Science of the Total Environment 624 (2018) 1561–1576

Contents lists available at ScienceDirect

Science of the Total Environment

journal homepage: www.elsevier.com/locate/scitotenv

Detecting the response of bird communities and biodiversity to habitat

loss and fragmentation due to urbanization
Xi Xu a,b, Yujing Xie a, Ke Qi a, Zukui Luo c, Xiangrong Wang a,⁎
a
Department of Environmental Sciences and Engineering, Fudan University, 220 Handan Road, Shanghai 200433, PR China
b
Institute of Green Investment, Detao Group, 2200 Wenxiang Road, Shanghai 201620, PR China
c
School of Environment and Life Sciences, Kaili University, Kaili, Guizhou, 556011, PR China

H I G H L I G H T S G R A P H I C A L A B S T R A C T

• Species richness of bird communities Fig. 10 Redundancy analysis biplots showing correlation between habitat loss and fragmentation and bird species
varied with the process of urban expan- biodiversity.
sion.
• The MSA could account for the process
of species replacement due to urbaniza-
tion.
• Birds inhabiting wetland and open
spaces were mostly affected by the hab-
itat change.
• The area and connectivity of wetlands
positively affected bird diversity.

Food type Habitat type Seasonal status

a r t i c l e i n f o a b s t r a c t

Article history: Birds are considered a good model for indicators of biodiversity response to habitat variations, as they are very
Received 20 July 2017 sensitive to environmental change. However, continuous observations of habitat alterations from undisturbed
Received in revised form 11 December 2017 landscapes to human-dominated ones, as well as the associated effects on bird biodiversity, are lacking. In this
Accepted 12 December 2017
study, New Jiangwan Town in Shanghai, China was selected to illustrate the response of bird species, and thus
Available online 28 December 2017
biodiversity, to habitat loss and fragmentation. Land use/land cover (LULC) data and bird records from 2002 to
Editor: Jay Gan 2013 were collected and analyzed. The results suggested that, due to urban sprawl, the area of wetland and
shrub land had dropped by 82.4% and 87.3% by the end of 2013. Four different urbanization stages were identified
Keywords: in terms of the spatio-temporal variations in the landscape. To measure bird biodiversity, species richness and
Bird diversity relative abundance were calculated, and they could account for the overall trend in biodiversity but might
Landscape change mask the process of species replacement. As an indicator of biodiversity accounting, the mean species abundance
Mean species abundance (MSA) of the original species would not include exotic or invasive species in its calculation, and its value de-
Biodiversity accounting creased from 100% to 76.8% to 52.2% to 24.5% in the four corresponding stages. Finally, suggested by redundant
Urbanization
analysis, the effects of habitat loss and fragmentation on bird biodiversity differed in various bird communities,
and the area and connectivity of wetlands were the most significant variables. Our findings could provide impor-
tant information to inform bird biodiversity protection and habitat restoration.
© 2017 Elsevier B.V. All rights reserved.

1. Introduction
⁎ Corresponding author.
E-mail addresses: [email protected] (X. Xu), [email protected] (Y. Xie), Biodiversity is a multi-dimensional and complex concept that in-
[email protected] (X. Wang). cludes the variations in genes, species, populations, communities, and

https://doi.org/10.1016/j.scitotenv.2017.12.143
0048-9697/© 2017 Elsevier B.V. All rights reserved.
1562 X. Xu et al. / Science of the Total Environment 624 (2018) 1561–1576

