Evolution of the Process of Urban Spatial and Temporal Patterns and Its Influencing Factors in Northeast China

Case Study

Evolution of the Process of Urban Spatial and Temporal Patterns and its Influencing Factors in Northeast China

Jiawen Xu1; Jianjun Zhao2; Hongyan Zhang3; and Xiaoyi Guo4

Abstract: The evolution of urban temporal and spatial patterns in Northeast China is complicated. In order to study the urbanization process in this area, explore the spatial and temporal laws of urban development in Northeast China, and find the main influencing factors affecting urban development in Northeast China, DMSP/OLS images are used as data sources. Urban built-up areas in Northeast China from 1993 to 2013 are extracted and temporal and spatial patterns of urban development are studied. Combining the economic, population, industrial structure, ecological and other statistical data, a geographical detector is applied to study the main influencing factors of urban development in Northeast China. According to a selection of 10 typical cities, the annual urban expansion speed and the urbanization intensity index are calculated to quantitatively analyze the development of typical cities. The present study indicates that the urbanization process in Northeast China was slow during 1995–1996. In fact, except for Daqing, the other typical cities developed slowly before 2003. While the urbanization process accelerated after 2003, it reached to its maximum rate in 2010. Moreover, it is observed that from 1993 to 2013, centers of cities gradually moved to their regional centers. On the other hand, it is concluded that from 2004 to 2013, the regional gross domestic product (GDP), GDP of the secondary industry, gross industrial product, GDP of the tertiary industry and the total investment in fixed assets were main indicators of the urbanization that affected change in the urban built- up area in Northeast China. Among them, the regional GDP had the greatest impact on urban development. As an old industrial base in China, the secondary industry mainly drove urban development before 2010. It is concluded that urban development began to change from 2010 and the driving force for urban development gradually changed from industry to the tertiary industry. DOI: 10.1061/(ASCE) UP.1943-5444.0000606. © 2020 American Society of Civil Engineers. Author keywords: Urbanization; Spatial and temporal patterns; Geographical detector; Northeast of China; DMSP/OLS.

Introduction policy, the government encouraged and supported foreign capital absorption to develop an open economy. Therefore, industrial and The evolution of urban development in Northeast China is compli- economic development in coastal areas signiﬁcantly increased. On cated. At the beginning of the founding of the People’sRepublic the other hand, the Northeast region gradually lost its advantage in of China, urban development of Northeast China ranked among the economic development so that urbanization in this region the top in the country. This is attributed to the good industrial gradually declined. In order to revitalize old industrial bases in base, appropriate urban construction, and abundant resources in Northeast China, the government has repeatedly issued special this region (Wang and Han 2018). Moreover, this urban develop- policies and opinions since 2003. These policies have been highly ment made important contributions to the country’seconomicde- valued by the state and attracted remarkable attention in the coun- velopment. However, with the implementation of the opening-up try (Wang 2018). The revitalization of old industrial bases in Northeast China and the implementation of China’s “One Belt, One Road” strategy are important turning points in the process 1Postgraduate Student, Key Laboratory of Geographical Processes and Ecological Security in Changbai Mountains, Ministry of Education, School of urbanization in the Northeast China. of Geographical Sciences, Northeast Normal Univ., Changchun 130024, Nighttime light data is widely used to assess the impact of China. Email: [email protected] human activities on the urbanization, gross domestic product 2Associate Professor, Key Laboratory of Geographical Processes and (GDP) estimation (Wu et al. 2013; Sutton et al. 2007; Marx and Ecological Security in Changbai Mountains, Ministry of Education, School Ziegler Rogers 2017), population density (Kasimu et al. 2009; of Geographical Sciences, Northeast Normal Univ., Changchun 130024, Song et al. 2019), energy consumption (Chand et al. 2009; Xie China (corresponding author). ORCID: https://orcid.org/0000-0002-0336 and Weng 2016; Shietal.2016), society health (Wen et al. -5764. Email: [email protected]

Downloaded from ascelibrary.org by University of Birmingham on 07/25/20. Copyright ASCE. For personal use only; all rights reserved. 3 2012), urban development and the evolution of the spatial pattern Professor, Key Laboratory of Geographical Processes and Ecological ’ Security in Changbai Mountains, Ministry of Education, School of Geo- (Ma et al. 2015; Fan et al. 2014; Zheng et al. 2016), and the city s graphical Sciences, Northeast Normal Univ., Changchun 130024, China. hierarchical structure and the spatial model (Wu et al. 2014). Email: [email protected] Using Defense Meteorological Satellite Program/Operational 4Associate Professor, Key Laboratory of Geographical Processes and Linescan System (DMSP/OLS) data to study the economic devel- Ecological Security in Changbai Mountains, Ministry of Education, School opment level of a region, nighttime light data can be used as an in- of Geographical Sciences, Northeast Normal Univ., Changchun 130024, dicator of economic growth in certain conditions. Studies show that China. Email: [email protected] the nighttime light data can be an effective means of estimation for Note. This manuscript was submitted on September 18, 2019; approved social and economic indicators such as GDP (Shi et al. 2014; Letu on May 6, 2020; published online on July 22, 2020. Discussion period open until December 22, 2020; separate discussions must be submitted for indi- et al. 2015; Propastin and Kappas 2012). Moreover, it is found that vidual papers. This paper is part of the Journal of Urban Planning and there is a signiﬁcant positive correlation between light intensity and Development, © ASCE, ISSN 0733-9488. regional GDP (Henderson et al. 2009; Fan et al. 2019), and

