Urbanization and sustainability under transitional economies: a synthesis for Asian Russia

Spanning a vast territory of approximately 13 million km2, Asian Russia was home to 38 million people in 2016. In an effort to synthesize data and knowledge regarding urbanization and sustainable development in Asian Russia in the context of socioeconomic transformation following the breakup of the Soviet Union in 1990, we quantified the spatiotemporal changes of urban dynamics using satellite imagery and explored the interrelationships between urbanization and sustainability. We then developed a sustainability index, complemented with structural equation modeling, for a comprehensive analysis of their dynamics. We chose six case cities, i.e., Yekaterinburg, Novosibirsk, Krasnoyarsk, Omsk, Irkutsk, and Khabarovsk, as representatives of large cities to investigate whether large cities are in sync with the region in terms of population dynamics, urbanization, and sustainability. Our major findings include the following. First, Asian Russia experienced enhanced economic growth despite the declining population. Furthermore, our case cities showed a general positive trend for population dynamics and urbanization as all except Irkutsk experienced population increases and all expanded their urban built-up areas, ranging from 13% to 16% from 1990 to 2014. Second, Asian Russia and its three federal districts have improved their sustainability and levels of economic development, environmental conditions, and social development. Although both regional sustainability and economic development experienced a serious dip in the 1990s, environmental conditions and social development continuously improved from 1990 to 2014, with social development particularly improving after 1995. Third, in terms of the relationships between urbanization and sustainability, economic development appeared as an important driver of urbanization, social development, and environmental degradation in Asian Russia, with economic development having a stronger influence on urbanization than on social development or environmental degradation.


Introduction
Integrated assessments of urbanization and sustainability in transitional economies are necessary to address the particular challenges facing those regions. With more than half of the global population already living in cities, the world is entering the urban era (United Nations 2014). Although many have examined the extent and spatial patterns of urban transformations around the globe (Seto et al 2011, Zhou et al 2015, Reba et al 2016, assessments of whether urbanized areas are evolving in a sustainable manner are rare (Nassauer et al 2014. This assessment is particularly relevant for countries with transitional economies, i.e., countries that have transitioned from central planning-based economies to market-based economic systems in Southeast and East Asia, Eastern Europe, and Russia, in the past several decades. In these countries, both the state and the market can affect the urbanization process, and their interplay may lead to a different pattern of urbanization from that of industrialized countries or nontransitional economies in the developing world. Furthermore, the differences in land use and institutions governing urban land use may lead to a variety of trajectories of urbanization and sustainability among transitional economies (Shatkin 1998, Leaf 2002, Sýkora and Bouzarovski 2012, Tian et al 2013. Yet, few comprehensive efforts have been made to explore the processes and consequences of the transformations of transitional economies, including post-Soviet countries, to understand how economic transition may have affected urbanization and sustainability , Fan et al 2017a, Park et al 2017. Asian Russia stands out as a unique system to help us understand the processes and challenges of urbanization under transitional economies due to the unusual changes that occurred in the region following the collapse of the former Soviet Union. Comprised of the eastern part of Russia, Asian Russia spans a vast territory of approximately 13 million km 2 from the Ural Mountains in the west to the Pacific Ocean in the east and was home to 38 million people in 2016. Here, we focus on Asian Russia in contrast to the smaller but more densely populated European Russia due to its very different economic and societal development over the 20th century (Forsyth 1994). While Asian Russia has a lower population density (2.9 people km −2 in 2016) compared to European Russia (27 people km −2 in 2016), 75.5% of the population in Asian Russia is concentrated in cities, primarily consisting of large cities spread over the southern edge of Asian Russia (figure 1). In contrast to European Russia, cities and towns in Asian Russia have long been associated with resource extraction and have served as transportation and industrial hubs for related activities since the start of the construction of the Trans-Siberian Railway in 1891. However, these human dwelling centers in Asian Russia (>70% of the population) have been experiencing new challenges since the collapse of the Soviet Union in 1990. For example, the population in Asian Russia decreased by 10.3% from 1990 to 2016, whereas the total population in Russia declined by only 2.7% during the same period. This change in population, coupled with a volatile economy, may hinder landscape change and urban planning toward the long-term sustainability of the region. Meanwhile, it provides us with a unique opportunity to study their dynamics. Despite the environmental vulnerability and the challenging conditions of human systems, current literature on urban systems and sustainable development in Asian Russia is very scarce, with the exceptions of works by Becker et al (2014) and Fan et al (2017b) on urbanization in Russia and Siberia, and a few on regional sustainability (e.g., Grigor'ev 2009, Bashalkhanova et al 2012, Suspitsyn 2012, Slepneva et al 2016, Zabelina and Klevakina 2016 (please see details of related key literature in appendix A).