ecosystems from local scales to the living world as a whole (Robinson et al., 2006; McKinney, 2008; Reis et al., 2012; Bregman et al., 2014).
et al., 1992). Recently, the global decline and the increase in pressures However, it appears to be an insufficient indicator by itself, as it some-
on biodiversity have been widely discussed (Gaston, 2000; Cockell times increases when new, human-favored species gradually replace
et al., 2006; Schluter and Pennell, 2017), and approximately 12% the original species, which is termed the “intermediate disturbance di-
of birds, 25% of mammals and more than 32% of amphibians are versity peak” (UNSD, 2014). This suggests that biodiversity cannot sim-
estimated to be under the threat of extinction (Millennium Ecosystem ply be characterized by a single measure, such as species richness or
Assessment, 2005b). Additionally, the amount of wildlife is half that of abundance, that tells us little about the composition and function of
40 years ago, according to the “The Living Planet Report 2014” the assemblages that persist in human-disturbed landscapes (Purvis
(McLellan et al., 2014). The worldwide loss of biodiversity is generally and Hector, 2000). Consequently, the index of change in abundance
caused by multiple drivers that operate in space and time. In terrestrial and the distribution of selected species have been chosen by the Con-
ecosystems, landscape change, especially urban expansion, which leads vention on Biological Diversity (CBD, VII/30) to track the process of bio-
to habitat loss and fragmentation, is recognized as one of the most sig- diversity loss (UNSD, 2014).
nificant direct drivers (Ortega-Álvarez and MacGregor-Fors, 2011; Reis However, biodiversity accounting, which follows the general form of
et al., 2012). It is even considered the primary pressure on biodiversity, asset accounting in the System of Environmental-Economic
as suggested by the UN's 3rd Global Biodiversity Outlook, and its influ- Accounting-Experimental Ecosystem Accounting (SEEA-EEA), might
ence continues to intensify (Millennium Ecosystem Assessment, shed new light on the measurement of biodiversity. In fact, the SEEA-
2005a; Marzluff et al., 2012; Aronson et al., 2014). EEA has been applied by a wide range of organizations (such as UNEP,
Birds offer a good system for understanding the effects of urbaniza- OECD and EEA) and in numerous studies (Edens and Hein, 2013;
tion on habitat structure and composition (Reis et al., 2012; Bregman Schröter et al., 2014; Remme et al., 2016). Ecosystem accounting is
et al., 2014), as the available details on the characteristics of bird species emerging as a promising approach for organizing comprehensive eco-
could provide the most comprehensive information about functional system data, tracking changes in ecosystems and linking those changes
traits (Vandewalle et al., 2010). Additionally, birds comprise a broad to economic and other human activities (UNSD, 2014; Hein et al., 2015;
range of species, including urban avoiders and urban-tolerant birds Schröter et al., 2015). Combining ecosystem accounts with landscape
that are very sensitive to environmental change (Gregory et al., 2005; information could provide important data that are directly linked to
Vandewalle et al., 2010). In particular, it is the habitat loss and fragmen- economic units and that might shed new light on the mitigation of eco-
tation caused by urbanization that significantly threatens birds system degradation (Haines-Young and Páramo, 2006). Two key points
(González-Oreja, 2011).Habitat loss and fragmentation can disrupt make ecosystem accounting innovative compared to traditional re-
key biological processes by reducing the breeding success of migrant search tools: it incorporates opening stocks and closing stocks, which
species, limiting dispersal, and decreasing resource acquisition can show a net change, and it allows the relative species abundance to
(Bregman et al., 2014). Generally, the process of urbanization can de- be compared between different periods, different species, and different
crease bird species richness as well as affect the species composition ecosystems (UNSD, 2014). Similarly, biodiversity accounting has been
gradient and increase the biomass and density of bird communities proposed as one of the tools for measuring and monitoring human im-
