1990-2015における中国都市の開発パターンの類型化 と中心地域間の比較

九州大学学術情報リポジトリ

Kyushu University Institutional Repository

1990-2015における中国都市の開発パターンの類型化 と中心地域間の比較

李, 鶴

九州大学大学院人間環境学府都市共生デザイン専攻 : 学術研究員

趙, 世晨

九州大学大学院人間環境学研究院都市・建築学部門 : 教授

https://doi.org/10.15017/2344816

出版情報：都市・建築学研究. 36, pp.1-10, 2019-07-15. Faculty of Human-Environment Studies, Kyushu University

バージョン：

権利関係：

f~m ·

9!~$:JiJf~ :trJl·J::k$::k$~Ma5~:t$E$EJf~~ff.c~ *36-l}, 2019~ 7 J=!

]. of Architecture and Urban Design, Kyushu University, No.36, pp.1~ 10, July. 2019

1990-2015~:;Bit

~

4r00t~rtf0)0096'J\

~

-- /'O)jj~1tt

4J,L":ffi!!trai O)Jt!Jx

Development Patterns of Chinese Cities and Comparison with Their Core Areas from 1990 to 2015

* • ^{*. m -tttHt* *}

He LI, Shichen ZHAO

The urbanization process in China is quite unique. To understand it well and make more reasonable plans for future, enough researches about the process is necessary. In this paper we apply Multivariate Analysis Method to classify 199 Chinese cities and their core areas into 8 groups based on their changes during 1990-2015 from the perspective of urbanization. Then we conclude 5 main development patterns based on the result of classification, compare the patterns between cities and core areas. The classification and comparison could reveal the effect of urban areas during the urbanization process and provide references to review and adjust their development goals and strategies.

Keywords: Multivariate Analysis Method, Classification, Development pattern, Comparison, City core area

?P-~~~~tfr,

)j"ffi, 0036/~7'~/, J:t~, ~rtfi=p{,,±llii)jt

1. Introduction

Since the implementation of the reform and opening-up policy, especially from 1990, when a frenzy of development zones and real estate construction swept across the country, Chinese cities have experienced dramatic growth¹). According to the China City Statistical Yearbook and the Statistical Communique of the People's Republic of China on the 2015 National Economic and Social Development, the number of cities increased from 300 in 1984 to 656 in 2015. By 2015, the urbanization level in China is about 56.10%. Depending on the general rule of global urbanization^2), China could be considered as being in a stage of accelerated development.

Because of the transformation from the planned economy to market economy, despite what is happening in China being superficially similar with what has already occurred in other countries, the process of urbanization in China is quite unique in terms of scale and speed as well as the increasing number and size of cities^3).

* ~ffiE}f~~

** f~rn · Yl~'¥:t~F~

To describe the unprecedented urbanization in china, many scholars' attention has been directed to issues of the whole country's urbanization pattern. They always consider the ratio of the urban population to total population as the proxy of urbanization and define urbanization pattern on basis of the relationship between level of urbanization and economic development or resource or environment3, 4, 5, 6, 7, 8). The provincial pattern of China's urbanization also has been a key issue among scholars^23). Chen et al. divided 31 provincial-level units in China into six types by applying improved methods of quadrant map approach^9). Ge and Liang analyzed the urbanization patterns in Zhejiang province¹⁰). Long applied emerging new data to redefine Chinese city system; provided a perspective of morphology, function and society for understanding urban development during 2009--2014 in China24).

As it is mentioned above, the study objective of most previous researches is large-scale area such as the whole country or province, they include a large ares of suburbs.

In this smaller-scale research, cities are considered as study objects. It could provide a different perspective to illustrate the picture of China' s urbanization from 1990 to 2015. In addition, as the core content of

urbanization11), the development of urban areas in cities is deserving of more attention from scholars. The comparison of development between cities and their urbanization process and provide references to review and adjust their development goals and strategies.

2. Study objects, data resource and indicator system 2.1 Study objects and data resource

In China, modem cities are defined from the administrative view and classified into several administrative categories ranging

directly under the central

from municipality government, vice- provincial city, prefectural-level city to county level city. In this study the former 3 categories are study objects. Generally, they govern districts under city and county-level cities, as shown in Fig. I. District under city has county level status; both of them are managed by city government but with different development policies. Generally speaking, the original site of the city locate in the oldest districts under city; the longer history the districts under city have, the more developed they are.

The most developed urban area of a city always locate in districts under city. In the process of urban expansion, counties always are converted to districts under city. By 2015, there are several cities only govern districts under city.

Their former county-level cites have been changed to districts under city. Such as the city of Beijing, Tianjin, Shanghai, Chongqing.

County-level City

Districts under City

County-level City

Fig. 1. Conceptual model of Municipality/ Vice-provincial city/

Prefectural-level city Table 1. Indicator system

No. Indicator Unit

Xl Per Capita Gross Regional Product (Current Yuan Price)

X2 Primary Industry as Percentage to GRP % X3 Secondary Industry as Percentage to GRP % X4 Tertiary Industry as Percentage to GRP % XS Per Capita Public Finance Income Yuan X6 Per Capita Expenditure for Science, Technology

Yuan and Education

X7 Per Capita Investment in Residential Buildings Yuan X8 Collections of Public Libraries per 100 Persons copy X9 Average Wage ofEmployed Staff and Workers Yuan

We take China City Statistical Yearbook which includes the consistent and reliable annual data of municipalities, vice- provincial cities, prefecture-level cities as our data resource.