To address the knowledge gap about changing urbanization patterns and their relationship with sustainability in Asian Russia following the collapse of the Soviet Union, our objective was set to synthesize the data and knowledge on urbanization and sustainable development in the context of socioeconomic transformation. Specifically, we aimed to answer the following research questions: (1) What are the trends in the spatiotemporal distributions of economic activity and urban built-up lands in Asian Russia after 1990?
(2) What are the spatiotemporal changes of regional sustainability, including its three pillars: economic development, environmental conditions, and social development? (3) What are the interrelationships between urbanization and different measures of sustainability?

Study area
Representing 77% of the Russian territory, Asian Russia is a large landmass of Northern Eurasia that has been endowed with extraordinary natural resources and plays a critical role in the wellbeing of the global environment. We define Asian Russia as an area that includes three federal districts (FDs, administrative level-1 units) of Russia: Ural, Siberia, and Far East (figure 1), with FD as an official division of Russia's regions. If Asian Russia were an independent country, it would be the largest country in the world. The elevation of Asian Russia varies greatly, with a relatively flat landscape below 100 m over hundreds of kilometers-the West Siberian Plain-and with more rugged terrain in the Central Siberian Plateau (a.s. l=1700 m) and 3100 m toward the eastern edges and in southern Siberia's Sayan Mountains (Afonin et al 2008). Similarly, the climate also varies considerably across Asian Russia. Asian Russia is an earthly gem because of its rich natural resources (e.g., minerals, oil, gas, soil, etc) and partially intact boreal forests -the largest terrestrial biome on Earth-which are home to some close-to-extinction species (e.g., the Siberian tiger, Panthera tigris tigri) and provide the necessary livelihoods for indigenous people who rely on the forests for their daily needs. In addition, Asian Russia, as the main terrestrial part of the arctic region, plays a vital role in the global ecosystem dynamics and biogeochemical cycles (World Wide Fund for Nature (WWF) 2007, Groisman et al 2017). However, the natural environment in Asian Russia currently faces many pressing challenges caused by intensified logging, mining, and construction, as well as rapid warming that is responsible for large and intense wild fires (Roshydromet 2014, Groisman et al 2017). The increasing global demands for wood products, gas, and natural gases have particularly escalated the land use intensity across the region.
Although human activities took place in the region as early as the 11th century, Asian Russia did not become an important economic unit of Russia until after the construction of the Trans-Siberian Railway. Until the mid 19th century, Asian Russia presented little economic interest to Imperial Russia. The region centered its main economic activities on fur trading from the 11th to the 18th century, with a few spots of agricultural production in the Far East (Okladnikov 1968). Mining in the region started in the 17th century, with silver, lead, and copper mining taking place in the 1700s and gold mining beginning in the 1830s. Due to its long and severe winters, the largest part of Asian Russia is not suitable for agricultural production, which also reflects the concentration of settlements in the southern belt of Asian Russia (e.g., populated in Yakutia and the Russian Far East, table 1) (Rosstat 2016). However, the land scarcity and low agricultural productivity in European Russia forced the Imperial Russian government to implement the Stolypin agrarian reform in the early 20th century, which was labeled as the Great Siberian migration of Russians towards Asian Russia (Treadgold 1957). The operation of the Trans-Siberian Railway facilitated the transport of agricultural commodities and goods (e.g., wheat and butter) from West Siberia to European Russia and Western Europe. After the 1917 revolution, Asian Russia experienced rapid industrialization and urbanization with forced labor in coal mining and the iron-steel complexes before World War II (WWII) (Blinnikov 2011). During WWII, some industries from western USSR regions were relocated to the east of the Ural Mountains. To meet the growing domestic food demand and improve national food security from 1954 to 1963, a Virgin Lands Campaign was implemented, resulting in ∼20 Mha of virgin steppes being ploughed to support immigration to Siberia from European Russia. After WWII, construction of large thermo-and hydro-electric power plants continued in Asian Russia (Prishchepov et al 2018).
