A gradient analysis with changing grain size in Shanghai, China

1
A gradient analysis with changing grain size in
Shanghai, China

• March 11, 2008

2
Introduction (a)

• Urbanization Globalization
• Urbanization one of the most important driving
  forces for LUCC
• Urban area 2 of earth land surface
• Urban population 50 of worlds population in
  2000, expected to be 60 in 2025
• Affects land use elsewhere through ruralurban
  linkage

3
Introduction (b)

• Urban ecology
• Definition in 1925 by Chicago scientists the
  study of the relationship between people and
  their urban environment
• A major goal to understand relationship between
  the spatial pattern of urbanization and its
  ecological processes

4
Introduction (c)

• Method
• Gradient analysis spatial view
• Landscape pattern metrics quantitive method
• Provide a way to relate the spatial pattern of
  urbanization with its ecological process

5
Introduction (d)

• Gradient analysis
• Landscape structure gradients can be descriptors
  of spatial variation over various
  landscapes(Linke, 2006 ).
• Since the concept of urban-rural gradients was
  introduced (McDonnell et al. 1997) it has been
  effectively used to study a variety of ecological
  issues in urbanized areas throughout the world
• By uncovering characteristics of urban
  fragmentation along the gradient of land use
  zones, researchers can understand the spatial
  distribution of urban fragmentation.

6
Introduction (e)

• A gradient analysis of urban landscape pattern a
  case study from the Phoenix metropolitan region,
  Arizona, USA
• (Matthew Luck and Jianguo Wu)
• The spatial pattern of urbanization
  could be reliably quantified using landscape
  metrics with a gradient analysis approach, and
  the location of the urbanization center could be
  identified precisely and consistently with
  multiple indices.

7
Introduction (f)

• Roads
• Roads have important ecological effects in
  landscape
• Impacts of forest roads dissecting the land,
  leading to habitat fragmentation, shrinkage, and
  attrition
• Ecological impacts of roads in urban landscape
  have rarely been reported

8
Introduction (g)

• Scaling
• Successful landscape pattern analysis depends
  critically on identifying the appropriate scale
• There is no single correct or optimal scale
  for characterizing spatial heterogeneity
• As a linear patch, roads land use is probably
  quite sensitive to changing scales

9
Logic of Introduction
Global change
LUCC
Urbanization
Urban LUCC
Roads
Gradient analysis
Landscape pattern metrics
10
Questions

• Will landscape pattern (patch density primarily)
  in the transect vary when road patches were
  merged with urban patches? How does the variation
  alter with increasing grain size?
• Do road patches have its unique spatial signature
  comparing with other land use types?
• How do road patches lead to changes in urban
  landscape?
• How do road patches relate urban landscape
  pattern and urbanization process?

11
Methods (a)

• Study area Shanghai
• Location3114'N 12129'E
• Mean annual temperature 16?
• Mean annual precipitation 1200mm
• Area 6340 km2
• South-north length 120km
• East-west distance 100km
• Population 13.34million
• Study transect519km2

12
Methods (b)

• Data inquisition
• ProducerXu Jiangang
• Data sourceIRS imagery
• Time2002
• Resolution5.8m
• Interpretationmanual interpretation with IKONOS

13
Methods (c)

• Data pretreatment ArcView
• Format transformation Arc Coverage ? Shape
• Land use reclassification
• 24?5 land use type Land use transect I
• 5 ? 4 (merge urban and roads) Land use transect
  II
• Transect clipping 519km2
• From vector to raster format Shape ? ArcGrid
• Grain size 7.5m,15m,30m,60m,120m

14
Land use reclassification scheme
15
Study Area
16
Study transects
17
Methods (d)

• Data analysisFragstats 3.0
• Landscape metrics selected
• Percent Coverage The proportion of total area
  occupied by a particular patch type
• Patch Density (PD) The number of patches of per
  100 ha
• Analysis typestandard moving window analysis
• Output statisticsclass landscape-level

