016;Tesfaye 2017;Giannecchini and Taylor 2018).Optimizing industrialisation and urbanisation and mitigating the challenges require cautious policy responses (UNECA 2017; Tegegne and Edlam 2019;Cheru and Aklilu 2020).Besides, national policies must incorporate (urban) space as a means of a habitable built environment, not a place for capital accumulation (Lefebvre 1991(Lefebvre , 1996)).
In the absence of a stable, long-term, coherent national growth policy and an equally stable spatial policy, the patterns of industrial location would cause economic activities to concentrate in the largest metropolitan regions (Renaud 1981).As cities and investments grow, governing them becomes more complex (Satterthwaite 2007).Specifically, where there is no long-standing planning and plan implementation practice in Ethiopia, it is impossible to avoid haphazard urban developments (Mohamed et al. 2020;Alem 2021).This implies that cities are a centre of anomalies that require intervention (Watson 2009).The critical question is how far macro-level policies are integrated with territorial elements (Healey 1997;Goodfellow 2013;Pieterse 2019).From this vantage point, it would be wiser to monitor industrial development and its possible impact on urban physical spaces and built-up environments across the south outlet of Addis Ababa, Ethiopia's capital and the largest industrial hub.
Several studies (Dadi et al. 2016(Dadi et al. , 2020;;Terfa et al. 2017;Giannecchini and Taylor 2018;Koroso et al. 2021) noted that as industrialisation and urbanisation progress, the issues of negative externalities associated with them should be given considerable attention.Addressing such issues and fostering cities' attractiveness and competitiveness requires an appropriate policy and urban planning framework.In Ethiopia, studies (Dadi et al. 2016(Dadi et al. , 2020;;Giannecchini and Taylor 2018;Beyer et al. 2021;Beyer and Hagemann 2021;Koroso et al. 2021) attempted to demonstrate the industrialisation and urban spaces nexus.Yet, these studies fail to consider and articulate problems on an urban corridor scale.Besides, Elden (2004) states that analysis of (urban) space is not static but must also be historical.Land Use and Land Cover (LULC) dynamics with the industrialisation process and at the scale of an urban corridor remain unexplored.
Galan, Dukem, and Bishoftu have experienced significant socioeconomic and rapid land-use changes (Giannecchini and Taylor 2018;Dadi et al. 2020;Mohamed et al. 2020;Terfa et al. 2020;Koroso et al. 2021).The study gives empirical-based evidence that may contribute to how the industrialisation process shapes urban spatial patterns in the context of a centrally planned structure plan.Specifically, the study focuses on three cities (Galan, Dukem, and Bishoftu) that are located along the southern corridor and adjacent to Addis Ababa and examines how the existing plans guide industrial developments and locations in the urbanisation processes and how the progress of industrialisation affects the built environment and urban corridors using Lefebvre's production of (urban) space as a conceptual framework (Lefebvre 1991).The assessment included the dynamics of urban land use and land cover changes from 2006 to 2021.

