Transcript Slides - 2nd International UGEC Conference

2nd International Conference
Urban Transitions and Transformations: Science,
Synthesis and Policy
Impact of urbanization on cultivated
land in China: a model-based
analysis in China
Xiangzheng Deng
November 6-8，2014
Urbanization is an inevitable development
process
 Urbanization is a natural process companying with
the development of human society at a certain stage
People inhabitance
-urban area 2%
-population >50%
Labor shifts
-Agricultural sector 
Non-agricultural sector
Global urbanization rate
Sources: The times of India, 2008
Sourced from NetEase 2012
Urbanization in China
 Urban population
– 1978-2013: risen from 18% to
54%
 Urbanization accelerated
– since middle 1990s
– 3/4 of the population would live
in cities by then end of 2050
 Population in the urban areas
– By 2020, over half of the people
will move into cities, according to
a planning made by NDRC
– About 39 percent of the
youngsters are employed in the
urban, while only 27 percent of
older workers work in small city
or county
Urbanization in China:
1949-2009
Source: Urban China, 2010
Urbanization
Source: ChinaDaily, 2013
Views on the impacts of urbanization on
agricultural land
Urbanization leading to cultivated land reduction
and reduced land production directly through
encroaching the land around urban core or the
leapfrogged urban cluster
Urbanization playing an active role in conserving
cultivated land by releasing the pressure from
land occupations accompanying with the
urbanization
Ideas on the urbanization models
Impacts of different models of urbanization on
the changes of cultivated land
– Small town model
– Lack of effective planning
– Consuming large areas of cultivated land
– City model
– Infrastructure built
– As for the scales of land consumption : urban built area >
rural built area
Research in Progress
Failed to build a measurement indicator on urbanization
models by combining the built-up area in both urban and
rural area and to do an integrated research
Need to fully control some factors with no matter
positive or negative effects on urbanization to get a more
robust estimation
Include different scales of cities with a diverse of
urbanization speed into the analyses
“Major” question under answer
Does urbanization of big cities in China consuming
more land compared with that of small towns as well as
the expansion of villages?
How has the urbanization affected the land productivity
and cultivated land area for the past three decades?
Is there any kinds of urbanization model saving land?
What is the regional characteristics of land consumption
of urbanization?
Goals of Presentation
Explored changes in China’s cultivated area and its
conversion to built-up area and other uses due to
urbanization, industrialization and rural settlement
expansion
Analyze the impacts of the urbanization models on the
changes of land productivity
Explore the various effects of urbanization models on
the changes of cultivated land over time and space
Answer questions from two aspects
The direct impacts of urbanization on
the land productivity
Impact of urbanization on cultivated
land changes in China
Estimate the Changes in Quantity and
Quality of Cultivated Land: Methodology
 Identify the quantity of China’s land use change
– Detection models of Land Use Change(LUC),1-km area
percentage data models
– Based on the prototype of the 1-km area percentage data model (1km APDM), developed a set of programs to generate 1-km area
percentage data according to map-algebra concepts
– i.e., the encroachment of urban land onto cultivated land
 Measure the quality of cultivated land conversions
– Agro-ecological Zones (AEZ) methodology
– Manipulate data with GIS technologies
– Use Agro-ecological zoning model (AEZ) with “other data” to
create a productivity index of cultivated land (essentially this index
is a way that geographers measure the bio-productivity of land; in
other words, it is a measure of the quality of land)
Estimate the Changes in Quantity and
Quality of Cultivated Land: Database
Database:
– Remote sensing data (RSD) on all of China (for land use)
– Other data includes information on climate; soils; slope and
elevation; etc. (e.g., from China Meteorological Bureau)
– 1x1 km GRID data
Temporal scale: 1988~2000; 2000~2008
Decoding the information on land use changes
from Landsat TM/ETM digital imagines
1988/1990,
1999/2000,
2005/2008
Landsat
TM digital
image
1988/1990,
Geometric 1999/2000,
correction 2005/2008
False color Landsat TM
composition registered
image
Arc/Info
Land-use change map of
predominant types
during the period
between 1988-2000 and
2000-2008
Mutual
interpretation
Vector map of land use in
1988、2000、2005、2008
Land-use change map
during the periods between
1988 to 2000；2000-2008
1km vector map
Zoning map of land-use
change during the period
between 1988-2000 and
2000-2008
Overlay
Land use conversion
maps during the period
between 1988-2000 and
2000-2008
Characteristics and Measures on Land-use
Change
The regional differentiation of land-use change rate
can be represented by the dynamic degree model
of land-use, i.e.