The dissolution of the Soviet Union had a significant impact on the population dynamics of the region. From 1990 to 2014, Asian Russia experienced a substantial population decrease of 9.9%, in contrast to a population decrease of 1.4% across the entirety of Russia. While the Far East witnessed the highest population decrease of 23% followed by Siberia with a decrease of 8.7%, Ural experienced the least decrease of 3.7%. At the federal subject (FS) level, i.e., an administrative level-II unit that is equivalent to the state level in the USA or provinces in many countries, 23 FSs in Asian Russia experienced a similar declining trend with the exception of only four, i.e., Tyumen (Ural FD), Altai Republic, Tuva, and Novosibirsk (Siberia FD). Despite the decrease or stagnation of the population within their respective FDs, all 6 major cities studied in this paper, except Irkutsk, experienced an increase in population from 1989 to 2016 (figure 2(b)). In 2016 less than 26% of Russia's population resided in Asian Russia (population density=7.8 persons km −2 ). The urbanization ratio (i.e., percentage of urban population) in the region was 75.5% in 2014, which was slightly higher than the Russian average of 74.0%.
Under this natural/historic background and in this particular socioeconomic context following the collapse of the Soviet Union, we conducted this study on urbanization and sustainability at three hierarchical levels: Asian Russia, FD, and federal subject (FS). Asian Russia includes three out of eight FDs and 27 out of 85 FSs of Russia and most of Russia's socioeconomic data are collected at the FD and FS level. We additionally incorporated six major cities as case cities to illustrate the dynamics of urban built-up area and population at the city level. Out of 47 cities with population over 100 000 inhabitants by 2016 in Asia Russia, thirteen cities have >500 000 inhabitants (figure 1, table A1). We acknowledge that large cities may not fully represent the region-wide urbanization trajectory because they tend to be the most resilient to change, due partially to their strong economic resources and large demographic bases. Nevertheless, cities rather than nations have become the centers of the global economy (Beeverstock et al 1999, Hall andPfeiffer 2013, Sassen 2013); large cities can serve as important nodes and connect their vast hinterland with the increasingly globalized economy. Becker et al (2014) argued that large cities, particularly those in the southern and western parts of Russia, seem to attract migrants from the far north and east. In this study, we focused on examining whether large cities in Asian Russian share similar fates as the large cities in Russia in general. This knowledge allows us to understand if they are in sync with the region in terms of population loss, and whether similar challenges in urbanization and sustainability exist in Asian Russia. The six cities were chosen due to their large sizes and importance in economic development, political administration, and cultural and educational structures. They are: Yekaterinburg in Ural, Novosibirsk, Krasnoyarsk, Omsk, and Irkutsk in Siberia, and Khabarovsk in the Russian Far East (figure 1). All of the cities are the capitals of their administrative level-II provinces. Additionally, Yekaterinburg, Novosibirsk, and Khabarovsk are the capitals of administrative level-1 FDs. Though Irkutsk is the fifth populous city in Siberia with a slightly lower population than Barnaul, it serves as an important node on the Trans-Siberian Railway and historically was an important trading hub in the eastern part of Siberia and its administrative capital during the Russian Empire epoch. We also conducted field trips to four of these cities, Yekaterinburg, Novosibirsk, Krasnoyarsk, and Irkutsk, in the summer of 2015.

Data and analysis
We relied on several data sources for this study, including remote sensing products (table A2), socioeconomic statistics, and on-site interviews with local experts. We examined the changes of economic statistics at the FD and FS scales through the analysis of 1 km resolution DMSP/OLS nighttime light (NTL) data from 1992 to 2012. To quantify the changes in built-up land for the six cities, we used data from the We performed geospatial analysis to explore the change of air pollutants of case cities. To understand the driving forces for the urbanization in Asian Russia, we collected data on demography, economic development, environmental conditions, and social conditions for FSs and population data for all cities with >100 000 people and conducted interviews with local experts of the case cities during our field trip (please see details in appendix B).