(Chace and Walsh, 2006; Zhou et al., 2012), and bird community struc- pacts on biodiversity (Cai et al., 2011; Hein et al., 2015). The mean spe-
ture might vary significantly between urbanized and non-urbanized cies abundance (MSA) of the original species has been suggested as an
areas (Reis et al., 2012; Zhou et al., 2012). Furthermore, natural habitat indicator of ecosystem accounting, as it excludes exotic or invasive spe-
loss and fragmentation would also pose a significant threat by affecting cies but their impacts can be examined by the decrease in the abun-
the composition, structure and ecological functions of biodiversity dance of original species being replaced (UNSD, 2014).
(Scolozzi and Geneletti, 2012). The importance of understanding habi- Here, we present an initial assessment of the effects of recent habitat
tat changes and bird biodiversity has been highlighted by many re- change in New Jiangwan Town, which is located in the northeast of the
searchers (Carrete et al., 2009; Scolozzi and Geneletti, 2012; Quesnelle central area of Shanghai City, China, focusing on the response of bird
et al., 2013; Blandón et al., 2016; Keinath et al., 2017). Biodiversity is communities and biodiversity to habitat loss and fragmentation trig-
commonly compared among different sampling areas, such as suburbs gered by urbanization. New Jiangwan Town was a relatively undis-
(MacGregor-Fors, 2008), urban-rural gradient (Yeh and Huang, 2009), turbed area before 2001, but it became urbanized due to extensive
new areas of urban growth (Zhou et al., 2012) and the green spaces in construction since the end of 2003, which makes it a good site for de-
different urban districts (Sandström et al., 2006). Several studies of tecting the impacts of rapid urbanization on bird biodiversity. More
the change in amphibian species through long-term observation has than ten years of land use/land cover (LULC) data and bird records
been conducted to identify the time-lag of the response of biodiversity from 2002 to 2013 were collected. We sought to address the following
to human activity (Löfvenhaft et al., 2004; Whitworth et al., 2017). research questions. 1) How have landscape and habitat structure
However, such research is still lacking, especially studies examining changed across space and time in our study area? 2) Have the number
the influence of habitat loss and fragmentation on bird biodiversity of species of various species groups and their relative abundance varied
from undisturbed landscapes to human-dominant ones through contin- with different levels of urbanization? 3) What information does the in-
uous observation (Jokimäki et al., 2011; Ciach & Fröhlich, 2016). dicator of MSA provide for demonstrating the process of species re-
Many difficulties still exist when measuring the process of biodiver- placement and biodiversity change? 4) How have bird communities
sity loss, as biodiversity is a broad concept that refers to all biotic varia- and thus biodiversity responded to habitat loss and fragmentation?
tions from the level of genes to species to ecosystems (Purvis and
Hector, 2000). Biodiversity can be measured based on the richness 2. Methods
and distribution of species, as well as the change in species “abun-
dance”, and it is usually characterized by the decrease in some original 2.1. Study area
species and the increase in some other, opportunistic species(Pfisterer
and Schmid, 2002). Although the decrease in original populations is a New Jiangwan Town (121°29′12″~121°31′47″E and 31°18′39″
significant signal of biodiversity loss, highly proliferating species may ~31°20′57″N) is in the northeast of Shanghai City, China, the coastal
sometimes invade, and their infestations might even become plagues area of which is in the middle of the Asia-pacific bird migration route,
(UNSD, 2014). making this triangular area a wildlife shelter of local, national and
Species richness is the most commonly recognized facet of biodiver- even international significance. New Jiangwan Town is approximately
sity (Purvis and Hector, 2000), and it has been widely applied as an in- 6 km from the mouth of the Huangpu River and 10 km from the
dicator of the effects of urban growth on biodiversity change (Ditchkoff mouth of the Yangtze River (Fig. 1), and it covers an area of 6.56 km2.
 X. Xu et al. / Science of the Total Environment 624 (2018) 1561–1576 1563