n

S   (Si  j / Si )  (1/ t )  Wi 100%
 ij

where, S is the land-use change rate, Si represents the total areas of i
(land-use category) at the former stage while is the weight of areas
proportion of i, represents the net change of area from i to j (landuse category) at the time scale of t. The basic unit to employ the
dynamic degree model is 1km GRID, and the statistical result
serves as basis to draw the land-use change and land-use conversion
maps classified by land-use categories.
Changes in cultivated land
 In 1988-2000, 2.7 million hectares of new cultivated land was
created. China’s farmers were cultivating 1.9% more land in 2000
than they were in 1988.
 In 2000-2008, the cultivated land area of China actually lose
considerable quantities of land, by 0.58 million hectares.
Panel a: 1988 to 2000
Panel b: 2000 to 2008
Conversions of Cultivated Land in China during 1988 to 2000 (Panel a) and 2000
to 2008 (Panel b)
Changes in cultivated land
 It should be noted that only in the case of Beijing, Shanghai
and Zhejiang did the conversions exceed 5% in 1986-2000.
 Apparently, the provinces that experienced the most
conversions are Shanghai and Shandong in 2000- 2008.
Panel a
Panel b
Land conversions from Cultivated Land to other uses, 1988-2000 (Panel a) and 20002008 (Panel b)
Changes in cultivated land
 During 1988-2000, In northeast China, there were large tracts of
forests that were converted to cultivated land; Some areas in
Sichuan also were converted from forests to cultivated
 During 2000-2008, there are less tracts of land that were converted
to cultivated land
Panel c
Panel d
Land conversions from other uses to cultivated Land, 1988-2000 (Panel c) and 2000-2008
(Panel d)
Changes in potential agricultural productivity
and production due to land conversions
 The average potential agricultural productivity fell by 2.2% during
1988-2000, and the total production potential fell by 5.9 trillion Kcal,
or by only 0.3%
Province
Beijing
Tianjin
Hebei
Shanxi
Inner
Mongolia
Liaoning
Jilin
Heilongjiang
Shanghai
Jiangsu
Zhejiang
Anhui
Fujian
Jiangxi
Shandong
Henan
Total
Total production
potential in 1988
4120
6220
72600
34600
837
204
1950
237
Net
change
-814
-191
-1554
31
Percentage
change
-19.75
-3.06
-2.14
0.09
4940
1630
3310
9.17
1470
1970
6210
0
240
313
471
543
537
162
1500
28971
505
441
524
1010
5000
3040
2110
772
1030
1430
1340
34826
965
1529
5686
-1010
-4760
-2727
-1639
-229
-493
-1268
160
-5855
2.84
4.87
10.67
-11.01
-4.18
-3.95
-1.2
-0.48
-0.47
-1.3
0.14
-0.3
Increase
Decrease
23
13
396
268
36100
34000
31400
53300
9170
114000
69100
137000
48100
106000
97600
111000
1965770
Total production
potential in 1988
Hubei
149000
Hunan
141000
Guangdong
90600
Guangxi
113000
Province
2320
1160
3460
852
Net
change
-1599
-827
-3193
488
Percentage
change
-1.07
-0.59
-3.52
0.43
Increase
Decrease
721
333
267
1340
Hainan
16100
191
352
-161
-1
Chongqing
Sichuan
Guizhou
Yunnan
Tibet
Shaanxi
Gansu
Qinghai
Ningxia
Xinjiang
Taiwan
56300
176000
63300
67900
1940
40800
32000
2780
8540
28700
13500
87
417
613
896
0
434
553
5.9
1200
2750
13
396
1390
99
1090
4
379
174
24
108
883
76
-309
-973
514
-194
-3
55
379
74
1092
1867
-63
-0.55
-0.55
0.81
-0.29
-0.16
0.13
1.18
2.66
12.79
6.51
-0.46
Change of total production potential associated with changes in cultivated land by provinces
for 1988-2000, measured in billion Kcal and percentage change (%).