We developed and calculated sustainability indexes of the socioeconomic-environmental systems for all 27 FSs in the three FDs to explore the interdependent dynamics of multiple dimensions from social, economic and environmental aspects. Based on our literature review, we constructed a regional sustainability index (RSI) that incorporated conditions of economic development (i.e., employment rate), environment condition (i.e., air pollution), and social wellbeing in regard to housing (i.e., living space per capita) and health services (i.e., number of physicians available per 10 000 people) (see details of calculation of RSI in appendix C.1).
We further studied the interrelationships among urbanization and its potential independent variables through structural equation models (SEM). Economic development (ECO) and urbanization (Urb p ) were directly characterized with the observable variables of employment rate and the ratio of urban population to the total population (Davis 1965, Henderson 2003. Environmental degradation (ENV) was directly modeled by an observable variable of air pollutants per capita from the stationary sources (Seldon and Song 1994), Carson et al 1997. Social development was modeled as a latent variable SD by the number of physicians per 10 000 people (PHY d ) and living space per capita LS pc ( ) (please see appendix C.2 for detailed method).

Economic development and urban built-up area
Economic development, indicated by the average nighttime light brightness, increased for all three FDs, with the Far East experiencing the fastest growth (figure 2). While the increase during 1995-2008 appeared gradual with occasional small dips in 1997, 2001 and 2007, the increase was more substantial before 1995 and after 2007. Similar changing trends were also observed at the FS level for Ural, Siberia, and Far East.
All major cities experienced urban expansion measured by total built-up area, with urban built-up land expanding by 13%-16% from 1990 to 2014. Novosibirsk had the largest urban built-up area of 764.6 km 2 in 1990 and maintained its rank in 2014 (879.6 km 2 ). Khabarovsk and Irkutsk had relatively small urban built-up areas of 319.8 km 2 and 381.1 km 2 respectively, in 1990, but expanded to 361.8 km 2 and 442.5 km 2 in 2014, respectively. Omsk led the pack in terms of total expansion at 16.3%, whereas Khabarovsk came in last at 13.1% (figure 3). The expansion of the urban built-up area of these cities matched well with the overall increasing trend of the population of these cities.

Regional sustainability and its three dimensions
The RSI of three FDs in Asian Russia followed closely the national trend, which showed a decrease until 2000 and an increase afterward with a small dip around 2008 (figure 4). While the changing trend of economic index (ECI) followed the dynamics of RSI closely, the environment index (EVI) fluctuated with small changes except for Ural, whereas the social development index (SDI) increased continuously. The three major socioeconomic crises during the study period (i.e., the 1992 'shock', the 1998 default, and the 2008 global financial crisis) appeared to have a major influence on Russia's economy. Temporal changes in the regional RSI and its three components, particularly the ECI, confirmed their influences (figure 4). Russia transitioned from the central planning system to the market economy during 1991-1998 and the region witnessed a dramatic decrease in RSI and ECI from 1993 to 1998. The Russian government and Russian Central Bank devalued the ruble and defaulted on its debt in 1998. As a result, Russia's economy bounced back from 1998 on and continued to grow until 2008. 3.3. Coupled relationship between urbanization and sustainability ECO affected urbanization, social development, and environmental degradation directly with path coefficient values (PCV) (i.e., standardized linear regression weights) of 0.33, 0.28, and 0.22, respectively. Economic development had a stronger influence on urbanization (PCV=0.33) than social development (PCV of 0.28) or environmental degradation (PCV=0.22). Urbanization exerted a stronger influence than economic development on environment degradation and social development. Urbanization had higher PCVs of 0.51 and 0.38 in relation to environmental degradation and social development, respectively, whereas economic development had PCVs of 0.22 and 0.28 (figure 6). Urbanization also affected environmental degradation (PCV=0.51) more than social development (PCV=0.38).