Fig. 1. Location of the study area.

From the 1930s to the 1990s, New Jiangwan Town was a relatively were employed to analyze the spatial-temporal patterns of landscape
enclosed military airport, and after the airport closed, the landscape change and were related to the analyses of bird species (Melles et al.,
remained undisturbed from 1994 to 2001 and provided natural habitat 2003; Sundell-Turner and Rodewald, 2008; Wood and Quinn, 2016).
for bird species and other rich flora and fauna that was even better than Nine landscape metrics were selected, including the proportion of the
that of the surrounding parks (Jin et al., 2004). However, at the end of land type (PLAND), patch density (PD), largest patch index (LPI), edge
2003, urban expansion was initiated by the construction of the density (ED), landscape shape index (LSI), Shannon's diversity index
Jiangwan Campus of Fudan University in the northwest corner of New
Jiangwan Town (Fig. 1). The Jiangwan Campus of Fudan University,
which has an area of 0.9 km2, lies in the central area of wetlands in
New Jiangwan Town. Continuous bird watching suggests that the com- Table 1
Land cover types in New Jiangwan Town.
position and number of bird species varied throughout the study period,
and in situ observations show that avian biodiversity has greatly de- Land cover type Detailed description
clined in the coastal areas of Shanghai during recent decades due to Built-up area With no/scattered vegetation (0–10% of total area, excluding
rapid urbanization (Cai et al., 2011). biotopes N100 m2)
With sparse vegetation (10–30%)
With dense vegetation (30–50%)
2.2. Data and methods
Forest/woodland Deciduous forest
Coniferous forest
2.2.1. Land cover and landscape metrics Mixed coniferous and deciduous forest
The land cover information was mainly gathered from historical sat- Young plantation
ellite imagery of New Jiangwan Town from Google Earth, which has Shrub land Semi-open area, wooded grassland, Boscage: shrub area with
high grass
been widely used in previous studies (Gulezian and Nyberg, 2010; Grassland Grassland, dry/mesic/moist
Pettit et al., 2011), for nine different years: 2000, 2002, 2004, 2005, Wetland Wetland with sparse tree cover
2006, 2008, 2009, 2012, and 2013. The land cover maps were rectified Wetland with dense tree cover
by field investigation and comparisons with historical maps, such as Wetland with boscage
Wet Forest
the 2002 land use and land cover (LULC) data, obtained from SPOT 5. Ac-
Grassy shallow water
cording to the habitat types of different birds, the land cover datasets Water with Floating vegetation
were classified into seven types, including built-up land, forest/wood- Seasonally flooded area
land, shrub land (semi-open areas), grassland (open areas), wetland, Water Perennial open water (lakes and streams)
water, and bare ground; their detailed characteristics are shown in Bare ground Building yard
Abandoned area
Table 1. Metrics at the landscape and land cover classification levels
1564 X. Xu et al. / Science of the Total Environment 624 (2018) 1561–1576

(SHDI), Shannon's evenness index (SHEI), effective mesh size (MESH), ecosystem under minimal human influence, which was 2002, before
and mean patch size (MPS). The calculation of landscape-level indexes the urbanization process began. The mean species abundance (MSA)
(PD, LPI, ED, LSI, SHDI and SHEI) aimed to give an overview of values of selected species were calculated and compared with the base-
the variations in landscape patterns. The variations in bird habitats line condition:
(i.e., wetland, shrub land, and forest land) were measured by class-
level metrics (PD, LPI, MESH, MPS) and PLAND of all the types were X
n
MSA ¼ SAi =n
calculated. The software package FRAGSTATS 4.2(McGarigal et al., i¼1
2002) was used to compute the selected metrics (see also in the
Appendix A). where SA i is the relative abundance of the ith species which is also
among the baseline species, and n refers to the total number of species
under the baseline condition.
2.2.2. Bird data collection and processing
The continuous bird records from 2002 to 2013 were provided by 2.2.3. The Statistical method
the China Bird Record Center (CBRC), which is the first public and pro- The Kolmogorov-Smirnov (K\\S) test was used to detect the nor-
fessional NGO for recording birds in China. Bird data were gathered by mality of distribution of the variables for bird species richness and land-
volunteer citizen-scientists, but each record was under the supervision scape metrics. To meet the requirements for statistical analyses, the raw
of a professional ornithologist. The data was collected every data of all variables were log-transformed. The PASW Statistic 18.0 was
8:00–10:00 am and 4:00–6:00 pm along the line transects. Every line used to calculate the metrics of bird biodiversity, including species rich-
transect was about 500–600 m long, with 50 m wide each side. Field ness, relative abundance. Pearson Correlation analysis was applied to
glasses (Kowa 8*BD42) were used during the investigations while walk- examine the relationship and response of bird biodiversity to urbaniza-
ing at a speed 1.0–1.5 km/h, and birds were recorded when seen, heard tion process.
or flying over. During the line transect investigation the sample point Using CANOCO 4.5 (Braak and Smilauer, 2002), a redundancy analy-
method was also applied to record birds seen or heard within 25 m ra- sis (RDA) was then employed to investigate how habitat loss and frag-
dius, each point staying 8–10 min. Every report covered 7–10 transects mentation explained the variability in bird diversity. Before the RDA,
distributed in the north, south, west, east and middle of the site. The line the bird diversity of different food and habitat types were imported
transects were decided by the professional ornithologist in order to rep- into the software to test if the DCA gradient shaft length was less than
resent the bird situation of the total New Jiangwan Town area. Approx- 3 (Shen et al., 2015). A manual variable selection process was chosen
imately 53 reports of bird observation were obtained in one study area, to identify the significant variables at multiple scales based on the re-
and more than 4 reports were collected in winter and spring in most sults of the Monte Carloper mutation method (n = 499).
years, despite 2010 and 2012 (2 reports in winter and spring). In
order to keep the data comparable, the data in winter and spring sea- 3. Results
sons was used.
Data included the species, number and activities of the birds as well 3.1. The changes in landscape pattern from 2000 to 2013
as information about the surrounding environment. Bird species were
categorized into different groups with respect to their food types, habi- 3.1.1. The temporal and spatial variations in land cover types
tat type and seasonal status according to previous studies (Luan, 2003; The temporal and spatial variations in landscape change in New
Jin et al., 2004; Cai et al., 2011; Zhou et al., 2012). The bird classifications Jiangwan Town from 2000 to 2013 are shown in Fig. 2 and Fig. 3. The re-
are shown in Table 2, and the bird species lists are shown in Appendix B. sults showed that before 2003, more than 50% of the land was covered
The indexes of bird biodiversity, including richness and relative abun- by wetland (53.9%) or water bodies (6.9%), and the built-up area only
dance, were calculated, and to avoid masking any change in species, occupied approximately 11.3% of the total area. A sharp increase in the
the biodiversity accounting method was applied. built-up area and a decrease in wetland area (to 34.4%) occurred in
In this study, the natural stock was the bird species, and the opening 2004–2005, especially in the core wetland area. Then the year of
stock (the reference condition) referred to the condition of the 2006–2008 represented a relatively stable phase for all land cover