A conceptual framework for the production of (urban) space
Theoretical conceptions have explained the interconnections of industrial developments and urban spatial consequences in many streams of literature, including the political economy of urban land use (McCarthy and Smit 1981), location, and urban economic theories (Alonso 1964), and new economic geography (Krugman 1998).Most are irrelevant to questions about the interplay between physical urban space and various spatial actors, to the spatiotemporal dynamics of urban form, nor do they provide details about the urban landuse change and social aspects (Alonso 1964).Yet, these theories have been influencing many policies towards cities and are used to explain cities' size, function, and sometimes location (Renaud 1981).
For this study, Lefebvre's theory of 'production space' (Lefebvre 1991) was chosen because it contributes to a more comprehensive understanding of the interplay between physical urban space and various spatial actors, as well as their conceptualisations of and claims to their built environment.Lefebvre's (1991) influential work, 'The Production of Space', has been dominant in reprioritizing space in interdisciplinary social science analyses.In various works on the city, space, and everyday life, he emphasises the importance of understanding space on several occasions.Elden (2004) explains that space production is a theme with explicit political aspects and is related to developing production systems within capitalism.
Several authors (Purcell 2002;Elden 2004;Merrifield 2006;Harvey 2008;Zieleniec 2018) recognised his work on space in the multidisciplinary sociospatial analysis of urban studies.For Lefebvre's (1991), space is both a product and a process of social activity that occurs within the structures and hierarchy of societies, increasingly subject to what he called 'an urban revolution' that continues to develop under capitalism.He identifies three necessary elements for the production of space: 'Spatial practices (perceived space), 'Representations of space' (conceived space), and 'Spaces of representations' (lived space).Elden ( 2004) defined the spatial practice as the physical form, real space, space that is generated and used, while representations of space are the instrumental space of social engineers and urban planners and the space that the experts interpret.The third element of space for Lefebvre's (1991) is defined as directly lived through its associated images and symbols, hence the space of 'inhabitants' and 'users'.The imagination seeks to change this dominated and passively experienced space.Zieleniec (2018) points out that to have actual knowledge of the production of space, one needs to understand the dynamic interaction and mutual interdependency between all three elements.Space is produced in a dynamic relationship between all three parts.There is, thus, a reciprocal relationship between the elements involved in its production.
The core idea behind Lefebvre's conceptual framework is that '(Social) space is a (social) product' and the production of space is a unity of the forces of production and their component elements are nature, labour, technology, knowledge; structures (property relations); superstructures (institutions and the state itself)' (Lefebvre 1991, p. 85).Zieleniec (2018) points out that understanding, and knowing Lefebvre's spatial thinking is appropriate and necessary for creating a more humane and inclusive sociospatial environment, in contrast to the increasing prioritisation of privatised and commodified public and social space.The linchpin of Lefebvre's (1991) (Social) space is a product of the social history itself.The social and political (state) forces are jeopardising this space.Lefebvre (1996) recognised that urban residents have the right to participate in the production of spaces and access, occupy, use, and create spaces that meet their needs.When Lefebvre first proposed the term 'right to the city,' he focused on the impacts of capital accumulation on the city (Lefebvre 1996).Later Harvey (2008) conceptualised the right to the city as far more than the individual liberty to access urban resources: it is a right to change ourselves by changing the city.As per Purcell (2002), Lefebvre's idea of the 'right to the city' underlines a fundamental power shift from capital to urban inhabitants.
Moreover, it is a common rather than an individual right since this transformation inevitably depends upon exercising a collective power to reshape urbanisation processes.Their focus is on how the neverending commodification of urban spaces undermined the city's role and its public spaces as sites.They called for the return of the predominance of the social functions of the city.Lefebvre (1996) argued that there is a double process with two aspects: industrialisation and urbanisation, growth and development, economic production, and social life.The two aspects of this inseparable process have unity, and yet it is a conflictual process.Historically there has been a violent clash between urban reality and industrial reality.Merrifield (2006) extended this argument as an inseparable and inexorable unity has been born, a terrible Janus-faced beauty coexisting, pitting industrial reality against urban reality, a mode of production against its built form.Lefebvre's spatial theory asks crucial questions about who produces, reproduces, owns, controls, and regulates urban space.Lefebvre's (1991) pointed out that the state and each of its constituent institutions call for spaces but spaces which they can then organise according to their specific requirements.Above all, Lefebvre (1996) asserts that we might be able to improve and democratise (urban) space if we understood how it was formed and shaped.As a result, Lefebvre considered the 'right' to make space vital to any 'right to the city,' and the governance, design, and planning of (urban) space enable us to create habitable spaces.
The process of urban spatial intervention is highly affected by a country's political process, which defines who gets what of the limited resources available (Lefebvre 1991;Merrifield 2006).Goodfellow (2013) noted that the constant political interference with city planners' activities affects the planning's effectiveness.Hence, the spatial decision process has no doubt been affecting the urbanisation and industrialisation trends.According to Beyer et al. (2021), largescale interventions constitute not only important vectors of spatial transformation and likely future urbanisation.The study cities have focused on recruiting investments and boosting their income.McCarthy and Smit (1981) noted that it is difficult to undermine the local political process that brought about marketinduced urban land-use changes.Tian et al. (2017) pointed out that the combination of state force and market forces contributed to the urban land-use changes in China.A study by Deng et al. (2008) confirmed the decisive role of income growth in China's urban expansion.As per McCarthy and Smit (1981), the new urban economics approach, both politics and economics, in toto, are responsible for urban land-use changes and urban spatial decisions.Healey (1997) acknowledged that planning practice is not a valueneutral activity.Hence, government interventions in urban spatial decisions are likely with multiple intentions.
Whatsoever the above-discussion implications, it is the fact that politics shape the processes of spatial decisions and influence the direction (Lefebvre 1991).Spatial planning and implementation require carefully designed multi-level governance arrangements and strategic decision-making (Healey 1999;Pieterse 2019).In summing, while the above debates can be assessed independently from a discipline standpoint, their interaction eventually results in a singular, integrated understanding of urban space.We employ the production of urban space as a conceptual framework to underline the need to consider urban space in a multifaceted urban spatial planning process.

Description of the study area
Oromia Regional State (see Figure 1) is one of the biggest regional states in Ethiopia.The region's population was 27.2 million in 2007 (CSA 2007) and 39.1 million in 2021 (CSA 2021).The region covers 363,399.8km2, accounting for 34.5% of the country's total area.Galan, Dukem, and Bishoftu cities are located in the Oromia Regional State 25, 37, and 47 km south of Addis Ababa, respectively.The cities are located along the highway between Addis Ababa and Djibouti.Information about population size, number of households, area, and the number of operational industries in the three study cities are shown in Table 1.
The study geographical unit was chosen as an industry development growth corridor in 2006 because of its closeness to Addis Ababa and access to the Addis-Djibouti import-export route (Dadi et al. 2020).Before this categorisation, the geographic unit was known for its agricultural production, and a significant portion of the area was occupied by agricultural land.
These cities were chosen purposively because they have been experiencing massive industrialisation developments, rapid land-use changes, and urbanisation.These cities also have strong interdependencies/ linkages with Addis Ababa, a primate city, where an unplanned urban corridor has evolved.Key informants were selected from Tullu-Gurracha, Gogecha, and Qajima kebeles from Galan, Dukem, and Bishoftu, respectively.The Kebeles were selected because of their industry concentration and peri-urban character.Kebeles are the lowest administrative units in Oromia region (Ayele 2014).The study cities are guided by highly centralised urban planning, which is overseen by the same entity, Oromia Urban Planning Institute (hereafter OUPI) (Dadi et al. 2016).OUPI was established in 2006 with the regulation no.067/2006 to prepare urban plans for cities in the Oromia region and advice zones and urban centres in areas of urban planning and implementation schemes.OUPI is responsible for conducting a structure and basic plans for cities in the Oromia region.The terms 'structure plans' and 'strategic plans' are closely related (UN-Habitat 2009).A structure plan is a 10-year framework that guides land development in an urban    (Mohamed and Worku 2019;Terfa et al. 2020).