Changes in potential agricultural productivity
and production due to land conversions
 The average potential agricultural productivity fell by 1.3% during
2000-2008, and the total production potential fell by 32.9 trillion
Kcal, or by around 1.7%
Province
Beijing
Tianjin
Hebei
Shanxi
Inner
Mongolia
Liaoning
Jilin
Heilongjiang
Shanghai
Jiangsu
Zhejiang
Anhui
Fujian
Jiangxi
Shandong
Henan
Total
Total production
Net
Increase Decrease
potential in 2000
change
3306
3.4
280.8
-277.3
6029
1.9
139.7
-137.8
71046
122.6
914.9
-792.3
34631
1.4
798.4
-797
Percentage
change
-8.39
-2.29
-1.12
-2.3
39410
460.6
118.8
341.7
0.87
34965
32929
58986
8160
109240
66373
135361
47871
105507
96332
111160
1959913
40.6
71.9
1380.2
0
28
15.7
535.9
18
633.2
166.4
99.3
6449
221.4
155.8
802.8
1993.9
8319.2
1673.6
3594.5
1220.3
1060.9
2967
1243.1
39348
-180.9
-83.9
577.4
-1993.9
-8291.2
-1657.9
-3058.6
-1202.3
-427.6
-2800.6
-1143.8
-32899
-0.52
-0.25
0.98
-24.43
-7.59
-2.5
-2.26
-2.51
-0.41
-2.91
-1.03
-1.68
Total production
potential in 2000
Hubei
147401
Hunan
140173
Guangdong
87407
Guangxi
113488
Province
31.9
2
37.5
31.5
2371.1
1215.1
2164.7
787.1
Net
change
-2339.3
-1213
-2127.2
-755.7
Increase Decrease
Percentage
change
-1.59
-0.87
-2.43
-0.67
Hainan
15939
23
131.1
-108.1
-0.68
Chongqing
Sichuan
Guizhou
Yunnan
Tibet
Shaanxi
Gansu
Qinghai
Ningxia
Xinjiang
Taiwan
55991
175027
63814
67706
1937
40855
32379
2854
9632
30567
13437
23.7
68.9
27.6
94.4
0
170.1
316.6
4.6
249.4
1781.6
6.9
2111.1
1553
1103.9
981.6
1.3
550.1
408.1
21.6
215.2
46.1
181.8
-2087.3
-1484.1
-1076.3
-887.2
-1.3
-380.1
-91.5
-17
34.2
1735.6
-174.9
-3.73
-0.85
-1.69
-1.31
-0.07
-0.93
-0.28
-0.59
0.35
5.68
-1.3
Change of total production potential associated with changes in cultivated land by provinces
for 2000-2008, measured in billion Kcal and percentage change (%)
Changes in potential agricultural productivity
and production due to land conversions
 During 1988-2000, the quantity of cultivated land rose by 1.9%.
The average potential productivity of land fell by only 2.2%
 During this period, the quantity of cultivated land deceased 0.58
million hectares and the average potential productivity of land
fell by 1.7%
Panel a: 1988 to 2000
Panel b: 2000 to 2008
Changes in total production potential (measured in million kcal) associated with changes in
cultivated area in China during 1988 to 2000 (Panel a) and during 2000 to 2008 (Panel b).
Summary
Indeed, net cultivated land actually increased during the
study period, 1986 to 2000. Our decomposition of
cultivated land changes show that nearly half of lost
cultivated land was due to cultivated land being
converted to grassland (30%) and forest (17%). Of the
remaining, nearly 40% was due to the shift to built-up
area.
There also was a considerable amount of newly
cultivated land created, some shifting into cultivation
from grassland and other from forestry areas
Although newly cultivated area rose, average potential
agricultural productivity actually fell
Answer questions from two aspects
The direct impacts of urbanization on
the land productivity
Impact of urbanization on cultivated
land changes in China
Three kinds of models of urbanization
Indicator measured by remote sense digital images
– Exploring the sizes of “villages”, “towns”, “cities” by a sampled
survey
– Raw data, Landsat TM/ETM, CBERS
– Four time period: late 1980s, mid-1990s, late 1990s, mid-2010s (Liu et
al, 2002; Liu et al, 2009 )
Aggregated based on neighborhood of the residential
polygons with the reference for the year 2005
– 18 counties/cities within nine provinces sampled and re-visited which
are located in the eastern, central and western regions of China
– The threshold to identifying the three levels of residential areas
• Villages, equal or less then one square kilometer
• Towns, one to five square kilometers
• Cities, bigger then five kilometers
Three kinds of urbanization models
Urbanization of Beijing
Urbanization of Fangshan
Village Model
Urbanization of Yancun
City Model
Town Model
Beijing Municipal Institute of City Planning & Design, 2007
Controlling and influencing factors
 Geophysical variables
– slope, percent of plain area, elevation, distances of counties’ (cities’) seats
to the capital cities and nearest port cities, and so on.