There are two series statistical data in our data resource: data of whole city, including districts under city and county-level cities under its jurisdiction; data of districts under city.

Because the urban area mainly concentrates in "Districts under City"; rural area mainly concentrates in "County-level Cities", when used in statistical data, the latter is available data which is most close to the real urban area. Therefore, in this study, the latter data is used to analyze the development pattern of cities' core areas, and the forn1er is for cities.

2.2 Indicator system

Classification analysis can let us know if there is any structure to the observations in the data set; can make the observations systematic and

because of abundant statistical

organized. Nowadays, data and advantaged computer technology, quantitative and objective Multivariate Analysis method 1s widely used in classification researches. Commonly used analytical techniques are principal component analysis and cluster analysis. In this study, Multivariate Analysis is used to classify the development of Chinese cities and core areas. To reveal the process of China' s urbanization, a nmltivariable indicator system is established from the perspective of urbanization.

Urbanization it is the complicated demographic transition from rural to urban, it is associated with shifts from an agriculture-based economy to mass industry, technology, and service. In this study, we focus on the dimensions of general economy, culture, education, and people's living conditions.

Primarily, 44 same or similar indicators covered by all of five China City Statistical Yearbooks 1991, 1996, 2001, 2006, 2011, 2016are picked out. Some of these 44 indicators' data are obtained

calculation. These indicators are by further

secondary selected according to the utilization in previous researches4· ^6•

s, 9, II, 12, 13, 14, 15, 16, 17, 18, 19). In this way, 20 of the 44 candidate indicators are selected. To reduce the overlap of information and neutral indicators, we cut 12 indicators.

Because of the close relationship between housing development

Per Capita Investment

and urbanization1), we in Residential Building.

add The eventual system includes 9 indicators, as shown in Table I.

Finally, 199 cities with complete data of the eventual 9 indicators are sorted out as study objects. They are 4 municipalities, 12 vice-provincial city, and 183 prefecture-level cities.

Table 2. Distances between the eight clusters

Cluster Cluster Cluster Cluster Cluster Cluster Cluster Cluster

1 2 3 4 5 6 7 8

Cluster 0 5.921 13.976 18.772 5.761 3.390 5.639 5.832 1

Cluster 5.921 0 19.705 13.506 9.614 8.549 10.546 10.171 2

Cluster 13.976 19.705 0 32.711 12.045 11.752 11.181 11.537 3

Cluster 18.772 13.506 32.711 0 22.611 21.441 22.864 22.909 4

Cluster 5.761 9.614 12.045 22.611 0 3.011 3.628 1.545 5

Cluster 3.390 8.549 11.752 21.441 3.011 0 2.555 2.601 6

Cluster 5.639 10.546 11.181 22.864 3.628 2.555 0 2.180 7

Cluster 5.832 10.171 11.537 22.909 1.545 2.601 2.180 0 8

3. Method

Firstly, we use Linear Trend at Point method to process missing values; calculate the reciprocal of Primary Industry as Percentage to GRP to make it positive. Then, to eliminate the influence of dimension and magnitude among raw variables, meanwhile, keep the differences between different yeas, raw variables of cities and districts under city in 2015 and 1990 were standardized together by Z-score normalization^{20, 21).}

Secondly, we apply Principal Analysis on standardized data of city in 2015 to extract comprehensive indicators; analyze the linear relationship between principal components and the original indicators, which is used to explain the meaning of the comprehensive index; and to obtain the formula for calculating the score of comprehensive indexes.

Thirdly, we separately plug the standardized data of cities and districts under city in 1990 and 2015 into the formula of comprehensive index to calculate the score of cities and districts under city in 2015 and 1990, then calculate the added score.

Fourthly, we view the added score as variable, employ Hierarchical Cluster Analysis and K-means Cluster Analysis to classify cities and districts under city together into several groups.

3.1 Principle component analysis

Firstly, raw data were standardized by using Z-score normalization, standardized data is presented by x[j, i

=

1,2 ... p(p is the a mount o fv ariables); j

=

1,2, .. q(q is the amount of cities). Secondly, the p x p correlation matrix R of standardized data was calculated. Thirdly, value of KMO was calculated. The range of the value is from O to 1. The greater the value, the more proper the method. When the value is smaller than 0.5, we should give up. Fourthly, eigenvalue Ai of R were figured out; then, eigenvector e' i' of Ai' were calculated, i'

=

^{1,2, ... m, m(m} s;, p) i s th ea moun t of extracted components, ei'i is the i th vector components of e'i'·

The loading of i' component Yi' on i variable:

li'i=.jX;;ei'i (Eq.1)

The score of the Yi' component ofj city:

p

Yi'i

= I

^{ei'i x[j (Eq. 2)}

i=l

Table 3. Total Variance Explained Initial Eigenvalues ConpnTit

Total %of Cumulative

Variance %

1 5.477 60.852 60.852

2 1.451 16.122 76.974

3 1.013 11253 88227

4 0.371 4.126 92.353

5 0.251 2.784 95.137

6 0.177 1.970 97.108

7 0.126 1.395 98.502

8 0.104 1.160 99.662

9 0.030 0.338 100

The value of eigenvalue A reflects the influence degree of component on original random variables, means how much the components can explain the original random variables. There two criteria for extracting principle components: Firstly,

the value of eigenvalues A ⁱ must be greater than 1; secondly, the greater the value of cumulative percentage, the better, the more information of original data included in extracted components;

the extracted principle components are considered as comprehensive indexes, the value of l i , i is the basis for explaining the meaning of comprehensive indexes, the greater the value, the more information of i indicator on the i ' comprehensive indexes²⁰, 21

, 22 ).