J. Urban Plann. Dev., 2020, 146(4): 05020017 nighttime illumination can be used as an explanatory indicator of Northeast China in 2004 and 2013. Fig. 1 shows that the study urbanization dynamics (Ma et al. 2012). A large number of studies area has jurisdiction over 38 prefecture-level cities, one autono- have confirmed that there is a close relationship between changes in mous prefecture and one district. nighttime light and economic activities (Bennett and Smith 2017; Northeast China has abundant natural resources and fertile land, Forbes 2013). Using a time series of nighttime light data can assess which makes it an ideal place for cultivation. Furthermore, this area determinants of urban expansion and their relative importance has remarkable underground reserves of iron ore, coal, oil and other (Zhang and Su 2016). The study found a linear relationship between minerals, which are necessary for industrial development. urban population changes and nighttime illumination (Zhang and Seto 2011; Zhou et al. 2015). Although some scholars have used nighttime light data to study Data Source factors that affect urban expansion, they have not pointed out DMSP/OLS Data whether the factors that affect urban development have changed The DMSP satellite is a dedicated military meteorological satellite over time. This study uses a geodetector model to reveal changes affiliated with the US Department of Defense. It is mainly used in in the main influencing factors that have affected urban develop- research fields, including extraction of urban spatial information, ment in Northeast China over time. urban system evolution, environmental monitoring, population es- The main objective of the present study is to explore the tem- timation, and economic development. Furthermore, the OLS sensor poral and spatial patterns of urban development in Northeast on the DMSP satellite has a strong ability to capture lights, fires, China based on long-term sequence nighttime light data, compre- and gas flares at night. Because of its reasonable temporal and spa- hensively analyze the influencing factors of the urban development tial resolution, it is considered as an effective complementary data in Northeast China, and reveal the main influencing factors that af- source for monitoring large-scale urban expansion processes. fect urban development in Northeast China. Furthermore, it in- The present study utilizes the nonradiation calibration stabilized tends to provide a theoretical basis for urbanization research in DMSP/OLS nighttime light data from 1993 to 2013. The data is ob- other regions. It is expected that the obtained results will improve tained from the National Geophysical Data Center in the United the general understanding of driving the mechanism of urban ex- States (https://www.ngdc.noaa.gov/). The spatial resolution is pansion and provide a theoretical basis for the implementation of 1 km and the pixel gray value is 0 to 63. It should be indicated effective measures to promote economic growth such as urban that the larger the pixel gray value, the stronger the light brightness. system construction. It is of great significance to realize the com- prehensive revitalization of old industrial bases in Northeast Statistical Data “ ” China and to promote the opening of the Belt and Road strate- Considering the availability of data, the present study utilizes dif- gic construction to the north (Wen and Wu 2017). ferent national statistical data from 2004 to 2013, including China Regional Statistical Yearbook data, Liaoning Statistical Yearbook data, Jilin Statistical Yearbook data, Heilongjiang Statis- Study Area and Methods tical Yearbook data, and Inner Mongolia Statistical Yearbook data, involving annual data on economy, population, industrial structure, and ecology of 40 cities in Northeast China. Study Area The study area of the present study is Northeast China, including Methods the eastern region of Inner Mongolia Autonomous Region, Hei- longjiang Province, Jilin Province, and Liaoning Province, located Based on the extracted data of urban built-up areas in Northeast between 115°05′∼135°02′E and 38°40′∼53°34′N(Wang et al. China, the spatial-temporal laws of urban development in Northeast 2019). The area of the cities in Northeast China is about 1.52 mil- China were analyzed, and the geographical detector model was used lion km2 and the population is about 130 million. Tables 1 and 2 to analyze the factors affecting urban development in Northeast show the resident inhabitants, urbanization area, province area, ur- China in combination with economic, population, industrial structure, banization density and population density of the provinces in and ecological statistics. According to the types of cities, 10 typical

Table 1. Overview of provinces in Northeast China in 2004 Resident inhabitants Urbanization area Province area Urbanization Population density Province (million person) (km2) (million km2) density (%) (person/km2) Liaoning 41.7290 2,606.35 0.1475 1.77 282.91 Jilin 26.6195 1,270.57 0.1891 0.67 140.77

Downloaded from ascelibrary.org by University of Birmingham on 07/25/20. Copyright ASCE. For personal use only; all rights reserved. Heilongjiang 37.6040 2,697.90 0.4548 0.59 82.68 Inner Mongolia 11.8424 483.89 0.4577 0.11 25.87

Table 2. Overview of provinces in Northeast China in 2013 Resident inhabitants Urbanization Province area Urbanization Population density Province (million person) area (km2) (million km2) density (%) (person/km2) Liaoning 42.3800 8,840.28 0.1486 5.95 285.20 Jilin 26.7856 5,253.94 0.1874 2.80 142.93 Heilongjiang 37.7940 9,095.88 0.4730 1.92 79.90 Inner Mongolia 11.5672 2,936.17 0.4624 0.63 25.02

J. Urban Plann. Dev., 2020, 146(4): 05020017 Fig. 1. Map of the study area.

as follows (Fan et al. 2019):

2 DNcorrect = a × DN + b × DN + c (1)

where DN and DNcorrect = gray values before and after the calibration, respectively. Moreover, a, b, and c are real constants.

Extraction of Urban Area Recently, defense meteorological satellite program DMSP/OLS nighttime light data has developed a more mature urban built-up area extraction method (Song et al. 2011). For example, based on the DMSP/OLS, the threshold method can be applied to extract urban built-up areas in different regions (Sutton et al. 2001; Imhoff et al. 1997). Moreover, stabilized DMSP nighttime light data and ra- diometric calibration images can be utilized to delineate the boundary of the built-up area of cities with different levels of urbanization and economic development and calculate the optimal threshold. In this way, the difference between the extracted urban boundaries can be minimized (Henderson et al. 2003). Based on the DMSP/OLS and urban land statistics, the optimal threshold was calculated to extract Fig. 2. Flowchart. the built-up area of the city and the clustering method was used to per- form optimal threshold estimation and city range mapping for different urban agglomerations (Zhou et al. 2014). cities were selected to quantitatively analyze the development of typical cities by calculating the annual urban expansion speed and urban- Weighted Average Model Downloaded from ascelibrary.org by University of Birmingham on 07/25/20. Copyright ASCE. For personal use only; all rights reserved. ization intensity index. The ﬂowchart is shown in Fig. 2. The weighted average model based on the gray value of the nighttime light data was applied in the present study to calculate the cen- DMSP/OLS Data Correction ter of the urban built-up area. This model can be mathematically Since the utilized DMSP/OLS nighttime light data is derived from described as the following (Wu et al. 2018): sensors of different satellites, data should be pre-processed in ad- n n vance. Because different sensors have different abilities in detect- = i=1 wixi = i=1 wiyi XG n YG n (2) ing light, in order to reduce the annual variation and difference i=1 wi i=1 wi between different sensors and to improve the continuity and com-