18
Fragstats (a)
19
Fragstats (b)
20
Results (a)

• Percent Coverage
• Class-level
• Standard analysis
• Moving window analysis

• Patch Density
• Class-level
• Standard analysis
• Moving window analysis
• Landscape-level
• Standard analysis
• Moving window analysis

21
Results (b)

• Percent Coverage
• Standard analysis
• A agriculture
• B green space
• C1 urban in LUT-I
• C2 urban in LUT-II
• D water
• E roads

22
Results (c)

• Percent Coverage
• Moving window analysis
• A agriculture
• B green space
• C1 urban in LUT-I
• C2 urban in LUT-II
• D water
• E roads

23
Results (d)

• Class-level
• Standard analysis
• (a) agriculture
• (b) green space
• (c) urban in LUT-I
• (d) urban in LUT-II
• (e) water
• (f) roads

24
Results (e)

• Class-level
• Moving window analysis
• (a) agriculture
• (b) green space
• (c) urban in LUT-I
• (d) urban in LUT-II
• (e) water
• (f) roads

25
Results (f)

• Landscape-level
• Standard analysis
• (a) LUT-I
• (b) LUT-II
• (c) LUT-I excluding road patches
• (d) LUT-I excluding road and water patches

26
Results (g)

• Landscape-level
• Moving window analysis
• (a) LUT-I
• (b) LUT-II
• (c) LUT-I excluding road patches
• (d) LUT-I excluding road and water patches

27
Discussion (a)

• Will landscape pattern (patch density primarily)
  in the transect vary when road patches were
  merged with urban patches? How does the variation
  alter with increasing grain size?
• Yes, see results

28
Rural-urban-rural gradient

• Hypotheses (Gordon Forman)
• patch density increases exponentially along a
  landscape modification gradient (i.e.,
  natural-managed-cultivated-suburban-urban)
• LUT-I agree with the hypotheses
• LUT-II disagree with the hypotheses

29
Discussion (b)

• Do road patches have its unique spatial signature
  comparing with other land use types?
• Yes, its sensitive to changing grain size.
• See results

30
Difference with Wu
Phoenix city (by Wu) Smallest grain 30m
Land use transect I Smallest grain 7.5m
31
Discussion (c)

• How do road patches lead to changes in urban
  landscape?
• Roads fragmented with increasing grain size
• See next slide

32
Changes in Land use transect I and II with
changing grain size within a 960960m2 square in
urban center
(I and II stand for vector Land use transect I
and II respectively, and the numbers behind I and
II represent grain size of raster map, e.g., I-15
stands for Land use transect I with grain size
1515m2)
33
Details of change in LUT-I
34
Discussion (d)

• How do road patches relate urban landscape
  pattern and urbanization process?
• Urban development ? build more roads ? urban
  landscape fragmented

35
Conclusions (a)

• Road patches played an important role in urban
  landscape pattern which cant be neglected. Urban
  landscape would be totally changed when road and
  urban patches were merged.
• Roads had a unique spatial signature with
  increasing grain size and displayed a different
  landscape pattern from other land use types in
  urban area.

36
Conclusions (b)

• Roads was sensitive to grain size 3030m2 because
  most of the roads in the study area were 10-30m
  wide.
• High percent coverage of roads indicated high
  patch density of landscape. A major ecological
  impact of roads in the process of urban land
  transformation was leading to habitat
  fragmentation.

37
Earth's City Lights
38
Earth's City Lights-China
39
(No Transcript)
40
Fractional Cover Map
41
Global change
LUCC
Urbanization
Urban LUCC
?
Quantification of urbanization level
42
Methods of assessing urbanization level

• Population method (Commonly used)
• LUp/Tp
• L urbanization level
• Up urban population
• Tp total population
• City lights method
• urbanization level is positive correlated with
  brightness of light at night
• New method?

43
Quantification of urbanization level with land
use pattern analysis

• Calculating landscape metrics
• Percent coverage
• Patch density
• Other land use metrics

44
(No Transcript)
45
Thank you!