3.2 Cluster analysis

According to formula Eq. (1), Eq. (2), the eigenvalues in Table 3 and standardized data, the score of three comprehensive indexes of cities and their core areas both in 1990 and 2015 are calculated. Then we calculate added score of each city and its core area from 1990 to 2015 and view the added score as variable to apply cluster analysis.

In this study, firstly we adopted Hierarchical Cluster Analysis to classify cities and districts under city together.

According to the Dendrogram for the Agglomerative Clustering, 8 groups are obtained; then let K equal to 8, apply K-means Cluster Analysis to optimize the classification. The finial distance between these 8 cluster is shown in Table 2.

Ward's Method and Squared Euclidean Distance are chosen during the Hierarchical Cluster Analysis; Centroid clustering and Squared Euclidean Distance are chosen during the K-means Cluster Analysis.

The whole computational process was accomplished by software, Statistic Package for Social Science and Microsoft Excel.

Table 4. The Loadings of Component on Variables

No.

Xl

X2 X3

X4 X5

X6

X7

X8

X9

Variable

Per Capita Gross Regional Product (Current Price)

Primary Industry as Percentage to GRP Secondary Industry as Percentage to GRP Tertiaiy fudustry as Percentage to GRP Per Capita Public Finance Income Per Capita Expenditure for Science, Technology and Education

Per Capita Investment in Residential Buildings Collections of Public Libraries per 100 Persons

Average Wage of Employed Staff and Workers

Explanation

4. Results and discussions 4.1 Comprehensive index

1

0.791

0.659 -0.184 0.659 0.954

0.945

0.833

0.920

0.778 ESDI

Component

2

0.235

0.237 0.919 -0.667 0.168

0.132

-0.028

0.041

-0.042 ISi

3

0.398

-0.673 0.257 0.145 -0.142

-0.136

0.281

-0.211

0.390 Pill

We extract 3 comprehensive indexes: Economic and Social Development Index (ESDI), Industrial Structure Index (ISI), Primary Industry-independence Index (PIII).

We apply Principal Component Analysis to standardized data of cities in 2015. The value of KMO is 0.801. That means it is proper to PCA. As shown in Table 3, according to Eigenvalues and Cumulative Percentage, the first 3 components are extracted as principle components. Their Cumulative Percentage is greater than 85%. That means the extracted principle components can explain the original variables well.

As shown in Table 4, all values of the top 6 loadings of the first principle component on variables are positive and greater than 0.7. In these 6 variables, there are 4 general economic indicators (Xl, X5, X6, X7), 1 culture and education indicators (X8), and 1 people's living conditions indicators (X9). Therefore, the first principle component reflects the economic and social development condition²⁰, 22

).

Consequently, the first comprehensive index is Economic and Social Development Index (ESDI).

As shown in Table 4, the loading of the second principle component on Secondary Industry as Percentage to GRP is positive and greatest, and the value is 0.919; the loading of Tertiary Industry as Percentage to GRP is negative and the absolute value is second greatest, and the value is -0.667.

Hence, this principle component is essentially the difference between the Tertiary Industry and Secondary Industry^{20, 22)-} So, the second comprehensive index reflects the industrial structure.

We name it as Industrial Structure Index (ISI).

As shown in Table 4, the loading of the third principle component on Primary Industry as Percentage to GRP is negative, and the absolute value is the greatest, 0.673.

Therefore, the third principle component reflects the relationship between cities' development and primary industry^{20, 22).} We name it as Primary Industry- independence Index.

4.2 Result of classification

Fig. 2 and Fig.3 show the distribution of study objects in ESDI-ISI coordinate system and ESDI-PIII coordinate system. Fig. 4, Fig. 5, Fig. 6 show the means of comprehensive indexes of each group.

There are 20 study objects in Group 1. The maximum of ESDI is 10.486; the minimum is 6.115, the means is 7.822. The maximum of ISI is 1.912, the minimum is -2.032, the mean is -0.285. The maximum ofPIII is 2.531; the minimum is -0.505, the means is 1.489. Among these 8 groups, the level of ESDI is high, the level of ISI is medium, the level of PIII is high.

There is only one study object in Group 2. The value of ESDI is 11.064; the value of ISI is 3.462, the value of PIII is -1. 7 51. Among these 8 groups, the level of ESDI is remarkably high, the level of ISI is remarkably high, the level of PIII is remarkably low.

There is only one study object in Group 3. The value of ESDI is -1.065; the value of ISI is -5.498; the value of PIII is 10.933. Among these 8 groups, the level of ESDI is low, the level of ISI is remarkably low, the level of PIII is remarkably high.