parability of the nighttime light data, the second-order regression where XG and YG = coordinates of the center of the urban built-up model was applied to calibrate the DMSP/OLS data prior to the ex- area; wi = the gray value of point i; xi, yi = geographic coordinates traction of the urban built-up area data. The calibration equation is of point i.

J. Urban Plann. Dev., 2020, 146(4): 05020017 The Geodetector Model city is shrinking. The equation is as follows (Shi et al. 2018): The geodetector model can be used to detect differences in the spatial distribution of a phenomenon and find the affecting factors. Re- BuAn+i − BuAn sults obtained by the factor detector reflect the influence degree of a UESn to n+i = (4) factor on the spatial differentiation of a phenomenon (Wang and Xu i 2017). The equation is as follows: where UESn to n+i, BuAn and BuAn+i = annual rate of change of the + L σ2 urban built-up area from n to n i years and total area of urban h=1 Nh h + 2 Qi = 1 − (3) built-up areas in n years and n i years respectively, unit is km ; Nσ2 i = the corresponding year.

… = where i, Qi and h (1, 2, 3, , L) the urbanization factor that af- Urbanization Intensity Index fects the change of the built-up area, inﬂuence degree of the i factor The urbanization intensity index (UII) was calculated to describe on the change of the built-up area and the classiﬁcation of urbani- the dynamic level of the urbanization quantitatively. It showed = zation factors, respectively; Nh and N the number of units of class that the larger the value, the faster the urbanization during this pe- σ2 σ2 = h and the whole zone, respectively; h and variances of the riod. The corresponding equation is as follows (Shi et al. 2018): built-up area of the class h and the whole area, respectively. It should be indicated that Qi varies from 0 to 1. BuAn+i − BuAn UII + = × 100% (5) n to n i B A Urban Expansion Speed u n The urban expansion speed is used to analyze the existing status of the typical urban expansion. When the obtained value is higher than where UIIn to n+i, BuAn and BuAn+i = urbanization intensity index zero, this indicates that the city is expanding during this period. On between n and n + i years and the total area of urban built-up the other hand, if the value is less than zero, this indicates that the areas in n and n + i years, respectively. Downloaded from ascelibrary.org by University of Birmingham on 07/25/20. Copyright ASCE. For personal use only; all rights reserved.

Fig. 3. Distribution of annual urban built-up areas in Northeast China in 1993.

J. Urban Plann. Dev., 2020, 146(4): 05020017 Linear Regression Model Temporal and Spatial Patterns of Urban Development Compared with a second-order regression model and log–log re- Fig. 6 illustrates the variation trend of the urban area in Northeast gression model, the linear regression model to fit the built-up China from 1993 to 2013. It is observed that the size of the area and regional GDP is more accurate and easier to implement. urban built-up area has a steady upward trend from 1993 to Therefore, the linear regression model was applied in the present 2006. However, the urban area increased significantly since 2006 study, where the corresponding equation is as the follows (Shi – et al. 2014): and the growth rate reached its maximum value in 2010 2011. Then the increase rate gradually slowed down. A = kG + b (6) According to administrative divisions, we selected 38 prefecture- level cities, one autonomous prefecture, and one region in Northeast where A = area of the built-up area of the city; G = regional GDP; China, giving a total of 40 cities. Based on the light intensity values k = regression coefficient; and b = intercept determined by the re- obtained from the nighttime light data, the weighted average model gression analysis based on the regional GDP. was applied to calculate the center of the built-up areas of 40 cities in the Northeast China from 1993 to 2013. The obtained city centers were then connected with a smooth curve in a chronological order. It Results was found that the center of most cities in the Northeast China remained basically unchanged during 1995–1996. Fig. 7 shows the distribution of the center of 40 cities in Boundary of Urban Northeast China in 1993 and 2013. These locations are obtained In the present study, dynamic thresholds were used to extract in accordance with the foregoing calculations. It is observed that boundaries of urban built-up areas in 40 cities of Northeast China centers of 15 cities, that is, Anshan, Baishan, Benxi, Daqing, from 1993 to 2013. Figs. 3–5 illustrates that the boundary light Daxinganling, Fushun, Fuxin, Liaoyang, Liaoyuan, Panjin, threshold was applied to calculate the average value of each county Qitaihe, Shenyang, Tieling, Yichun, and Yingkou have changed and find the boundaries of the built-up area of each county. slightly, while the centers of Baicheng, Chaoyang, Changchun, Downloaded from ascelibrary.org by University of Birmingham on 07/25/20. Copyright ASCE. For personal use only; all rights reserved.

Fig. 4. Distribution of annual urban built-up areas in Northeast China in 2003.

J. Urban Plann. Dev., 2020, 146(4): 05020017 Fig. 5. Distribution of annual urban built-up areas in Northeast China in 2013.