Discussion
Asian Russia underwent a drastic socioeconomic change after the collapse of the Soviet Union, reflected by the reverse trends of population dynamics and economic development of the region and the further enhanced role of large cities, illustrated by their increasing populations and urban built-up land. Despite the generally declining population of Asian Russia, the region has experienced enhanced economic growth as indicated by nighttime light brightness. In particular, the six focal cities experienced a reverse trend of the increased population and expansion of built-up area. These findings imply two parallel processes: (1) the rise of economic growth against the overall decline of the population of the region, and (2) population growth and urban sprawl of large cities to accommodate retail, warehouses, and new economic activities as well as rural-urban migrants. Current literature, however, provides different views on whether or not large cities grow faster than small cities, because the growth of the city depends on various factors, including the trade-off between economies of scale and congestion, which increase as cities grow (Batty 2008). For example, Glaser (2000) argued that large cities grow at a similar rate as small cities. In contrast, Xu and Zhu (2009) found that smaller cities grew faster than large cities in 1990s in China. However, Henderson (2003) pointed out that the urbanization process was much more complicated than initially thought by Glaser (2000). While major cities may not be good representatives of the region, our results indicate the resilience of major cities even during a down time for Asian Russia. We remain unsure about the specific driving forces and mechanisms for this phenomenon, although we believe that large cities may provide better economic prospects and social development (i.e., high resilience) in both prosperous and difficult times (Becker et al 2014). In addition, globalization may have augmented the hierarchical structure of cities (Sassen 2013) and exerted more positive impacts on large cities than smaller cities and rural areas of the region.
The trend in regional economies underlines the dramatic impact of domestic institutional change and the global economy, as illustrated by the dynamics of the components of the sustainability index at the federal district level. As presented in section 3.2, three major socioeconomic crises in 1992, 1998, and 2008 had major negative influences on the regional economy, especially the 1992 'shock' and the global financial crisis in 2008 that appeared to cause a prolonged decline of the economy. From a temporal perspective, the declining ECI also reflected a decade of stagnancy and economic hardship after the collapse of the Soviet Union in the 1990s (figure 4), but its overall increase in the 2000s (although some FSs experienced some decline after 2012) revealed that the economy grew remarkably, likely due to the higher domestic demand, greater political stability, and continuous increase in the price of oil until the dramatic drop in prices of oil and gas in 2008 (figure 6). We consider that the bounce-back of the ECI in 2011 may be due to the joining of the World Trade Organization in December 2011 (figure 6). The ECI appeared to exhibit more volatile patterns and thus could lead to a similar change in RSI, whereas the EVI and SDI continuously improved.
The overall increase in the EVI and SDI, especially after 2002, indicates that the region's sustainability on environment conditions and social development did not suffer much despite the economic hardship. This finding may be contradictory to the findings of some researchers who found a decline in social development as indicated by social welfare of some regions of East Siberia (Bezrukov and Bonadysenko 2010). While our findings may be limited due to the selection of variables representing the environmental conditions and social development, we consider that our indicators provide meaningful insights into the sustainability of the region from a different perspective. The SEM results illustrated that urbanization and sustainability evolved as coupled relationships for Asian Russia. In particular, ECO affected urbanization, social development, and environmental degradation, which further emphasized the economy as the main driver for urban growth and other components of sustainability.
Nevertheless, the fact that urbanization exerted a stronger influence than economic development on environment degradation and social development implies that urban population growth may have exerted a large pressure for the urban environment, thus causing environmental degradation. However, urban population growth may also have actually boosted social welfare provision, as cities may gather more necessary resources from the increased population.
While we provide useful insights on urbanization and sustainability of Asian Russia, there is further research that can be conducted to explore the influence of other major drivers, i.e., natural external drivers such as global climate change, and socioeconomic drivers such as globalization and regional geopolitics, for urbanization and sustainability of Asian Russia. Broadly speaking, Asian Russia is trending toward a warmer and wetter climate, with many of these trends having been statistically significant over the period of 1950-2010. In response to these changes, Figure 5. Spatial distribution of RSI and its three components: ECI, EVI, and SDI. Note: in the official statistics, many socioeconomic variables (e.g., population) of Tyumen include FS of Khanty-Mansiy and Yamal-Nenets. Therefore, for RSI, ECS, EVI, and SDI, Tyumen's value represents the values of these three FSs.