• DEM data and topographic map
• Thematic maps on residence and road network
– precipitation and average temperature data
• China Meteorological Administration during 1950-2000
 Economic variables
– economic data and population of counties (cities)
• National and provincial bureaus of statistics, various years
– FDI
– development zones
 Policy variables
• Household registration policy
• County updated to city
Descriptive statistics of the main variables
1996
Variables
Unit
2000
2008
mean
std.
deviation
mean
std.
deviation
mean
std.
deviation
cultivated land area
hectare
69277
66091
68250
67952
66062
68643
cultivated land area (1989)
hectare
66630
61212
66630
61212
-
-
cultivated land area (1995)
hectare
-
-
69277
66091
69277
66091
village-model land ratio
percentage
67.16%
22.01%
65.71%
21.78%
60.54%
22.36%
town-model land ratio
percentage
12.00%
11.57%
12.66%
11.87%
15.96%
12.97%
city-model land ratio
percentage
20.84%
20.84%
21.63%
22.21%
23.50%
23.50%
percentage
23.76%
20.38%
25.19%
20.43%
45.51%
34.83%
0.24
0.43
0.26
0.44
0.29
0.45
2212
2314
3911
3894
5200
5019
0.36
0.48
0.37
0.48
0.41
0.5
built-up area：
policy factors：
non-agricultural population
registered (t-1)
County upgraded to city (yes =1)
foreign direct investment per capita
development zone (exist =1)
yuan per capita
Descriptive statistics of the main variables
(continued)
1996
variable
unit
mean
2000
2005
std.
deviation
mean
std.
deviation
mean
std.
deviation
9603
6567.68
16370
9467
18823
Economic factors:
GDP(t-1)
million yuan* 4331.98
Agriculture GDP(t-1)
million yuan
825.86
591.79
986.72
696.27
1338.68
962.11
Industry GDP(t-1)
million yuan
2004.27
4980.47
3043.47
8167.19
4324.11
11028
Service Industry GDP(t-1)
million yuan
1501.85
4647.55
2537.49
8225.31
3804.21
13123
Person
631933
607333
653616
628337
731665
821621
Degree
2
2
2
2
2
2
distance to the nearest port
Kilometer
467
342
467
342
467
342
distance to the capital city
Kilometer
164
96
164
96
164
96
Meter
233
255
233
255
233
255
plain area proportion
percentage
0.53
0.38
0.53
0.38
0.53
0.38
annual precipitation
mm
1016
510
1016
510
1016
510
average temperature
℃
13
6
13
6
13
6
population(t-1)
Environmental factors:
slope
DEM
Observations
870
870
879
Equations
Cultivated land area = f (ratio of urbanization
models, social and economic variables, geophysical
variables, other control factors, random error term)
Build-up areas of urbanization models = f (social
and economic variables, geophysical variables, other
control factors, random error term)
Estimation results, for the eastern region, the decision
factors of urbanization models and cultivated land,
1995-2000, (Pooled OLS)
Explanatory variables
"Small-town"
model proportion
"City" model
proportion
Explained variables in 1989
"Small-town" model proportion
0.084
(77.85)***
0.853
(64.85)***
"City" model proportion
Policy instrumental variables
Non-agri population registered (t-1)
County updating to city (yes=1)
Foreign direct investment per
capita
Development zone (exist=1)
-0.025
(2.36)**
0.004
-1.22
0.001
-0.82
-0.01
(2.83)***
0.087
(5.96)***
0.013
(3.02)***
-0.003
(1.30)
0.011
(2.50)**
Cultivated land
area(3SLS)
1.018
(199.39)***
0.117
(4.25)***
0.036
(1.80)*
Explanatory variables
Town-area land
expansion
City-area land
expansion
cultivated land
area
-0.006
(1.98)**
0.002
-0.65
0.003
-0.87
-0.003
(-0.71)
0.002
(0.42)
0.005
(1.45)
0.014
(3.00)***
-0.022
(4.85)***
0.021
(3.25)***
-0.009
(1.66)*
-0.025
(3.66)***
-0.006
(0.80)
-0.002
(2.15)**
0.001
(-0.87)
-0.000
(0.02)
-0.000
(0.33)
-0.001
(0.22)
-0.007
(1.17)
0.000
(1.06)
0.78
1738
-0.002
(1.88)*
-0.004
(2.70)***
-0.007
(2.96)***
0.002
(2.66)***
-0.012
(1.53)
-0.057
(7.97)***
0.005
(9.34)***
0.91
1738
-0.001
(0.30)
0.007
(2.83)***
0.02
(5.48)***
-0.000