Data collection and analysis
The classifications of land use classes by these satellite imageries were complemented with aerial photographs for 2006 and Google Earth images for 2006 and 2021 to measure the accuracy label and for verification of each land-use change detection pattern analysis.Furthermore, industrial locations in 2006 and 2021 have been identified using the structure plans of the three cities.The administrative boundaries of these cities have expanded rapidly.Hence, the recent (2017) shapefiles indicating the administrative boundaries of the three cities were collected from their respective cities' administrations and the Oromia Urban Planning Institute (OUPI).
Image classification to quantify the change matrices and data processing techniques such as cloud and strip cleaning, and geometric correction were made using ArcGIS software version 10.2 package.Supervised image classifications were made using the Maximum Likelihood Classifier Algorism.An accuracy assessment was made using the Google Earth images to determine the quality of the information derived from the images.More than 180 training samples were collected for each land cover class identified.Confusion Metrics and Kappa Coefficient were used to assess overall categorisation accuracy so that 50 points from each class and a total of 250 (50 × 5 = 250), points for accuracy assessment were employed.Kappa Coefficient allows for an empirical assessment of the accuracy level of the classified land classes compared to the randomly selected locations from each category (Ganasri and Dwarakish 2015;Das et al. 2021).
The level of accuracy varies between the study city and investigation years, albeit a confusion matrix of land use/cover types using ground control points revealed that an average overall accuracy of landuse change analysis was 86% and 87% as assessed in the Landsat Enhanced ETM+2006 and 2021 image classification, respectively.This indicates that the produced land use/cover maps were realistic, and the image processing method was effective.
Depending on the prior knowledge of the study area, personal experience, land use plans of each city, spectra response of the Landsat, the resolution of aerial photo and Google Earth imagery, and physiographical nature of the land use classes, five categories of land cover classes, namely, built-up area, agricultural land, vegetation, water bodies, and open spaces were mapped (see Table 3).
Besides, image enhancement and composition were applied to determine land cover classes better.Accordingly, land cover maps were generated based on pixel-based supervised classification techniques.Areas for land use classes were calculated in hectares (ha) to identify the percentage changes.The rate of percentage changes between the years 2006 and 2021 was calculated against the resulting land-use types using theEquation:

LULC changes in Addis Ababa-Bishoftu growth corridor
The three study cities, land use, and land cover changes have significantly been altered.The spatiotemporal study shows varying degrees of change in land use/cover features across the 15 years from 2006 to 2021 (see Figure 2).The results confirm massive land use/cover alterations during the study period in the three cities.As a result, the built-up areas of Galan, Dukem, and Bishoftu cities have been drastically transformed by 79%, 60%, and 94%, respectively.
Most built-up area changes are detected along the urban corridor (transportation lines).The spatial expansion of built-up areas occurs at the expense of agricultural lands.The built-up and agricultural lands show rapid changes among all land use classes.Likewise, the percentage changes of agricultural land areas into other land use classes in all the study cities were also very high.The spatiotemporal analysis of LULC changes shows that built-up areas expanded by 79%, 60%, and 94% in Galan, Dukem, and Bishoftu, respectively, in the study period (2006-2021) (Figure 3), while agricultural areas decreased by 17%, 47%, and 5% in the three cities respectively.On the other hand, open spaces decreased in Bishoftu by 38%, albeit it shows an increase in Galan and Dukem cities by 4% and 42%, respectively.The water body class in Bishoftu city is represented by Lakes Bishoftu, Hora, Cheleleka, and Horo Hora.This class has remained unchanged in size.The net loss of agricultural land was 5% in the case of Bishoftu, yet the highest agricultural land loss (47%) was registered in Dukem (Figure 3).On the other hand, the loss of open spaces was found to be high (38%) in Bishoftu.The built-up areas persistently gained additional areas by 79%, 60%, and 94% in Gelan, Dukem, and Bishoftu, respectively.Bishoftu has the most built-up changes (94%) and Galan (79%).
Besides, the study cities' proximity to Addis Ababa has contributed to urbanisation through investment expansion.
The industrial developments within the built-up areas were identified from the current structure plans of the study cities.Currently, 624 ha, 640 ha, and 1860 ha of land are occupied by industrial developments in Galan, Dukem, and Bishoftu, respectively (OUPI 2017a;2017b, 2019).As a result, industrial expansion and intensification have played a significant role in the growth of the built-up areas.In the next 10 years, each city plans to boost industries' contribution to the cities' economy and allocate a significant portion of the land for industrial development.Galan city has allocated 2200 ha of land for industrial developments for the next ten years (end of 2028).As per the KII conducted with OUPI experts, OUPI prepares urban plans based on 'scientific methods for the cities they are mandated, although political decisions from both Federal and Regional governments violate what has been planned and land uses (KII, 2021).
As shown in the results of Landsat image analysis, there have been significant changes in the agricultural land areas in the study cities.The percentage changes in agricultural lands in Galan, Dukem, and Bishoftu cities showed a decline of 17%, 47%, and 5% between 2006 and 2021.The incompatible location within residential areas of manufacturing industries has been a serious concern for the residents.The interview results with a community leader in Bishoftu city (Qajima kebele) attested that industry-related services are posing problems related to noise, air, and water pollution, as well as traffic congestion (KII, 2021).OUPI urban planners have also shared this viewpoint and acknowledged that these industries, located in the inner city, pose health and other socioeconomic problems to many residents (KII, 2021).In a nutshell, the respondents in the interview identified that inadequate infrastructure supply, land use violation, the absence of conducive policies, and the lack of clear industrial development and employment generation strategy have hindered accelerated progress in industrial investments.
As depicted in Table 6, built-up areas increased in Gelan city by 79% between 2006 and 2021.More precisely, during the investigations period, about 604 ha, 1052 ha, and 2445 ha of new urban land in Galan, Dukem, and Bishoftu have been formed.Of the total area, based on the structure plan of each city, nearly 624 ha, 640 ha, and 1860 ha in Galan, Dukem, and Bishfotu, respectively, were found to be due to demarcation for new industry production set-up.This shows that industry location occupied a significant proportion of other land use classes within the builtup areas, including social services, utilities, and residences.More specifically, about 110 operational industries in Galan city have occupied 624 ha of land, or, i.e. 46% of the total built-up area.During the investigation period, about 415 ha, 310 ha, and 610 ha of gross change land for industries were recorded in Galan, Dukem, and Bishfotu, respectively.Galan has shown the highest percentage change (197%) of the three cities in industrial land allocation.Most of them are occupied along the main roads of the city.An interview with the Galan city administration reveals that there have been massive establishments of industries in the city without clear guidelines since the city's establishment in 2008.The KII conducted with Galan residents largely agrees that the green areas and agricultural lands are frequently handed to investors without consultation with local communities (KII, 2021).Whereas, in Dukem, industries occupied 640 ha and 22% of the total built-up land.Yet, about 200 ha of land is occupied by the Eastern Industry Zone, and 200 ha more land is reserved for the second phase of expansion.The interview with Gogecha kebele community leader showed that new developments have been most intense on either side of major roads and need a prior consultation with them.This is particularly visible in Gogecha kebele, where many industries are located.Gogecha kebele community leader attests that a linear type of industrial development has been formed, unlike in Tullu Gurracha kebele, where industries are being expanded and concentrated horizontally (KII, 2021).
Large industrial sites occupied locations that were once fully or partially agricultural grounds, as shown in Figure 4. On average, about 40, 70, and 149 ha of the urban built-up land in Galan, Dukem, and Bishoftu cities, respectively, increased annually in the study period (2006 and 2021).This is equivalent to adding 3.3 ha, 5.8 ha, and 12.4 ha of built-up areas each month in Galan, Dukem, and Bishoftu, respectively.Since the city of Galan was established in 2008, development activities in the city have exploded, particularly in three areas -at the exiting and entry points to the city from Addis Ababa, the exiting and entrance to the city from Dukem, and the eastern portion.Huge industries, e.g.Euro Cables Private Limited Company and others, are found along this road, and enormous trucks damage the road (Figure 5).This indicates the industrial developer's lack of corporate social responsibility 1 in the area.On the contrary, land use for pure/mixed residences and administration has areas less than the minimum by 38.24% of the standard (55-70%) (MoUDH 2018).Similarly, such discrepancies are also common in the case of Dukem and Bishoftu.
The residents have raised the need for inclusive growth in Dukem, Gogocha kebele.For instance, around 200 households in Gogocha kebele are settled far from the main road.They feel marginalised and disconnected from the main centre due to investment expansion.One of the Gogecha kebele residents, aged 65, explained: 'I feel living a very far remote area while living only 30 km away from the capital Addis Ababa and industrial developers in the kebele do not engage in Corporate social responsibilities'.
Ethiopian urbanisation has been following accelerated economic growth.A study by MoUDH (2015) revealed that the urban population of Ethiopia is growing at over 5% per annum.Between 2006 and 2021, for example, Galan city grew from 12,300 to 36,500 inhabitants (CSA 1994;2007, 2021), a considerable population gain (191%).During the same period, however, its built-up area expanded by 79%.This pattern is roughly the same in the case of Dukem and Bishoftu cities.
As per the Galan Structure plan (2017), the current 21,333 household number is expected to grow to around 41,404 by the end of the planning period or 2028.In the case of Gelan, for instance, average individuals used to have 1.25 ha-5 ha before the city establishments in 2008.Yet, since then, the landholding of households has decreased to less than 1 ha (OUPI 2017b).The intensification of industry develop-   which is not profitable in comparison to their previous farming system (KII, 2021).So far, land granted for industry development has reached 640, 620, and 1800 ha in Galan, Dukem, and Bishfotu, respectively (OUPI 2017a;2017b, 2019).This figure will also be expected to be doubled in the next 10 years when the respective cities' planning period ends.