100%

Table 2 90%
The bird classification criteria and their attributes.
80%
Bird Classification Subclassification Value
Criteria 70%
Food type Granivores GRA
Insectivores INS 60%
Omnivores OMN
Frugivores FRU
50%
Carnivores CAR
40%
Piscivores PIS
Benthivores BEN
30%
Habitat type Open area species O1
Species that prefer open areas but also use forested O2 20%
areas
Forest species that only use forested areas F1 10%
Forest species that also use open areas F2
Forest species that use boscage areas F3 0%
Swimming birds that use open water W1 2000 2002 2004 2005 2006 2008 2009 2012 2013
Waders W2
Built-up land Bare ground Grassland Shrub land
Species that conceal themselves in marshes and W3
aquatic areas with high grass Forest land Wetland Water
Seasonal status Resident R
Migratory M
Fig. 2. Proportions (%) of different land cover types (2000−2013).
 X. Xu et al. / Science of the Total Environment 624 (2018) 1561–1576 1565

Fig. 3. Spatial distributions of land cover types from 2000 to 2013.

types. However, built-up areas, bare ground and forest land increased shown in Fig. 3. During the first stage (2000–2003), the northwest
during the period of 2009–2013, and by the end of 2013, the proportion part of the study area was dominated by wetlands and water bodies,
of wetland dropped by 9.5%, and the amount of built-up land increased and the southeast part was dominated by shrub land and bare ground.
by 50.3%. During the study period, the proportion of grassland increased During the second stage (2004–2005), the center of the northern wet-
from 0 to 15.8%, but shrub land decreased from 25.3% to 3.2%. land area was drained and built up for the Jiangwan Campus, and as
Accordingly, four stages of urbanization were defined, including the road system was established, the construction of the campus inten-
the stage before construction (2000 − 2003), the early stage of sified. During the third stage (2006–2008), bare ground increased near
rapid construction (2004–2005), the middle stage of gradual construc- the southeast entrance of the campus, and during the final stage
tion (2006–2008) and the final stage of intensive construction (2009–2013), the area around the campus was almost entirely replaced
(2009–2013). The spatial distributions of the land cover changes are by built-up area, with only a small part of the wetland remaining.
1566 X. Xu et al. / Science of the Total Environment 624 (2018) 1561–1576