There are 2 study objects in Group 4. The maximum of ESDI is 21.062; the minimum is 19.905, the means is 20.484. The maximum of ISI is 5.175, the minimum is 4.940, the mean is 5.057. The maximum of PIII is -10.752; the minimum is -11.844, the means is -11.298. Among these 8 groups, the level of ESDI is remarkably high, the level of ISI is remarkably high, the level of PIII is remarkably low.

There are 70 study objects in Group 5. The maximum of ESDI is 5.403; the minimum is 0.958, the means is 2.555.

The maximum of ISI is 4.254, the minimum is 0.152, the mean is 1.310. The maximum of PIII is 2.528; the minimum is 1.054, the means is 1.667. Among these 8 groups, the level of ESDI is low, the level ofISI is high, the level of PIII is high.

There are 83 study objects in Group 6. The maximum ofESDI is 6.197; the minimum is 3.195, the means is 4.491.

6.00

4.00 - - - - 1 - - - - - - - 1 - - - -

1

•

¹

- I

?< 2.00

(I)

- - - -I

"a

,.s

(I) 1--<

.a

_u

5

^0.00

VJ

]

ti _r:r,

..§ ,.s _-2.00 - - - -

1

-4.00 - - - ~ - .,_ _____ L ______ ^~ _______ L _____ _

I A I I I

I I

-6.00

•

-2.50 2.50 7.50 12.50 17.50 22.50

Economic and Social Development Index

Fig. 2 Scatter diagram of study objects in ESDI-ISI coordinate system

12.50

•

7.50 --- - - - i - - - - -1- - - - i - - - i - - - -

I I I I

I

I I I I

2.50 -- - - -

,~~v=.::1:···-

^~

----:---:---

I .uJI. ;. I ~ I I

0.00

,,_ I I .a. I I

I I I

-2.50 -- - - - - - - -+ - - - - -1- - - -

•

l- - - - -+ - - - -

I I I

I

-7.50 -~--- - - - - l - - - 1 - - - 1 - - - 1 - - - -

I I I I

I I I

-12.50 I I I I

-2.50 2.50 7.50 12.50 17.50 22.50

Economic and Social Development Index

Fig. 3 Scatter diagram of study objects in ESDI-PIII coordinate system

:§ ~

]]

'"O...., i:: i::

ro <l.l

<) s

· -8 0 0.

§]

~Q 25.00 20.00 15.00 10.00 5.00 0.00 -5.00

2 4 5 6 7 8

Group

Fig. 4 Mean ofESDI

2 4 5 6 7 8

Group

Fig. 5 Mean ofISI

Fig. 6 Mean of PIII

... Groupl

• Group2

• Group3

• Group4

- Group 5

* Group 6 .,_ Group 7 Group 8

7 5 0 0 0 - - - - - 70000

65000

60000 --•---

0 55000 ---

§ 50000 ---

] o 45000 ---

.1

40000 ---

~ 35000 .---

1 4 0 0 0 - - - - ~

·113000

.§ 12000 ] 11000

@ 10000

J

0 ⁹⁰⁰⁰

8000

"

.1

⁷⁰⁰⁰ •---

"'

~ 30000 ---~--- cB

p... t j

. .§! 25000 • --- ~

6000

.

--- 5000 f---:--- 4000 •

---~-1

^*-~--

3000 .. ---:::- ---·

~

^{• *}

l

U 20000 • --- _::;

~ : -

.

P... 15000 .. ---=.-•----

·i

10000

----.--tl--1-- ~

500~ ---;- - - 12345678

Group

2000

---~---1- ..

-1*-l'

1000 ---~--- ... - o,...,.._~~~-r'

12345678 Group

8 5 0 - - - - - 800

750 ---~---

§ ⁷⁰⁰

~ 650

~ 600 8 550

g_ 500

.s,l 300 --- .. ---

~ 250 r---~---

'ti 200 • ---r---1-

§ 150 • -~---:.-•----1-

·- 100 ;--- .. - --.,.--,,-

] ~

- 50 ·---1-

8

⁰ ~

_;~i ::::::}:rt:

-150 ...

12345678 Group

3 5 0 0 0 ~ - - - ~ 32500

f5h 30000 ---- --- - -- .5 ~ 27500 --- .. ----

] 25000 .---.---*----

1=] 22500 :---"---"----

:g t ~

~ 20000 ---:---- .5 17500 .---;;---- E ~ ! ...

~ 15000 :--.---~-.---- :{l 12500 • ______ ::J ___ ~

~

I ...

^I

'; 10000 ---~-,-~--^{~ - I}

8

⁷⁵⁰⁰

---~---n-r

~ :::: :::::::p:r

01...J,-~~--.---l'---r

12345678 Group

212500 200000 187500 175000 0 162500

.g 150000

£

«! 137500

::l 125000 -~

Q) 112500

~

~ 100000

c

87500

.~ ₇₅₀₀₀ u ~

ti ⁶²⁵⁰⁰

p... 50000

37500 25000 12500 0

-+---* --- ---•--- i ----'

--! ---

Group

1 1 5 0 0 0 - - - - 110000

105000

~ 100000 .---

;:o

::> 95000 ---

] 90000 ---

¢: 85000 ________ * __ _

~ 80000 • ____________ '

V J & *

"" 75000 ~---·-_,_ ... _

0~ t - i I

70000

---c-

l

⁶⁵⁰⁰⁰ .---=-1-A- 'ti 60000 ---a- +

§ :I:

]i

55000

.---i T

;> 50000

---•---e---1-

f

⁴⁵⁰⁰⁰

^{---h -}

..., 40000 ---_-*_±_

35000 ---,:- 30000 --- - __ _.._

25000-'---~---"

12345678 Group Fig 7 Scatter Diagram of study objects on Added Value-Axis (economy and society)

The maximum ofISI is -0.732, the minimum is -2.363, the mean is -0.775. The maximum of PIII is 2.597; the minimum is 0.033, the means is 1.747. Among these 8 groups, the level of ESDI is medium, the level ofISI is medium, the level of PIII is high.