Shuangyashan, Suihua, Tonghua, and Tongliao have the largest change. Moreover, it was observed that the center of Liaoyang remained basically unchanged from 1993 to 2013. On the other hand, it was found that the city centers of Anshan, Chaoyang, Daqing, Heihe, Huludao, Panjin, Songyuan, Xing’an League, and Yanbian Prefecture moved to the southwest, while the city centers of Benxi, Changchun, Dalian, Dandong, Fushun, Fuxin, Hegang, Hulunbeier, Jixi, Jiamusi, Jinzhou, Mudanjiang, Qitaihe, Qiqihar, Shuangyashan, Siping, Suihua, and Tieling moved to the northeast. Fig. 5 indicates that the city centers of Baicheng, Daxinganling, Liaoyuan, Jilin, and Tonghua moved to the southeast, while the city centers of Chifeng, Shenyang, Tongliao, and Yichun moved to the northwest. Finally, it was observed that the center of Harbin Downloaded from ascelibrary.org by University of Birmingham on 07/25/20. Copyright ASCE. For personal use only; all rights reserved. city moved eastward, the center of Yingkou city moved southward, and the center of Baishan moved northward. The direction and distance of the shift are shown in Table 3. In general, it is concluded that the centers of cities gradually moved to their regional centers Fig. 6. Area change of the urban built-up area in Northeast China. from 1993 to 2013.

Dalian, Dandong, Harbin, Hegang, Huludao, Jilin, Jinzhou, Influencing Factors of the Urban Development Mudanjiang, Siping, Songyuan, Xing’an League, and Yanbian In this section, numerous urbanization indicators, including the have changed signiﬁcantly. It should be noted that the centers total population at the end of the year, the regional GDP, per capita of Chifeng, Heihe, Hulunbeier, Jixi, Jiamusi, Qiqihar, GDP, GDP of the primary, secondary, and the tertiary industries,

J. Urban Plann. Dev., 2020, 146(4): 05020017 Fig. 7. Distribution of the center of cities in Northeast China from 1993 to 2013.

Table 3. The direction and distance of the shift of the centroid of urban area City Direction Distance (km) City Direction Distance (km) Shenyang Northwest 5.77 Songyuan Southwest 18.57 Dalian Northeast 19.47 Baicheng Southeast 27.71 Anshan Southwest 10.76 Yanbian Southwest 19.71 Fushun Northeast 6.40 Harbin East 28.64 Benxi Northeast 3.71 Qiqihar Northeast 34.92 Dandong Northeast 8.88 Jixi Northeast 65.48 Jinzhou Northeast 18.60 Hegang Northeast 25.17 Yingkou South 8.64 Shuangyashan Northeast 42.91 Downloaded from ascelibrary.org by University of Birmingham on 07/25/20. Copyright ASCE. For personal use only; all rights reserved. Fuxin Northeast 12.15 Daqing Southwest 10.17 Liaoyang Northeast 1.77 Yichun Northwest 10.25 Panjin Southwest 9.22 Jiamusi Northeast 84.99 Tieling Northeast 5.88 Qitaihe Northeast 3.61 Chaoyang Southwest 38.74 Mudanjiang Northeast 20.25 Huludao Southwest 20.38 Heihe Southwest 45.79 Changchun Northeast 13.82 Suihua Northeast 34.40 Jilin Southeast 14.52 Daxinganling Southeast 5.71 Siping Northeast 17.12 Hulunbeier Northeast 27.43 Liaoyuan Southeast 8.38 Xinganmeng Southwest 15.21 Tonghua Southeast 47.00 Tongliao Northwest 42.29 Baishan North 5.86 Chifeng Northwest 38.43

J. Urban Plann. Dev., 2020, 146(4): 05020017 Table 4. Urbanization factor detection result Primary Secondary Tertiary GDP per Investment in Population GDP industry industry Industry industry capita ﬁxed assets

Year Q P QPQ P QPQPQ P Q P Q P 2004 0.37 0.01 0.75 0 0.33 0.09 0.76 0 0.76 0 0.49 0.01 0.49 0 0.58 0 2005 0.38 0.01 0.84 0 0.28 0.49 0.75 0 0.75 0 0.59 0 0.50 0 0.70 0 2006 0.39 0.01 0.86 0 0.32 0.19 0.85 0 0.73 0 0.61 0 0.44 0.01 0.47 0.05 2007 0.40 0.01 0.85 0 0.33 0.18 0.84 0 0.79 0 0.63 0 0.45 0 0.50 0.04 2008 0.40 0.01 0.84 0 0.32 0.19 0.67 0 0.76 0 0.60 0 0.40 0.01 0.48 0.05 2009 0.40 0.01 0.69 0 0.29 0.22 0.65 0 0.65 0 0.63 0 0.37 0.02 0.56 0 2010 0.47 0 0.80 0 0.34 0.16 0.62 0 0.62 0 0.71 0 0.35 0.03 0.62 0 2011 0.48 0 0.79 0 0.25 0.16 0.59 0 0.59 0 0.72 0 0.31 0.08 0.64 0 2012 0.48 0 0.78 0 0.24 0.18 0.63 0 0.63 0 0.76 0 0.39 0.08 0.59 0 2013 0.48 0 0.77 0 0.36 0.03 0.61 0 0.61 0 0.72 0 0.30 0.08 0.63 0