3.1.2. The overview of landscape change as measured by landscape-level increase and final decrease in the PD values of shrub land and forest land
metrics suggested that the large patches were fragmented into small patches at
As shown in Fig. 4, the variations in PD and ED appeared as inverted the beginning of the urbanization process, and some small patches even
U-shape curves and increased to their highest values during the period disappeared. Combined with Fig. 2, the increase of grassland in PD, LPI,
of 2004–2008. In 2013, the value of PD decreased to the baseline level, MPS and MESH was due to the growth of grassland area during the pe-
but the value of ED was much higher than that in 2000. Considering riod of 2000–2013.
the change in the land cover percentages described in the above section,
it could be concluded that numerous man-made patches replaced natu-
ral patches while patches became fragmented, especially during the pe- 3.2. Variation in bird species with respect to food type and habitat type
riod of 2004–2008. The variation in LSI showed a trend similar to that of
ED. The lowest LPI value in 2004 suggested some of the largest patches According to the bird records from 2002 to 2013, 53 reports were col-
had been converted to urban patches, whereas the highest value in 2013 lected, and approximately 98 species belonging to 31 families were ob-
indicated that the relatively large built-up patches dominated the study served. The variations in bird species composition with respect to food
area, as shown in Fig. 4. The SHDI and SHEI index values fluctuated dur- type are shown in Fig. 6. And overall, the total number of bird species
ing the urbanization process of 2000–2013, appearing as an inverted U- gradually dropped during our study period. The highest number was
shape curve, but the values of SHDI and SHEI at the landscape level gen- 73 in 2003, and the lowest value was 8 in 2010. Omnivores, insectivores
erally ignored the details of the different land cover types by only calcu- and frugivores were the dominant species in terms of bird richness from
lating the overall quantity of landscape diversity. Therefore, the 2002 and 2013, and omnivores exceeded 50% of the total birds during
variations in landscape diversity might only indicate the number of ac- the 2010–2013 period. The benthic fauna (benthivores) and fish eaters
tivity types that the landscape could support but not whether they were (piscivores) were water birds that generally relied on large areas of wet-
for humans or wildlife. lands and thus rarely appeared in the city center (Mackinnon et al.,
2000). The proportions of benthivores and piscivores decreased from
15% to 6.9%, which might be due to the extensive loss and fragmentation
3.1.3. Habitat loss and fragmentation measured by class-level metrics of wetlands under urban construction. The four different food type
As shown in Fig. 5, the most obvious variation occurred in wetlands; groups (BEN, PIS, INS and FRU) have temporal-autocorrelations of bird
from 2000 to 2003, the PD and LPI values of wetlands gently increased richness. The strength of associations varied from strong (BEN, R2 =
and then dramatically declined due to the extensive urban development 0.652, p = 0.003; PIS, R2 = 0.787, p = 0.000), to moderately strong
and construction in 2004. During the period of 2004–2008, the indicator (INS, R2 = 0.530, p = 0.001; FRU, R2 = 0.567, p = 0.007) which
of PD significantly increased and remained high, whereas the LPI approached statistical significance (Fig. 6.). While the OMN (R2 =
remained at a low value, which indicated that the large patches of wet- 0.246, p = 0.121) and CAR (R2 = 0.302, p = 0.08) did not respond
land were fragmented into small ones. The change in the MESH value significantly.
suggested that patch connectivity declined from 2002 to 2013. As indi- The variations in bird species richness according to habitat type are
cated by Fig. 4 and Fig. 5, the loss and fragmentation of wetlands oc- shown in Fig. 7, and it can be seen that birds inhabiting wetlands rapidly
curred throughout the entire period, so the PD and LPI of wetlands decreased during the urbanization process, especially birds inhabiting
dropped to their lowest levels in the last stage of our study period. In W2 and W3. Birds inhabiting forest fluctuated but remained slightly de-
contrast, the increase in the LPI and MPS of built-up land during the en- creased, and birds inhabiting F1 decreased to 0 in 2009. Additionally,
tire study period indicated that numerous habitats have been converted birds inhabiting open areas greatly decreased during the study period,
to urban area and have become scattered around the large urban especially birds living in O2. In general, birds preferring wetland
patches. For the shrub land and forest land, the decrease in LPI and declined fastest at an annual rate of decrease of 7.7% followed by
MPS implied the loss of large patches during the study period. The initial species preferring open areas at an annual rate of decrease of 4.3% and