There are 76 study objects in Group 7. The maximum of ESDI is 3.852; the minimum is 1.433, the means is 2.521.

The maximum ofISI is -0.929, the minimum is -4.136, the mean is -2.103. The maximum of PIII is 1.967; the minimum is -0.180, the means is 0.900. Among these 8 groups, the level of ESDI is low, the level of ISI is low, the level of PIII is low.

There are 145 study objects in Group 8. The maximum of ESDI is 3.175; the minimum is 0.069, the means is 1.930. The maximum ofISI is 1.331, the minimum is -1.101, the mean is 0.001. The maximum of PIII is 1.902; the minimum is 0.551, the means is 1.229. Among these 8 groups, the level of ESDI is low, the level ofISI is medium, the level of PIII is low.

Fig. 7 shows the study objects' added value of important original variables about economic and social development;

Fig. 8 shows the distribution of study objects in Added Value of Secondary Industry-Tertiary Industry coordinate system; Fig.9 shows the study objects' added value of Primary Industry.

4.3 Development patterns and comparison between cities and core areas

Since added score of each city and its core area from 1990 to 2015 is viewed as variable in cluster analysis, according to the meaning of comprehensive indexes, the attribute of cluster can be considered as development pattern. On this basis, according to the analysis of added value of original variables, 5 main development patterns of study objects from 1990 to 2015 are concluded.

Pattern 1, rapid economic and social development;

development focus shift to Tertiary Industry, obvious decrease of dependence on Primary Industry. This pattern is concluded from group 1. As shown in Fig. 7, excluding group 2, 3, and 4, for group 1, study objects' added value of important original variables about economy and society mainly distribute within the relatively high range. In other word, they experienced relatively rapid economic and social development. As shown in Fig. 8, the study objects mainly distribute in the second quadrant. It means that their Secondary Industry as percentage to GRP decreased during 1990 to 2015, while the percentage of Tertiary Industry increased. That is their development focus shift to Tertiary Industry. As shown in Fig. 9, for group 1, study objects' added value of Primary Industry mainly distributes within the range from -20.00 to 0.00. That is, during 1990 to 2015, study objects' dependence on Primary Industry obviously decreased.

~

.8

(!) bl}

c<:l

1:: _(!)

u i-..

(!)

A..

Cl)

c<:l

g

.g Cl)

,.s

~

s

·i::

A..

40.00 ~ - - - , 30.00

20.00 10.00 0.00 -10.00 -20.00 -30.00 -40.00 -50.00 -60.00

J.

J.

---r- ---}----J. -

•

=

-i ----.---~-- i

---:----~----~----

^-

~ *

---=---;.---t

2 3 4 5 6 7 8

Group

Fig. 9 Scatter Diagram of study objects on Added Value-Axis (Primary Industry)

~ 0

.8

(!) bl) e<:!

1:: _(!)

...

u _(!) 0...

00 e<:!

£

₀₀

.g

..s

·-e

~

(!)

E-<

50.00

40.00-

30.00-

20.00-

10.00-

0.00

-10.00-

-20.00-

Ai I I

;,.I I I

---0---~---~---: ... ________ ~---

: & Ji.. I J.. J..

I ..C- J.. J..1

- - - -

~

^{- - : ;}

~

^-

~

^{- "£}

~ ~

^{- - - -- -}

~

- - - -

~

- - - -

I J.. J.iJ.. J . . ~ ~ ; .R('&J..

*

~

'*

I

I 1'.it :/f

* ¥*

*~.dS'J.~ ^I

I • )..I "'j

*

J.. I "% I I I

--- ^~ -- ^~ -/-~=~{- ^{f T} ~:._:!~\*II\~ L ~ --+~_:- - - -

1 J.. Ji. JS_* 7)

fl I I I

^~

I ^'r1--

^{JI -}

I J.. I -k 11 fit.

~* ,*llJ

^I

'l

^{I - -}

IJ.. J.. I J.. I I I 11 1IL -:=_1 -

- - - -

~

_ l-_ - - - : - ~'

'*~H,t l ~~~'~,:_-=:-~~;: -- ^~ --

I I fjj-1l~ll""'jl

=--

^{I - - _}

I~)'*-*•

I

I I -r

I I ~ -

- - - r - - - r - - - -r - - - ... - ,- - - -_- - -r - - - -

: : I - :

________ I ________ I --->---1 - - - -

1 I I

I

-30.00-+---+, - - - l , . - - - + - - - < ; . - - - '

-60.00 -40.00 -20.00 0.00 20.00

... Groupl

• Group2

Secondary Industry as Percentage to GRP

• Group3

• Group4

- Group 5

* Group 6

40.09

.,_ Group 7 Group 8 50.00 ~ - - - ~ 50.00.---r---, 30.00

10.00

---- ... _ * *; * * * * * *

---~-:lt''* *\.:

~w*:-- ""*-~

~*:* * ** ---

* * * -10.00

-30.00

0\ 1. N ⁰ ₀ ^N ₀

0 0 0

0 0 0 ⁰ 0

0 0 0 0

SIP Group 6

.i::,.