Table 5. Fitting function of area of the typical urban built-up area and the including Dalian and Jilin, the second largest city in Jilin Province, regional GDP were selected in this regard (Wang et al. 2019). City Formula R2 The nighttime total illumination and urban built-up area of typical cities from 2004 to 2013 were obtained through nighttime light = − Changchun y 0.3553x 181.76 0.9319 data. Moreover, Table 5 shows that the linear regression method Jilin y = 0.3907x − 53.171 0.9445 was applied to obtain the fitting function between the typical Songyuan y = 0.2998x − 38.799 0.9691 = + urban built-up area and the regional GDP. It was observed that Shenyang y 0.2557x 27.864 0.9470 2 Dalian y = 0.1792x + 76.333 0.9473 the fitting function of Yichun corresponds to the smallest R , 2 Harbin y = 0.4182x − 189.82 0.9477 while the R of other nine typical cities are higher than 0.9. Jixi y = 1.1088x − 176.44 0.9260 The urban expansion speed was applied to calculate the annual Hegang y = 1.2586x − 78.307 0.9021 rate of change of the built-up area of typical cities from 1993 to Daqing y = 0.3338x + 338.07 0.9474 2013. The UII was calculated from 1993 to 2013 to quantitatively = − Yichun y 2.088x 50.307 0.8667 describe the dynamic level of the urbanization. Fig. 8 shows the annual change rate and the urbanization intensity index curve of typical cities. gross industrial production value, proportion of the secondary and It was observed that the annual change rate of typical cities tertiary industries, total investment in fixed assets and the green from 1993 to 2013 was similar to the trend of the urbanization fl coverage rate in built-up areas were selected in 40 cities in North- intensity index. The index uctuated before 2005, while it gener- east China from 2004 to 2013. Moreover, the geodetector method ally increased from 2005 to 2010 and declined from 2010 to was used to calculate the degree of spatial differentiation of the 2013. Moreover, it was found that the maximum annual rate of urban built-up area. It showed the larger the value, the stronger change in each typical city occurred in 2010. It should be indi- ’ the explanatory power of the factor for the change of the built-up cated that Daqing s urbanization intensity index reached its high- area, and the greater the corresponding impact on the development est value in 1994, and reached the second highest value in 2002. of the city. Table 4 illustrates the detection results. However, the urbanization intensity index of Daqing in 2010 was It should be noted that Q and P are the influence degree of the significantly smaller than other typical cities. On the other hand, factor on the built-up area and a parameter for evaluating the hy- Songyuan’s urbanization intensity index reached its highest pothesis test results, respectively. Moreover, the smaller the P value in 1995 and reached the second highest value in 2010. value, the more significant the results. However, the difference was not obvious. Moreover, the urbani- Table 4 shows that the regional GDP, GDP of the secondary in- zation intensity index of the remaining typical cities shows max- dustry, gross industrial product, GDP of the tertiary industry, and imum value in 2010. the total investment in fixed assets are main indicators of the urban- The annual change rate and the urbanization intensity index of ization that affect the urban built-up area in Northeast China. the built-up area of typical cities in 1993–2003 and 2003–2013 Among them, regional GDP has the greatest impact on the urban were calculated and Table 6 illustrates the obtained results. development. From 2004 to 2009, the impact of GDP of the sec- It was observed that the annual change rate of the built-up – Downloaded from ascelibrary.org by University of Birmingham on 07/25/20. Copyright ASCE. For personal use only; all rights reserved. ondary industry and the industrial production on the urban develop- area of each typical city in 2003 2013 was greater than that in ment was only less than the regional GDP. It is observed that since 1993–2003. Moreover, the urbanization intensity index of Daqing 2010, the impact of the tertiary industry on the urban development from 1993 to 2003 was greater than that from 2003 to 2013, while has gradually surpassed the impact of the secondary industry on the other cities have contrasting pattern for this issue. It was found urban development. that the urbanization intensity index of Songyuan and Yichun from 1993 to 2003 was less than 1, while such index of other typical cities exceeded 1 and Daqing city had the largest urbanization Analysis of Typical Cities intensity index. It was found that the urbanization intensity index In this section, cities from different categories were selected for the of each typical city in 2003–2013 was higher than 1. It should be analysis. Oil-based cities including Daqing, Songyuan, coal-based noted that Daqing had the smallest urbanization intensity index, cities including Jixi, Hegang, industrial cities including Harbin, while the index for Jilin, Songyuan, Jixi, and Hegang cities Changchun, Shenyang, the forestry city of Yichun, coastal cities were larger.

Fig. 8. Variations of annual change rate of the urban built-up area and urbanization intensity index of typical cities.