there is evidence of earlier spring phenology, later autumn senescence, and community level changes in vegetation (Ovaskainen et al 2017). Typical Siberian crops could be grown as far as 500 km further north by the end of the century (Tchebakova et al 2011) and there could be a similar expansion of the boreal forest (Tchebakova et al 2016). Potentially, these changes could significantly increase agricultural productivity across the region, boosting economic development of the nearby cities and thus attracting a larger population to the region, including the cities. However, some significant negative consequences could arise at the same time, including methane venting from permafrost melting and the disappearance of arctic lakes and alterations of food webs, such as that of Lake Baikal (Hampton et al 2008). It remains to be evaluated how these changes as a result of climate may affect economic development and urbanization. Globalization and regional geopolitics, including foreign direct investment, imports and exports, foreign sanctions, migrations from the former USSR countries, and foreign tourists, can all exert significant impacts on urbanization and sustainability. As discussed, the ECI of each FS reflected the overall trend of national economic development of Russia. Global influence could also have additional influence on regional economic development. For example, Asian Russia includes several leading regions in Russia in terms of the scale of activities of organizations with foreign capital, such as Sverdlovsk, Tyumen, and Chelyabinsk in Ural, Krasnoyarsk in Siberia, Sakhalin in the Far East, measured by number of employed, turnover, and the average three-year volume of fixed capital investments in 2013 (Kuznetsova 2016). Along with the FSs of Moscow and St. Petersburg, Sverdlovsk and Krasnoyarsk are two of the top four Russian FSs that are included in the top ten for all three indicators. Factors such as territory proximity and cultural and historic linkages may have played a significant role in determining each FS's attractiveness for foreign capital of a certain country (Kuznetsova 2016). For example, China's investment has been particularly significant in Asian Russia, accounting for 31% and 32% of the foreign capital in Siberia and Far East, respectively, in 2013 (Kuznetsova 2016). The influence of global investments may be particularly apparent in large cities of the region as cities, especially large cities, serve as important nodes of networks for flows of global capital (Castells 2000). Furthermore, the region has unique natural and cultural assets such as Lake Baikal, the Altai Mountains, valleys and volcanoes in Kamchatka, and cities along the Trans-Siberian Express Railway, that can be of particular interest to foreign tourists. In fact, the Russian authority announced in 2017 that to revive the regional economy through attracting global investment and tourism, free e-visas for fewer than eight days will be issued to visitors from 18 countries who head to the Primorye, Khabarovsk, Sakhalin, Chukotka and Kamchatka regions in Far East. This new rule was expected to increase international tourism by 30% (Russian Times 2017).

Conclusions
We examined the spatiotemporal changes of and the interrelationships between urbanization and sustainable societal development of Asian Russia after the collapse of the former Soviet Union. Three major findings are: (1) despite its population decline, Asian Russia experienced an increase in economic activity as reflected by increased nighttime light brightness and most case cities of the region we selected experienced population increases and expansion of urban built-up areas.
(2) Asian Russia and its three sub-regions (federal districts) have improved their sustainability and its three major components: economic development, environmental conditions, and social development. While regional sustainability and economic development declined substantially in the 1990s, levels of environmental conditions and social development continuously increased from 1990 to 2014, especially social development after 1995.
(3) Economic development was an important driver of urbanization, social development and environmental degradation in Asian Russia, with economic development having a stronger influence on urbanization than on social development or environmental degradation. Nevertheless, urbanization exerted more direct influence than economic development on environmental degradation and social development in Asian Russia. Urbanization also affected environment degradation more than social development.