Industrialisation and formation of an urban corridor next to a primate city
The result of the interview with communities in Tullu Gurracha, Gogecha, and Qajima kebeles in Galan, Dukem, and Bishoftu, respectively, also portrayed that the land allocated along the southern axis (along the main road and the immediate expansion areas) of Addis Ababa for medium and heavy industries has been significant.The recent development of the Addis Ababa-Adama Expressway, Ethiopia's first toll road, includes the study's spatial units, which connect Addis Ababa, Galan, Dukem, and Bishoftu, which were built in 2014.Without a doubt, the highway aids in the acceleration of economic activities further.
From 2006 to 2021, Addis Ababa-Bishoftu urban corridors have shown many industrial development changes.Growth has been particularly strong along the southern outlet of a metropolitan corridor connecting Addis Ababa, where significant industries are located.The Landsat image analysis depicts a linear form of built-up form during the investigation period.The interview and observations support this, as many industrial developments were established along the main road from Addis Ababa to Bishoftu urban corridor.
The Land use plan lacks consideration of the formation of a conurbation.The study cities Gelan and Bishoftu share boundaries with Dukam cities.However, the urban spatial planning process needs to include such a consideration.As Galan, Dukem, and Bishoftu city administrators testified in the interviews, the land-use proposal lacks taking into consideration the economic, infrastructure, administrative, environmental, and physical linkage between these cities.
Because the study cities (Galan, Dukem, and Bishoftu) are located near the primate city, which is a market source, industrial developers want to invest in and around the area.Galan municipal administration officials were interviewed, and it was discovered that a large number of applicants for land were received daily.However, there is no guarantee that they will not speculate and use the land for other purposes.Many investors already took the land for industrial purposes but have yet to begin construction/operation.Ostensibly the study of cities' current urban morphology was partly shaped by the Ethio-Djibouti Railway line.Since then, the study cities have had access to the Addis Ababa-Djibouti import-export route, and the concentration of industries has intensified.OUPI planners also believe that the construction of the Ethio-Djibouti Railway (1897-1917) influenced the pattern of Ethiopia's agglomeration of industries and the growth of new urban centres along the rail corridor.Prominent factories have been built along the roadsides from Addis Ababa to Bishoftu.However, problems in the urban fabric, such as the lack of a coordinated transport network system, have been significant constraints.
Officials from the study cities stated that their proximity to the capital, Addis Ababa, is not beneficial to their development but harms land utilisation and results in haphazard expansion.Due to its proximity to Addis Ababa, the Addis Ababa-Bishoftu urban corridor has been attracting significant industrial development investments (KII, 2021).Interviewed OUPI urban planners identified some of the negative repercussions of the haphazard formation of the Addis Ababa-Bishoftu urban corridor, including certain places cut off from urban services, which contributed to a lack of urbanrural connectivity and an uneven distribution of urban services (KII, 2021).The structure plan of each city is carried out by OUPI professionals from as far away as the capital, Addis Ababa.Interviewed kebele leaders also confirmed that the experts had no idea what was happening on the ground.Furthermore, they do not contemplate the emergence of a conurbation or the necessity of integrated planning.OUPI officials, on the other hand, say that a lack of experienced planning specialists and significant resource requirements were the biggest problems they faced while preparing structural plans and the need to consider conurbation formation.

LULC changes in Addis Ababa-Bishoftu growth corridor
As the findings revealed that the study's unit of analysis of cities grew horizontally along crossing highways, the majority of the built-up and industrial areas were concentrated there (see Figure 6).Planning, policy, and investment decisions may have both positive and negative impacts on the quality of life (Zieleniec 2018).Without a clear land-use plan, unplanned, and frequent land-use changes occur, primarily from agriculture to built-up and industrial growth.According to an interview with OUPI urban planners, there has been a mandate issue because cities approve plans while land-use changes are allowed by the region.A study by Beyer et al. (2021) pinpointed that municipalities in Ethiopia often lack the requisite resources and are poorly mandated to lessen the effects of industry site development.Many development organisations have considered implementing urban plans as a technical, apolitical exercise (Goodfellow 2013, p. 6).Yet, implementing plans is far from a technical exercise but political (Healey 1997;Goodfellow 2013).Aside from that, private demand and planners' technical viewpoints are essential factors to consider when changing land use.An interview with the Bishoftu (Qajima kebele) leaders confirmed that areas allocated for green spaces were turned into investment places without prior notice and public discussion (KII, 2021).Consultation and collaboration are indispensable for successful spatial planning (Healey 1997).Ethiopia aspires to make major urban centres industrial hubs and emphasis the promotion and expansion of manufacturing industries in the country (NPC 2020).
In all of the study cities, agricultural lands have decreased considerably, showing high susceptibility and a lack of agricultural area protection.Hence, inefficient land usage and management have been the serious bottlenecks to steering the industrialisation process correctly.This is also similar to a study by (Koroso et al. 2021).This study attested that industrial development of the study cities has dramatically increased but lacks prior planning.Land provision for industrial developers has no clear rules and regulations; in most cases, these cities have no right to decide and comment on what the federal government decides.For instance, Eastern Industry Development in Dukem city has been allocated without consultation with the community.Similar viewpoints have been put forward by Beyer and Hagemann (2021).Investment location decisions do not take into account a broader urban development perspective.This practice indicates the need for public opinion in the development process of their localities.As per Lefebvre's (1991), this is a process and procedure by which power, capital, and class were imposed.Healey (1997) also argues that knowledge and understanding can be produced, and consensus can be reached on planning issues through collaborative social learning processes involving all stakeholders.The Ethiopian constitution has stated the right to public participation in the development process (FDRE 1995).There ought to be greater public participation, including the right to identify the problem and get information concerning the plan ( UN-Habitat 2009;Feltynowski 2015).That is why authors like Watson (2009) noted that inclusiveness has yet to be given much attention in global south planning systems.Public participation is a decision-making process (Zheng & Zheng, 2022) and helps to prevent spatial conflicts (Feltynowski 2015).
McCarthy and Smit (1981) conclude that politics and economics are responsible for urban land use and landuse change.Lefebvre (1996) and Harvey (2008) contend that politics controls all economic, social, or otherwise urban activities.Ethiopia is one of the few African countries to have formulated a full-fledged industrial policy, which has been aggressively pursued over the last decade (Mulu 2013;Oqubay 2015Oqubay , 2018)).However, there is a widespread occurrence of schizophrenic behaviour, as what is written in policy documents and what is practiced frequently contradict each other (Melaku et al. 2021).A case in point, unrestricted horizontal expansion of Addis Ababa often contradicts the urbanisation policy of the country.The policy encourages balanced urban development and discourages horizontal expansion (NPC 2020).
The trend of urban expansion is well over cities' real population capacity and economic growth.According to the World Bank (2015), Ethiopian cities' built-up areas have grown faster than the urban population.For instance, from 2007 to 2014, the population of Addis Ababa increased by 17%, while its urban extent increased by 51%, nearly three times faster.Secondary cities are also experiencing rapid horizontal expansion, as much as 2.3-6.2 times the city's current territory (Koroso et al. 2021).Expansions are characterised by expanding the total built-up areas of the cities outside their administrative boundaries.This pattern is common in all of the study cities.Angel et al. (2012) argued that room is required for urban expansion in rapidly growing cities and should consider their expected built-up areas.Haregeweyn et al. (2012) suggested that a deeper understanding of the processes of urban growth and the impacts can assist to build the foundation for longterm planning of future developments in urban regions.Tian et al. (2017) pointed out that the market forces have contributed to changes in agricultural areas to nonagriculture areas in China.Indeed, industrial expansion was a major factor in the growth of the built-up area.This result is in line with the works of Deng et al. (2008), Kassahun and Tiwari (2012), Dadi et al. (2016), and Giannecchini and Taylor (2018).Unquestionably, industrial policy is back on the African policy agenda, with many countries following new strategies for rapid industrialisation and none have done so more eagerly than Ethiopia (Brautigam et al. 2018).Christopher et al. (2020) state that there is never a straightforward causal relationship between ideas and policy implementation.