J. Urban Plann. Dev., 2020, 146(4): 05020017 Table 6. Annual change rate of the urban built-up area and the so that the economic structure of Northeast China is still dominated corresponding urbanization intensity index by industry and resource-based industries, and the tertiary industry has developed slowly (Wang and Han 2018). This is inaccurate, UES1993–2003 UES2003–2013 2 2 and the results we calculated using the geographical detector are City (km ) (km ) UII1993–2003 UII2003–2013 more convincing. Changchun 25.2828 107.5994 1.6026 2.6206 Jilin 12.3877 72.6523 1.1220 3.1010 Songyuan 3.3782 37.3720 0.8990 5.2373 Analysis of Typical Cities Shenyang 33.2601 124.2551 1.4644 2.2199 Dalian 26.4991 95.8403 1.6976 2.2760 The foregoing section showed that the R2 of the fitting function of Harbin 32.8060 130.5259 1.3485 2.2846 the built-up area of Yichun and the regional GDP are the small- Jixi 3.5730 38.3412 1.1823 5.8136 est, which may be attributed to the type of city. In other words, Hegang 5.2222 28.2982 1.9287 3.5686 Yichun is a forest-based city and the secondary industry is rela- Daqing 55.2516 95.7695 3.1915 1.3198 tively weak. From 2004 to 2013, the change of the built-up area Yichun 4.7978 34.1807 0.4687 2.2734 of Yichun was relatively small so that the urban expansion rate was relatively slow. This is consistent with the results obtained Discussion by Zhao et al. (2019), which showed that Yichun’s urbanization rate is slower than that of industrial-dominated Harbin and energy-dominated Daqing. Temporal and Spatial Patterns of the Urban Development It was observed that the annual change rate and the urbaniza- In terms of time, the area of urban built-up areas in Northeast China tion intensity index of typical cities fluctuated before 2005. How- has been on the rise since 1993–2013, and the rate of increase in ever, from 2005 to 2013, both the annual change rate and the urban area has first increased and then decreased. Moreover, it urbanization intensity index of typical cities first increased and was found that the growth rate reached the maximum in 2010– then decreased. Moreover, the annual rate of change of each typ- 2011, indicating the fast urbanization process in 2010–2011. Con- ical city reached its peak in 2010 and the urbanization intensity sistent with the results obtained by Wang and Han (2018), the de- index of other typical cities except Daqing and Songyuan maxi- velopment of the old industrial bases in Northeast China has mized in 2010. The annual change rate of the built-up area of – – improved significantly since 2008, which means that the state’s each typical city in 2003 2013 was greater than that of 1993 support policies for old industrial bases in Northeast China have 2003. The urbanization intensity index of all typical cities except – – played a positive role in promoting regional economic growth. Daqing in 2003 2013 was greater than that of 1993 2003 and – In terms of space, it was found that the center of cities in Liao- such index of each typical city in 2003 2013 was higher than ning Province slightly changed from 1993 to 2013, while the center 1. This shows that most typical cities developed slowly before of cities in Jilin Province changed a great deal. It was observed that 2003 and the urbanization process accelerated after 2003. More- the cities in Heilongjiang Province and Inner Mongolia Autono- over, the fastest urbanization process was in 2010, which may mous Region had the most variations in this issue. This may be at- originate from the national revitalization policy. ’ tributed to the area of each city, where the area of the cities in Daqing s urbanization intensity index reached its highest value in Liaoning Province is generally small, while the area of the cities 1994, and the second highest value occurred in 2002. Moreover, it in Heilongjiang Province and Inner Mongolia Autonomous Region was found that the urbanization intensity index of Daqing in 2010 was significantly smaller than that of other typical cities. The UII is generally larger. During 1995–1996, the center of most cities in of Daqing from 1993 to 2003 was greater than that from 2003 to Northeast China remained basically unchanged, indicating that the 2013. Compared with other typical cities, Daqing’s urbanization in- urbanization process in Northeast China was slow in 1995–1996. tensity index was the largest in 1993–2003, while it was the smallest From the perspective of the urban center migration process, the in 2003–2013. This shows that the urbanization process in Daqing centers of cities gradually moved to their regional centers. was faster before 2003, while the urban development slowed down after 2003. This is consistent with the results obtained by Zhao Influencing Factors of the Urban Development et al. (2019). Daqing is an important oil- and energy-based city, relying mainly on the industry to drive the economic development. Con- The regional GDP, GDP of the secondary industry, gross industrial sidering the previously mentioned analysis of factors affecting urban product, GDP of the tertiary industry, and the total investment in development in Northeast China, it was concluded that the driving fixed assets are the main indicators of urbanization that affect the force affecting urban development is changing. change of the urban built-up area in Northeast China. Among Based on the results obtained from the analysis of typical cities, these factors, the regional GDP has the greatest impact on urban de- we infer that the urban development in the northeast region was velopment. From 2004 to 2009, the impact of the GDP of the sec- slow before 2003, the urbanization process accelerated after

Downloaded from ascelibrary.org by University of Birmingham on 07/25/20. Copyright ASCE. For personal use only; all rights reserved. ondary industry and industrial production on the urban 2003, and the urbanization process was the fastest in 2010. After development was only less than the regional GDP. It shows that 2010, urban development slowed down, and the driving force af- the secondary industry occupies a dominant position in the urban fecting urban development changed from secondary industry to ter- development. As an old industrial base in China, secondary indus- tiary industry. Marketization and opening to the outside world are try has an important role in the development of cities in the North- effective mechanisms for promoting economic growth. However, east. However, since 2010, the impact of the tertiary industry on the the degree of marketization in the northeast is low and it is difﬁcult urban development has gradually surpassed that of the secondary to fundamentally change the industrial structure. In order to pro- industry. It shows that the urban development pattern has changed mote the development of the northeast economy, we should give since 2010 and the driving force for the urban development has full play to the characteristics of the northeast’s abundant resources gradually changed from the industry to the tertiary industry. to vigorously develop the processing and manufacturing industries, Based on the fact that the proportion of the secondary industry in and promote the optimization of the industrial structure with new GDP in Northeast China is greater than that of tertiary industry, technologies and intelligent manufacturing.