Acknowledgments
We would like to acknowledge the financial support from the National Aeronautics and Space Administration (NASA)'s Land-Cover and Land-Use Change (LCLUC) Program through its grant to Michigan State University (MSU) (NNX15AD51G). In addition, the Center for Global Connections (CGC) at MSU funded Peilei Fan's field trip to Russia in 2015. The work is also supported by the Russian Government Program of Competitive Growth of Kazan Federal University and US NSF Grant 1717770. We thank all interviewees for providing insights on cities in Asian Russia. Peilei Fan thanks faculty members of Russian Studies Program at MSU, Drs Zarema Kumakhova, Shannon Spasova, and Jason Merrill for their help in understanding Russian language, culture, and history. We thank Connor Crank for editing the manuscript and the anonymous reviewers for their valuable comments. This paper was presented at the Session M-IS01 'Environmental, socioeconomic and climatic changes in Northern Eurasia' at the joint JpGU-AGU meeting in May 2017 in Makuhari, Chiba, Japan and a synthesis workshop on 'Coupled Human Nature System in Mongolian Plateau' and 'Urbanization and Sustainability Under Global Change and Transitional Economies in Southeast, East, and North Asia (SENA)' in June 2017 in Ulaanbaatar. Any opinions, findings, and conclusions or recommendations expressed in this paper are those of the authors and do not necessarily reflect the views of NASA, NSF, the Russian Government, or CGC of MSU.

Appendix A. Key literature on Asian Russia relating to urbanization and sustainability
There are only a few key literatures on urbanization and sustainability in Asian Russia, providing us a brief background of urbanization, population dynamics, and land transitions in the region, particularly during the transitional economy. Through a comprehensive review on urbanization and urban growth patterns in Russia from 1897 through 2010, Becker et al (2014) underlined a significant shift in population within the region following the transition from central planning to the market economy, as people moved away from the far north and east to the south and west: 'from industrial cities with few amenities, towards larger cities and emerging service centers' (pp 100). Historical changes in land use in Asian Russia from the 12th century to 2017 were described by Prishchepov et al (2018). Similarly, they highlighted the rapid response of agricultural land use to the major socio-political disruption brought on by the dissolution of the Soviet Union in 1991. Focusing on the transitional period and based on data derived from satellite images of nighttime light from 1992 to 2014, Fan et al (2017b) found that urban land with active economic activity in Siberia decreased in the 1990s and then increased in the 2000s. Despite the useful background information provided by the above-mentioned studies, we know very little about urban dynamics and the linkages with different dimensions of sustainability for Asian Russia.
In addition, some studies have focused on the regional economy, such as the level of economic development of the region and sub-regions and possible drivers. For example, using ten indicators, Slepneva et al (2016) assessed the level of socioeconomic development of regions in Russia and identified the regions of Krasnoyarsk and Omsk as having high levels of socioeconomic development within the Siberia Federal District in 2013. Furthermore, inequality of income distribution has been found to have a significant impact on Russia's regional economic growth from 2006 to 2010 (Grigor'ev 2009). In addition, the regional economy has been found affected by globalization, both in the past and in near future. For example, foreign investments from China had been found significant in the Asian Russia based on data from 1999 to 2013 (Kuznetsova 2016). After estimating the spatial structure of the economies of Siberia under two different hypothetical scenarios: the entry of Russia into the ranks of developed countries, and a scenario that relies on the potential for self-development of the regions, it was emphasized that the future of Siberia will be determined by how well it is integrated into international economic development (Suspitsyn 2012).
Other studies have focused on non-economic aspect of development by examining the overall quality of life within the regions, and how they have been affected by various natural and socioeconomic shocks (Yaremenko 2005, Rimashevaskava 2010, Bashalkhanova et al 2012. For example, from 1990 to 2009, the quality of life in municipal districts of East Siberia declined under conditions of extreme climate severity (coldness) and inadequate infrastructure for transport and services, with a correlation between the low subsistence level of the northern territories and an

B.2. Urban built-up land
We used the GHS built-up grid to estimate urban built-up land of the six case cities because of its wide usage as a data source by the scientific community and its accuracy (Pesaresi et al 2015). GHS built-up grid was derived from Landsat images that provide multitemporal coverage (1975, 1990, 2000, and 2014). Using 3826 sample raster tiles collected from cartography data (which covers an area of 133 909 km 2 and a total built-up area of 4656 km 2 ), the GHS built-up grid was assessed to have an overall accuracy of 0.899 (Pesaresi et al 2015). In this study, we extracted the built-up land distribution within the administrative boundary of each city for 1990, 2000, and 2014. An exception is Khabarovsk, for which we extracted the built-up area within a square polygon that incorporated >95% built-up area because the administrative boundaries of Khabarovsk were too large and not comparable with other cities.