PD ED LPI
200 55
45
45
35
150 35
25
25
15 100 15
2000 2004 2006 2009 2013 2000 2004 2006 2009 2013 2000 2004 2006 2009 2013

LSI SHDI SHEI

16 1.8 0.9
14
1.6 0.84
12
10
1.4 0.78
8
6 1.2 0.72
2000 2004 2006 2009 2013 2000 2004 2006 2009 2013 2000 2004 2006 2009 2013

Fig. 4. Landscape change as measured by landscape-level metrics from 2000 to 2013.

 X. Xu et al. / Science of the Total Environment 624 (2018) 1561–1576 1567

20 60

16

40
12
PD

LPI
8
20
4

0
0
2000 2002 2004 2005 2006 2008 2009 2012 2013
2000 2002 2004 2005 2006 2008 2009 2012 2013

80
200

60 160

MESH(ha)
MPS(ha)

120
40
80

20
40

0 0
2000 2002 2004 2005 2006 2008 2009 2012 2013 2000 2002 2004 2005 2006 2008 2009 2012 2013

Built-up land Grassland Forest land

Shrub land Wetland

Fig. 5. Habitat loss and fragmentation as measured by class-level metrics from 2000 to 2013.

those preferring forest at 3.2%. The bird composition respect to grassland (O). And the W and O had temporal autocorrelation
habitat requirements distinctly reflected that birds inhabiting in wet- responded statistical significance while the forest birds (R2 = 0.325,
lands (W) decreased more sharply than in forest land (F) and shrub/ p value = 0.067) did not. The strength of association varied from strong

80 20 BEN
15 PIS

R2=0.652 10 R2=0.787
70 10 p =0.000
p =0.003
5
60
0
2002 2013 0
Bird richness

50 2002 2013
-10
40

30 15
INS 15 FRU

20 10 R2=0.530
10
R2=0.567
p =0.01 p =0.007

10 5
5

0 0 0
2002 2003 2004 2005 2006 2007 2008 2009 2010 2012 2013 2002 2013 2002 2013

Year
BEN CAR PIS FRU INS OMN

Fig. 6. Bird species richness categorized by food type from 2000 to 2013.
1568 X. Xu et al. / Science of the Total Environment 624 (2018) 1561–1576

(a)
40
40
35 R2=0.789
35

Bird species richness

p =0.000
Birds species richness
30 30
25 25
20 20
15 15
10 10
5 5
0 0
2001 2003 2005 2007 2009 2011 2013
Year
Year
W1 W2 W3 (b)
18
18
Bird speciess richness

16
R2=0.468
14 16
p =0.020

Bird species richness

12 14
12
10
10
8
8
6
6
4 4
2 2
0 0
2001 2003 2005 2007 2009 2011 2013
Year
Year
O1 O2

35
(c)
30 35
Bird species richness

30 R2=0.325
Bird speices richness

25 p =0.067
25
20
20
15
15
10
10
5
5
0 0
2001 2003 2005 2007 2009 2011 2013
Year
Year
F1 F2 F3

Fig. 7. Bird species richness with respect to habitat type: a) wetland (W); b) shrub/grass land (O); and c) forest land (F) from 2000 to 2013.