J. Urban Plann. Dev., 2020, 146(4): 05020017 Conclusions the Fundamental Research Funds for the Central Universities (Grant No. 2412019BJ001), the Foundation of the Education De- The main conclusions of the present study are as follows: partment of Jilin Province in the 13th Five-Year project (Grant 1. The urban built-up areas in Northeast China have risen since No. JJKH20190282KJ), and the Science and Technology Develop- 1993–2013, and the rate of increase in the urban area first in- ment Project of Jilin Province (Grant No. 20190802024ZG). creased and then decreased. Moreover, the growth rate reached the maximum in 2010–2011, indicating the fastest urbanization process in this period. References 2. Centers of cities gradually moved to their regional centers from 1993 to 2013. Centers of cities in Liaoning Province had slight Bennett, M. M., and L. C. Smith. 2017. “Advances in using multitemporal variation, while the centers of cities in Jilin Province changed a night-time lights satellite imagery to detect, estimate, and monitor soci- great deal. Moreover, cities in Heilongjiang Province and Inner oeconomic dynamics.” Remote Sens. Environ. 192: 176–197. https://doi Mongolia Autonomous Region had the largest variations. Dur- .org/10.1016/j.rse.2017.01.005. ing 1995–1996, centers of most cities in Northeast China re- Chand, T. R. K., K. V. S. Badarinath, C. D. Elvidge, and B. T. Tuttle. 2009. mained basically unchanged, indicating the slow urbanization “Spatial characterization of electrical power consumption patterns over process in Northeast China in this period. India using temporal DMSP-OLS night-time satellite data.” 3. From 2004 to 2013, the regional GDP, GDP of the secondary in- Int. J. Remote Sens. 30 (3): 647–661. https://doi.org/10.1080 dustry, gross industrial product, GDP of the tertiary industry, and /01431160802345685. the total investment in fixed assets were the main indicators of ur- Fan, J., M. Ting, Z. Chenghu, Z. Yuke, and X. Tao. 2014. “Comparative estimation of urban development in China’s cities using socioeconomic banization that affected the change of the urban built-up area in ” – Northeast China. Among them, regional GDP had the greatest im- and DMSP/OLS night light data. Remote Sens. 6 (8): 7840 7856. https://doi.org/10.3390/rs6087840. pact on the urban development. Moreover, as an old industrial base Fan, L., J. Zhao, Y. Wang, Z. Ren, H. Zhang, and X. Guo. 2019. in China, the secondary industry occupied a dominant position in “Assessment of night-time lighting for global terrestrial protected and urban development before 2010. In 2010, urban development wilderness areas.” Remote Sens. 11 (22): 2699. https://doi.org/10 began to change and the driving force for urban development grad- .3390/rs11222699. ually changed from industry to the tertiary industry. Forbes, D. J. 2013. “Multi-scale analysis of the relationship between eco- 4. Except for Daqing, other typical cities developed slowly before nomic statistics and DMSP-OLS night light images.” GISci. Remote 2003. However, the urbanization process accelerated after 2003 Sens. 50 (5): 483–499. https://doi.org/10.1080/15481603.2013.823732. and reached its fastest rate in 2010, which may be attributed to Henderson, J., A. Storeygard, and D. Weil. 2009. “Measuring economic ” – the national revitalization policy. Before 2003, Daqing’s urban- growth from outer space. Am. Econ. Rev. 102: 994 1028. https://doi ization process was fast, but slowed down after 2003. Daqing is .org/10.2307/23245442. Henderson, M., E. T. Yeh, P. Gong, C. Elvidge, and K. Baugh. 2003. an important oil- and energy-based city, relying mainly on the “Validation of urban boundaries derived from global night-time satellite industry to drive the economic development. This indicates imagery.” Int. J. Remote Sens. 24 (3): 595–609. https://doi.org/10.1080 that the driving force affecting urban development is changing. /01431160304982. 5. Cities in Northeast China developed slowly before 2003, and the Imhoff, M. L., W. T. Lawrence, D. C. Stutzer, and C. D. Elvidge. 1997. “A urbanization process accelerated after 2003, and reached its fastest technique for using composite DMSP/OLS “city lights” satellite data to in 2010. After 2010, urban development slowed down, and the map urban area.” Remote Sens. Environ. 61 (3): 361–370. https://doi.org driving force affecting urban development changed from the sec- /10.1016/S0034-4257(97)00046-1. ondary industry to the tertiary industry. In order to promote the de- Kasimu, A., R. Tateishi, and N. T. Hoan. 2009. “Global urban characteri- ” velopment of the Northeast economy, we should give full play to zation using population density, DMSP and MODIS data. In Proc., – the characteristics of the Northeast’s abundant resources to vigor- 2009 Joint Urban Remote Sensing Event,1 7. Piscataway, NJ: IEEE. Letu, H., M. Hara, G. Tana, Y. Bao, and F. Nishio. 2015. “Generating the ously develop the processing and manufacturing industries, and nighttime light of the human settlements by identifying periodic compo- promote the optimization of the industrial structure with new tech- nents from DMSP/OLS satellite imagery.” Environ. Sci. Technol. nologies and intelligent manufacturing. 49 (17): 10503–10509. https://doi.org/10.1021/acs.est.5b02471. The main innovation and contribution of this research are: a qual- Ma, T., C. Zhou, T. Pei, S. Haynie, and J. Fan. 2012. “Quantitative esti- itative and quantitative study of the evolution of the spatial and tem- mation of urbanization dynamics using time series of DMSP/OLS poral patterns of cities in Northeast China, revealing the spatial and nighttime light data: A comparative case study from China’s cities.” temporal laws of urban process in the northeast. By calculating the de- Remote Sens. Environ. 124: 99–107. https://doi.org/10.1016/j.rse gree of influence of various urbanization indicators on urban develop- .2012.04.018. ment, the main factors affecting urban development were found, and it Ma, T., Y. Zhou, C. Zhou, S. Haynie, T. Pei, and T. Xu. 2015. “Night-time light derived estimation of spatio-temporal characteristics of urbaniza- provided a theoretical basis for the implementation of effective mea- ” sures to promote economic growth such as urban system construction. tion dynamics using DMSP/OLS satellite data. Remote Sens. Environ. 158: 453–464. https://doi.org/10.1016/j.rse.2014.11.022.

Downloaded from ascelibrary.org by University of Birmingham on 07/25/20. Copyright ASCE. For personal use only; all rights reserved. Marx, A., and M. Ziegler Rogers. 2017. “Analysis of Panamanian DMSP/OLS nightlights corroborates suspicions of inaccurate ﬁscal Data Availability Statement data: A natural experiment examining the accuracy of GDP data.” Remote Sens. Appl.: Soc. Environ. 8: 99–104. https://doi.org/10 All data, models, and code generated or used during the study ap- .1016/j.rsase.2017.08.005. pear in the published article. Propastin, P., and M. Kappas. 2012. “Assessing satellite-observed nighttime lights for monitoring socioeconomic parameters in the Republic of Kazakhstan.” GISci. Remote Sens. 49 (4): 538–557. https://doi.org /10.2747/1548-1603.49.4.538. Acknowledgments Shi, G., N. Jiang, Y. Li, and B. He. 2018. “Analysis of the dynamic urban expansion based on multi-sourced data from 1998 to 2013: A case study This research was funded by the National Natural Science Founda- of Jiangsu province.” Sustainability 10: 3467. https://doi.org/10.3390 tion of China (Grant Nos. 41771450, 41871330, and 41630749), /su10103467.