(0.06)
-0.043
(3.44)***
0.041
(3.89)***
-0.004
(4.58)***
0.99
1738
Socio-economic factors
Agriculture GDP(t-1)
Industry GDP(t-1)
Service Industry GDP(t-1)
Population(t-1)
Physiographic factor
Slope
Distance to the nearest port
Distance to the capital city
DEM
Plain area proportion
Average precipitation
Average temperature
R2
Observations
Estimation results for the eastern region, the decision
factors of urbanization model and cultivated land,
2000-2008, (Pooled OLS)
Explanatory variables
"Small-town"
model proportion
"City" model
proportion
Explained variables in 1989
"Small-town" model proportion
0.826
(54.23)***
0.793
(64.47)***
"City" model proportion
Policy instrumental variables
Non-agri population registered (t-1)
County updating to city (yes=1)
Foreign direct investment per
capita
Development zone (exist=1)
0.021
(2.10)*
-0.005
(-0.95)
0.003
(1.14)
0.000
(0.03)
0.037
(3.51)***
-0.001
(-0.13)
-0.008
(-2.88)***
-0.005
(-0.95)
Cultivated land
area(3SLS)
1.023
(128.84)***
-0.146
(2.92)***
-0.067
(2.01)*
Explanatory variables
Town-area land
expansion
City-area land
expansion
cultivated land
area
0.015
(3.59)***
0.004
(1.04)
-0.015
(-2.69)***
-0.012
(-2.07)*
-0.018
(-4.24)***
0.012
(2.75)***
0.022
(3.82)***
-0.007
(-1.25)
0.014
(1.38)
-0.003
(0.34)
-0.037
(3.25)***
0.038
(3.06)***
0.004
(3.03)***
-0.011
(-3.49)***
-0.000
(-0.13)
-0.001
(0.89)
0.017
(1.77)
0.008
(1.01)
-0.000
(-0.60)
0.56
1738
-0.001
(-0.46)
-0.001
(-0.36)
0.004
(1.28)
-0.000
(-0.30)
0.001
(0.13)
-0.012
(-1.47)***
0.004
(5.37)***
0.85
1738
0.005
(1.38)
0.002
(0.39)
-0.003
(-0.41)
-0.008
(-3.42)***
0.009
(0.43)
-0.091
(-5.42)***
0.000
(0.06)
0.97
1738
Socio-economic factors
Agriculture GDP(t-1)
Industry GDP(t-1)
Service Industry GDP(t-1)
Population(t-1)
Physiographic factor
Slope
Distance to the nearest port
Distance to the capital city
DEM
Plain area proportion
Average precipitation
Average temperature
R2
Observations
The estimation result of urbanization
models
 In 1995-2000, the household registration policy has significant
different impacts on different urbanization models，the influence
coefficient is -0.025；However, the household registration policy
has significant positive influence on “City” model，the
coefficient is 0.087
 In 1995-2000, the implementation of county to city (or district)
has a negative effect on “Small-town” model, but has a significant
positive effect on “City” model, the influence coefficient of
estimation is 0.013, and statistical tests of coefficient are
significant at the level of 1%. However, in 2000-2008, the
estimation results show that the development of county to city
(district) has an effect on “Small-town” model and “City” model
The estimation results of urbanization models
Further, In 1995-2000, the effects of the foreign direct
investment on both “Small-town” model and “City”
model are not significant, but in 2000-2008, the effect
was significant in “City” model, the influence
coefficient is 0.008
The regional development policy has a significant
negative effect on “Small town” model, but has a
positive effect on “City” model, this is because the
development zones are generally set up around the town
and consequently promote its expansion
The estimation results of the cultivated
land model
In 1995-2000, the cultivated land which was used in the
small town model and city model is more economical
than in village model, and the less percent of occupied
the cultivated land is 0.12% and 0.04%, respectively
In 2000-2008, the cultivated land which was more used
in the small town model and city model than in village
model, and the urbanization level increases every one
percent will lead to the cultivated land which was
occupied by construction land increase more than 0.15%
and 0.07%.