0

0 0

30.00

E:: 10.00 E-<

-10.00 -30.00

...

... .., ... ... ...

" ... .,!,± ...

- - - . . 1 . . . ~ , . 1 , . . 1 . . I . J . .

)./·;.1f1~--A-- ---

). J..~ A A,J,., A A ) .

I I ^I

0\ ^.i::,. N ^{0 N .i::,.}

0 ^{0 0}

0 0 0

0 0 0 ⁰ 0 0

0 0 0 0 0

SIP Group 7

5 0 . 0 0 ~ - - - -

30.00 E:: 10.00 E-<

-10.00 -30.00 I

O'\

0

0 0 ,,

..

·'~,,

.

^',,,

..

1. 0

0 0

:-..:,

..

^~, ..

N 0

0 0

.... ^,~

.. ^', _',,.

a',

0

0 0

SIP Group 1

N ^.i::,.

0 0

0 0

0 0

5 0 . 0 0 ~ - - - ~

30.00

E:: 10.00 E-<

-10.00- - -30.00 _I

O'\

0

0 0

50.00 30.00

0...

10.00

-

^E-<

-10.00 -30.00 ^I

O'\ 0

0 0

1. 0

0 0

1. 0

0 0

==- -£'..::-=-

--_:-:;...---::-""----.:~

N ₀ ^{0 N} 0 0 ^:,.

0 0 0

0 ⁰

0 0 0

SIP Group 5

I I

', _1'.,,:¹¹¹11

\ -

1111

11¹¹ l,/11\',f-·---

/,11 Ir ~11 q-

~ ,,,,,1 •'1 ___ ,--,r,,,,

- - I 111

/1 ,,

I I I

I ^, ,

I

N ⁰ _{0 0} ^{N .i::,.} ₀

0 0 0

0 ⁰

0 0 0

SIP Group 8 Fig. 8 Scatter Diagram of study objects in Added Value of Secondary Industry-Tertiary Industry coordinate system Pattern 2, Relatively low speed economic and social

development; development focus on Secondary Industry, obvious decrease of dependence on Primary Industry.

This pattern is concluded from group 5. As shown in Fig. 7, excluding group 2, 3, and 4, for group 5, study objects' added value of important original variables about economy and society mainly distribute within the relatively low range. In other word, they experienced relatively low speed economic and social development. As shown in Fig. 8, the study objects mainly distribute in the second and the fourth quadrant. It means that their Secondary Industry as percentage to GRP increased

during 1990 to 2015. That is their development focus on Secondary Industry. As shown in Fig. 9, for group 5, study objects' added value of Primary Industry mainly distributes within the range from -50.00 to 0.00. That is, during 1990 to 2015, study objects' dependence on Primary Industry obviously decreased.

Pattern 3, Medium speed economic and social development;

development focus shift to Tertiary Industry, obvious decrease of dependence on Primary Industry. This pattern is concluded from group 6. As shown in Fig. 7, excluding group 2, 3, and 4, for group 6, study objects' added value of important original variables about economy and society mainly distribute within

the medium range. In other word, they experienced medium speed economic and social development. As shown in Fig. 8, the study objects mainly distribute in the second quadrant. It means that their Secondary Industry as percentage to GRP decreased during 1990 to 2015, while the percentage of Tertiary Industry increased.

That is their development focus shift to Tertiary Industry.

As shown in Fig. 9, for group 6, study objects' added value of Primary Industry mainly distributes within the range from -40.00 to 0.00. That is, during 1990 to 2015, study objects' dependence on Primary Industry obviously decreased.

Pattern 4, Relatively low speed economic and social development; development focus shift to Tertiary Industry, no obvious decrease of dependence on Primary Industry. This pattern is concluded from group 7. As shown in Fig. 7, excluding group 2, 3, and 4, for group 7, study objects' added value of important original variables about economy and society mainly distribute within the relatively low range. In other word, they experienced relatively low speed economic and social development. As shown in Fig. 8, the study objects mainly distribute in the second quadrant. It means that their Secondary Industry as percentage to GRP decreased

Table 5. Comparison of development pattern between cities and core areas Pattern

c c

^~~

A

1 1

2 2

3 3

4 4

5 5

2 3 2 5

3 4 3 5 1 5 2 5 3 5 4

Beijing Tianjin Shanghai Suzhou Hangzhou Xiamen Guangzhou Zhuhai

Tongliao Hulunbuir Bayannur Putian Ningde Yingtan Fuzhou Rizhao Liaocheng Huizhou Wuzhou Baise Deyang Neijiang Nanchong Xianyang Yulin

Langfang Taiyuan Hohhot Baotou Shenyang Panjin Nanjmg Wuxi Changzhou Maanshan Nantong Yangzhou Zhenjiang Jiaxing Huzhou Jinhua Zhoushan Hefei