Industrial developments and urban spatial planning trade-offs
One of the industrialisation process challenges is the lack of coordination of national-level policies with regional states.The country's industrial policy is fuzzy on urban development and planning outcomes, and 'space blindness' is the policy's 'Achilles heel.'In the Ethiopian context, the importance of politics in plan-making and implementation is not recognised in a language-based federation, although planning is both a technocratic and a political endeavour (Melaku 2018).Serin et al. (2020) investigated the state's role in commodifying urban spaces in Turkey.They found that the state has played a crucial role as a restructuring agency by changing legislation favouring urban development and fostering the commodification of urban spaces.This evidence demonstrates that political intervention can shape urban spaces (Lefebvre 1991;Healey 1997;Merrifield 2006;Robinson 2008;Watson 2009;Goodfellow 2013;Moberg 2015).
Policymakers must overcome knowledge and incentives problems.They must thrive in dealing with the knowledge problem to circumvent misdirected industrial development.The incentive problem, on the other hand, manifests in behaviours like corruption, nepotism, and favouritism (Moberg 2015).This is typical in the Ethiopian industrial development sphere (Befekadu and Nega 2000;Mulu 2013;Oqubay 2015).The national development policies have recognised that urban centres are entitled to implement the policy by devising tailored strategies, plans, and appropriate institutional arrangements in their geographical boundaries (NPC 2015;2020).Nevertheless, the study cities have executed plans and development initiatives regionally organised through OUPI.During the interview with Galan, Dukem, and Bishoftu city administrations confirmed that their plans have been conducted by planners who reside at the regional level (KII, 2021).The national policies gave less emphasis to urban space.Related to this, a key informant from Bishoftu explains, 'The decision of politicians to amass industries in cities next to Addis Ababa without adequate urban and regional integrated planning knowledge has led to huge gaps between expectations and reality.Ethiopia's industrialisation initiative does not take into consideration geography.For instance, politicians ignore Bishoftu's inherent tourism potential and are rather preoccupied with selling plots for manufacturing industries' (KII, 2021).Macro-level policies and investment procedures in connection to space organisation and industrial development have yet to be smoothly integrated (UN-Habitat 2009).In public policy, lack of place consideration is strikingly evident in highly centralised countries (Healey 1999).
In a weak planning system, central government policies are often spatially blind, and the territorial reality is lost.Hence, implementing agencies find difficulties in translating policy objectives to the local conditions (Spaliviero et al., 2019).Goodfellow (2013) pointed out that the effective implementation of urban spatial plans is an increasingly urgent development concern with significant implications for urban development.Renaud (1981) corroborated this argument that national trade policies that protect the manufacturing sector commonly generate unintended spatial biases.Hence, urban space and land husbanding for industrial development have yet to be coordinated.Beyer and Hagemann (2021) indicated that industrial investment showed minimal potential for interaction with the surrounding urban fabric.
Urbanisation and urban growth have often been out of sync with industrialisation, especially in the recent demographic transitions experienced in sub-Saharan Africa (Duranton 2009).Studies by Dadi et al. (2016), Giannecchini and Taylor (2018), and Beyer and Hagemann (2021) are critical of the catalytic role of the industrialisation process.Lefebvre's "the right to the city'' dialectical process of economic production and social life are inseparable processes that have unity, yet it is a conflictual process (Lefebvre 1996).Harvey (2008) further noted that urbanisation has always been a class phenomenon since surpluses are extracted from somewhere and somebody, while the control over their disbursement typically lies in a few hands.The interconnectedness of industrial growth and national policies encourages discussion of tradeoffs between investment expansion and sustainable urban spatial consequences.Hence, the right to inhabit is required to be incorporated into the urban planmaking process (Lefebvre 1996).
In cities where future growth is dependent on industrial development, consideration should be given to where to situate industry, and the compatibility of urban land uses.As Lefebvre's (1991) points out, space is no longer plentiful, and the political economy of (urban) space requires attention based on the essence of economic scarcity.In principle, a master/structure plan is a legally binding document that regulates land use, land, and infrastructure development.Nonetheless, in many instances, the study cites developed urban land before or during the plan approval process.This is partly because the plan preparation has taken lengthy processes and time to finalise.A case in point, the current structure plan of Bishoftu city took nearly four years from the start and end of the project (June 2015-2019).As a result, several land-use violations prevailed before the plan became a legally binding document.Policywise, article 43 of the Ethiopian constitution states the right to the development of citizens (FDRE 1995).Yet, things on the ground have been going against the rights of citizens and cities (Lefebvre 1996).