The estimation results of the cultivated
land model
 Natural factors are also the important factors to explain regional
differences of cultivated land. In the seven natural factors which
are considered, there are five variables are reached 1% significant
level, they are the nearest distance to the province capital, the
nearest distance to the port, the ratio of plain area, precipitation
and average temperature
 In 2000-2008, when the agricultural GDP growth by 1%, the
cultivated land will increase (or save) about 0.02%. This is
because agriculture development needs a large number of
cultivated land, the more agriculture develop, the more cultivated
land used for farming
 In 2000-2008, as industrial GDP or service industry GDP growth
by 1%, the cultivated land will reduce about 0.003% and 0.037%,
respectively
Decomposition analysis of cultivated land
change;(1996-2000)
Estimated
parameter
[1]
Variation（%）[2]
Influence
[3]=[1]×[2]
Rate of contribution
（%）[4]=[3]/
（-1.48）*100
“Village”(Rural residential)
Town
City
Agriculture GDP*（t-1）
Industry GDP*（t-1）
0.117
0.036
0.021
-0.009
0.66
0.79
18.08
40.45
0.08
0.03
0.37
-0.35
-5
-2
-25
24
Service industry GDP*（t-1）
-0.025
50.07
-1.27
86
Population *（t-1）
Other variables
-0.006
3.12
-0.02
-0.32
1
21
Variable
Urbanization (construction land area
ratio)
Change of cultivated land area（%）
-1.48
Note: Italicized data represents that the coefficient based on the decomposition analysis is not significant
100
Decomposition analysis of cultivated land
change;(2000-2008)
Estimated
parameter
[1]
Variation（%）
[2]
Influence
[3]=[1]×[2]
Rate of contribution
（%）[4]=[3]/
（-1.48）*100
“Village”(Rural residential)
Town
City
Agriculture GDP*（t-1）
Industry GDP*（t-1）
-0.146
-0.067
0.014
-0.003
3.33
1.87
35.67
42.08
-0.486
-0.125
0.499
-0.126
23.2
6.0
-23.8
6.0
Service industry GDP*（t-1）
-0.037
49.92
-1.847
88
Population *（t-1）
Other variables
0.038
11.94
0.454
1.150
-21.6
77.7
Variable
Urbanization (construction land area
ratio)
Change of cultivated land area（%）
-2.1
Note: Italicized data represents that the coefficient based on the decomposition analysis is not significant
100
Summary
Assuming that other factors remain constant, in 19952000 of eastern region, the urbanization alleviates the
loss of cultivated land by 7%, compared with the
expansion of villages or the development of small towns
In the period of 2000-2008, the rapid urbanization
resulted in the cultivated land loss by 29.2%. The
policies designed to protect cultivated land by
encouraging people move to small towns may actually
accelerate the occupation of cultivated land
Concluding remarks
 We saw net cultivated land actually increased during the
study period 1986 to 2000. Although newly cultivated area
rose, average potential agricultural productivity actually fell.
 Despite this, when examined in the aggregate for the entire
period, the effect on total agricultural potential output was
negligible.
 Economic growth is the major determinant of any changes in
cultivated land use
 social, economic, and geophysical factors, such as industrial
structure, population growth…played an important role in
influencing urbanization
 Although urbanization has an effect on the changes of
cultivated land, its effect is marginal
2nd International Conference
Urban Transitions and Transformations: Science,
Synthesis and Policy
Thank you for your attention
Xiangzheng Deng
November 6-8，2014