B.3. Socioeconomic, demographic, and other statistical data
We relied on the official statistics of Russian government for socioeconomic, demographic, and other statistical data due to widely acknowledged reliability and availability. The main data sources came from the Socioeconomic Indicator of Russian Regions and Census data of the Federal State Statistics Service of Russian Federation for 2004-2015. In addition to secondary data source, we conducted field interviews in Yekaterinburg, Novosibirsk, Krasnoyarsk, and Irkutsk in May 2015 to offer additional insights on drivers, patterns, and impacts and urbanization and sustainability. In each city, we conducted five to six semi-structured interviews of local experts in urbanization, economic development, urban environment, or scholars with relevant expertize. We developed an interview guide for the semi-structured interviews: a list of questions and topics that needed to be covered. The interviewees followed the guidelines, but they were encouraged to expand the discussion during the interviews when appropriate. Each interview lasted 1-2 h, during which time we asked the local experts to describe the urban development stages and main driving forces of each stage, to draw an illustrative map of the city's spatial structure, and to enumerate the main environmental problems and social challenges.   Elvidge et al (2013).

Satellite
Year These variables (EMP, AP , pc PHY , d and LS pc ) were normalized into unit-free variables XI, i.e., EMPI, API , pc PHYI , d and LSI , pc so that their respective values would be between 0 and 1, using the following formula: where min(X) and max(X) are the minimum and maximum values of variable X, respectively.

C.2. Structural equation modeling (SEM)
SEM is a statistical method to analyze multiple relationships among variables (Grace 2006). Because of its capacity to include both observable and unobservable variables that represent underlying relationships in the model, it has been widely used in various disciplines such as science, business, and education (Hair et al 2014. The SEM model in this paper reflects the relationship between urbanization, economic development, environmental degradation, and social development. Our underlying hypothesis is that economic development affects urbanization and social development positively, i.e., urbanization will lead to the improvement of social development (Davis 1965, Ranis et al 2000, Deaton 2008) but will also lead to environment degradation (Dinda 2004, Stern et al 1996. The SEM has the following structure: economic development is a potential driver affecting urbanization, environmental degradation, and social development, whereas urbanization also exerts its influence on environmental degradation and social development. All the variables used in the SEM were also used in the calculation of RSI and its three subcomponents. In our SEM, ECO and urbanization (Urb) were directly modeled by the observable variables of employment rate and the ratio of urban population to the total population (Davis 1965, Henderson 2003. ENV was directly modeled by an observable variable of air pollutants per capita from the stationary sources (Seldon and Song 1994, Carson et al 1997). Social development was modeled as a latent variable SD by the number of physicians per 10 000 people (PHY d ) and living space per capita (LS pc ). Education, health care, and quality of life have been recognized as the main areas of social development (Atkinson 1996). Due to data limitations, we chose two indicators to represent health care provision and quality of life, illustrated by the number of physicians per 10 000 people and living space per capita, which have been used frequently by other researchers as indicators of social development (e.g., Wu 2002, Macinko et al 2003, Chan et al 2002, Park et al 2017. We fit the model with all data during 1992-2014 in Mplus (version 8.11)-computer software designed for SEM. The default maximum likelihood estimator was used for estimating model coefficients, while robust standard errors for these coefficients were computed using the sandwich estimator known as the Taylor expansion of Huber-White (Muthen and Satorra 1995). To test model fit, we used the comparative fit index (CFI), the Tucker-Lewis index (TLI), and root mean square error (RMSE). CFI and TLI measure the improved fitness of a target model relative to the null model (a model in which the variables are assumed to be uncorrelated), while RMSE measures the actual differences between corresponding elements of the observed and predicted covariance matrix. We did not adopt a X 2 test because it is sensitive to sample size (Gerbing and Anderson 1992). Values of CFI and TLI larger than 0.97 and 0.95 were regarded as acceptable fit, respectively (Gerbing andAnderson 1992, Schermelleh-Engel et al 2003). RMSE between 0.05 and 0.08 further suggested a good model (Cangur and Ercan 2015). Our model was tested to have CFI of 0.99, TLI of 0.97, and RMSE of 0.06, indicating satisfactory model fitness.