J. Urban Plann. Dev., 2020, 146(4): 05020017 Shi,K.,Y.Chen,B.Yu,T.Xu,C.Yang,L.Li,C.Huang,Z.Chen,R. Wu, J., H. Liu, J. Peng, and L. Ma. 2014. “Hierarchical structure and spa- Liu, and J. Wu. 2016. “Detecting spatiotemporal dynamics of global tial pattern of China’s urban system: Evidence from DMSP/OLS electric power consumption using DMSP-OLS nighttime stable light nightlight data.” J. Geog. Sci. 69: 759–770. https://doi.org/10 data.” Appl. Energy 184: 450–463. https://doi.org/10.1016/j.apenergy .11821/dlxb201406004. .2016.10.032. Wu, J., Z. Wang, W. Li, and J. Peng. 2013. “Exploring factors affecting the Shi, K., B. Yu, Y. Huang, Y. Hu, B. Yin, Z. Chen, L. Chen, and J. Wu. relationship between light consumption and GDP based on DMSP/OLS 2014. “Evaluating the ability of NPP-VIIRS nighttime light data to es- nighttime satellite imagery.” Remote Sens. Environ. 134 (7): 111–119. timate the gross domestic product and the electric power consumption https://doi.org/10.1016/j.rse.2013.03.001. of China at multiple scales: A comparison with DMSP-OLS data.” Wu, L., T. Zhou, W. Li, X. Liu, and X. Zhu. 2018. “Study on spatio- Remote Sens. 6 (2): 1705–1724. https://doi.org/10.3390/rs6021705. temporal relationship between PM2.5 and urbanization based on remote Song, J., X. Tong, L. Wang, C. Zhao, and A. V. Prishchepov. 2019. sensing data—A case study of Chengdu-Chongqing urban agglomera- “ ﬁ Monitoring ner-scale population density in urban functional zones: tion.” Resour. Environ. Yangtze Basin 27 (9): 244–254. https://doi.org ” A remote sensing data fusion approach. Landscape Urban Plann. /10.11870/cjlyzyyhj201809024. 190: 103580. https://doi.org/10.1016/j.landurbplan.2019.05.011. Xie, Y., and Q. Weng. 2016. “World energy consumption pattern as re- “ Song, S., B. Yu, J. Wu, and H. X. Liu. 2011. Methods for deriving urban vealed by DMSP-OLS nighttime light imagery.” Mapp. Sci. Remote ” built-up area using night-light data: Assessment and application. Sens. 53 (2): 265–282. https://doi.org/10.1080/15481603.2015 – Remote Sens. Technol. Appl. 26 (2): 169 176. .1124488. “ Sutton, P., C. Elvidge, and G. Tilottama. 2007. Estimation of gross domes- Zhang, Q., and K. C. Seto. 2011. “Mapping urbanization dynamics at re- ” tic product at sub-national scales using nighttime satellite imagery. gional and global scales using multi-temporal DMSP/OLS nighttime Int. J. Ecol. Econ. Stat. 8: 5–21. light data.” Remote Sens. Environ. 115 (9): 2320–2329. https://doi.org Sutton, P., D. Roberts, C. Elvidge, and K. Baugh. 2001. “Census from /10.1016/j.rse.2011.04.032. heaven: An estimate of the global human population using night-time Zhang, Q., and S. Su. 2016. “Determinants of urban expansion and their satellite imagery.” Int. J. Remote Sens. 22 (16): 3061–3076. https:// relative importance: A comparative analysis of 30 major metropolitans doi.org/10.1080/01431160010007015. in China.” Habitat Int. 58: 89–107. https://doi.org/10.1016/j.habitatint Wang, J., and C. Xu. 2017. “Geodetector: Principle and prospective.” Acta .2016.10.003. Geog. Sin. 72: 116–134. https://doi.org/10.11821/dlxb201701010. Zhao, X., Z. Meng, X. Zheng, J. Fan, and Q. Wang. 2019. “Research on Wang, S., and Y. Han. 2018. “Exploration of urbanization in Northeast China: Its development status and path.” J. Northeast Normal Univ. landscape spatial pattern estimation of cities in Northeast China using ” – (Philos. Social Sci.) 295 (5): 92–99. https://doi.org/10.16164/j.cnki.22 nighttime luminosity data. Sci. Surv. Mapp. 44 (4): 72 80. https:// -1062/c.2018.05.014. doi.org/10.16251/j.cnki.1009-2307.2019.04.012. “ Wang, Z. 2018. “Marketization, heavy industrialization and new phenom- Zheng, W., J. Run, R. Zhuo, Y. Jiang, and X. Wang. 2016. Evolution pro- enon of Northeast China—An analysis based on 37 prefectural panel cess of urban spatial pattern in Hubei Province based on DMSP/OLS ” – data of Northeast China.” Mod. Econ. Sci. 40 (5): 90–102. nighttime light data. Chin. Geog. Sci. 26: 366 376. https://doi.org/10 Wang, Z., J. Zhao, J. Xu, M. Jia, H. Li, and S. Wang. 2019. “Inﬂuence of .1007/s11769-016-0814-1. “ straw burning on urban air pollutant concentrations in Northeast Zhou, N., H. Klaus, and M. Roberts. 2015. Analysis of spatial patterns of China.” Int. J. Environ. Res. Public Health 16 (8): 1379. https://doi urban growth across South Asia using DMSP-OLS nighttime lights .org/10.3390/ijerph16081379. data.” Appl. Geogr. 63: 292–303. https://doi.org/10.1016/j.apgeog Wen, S., and T. Y. Wu. 2017. “How does Northeast China advance the belt .2015.06.016. and road initiative.” J. Jiamusi Vocat. Inst. 2: 481–481. Zhou, Y., S. J. Smith, C. D. Elvidge, K. Zhao, and M. Imhoff. 2014. “A Wen, W., C. Hui, and Z. Li. 2012. “Poverty assessment using DMSP/OLS cluster-based method to map urban area from DMSP/OLS nightlights.” night-time light satellite imagery at a provincial scale in China.” Adv. Remote Sens. Environ. 147: 173–185. https://doi.org/10.1016/j.rse.2014 Space Res. 49 (8): 1253–1264. https://doi.org/10.1016/j.asr.2012.01.025. .03.004. Downloaded from ascelibrary.org by University of Birmingham on 07/25/20. Copyright ASCE. For personal use only; all rights reserved.

J. Urban Plann. Dev., 2020, 146(4): 05020017