Fuzhou Jinan Zibo Weifang Weihai Changsha Foshan Haikou Kunming Xi'an Urumqi

Xingtai Datong Yangquan Jinzhong Anshan

Harbin Jixi Hegang Shuangyashan Yichun(Heilongjian) Shiyan Zhangjiajie Chenzhou Y ongzhou Shaoguan

Qinhuangdao Baoding Chengde Chifeng Yingkou Fuxin Ganzhou Ji'an Zaozhuang Tai'an Linyi Dezhou Xinyang Zhoukou Zhumadian Huangshi Ezhou Shaoyang Shangqiu Guilin Qinzhou Yulin Zigong Yibin Baoshan Weinan Hanzhong Tianshui Zhangye Jiuguan

Fushun Jiamusi Zunyi

Suihua Bizhou Yiyang Dazhou

Chaoyang Mudanjiang Zhaotong

Jilin Huainan Lanzhou Huaibei

Heze Loudi Anshun

Nanping Nanyang Shantou Qujing Tangshan Lishui Wuhu Quanzhou Zhangzhou Jiujiang Xuchang Zhuzhou Huaihua Yinchuan Lianyungang YichunGiangxi) Xiangyang Panzhihua Luzhou Leshan Ya'an Yan'an Ankang Shangluo Qingyang

Dalian Qingdao Y antai Yuxi

Cangzhou

Siping Tonghua Bengbu Chuzhou Chizhou Jingdezhen Pingxiang Jingmen Huanggang Xianning Hengyang Mianyang

Changchun Xuzhou Xiangtan Yueyang Shijiazhuang Handa Tieling Baicheng Jiaozuo Changde

Yancheng Huangshan Sanming Longyan Jiangmen Nanning Liuzhou

Changzhi Yuncheng Xinzhou Linfen Qiqihar Huai'an Quzhou Fuyang Chongqing Pingliang Xining

Jining Luoyang Xinxiang Yichang Dandong

Kaifeng

Jinzhou Pingdingshan

Liaoyang Anyang Note: Ccity CA: core area

during 1990 to 2015, while the percentage of Tertiary Industry increased. That is their development focus shift to Tertiary Industry.

As shown in Fig. 9, for group 7, study objects' added value of Primary Industry mainly distributes within the range from -15.00 to 5.00. That is, during 1990 to 2015, study objects' dependence on Primary Industry didn't obviously decrease.

Pattern 5, Relatively low speed economic and social development; development focus on tertiary Industry, no obvious decrease of dependence ^on Primary Industry. This pattern is concluded from group 8. As shown in Fig. 7, excluding group 2, 3, and 4, for group 8, study objects' added value of important original variables about economy and society mainly distribute within the relatively low range. In other word, they experienced

relatively low speed economic and social development.

As shown in Fig. 8, the study objects mainly distribute in the first and the second quadrant. It means that their Tertiary Industry as percentage to GRP increased during 1990 to 2015. That is their development focus on Tertiary Industry. As shown in Fig. 9, for group 8, study objects' added value of Primary Industry mainly distributes within the range from -35.00 to 15.00. That is, during 1990 to 2015, study objects' dependence on Primary Industry didn't obviously decrease.

As show in Table 5, totally, 120 cities and their core areas' development pattern are same: 8 cities are pattern 1;

17 cities are pattern 2; 29 cities are pattern 3; 24 cities are pattern 4; 42 cities are pattern 5. As well as their core area.

7 6 cities and their core areas' development pattern are different. Among them, 10 cities are pattern 2, while their core areas are pattern 3; 11 cities are pattern 2, their core areas are pattern 5; 3 cities are pattern 3, their core areas are pattern 1; 1 city is pattern 4, its core area is pattern 3, 1 city is pattern 5, its core area is pattern 1; 12 cities are pattern 5, their core area are pattern 2; 15 cities are pattern 5, their core areas are pattern 3; 23 cities are pattern 5, their core area is pattern 4.

Table 6. Feature of development pattern

Pattern Speed of Emphasis Dependence

economic of on

and social industry Primary development development Industry Rapid Shift to tertiary Obvious Industry decrease

2 Low Secondary Obvious

Industry decrease 3 Medium Shift to tertiary Obvious Industry decrease 4 Low Shift to tertiary No Obvious

Industry decrease

5 Low Tertiary No Obvious

Industry decrease

5. Conclusion and Discussion 5.1 Summary of the study

In this paper we apply Multivariate Analysis Method to classify 199 Chinese cities and their core areas into 8 groups based on their changes during 1990-2015 from the perspective

0 I I

- - - ~ - - - r - - - , - - - r - - - -

1 I I J

X I I t I

Q)

"d

.s

_2.0

Q)

$.-<

.8

cl': . : : :

•"' I I I

- ~ ~ ~ ~ ~ - r - - - - ~ - ~ - - - i - - - ~ - - - -

1 I I

I I

E

u ^.0

VJ

-4.0

-6.01 H - - - ' - - - + - - - 1 - - - 1 - - - 1 - - - 1

-2.50 2.50 7.50 12.50 17.50 22.50

Economic and Social Development Index

~ City · Core Area

Fig. 10 Scatter diagram of study objects in ESDI-PIII coordinate system

patterns based on the result of classification, compare the patterns between cities and core areas. The feature is shown in Table 6. The classification and comparison could reveal the effect of urban areas during the urbanization process and provide references to review and adjust their development goals and strategies.