Industrialisation and formation of an urban corridor next to a primate city
The corridor between Addis Ababa-Bishoftu is Ethiopia's main economic corridor, which is reflected by the main route that connects the cities (Addis Ababa-Galan-Dukem-Bishoftu) is one of the most heavily trafficked roads in the country.Whebell (1969) defines a corridor (synonymously labelled as linear urban systems) as a linear pattern of major cities joined by highly developed 'bundles' of transport routes.Finding an approved definition of corridor spatial development in the Ethiopian context is difficult.This is partly because the planned corridor spatial development concept rarely appeared in the country, but unruly urban corridor formation has not been uncommon.A study by Terfa et al. (2020) depicted that cities that shared a common boundary with Addis Ababa, tended to expand towards it, following the main road, suggesting a lack of rigorous implementation of the master plan.Hence, some areas disconnected from urban service and contributed to weak urban-rural linkage.Spaliviero et al., (2019) explain that in countries with weak planning systems, the powers of local planning authorities are limited to jurisdictional boundaries instead of the functional cities within their regional contexts.An interview with an expert from OUPI attested that unplanned and haphazard industrial development, inadequate road network, and inefficient management of urban land had created societal problems in the study cities along the urban corridor.
In the absence of any effective land-use plan, land development occurs due to various actors' legal and illegal actions (Satterthwaite 2007).An expansion of industry development calls for various services and amenities including housing.The industrial parks in Addis Ababa, and the Eastern Industry Zone (EIZ) in Dukem, all attract thousands of migrants to the cities for jobs, and they need housing.The KII conducted in 2021 with Dukem city administration officials describes, 'There is a huge housing shortage in Dukem due to the presence of the Eastern Industry Zone.Consequently, the housing demand has not been matched by a corresponding housing supply and urban infrastructure, resulting in a massive backlog'.With industrialisation and urbanisation, more people have moved to already crowded cities, especially Addis Ababa, putting pressure on the housing market.Hence, urban developments in Ethiopia are characterised by spontaneity and usually emerge from informal settlements (Terfa et al. 2017).There is a housing strategy generating housing on the periphery of Addis Ababa (see Figure 7), but unconnected to and not planned with industrial development.
The GoE envisioned reducing the imbalance between the demand and supply of residential housing by involving all stakeholders and launching various housing programs (NPC 2020).As shown in Figure 7, the Addis Ababa-Bishoftu growth corridor experienced accelerated growth in the built-up areas (Terfa et al. 2020;Koroso et al. 2021).A study by Mohamed et al. (2020) stated that haphazard land development is a crucial urban development problem in Addis Ababa and the surrounding cities.More notably, these study cities are located next to a primate city, Addis Ababa which resulted in soaring urban land values.This is consistent with Goodfellow (2013), in East Africa, the levels of urban primacy are tremendously high, there is huge pressure on urban land, and ongoing increases in value result in a kind of 'scramble'.
There was an integrated urban spatial planning preparation attempt that includes several cities including this study unit of analysis, but it failed (Melaku 2018), mainly because of the social unrest raised in Oromia region (Mohamed et al. 2020).Integrating urban planning and development processes is indispensable for sustainable urban development (Haregeweyn et al. 2012;Alem 2021;Matsumoto and Crook 2021).

