(Manavalan et al. 1995), detecting land cover change (Kleynhans et al. 2011;
Kaufmann and Seto 2001), detecting mining process and land use change (Prakash
and Gupta; 1998), and monitoring landscape change of coastal area (Alphan 2011).
Using PCA method, land cover change (Byrne et al. 1980; Parra et al. 1996), forest
conversion (Jha and Unni 1994) can be detected. CVA also can be used in
vegetation degradation detection (Lunetta et al. 2004), desertification monitoring
(Dawelbait and Morari 2012), and LULC change detection (Song et al. 2012). As
for PCCD method, thematic maps and valuable “from-to” change information can
be obtained from PCCD (Jensen 2005). Therefore, many applications are focusing
on LULC change and urban growth employed PCCD method to identify specific
categories of LULC. Thus to explore the change pattern and change effect on
surrounding environment (Abd El-Kawy et al. 2011; Yuan et al. 2005;
Sundarakumar et al. 2012; Peiman 2011) has been stressed.
To monitor nation-wide LULC change of the U.S. and evaluate and manage the
consequences of change, USGS had developed a Land Cover Trends (LCT) project
to detect LULC changes at ecoregional scale for the 1972–2000 period using
Landsat data (USGS 2013a, b, c, d, e, f). The PCCD method has been employed
to obtain specific “from-to” information (which LULC classes are changing, what
they are changing to, and how much they change) and monitor LULC change
dynamics (Sleeter et al. 2012). The study of Mojave Basin and Range Ecoregion
is a typical example of LULC change detection. Since Las Vegas is one of the
fastest growing cities in the U.S., significant urban growth in place of grassland has
been detected. It showed that the most rapid growth happened during 1986–1992. In
2011, Huang et al. applied PCCD method using the Iterative Self Organizing Data
Analysis (ISODATA) classifier to analyze urbanization process and its effect on
irrigation districts of the Lower Rio Grande Valley in the south of Texas. Using the
same PCCD method, Tan et al. (2009) evaluate the impact of land surface temperature by monitoring urban expansion based on LULC maps which were classified
by maximum likelihood classifier (MLC) in Penang Island, Malaysia. For spatial
progressive urban growth mapping of Atlanta metropolitan area Yang (2002) and
Yang et al. (2003) designed a change detection scheme based on multi-temporal
map-by-map comparison. Similarly, Yin et al. (2011) detected urban growth
dynamics applying multi-temporal change detection scheme. They evaluated how
Shanghai metropolitan area conformed to the “reform and opening-up” policy. In
addition, Yin et al. (2011) generated radar graphs to illustrate spatial orientation of
LULC change. Moreover, other studies, conducted by Yuan et al. (2005),
Sundarakumar et al. (2012), Tang et al. (2008), Afify (2011), and Abd El-Kawy
et al. (2011) proved that PCCD is a very useful and popular approach for LULC
change detection.
According to previous studies, most of the urban area change detection analyses
were conducted based on bi-temporal scheme (Afify 2011) or coarsely multitemporal scheme (Abd El-Kawy et al. 2011; Yuan et al. 2005; Sundarakumar
et al. 2012; Peiman 2011; Tian et al. 2011; Zha et al. 2003; Zhao et al. 2005).
With easy accessibility of data availability recently, more and more studies used
multi-temporal datasets to detect change dynamics of urban area. However, as
70
A. Fu et al.
Kaufmann and Seto 2001), detecting mining process and land use change (Prakash
and Gupta; 1998), and monitoring landscape change of coastal area (Alphan 2011).
Using PCA method, land cover change (Byrne et al. 1980; Parra et al. 1996), forest
conversion (Jha and Unni 1994) can be detected. CVA also can be used in
vegetation degradation detection (Lunetta et al. 2004), desertification monitoring
(Dawelbait and Morari 2012), and LULC change detection (Song et al. 2012). As
for PCCD method, thematic maps and valuable “from-to” change information can
be obtained from PCCD (Jensen 2005). Therefore, many applications are focusing
on LULC change and urban growth employed PCCD method to identify specific
categories of LULC. Thus to explore the change pattern and change effect on
surrounding environment (Abd El-Kawy et al. 2011; Yuan et al. 2005;
Sundarakumar et al. 2012; Peiman 2011) has been stressed.
To monitor nation-wide LULC change of the U.S. and evaluate and manage the
consequences of change, USGS had developed a Land Cover Trends (LCT) project
to detect LULC changes at ecoregional scale for the 1972–2000 period using
Landsat data (USGS 2013a, b, c, d, e, f). The PCCD method has been employed
to obtain specific “from-to” information (which LULC classes are changing, what
they are changing to, and how much they change) and monitor LULC change
dynamics (Sleeter et al. 2012). The study of Mojave Basin and Range Ecoregion
is a typical example of LULC change detection. Since Las Vegas is one of the
fastest growing cities in the U.S., significant urban growth in place of grassland has
been detected. It showed that the most rapid growth happened during 1986–1992. In
2011, Huang et al. applied PCCD method using the Iterative Self Organizing Data
Analysis (ISODATA) classifier to analyze urbanization process and its effect on
irrigation districts of the Lower Rio Grande Valley in the south of Texas. Using the
same PCCD method, Tan et al. (2009) evaluate the impact of land surface temperature by monitoring urban expansion based on LULC maps which were classified
by maximum likelihood classifier (MLC) in Penang Island, Malaysia. For spatial
progressive urban growth mapping of Atlanta metropolitan area Yang (2002) and
Yang et al. (2003) designed a change detection scheme based on multi-temporal
map-by-map comparison. Similarly, Yin et al. (2011) detected urban growth
dynamics applying multi-temporal change detection scheme. They evaluated how
Shanghai metropolitan area conformed to the “reform and opening-up” policy. In
addition, Yin et al. (2011) generated radar graphs to illustrate spatial orientation of
LULC change. Moreover, other studies, conducted by Yuan et al. (2005),
Sundarakumar et al. (2012), Tang et al. (2008), Afify (2011), and Abd El-Kawy
et al. (2011) proved that PCCD is a very useful and popular approach for LULC
change detection.
According to previous studies, most of the urban area change detection analyses
were conducted based on bi-temporal scheme (Afify 2011) or coarsely multitemporal scheme (Abd El-Kawy et al. 2011; Yuan et al. 2005; Sundarakumar
et al. 2012; Peiman 2011; Tian et al. 2011; Zha et al. 2003; Zhao et al. 2005).
With easy accessibility of data availability recently, more and more studies used
multi-temporal datasets to detect change dynamics of urban area. However, as
70
A. Fu et al.