The strength of our study mainly lie in the following aspects. First, compare to most previous researches, our smaller-scale research provide a perspective of city development for illustrating the picture of China' s urbanization from 1990 to 2015. Second, since emerging new data of early years is not available, we apply the conventional statistic data from Yearbook, so our study can focus on a longer timespan then those researches with emerging new data.

5.2 Potential bias and future steps

As shown in Fig. 10, the development patterns of cities and core areas are similar. The development of real urban area should be obviously different from the city included large rural area. This situation could be due to this:

for many cities, even Districts Under City, they still include considerable suburban area. While because Chinese cities are defined from administrative view, it is hard to collect the data of real urban area. To improve this situation, scholars are making their efforts. Long et al make the effort to understand uneven urban expansion with natural cities using open data25). For next step, to explain the phenomenon concluded from this paper that "the development patterns of cities and core areas are similar but different", we will explore the reasons from several perspectives: the contribution of suburban area, development policy, geographical location. In addition, we will conduct the interrelationship of multiple cities during high-speed development and the cause of several unique cases excluded by main development patterns (Shenzhen, Zhangjiakou, Erdos).

References

1) Chen, J., Fei, G. and Ying, W.: Chinese urbanization and urban housing growth since the mid-1990s, J Housand the Built Environ. 26, 219-232, 2011

2) Northam, R. M.: Urban Geography, New York: John Wiley & Sons, 1975

3) Chan, R.C.K., Yao, S.M.: Urbanization and sustainable metropolitan development in China: Patterns, problems and prospects, Geo Journal 49, 269-277, 1999

4) Chang, G.H., Brada, J.C.: The paradox of China's growing under-urbanization, Economic Systems.

30, 24-40, 2006

5) Friedmann, J.: Four theses in the study of China's urbanization, International Journal of Urban and Regional Research, 30(2), 440-451, 2006

6) Chen, M.X., Lu, D.D., Zha, L.S.: The comprehensive evaluation of China's urbanization and effects on resources and environment, Journal of Geographical Science, 20(1), 17-30, 2010

7) Li, Y.F., Li, Y., Zhou, Y., Shi, Y.L., Zhu, X.D.: Investigation of a coupling model of coordination between urbanization and the environment, Environmental Management, 98, 127-133, 2012

8) Chen, M.X., Liu, W.D., Tao, X.L.: Evolution and assessment on China's urbanization 1960-2010: Under-urbanization or over- urbanization, Habitat International, 38, 25-33, 2013 9) Chen, M.X., Lu, D.D., Liu, H.: The provincial pattern of the relationship between China's urbanization and economic development, Acta Geographica Sinica, 65(12), 1443-1453,2010

10) Ge, Y., Liang, Q.: Empirical analysis on urbanization patterns in Zhejiang province based on spatial economy and GIS. Chinese Geographical Science.

16(3), 193-198, 2006

11) Wei, H.K., et al.: China's urbanization quality after conducting extensive research and surveys, Review of Economic Research, 2519, 3-32, 2013 12) Duo, W.Z., Li, G.P., Gan, Y.J., Wang, L.M., Yang, K.Z.: Urban competitive pattern and its changes in China, Chinese Geographical Science, 10 (2), 105-112, 2000

13) Fang, P.L. and Qi, J.G.: Assessing the sustainability of major cities in China, Sustain Sci., 5, 51-68, 2010

14) Henderson, J.V.: How urban concentration affects economic growth, Policy Reasearch Working Papers, 2326, 2000

15) Shen, L.Y., Yi, P., Zhang, X.L., Wu, Y.Z.:

An alternative model for evaluating sustainable urbanization, Cities, 29, 32-39, 2012

16) Wang, Z.P., Shi, C.L., Li, Q., Wang, G.:

Coupling trending analysis about urbanization and urban resource in Beijing, Energy Procedia, 5, 1589-1596, 2011

17) Yang, S.G., Wang, M.Y.L., Wang, C.L.: Revisiting and Rethinking Regional Urbanization in Changjiang River Delta, China, Chin. Geogra. Sci., 22(5), 617-625, 2012 18) Yang, J.Y., Chen, W.: Research on the evaluation model of the modernization level of urban construction inJiangsu Province, China, Science China Technological Sciences, 53, 2510-2514, 2010

19) Yuan, W. and James, P.: Evolution of the Shanghai city region 1987-1998: an analysis of indicators, Journal of Environmental Management, 64, 299-309, 2002 20) Hardle, W. and Simar, L.: Applied Multivariate Statistical Analysis--Second Edition, New York: Springer, 2007

21) Wang, X.M.: Applied Multivariate Statistical Analysis (Second Edition) (In Chinese). Shanghai: Shanghai University of Finance & Economics Press, 2004

22) Zhang, W.T.: Statistical Analysis Coursebook of SPSS 11 (In Chinese). Beijing: Beijing Hope Electronic Press, 2006

23) Shen, J.: Estimating Urbanization Levels in Chinese Provinces ^Ill 1982-2000, International Statistical Review, 74(1), 89-107, 2006

24) Long, Y.: Redefining Chinese City System with Emerging New Data, Applied Geography, 75, 36-48, 2016

25) Long, Y., Zhai, W., Shen, Y., Ye, X.:

Understanding Uneven Urban Expansion with Natural Cities Using Open Data, Landscape and Urban Planning, 177, 281-293, 2018