Conclusion
Our findings contribute to a better understanding of the complicated relationship between investment flows and land use, as well as gaps in policymaking processes with macro objectives, such as Ethiopia's industry policy.A more nuanced integrated planning approach is necessary to fully capitalise on the potential of the study cities' available resources and proximity to the capital, Addis Ababa.Hence, urban spatial planning should be flexible, and its content should include appropriate industrial locations.Furthermore, the rising rate of investment flows has created a demand for urban land, necessitating planning to govern urbanisation.Urban development policies and plans, including land development, should be aligned with industrial development strategy.
Addis Ababa shows inter-spatial dominance of the surrounding urban system in general and Galan, Dukem, and Bishfotu in particular.Infrastructure accessibility and proximity to Addis Ababa may be key factors in determining industry concentration in the study cities, although, it needs further studies to pinpoint industry locations determining factors.In a nutshell, the prey-predatory (Addis Ababa and small cities) relationship must be converted into a mutual dependency.A sustainable urbanisation process necessitates a careful balance of expansion between Addis Ababa, secondary cities, and small cities.The situation is particularly acute in Addis Ababa-Bishoftu urban corridor.Thus, planned urban growth corridor development becomes essential for both the capital city and the southern urban corridor -Galan, Dukem, and Bishoftu.The administrations of Addis Ababa and the three studied cities in the Oromia region must collaborate closely to maximise mutual benefits while circumventing adverse effects.
From the discussion above, it is clear that industrialisation has been a major spatial intervention with significant spatial impact.Hence, collaborative efforts between federal, regional, and cities are required to ensure harmony between industrialisation and urban spaces.Most of the problems in the urban fabric, such as the lack of a coordinated transport network system between/among cities along the urban corridor, conversion of agricultural land to other uses, and unplanned growth trends of these cities, can only be overcome through coordination and harmonious unitary work.This will require a sustainable, strong political will and human capacity building.Furthermore, our data demonstrate that community engagement is lacking in industrialisation and urbanisation processes, and we conclude that sustainable city growth will be a nightmare without community engagement.All in all, the study cities are in a unique position to guide urbanisation in the right direction before it is too late.There needs to be a straightforward industry zone designation scheme for each city's uncoordinated and scattered industrial development.Urban planning practices must incorporate Lefebvre's ideas of (urban) space as a means of a habitable built environment, not as a place for capital accumulation.To avoid being 'space-blind', policymakers must address both the economic and spatial components of planning.We 'cry for and demand' the harmony of industrialisation and urban spaces that could result in many benefits and reduce the trade-offs.

Note
1. Ethiopia has no clearly defined policies and strategies that encourage businesses and investors to fulfill their Corporate Social Responsibility (CSR).However, the country's investment laws support social and environmental sustainability values in various investment activities.

Figure 1 .
Figure 1.The study cities and their order of location from Addis Ababa.

Figure 3 .
Figure 3. Percentage of land use/cover class changes from 2006 to 2021.
ment and built-up areas resulted in the landholding size of the local people diminishing.During the interview with the elders in Tullu Gurracha Gogecha and Qajima kebeles in Galan, Dukem, and Bishoftu, respectively, revealed that respondents have been losing their land in the name of investment and have been compelled to shift their livelihoods system to the various off-farm activities such as petty trade,

Figure 7 .
Figure 7. Addis Ababa-Bishoftu Economic Growth Corridor with major developments and built-up areas.

Table 1 .
(FDRE 2008)t is a legally binding plan(FDRE 2008).The structure plan preparation and implementation manual determined that urban land allocation for cities should be based on the 30-30-40 ratio for road and other infrastructure, greenery and open spaces, and residences and administration, respectively (MoUDH 2018).
, household,2017a coverage, and number of operational industries in the study cities(OUPI 2017b;2017a, 2019).Study citiesTotal population size Number of households Area (in ha) Operational industrial plants Time-series Landsat images are important data sources for detecting land-use changes.This study used remote sensing data, namely, Enhanced Landsat Thematic Mapper Plus (ETM+) satellite imageries, for 2006 and 2021 (see Table2) to classify and detect land-use changes in the three cities -Galan, Dukem, and Bishoftu.The same dates and months of the years, i.e.
April 12 and June 11, were used for the acquisition of the 2006 and 2021 Landsat images, respectively, considering the lowest moisture and cloud cover content at these times of the year in Ethiopia to minimise discrepancies in reflectance.The images were collected from the United States Geological Survey (USGS) official website and the Ethiopian Mapping Agency (EMA), which was similarly applied by researchers

Table 2 .
Descriptions of the satellite imageries used for land-use change detection analysis.
Alem (2021)al.(2021)approach was recently employed byNikolić et al. (2021)andAlem (2021).Regarding the secondary data sources, various structure plans of the three cities, economic and physical plans, manuals, standards, directives, and national plans and policies were reviewed and analysed.
The water bodies in Bishoftu city show relative stability over the investigation period.Similarly, the open spaces have increased in Galan and Dukem cities.In quantitative terms, open spaces increased from 8% in 2006 to 15% in 2021 in Galan city and also increased from 18% in 2006 to 29% in 2021 in Dukem city, thus representing an increase of 94 ha and 1231 ha area in Galan and Dukem city, respectively (see Table 5).Whereas the open space land use class was reduced by 1851 ha in Bishoftu from 2006 to 2021.Areas covered by vegetation were relatively low in all corresponding cities, showing 200 ha and 119 ha reduction from the base year in Dukem and Bishoftu, respectively.

Table 3 .
Description of land use/land cover change (LULC) classes used in the study.LULC classes Descriptions Built-up areas Residential, commercial, industrial, transportation, services, communication, and utilities.Agricultural land Cropland and pasture, horticultural areas, confined feeding operations, other agricultural lands Vegetation cover Deciduous, evergreen, and mixed forest land, woodland, shrubs, scattered trees and mixed with grass, Waterbody Streams and canals, lakes, reservoirs, rivers Open-space Quarry sites, sandy areas, bare exposed rock strip mines, quarries, and gravel pits, mixed barren land

Table 5 .
Statistical distribution of land use/cover classes in the three cities by years and their change areas in hectares (ha).

Table 6 .
Total built-up area (ha) and proportion (%) of industrial increase from 2006 to 2021.
bPercentage of industry area calculated from the built-up land and structure plans.