(W, R2 = 0.789, p = 0.000) to moderate (O, R2 = 0.468, p = 0.020). adaptive birds, such as Eurasian Tree Sparrow and Vinous-throated
However, the change in the richness of species in different categories Parrotbill, as the damage to the wetlands during the process of intensive
may ignore the process of species replacement, such as the turnover construction made the food resources available to these birds much
from original species to invasive species. richer than before. However, relative abundance quickly declined and
never recovered. Generally, with an increase in human disturbance,
3.3. The dynamics of bird biodiversity from 2002 to 2013 many endemic species would decrease in abundance, while some spe-
cies, those that benefit from disturbed habitats in particular, would in-
3.3.1. Variation in bird richness and relative abundance crease. In our study, the endemic species were gradually replaced by
As shown in Fig. 8, the values of bird richness and relative abundance species, either endemic or exotic, favored by human influences. The var-
fluctuated but declined during the 2000–2013 study period. The bird iation in species richness could imply an overall trend of biodiversity
richness (R2 = 0.567, p = 0.007) was temporal autocorrelation which loss in the study area, but the increase in species richness during the
responded significance while the relative abundance (R2 = 0.205, p = 2010–2013 period might have been caused by the significant increase
0.162) did not. It should be noted that bird relative abundance peaked in invasive birds, which will be discussed further in the following anal-
in 2005, which might have resulted from the increase in some urban- yses. Therefore, the two indexes of richness and relative abundance
 X. Xu et al. / Science of the Total Environment 624 (2018) 1561–1576 1569

80 20
70 18
16

Relative Abundance
60

Species Richness
R2=0.567 14
50 R2=0.205
p =0.007 12
p =0.162
40 10
30 8
6
20
4
10 2
0 0
2002 2007 2012 2002 2004 2006 2008 2010 2012
Year Year

Fig. 8. Variations in bird species richness and relative abundance from 2000 to 2013.

seemed to hardly make clear what happened to species composition grassland which approached statistical significance (Table 3). The con-
and structure under rapid urbanization as bird species were evaluated clusion could be drawn that the habitat loss of New Jiangwan Town
equally whether they were endemic or invasive. Therefore, these two was significantly caused by the increasing of the proportion of built-
indexes might not fully account for the process of biodiversity loss and up area and grassland and the decreasing of shrub land and wetland
ecosystem degradation. (Fig.2 and Fig. 3).
As the PLAND of water bodies and bare ground did not vary sub-
3.3.2. The process of bird replacement and the change in MSA stantially and grasslands were mainly artificial showing negative
According to the four urbanization stages mentioned above, four correlation with bird richness during the entire study period, their
corresponding stages were defined for biodiversity accounting. The de- variations have not been included into the RDA. In addition to habitat
tailed information on bird species in the four stages and the species loss, habitat fragmentation was another important factor that influ-
moving in and out during each stage are shown in Appendix C. During enced bird diversity in this study. Habitat fragmentation generally
the second (2004–2005), third (2006–2009) and forth (2010− 2013) includes the following processes (Scolozzi and Geneletti, 2012):
stages, 35, 14, and 23 bird species moved out, while 9,19and 11speices 1) a reduction in the total habitat area (measured by PLAND of differ-
moved in. Considering the detailed information in Appendix D, it can ent habitat types), 2) an increase in the number of habitat patches
be seen that the urban avoiders moved out and the urban-adapted (measured by PD), 3) an increasing in the isolation of patches (mea-
birds moved into the study area. During the second stage of sured by MESH), and 4) the breaking up of one patch of habitat into
2004–2005, the original water birds, such as Mallards, Spot-billed several smaller patches (measured by LPI and MPS). As shown in
Ducks and Philippine Ducks, were gone, as were the shorebirds, such Table 4, patch size and connectivity of habitats significantly affected
as the Common Greenshank, Wood Sandpiper, Temminck's Stint, and bird biodiversity from the RDA. Loss and fragmentation of wetland,
Pacific Golden Plover. These birds are urban avoiders and could rarely forest land and shrub land could explain more than 80% of the bird
be seen in the city center, only existing in the rural and coastal regions. diversity variation, with largest explanatory ability of 92.4% for the
Meanwhile, more urban-adapted birds (e.g., the Egrets, Eurasian Black- habitat type in the full model.
birds and Sparrows) moved into the study area, which might have led to However, various bird communities with respect to habitat type,
the peak relative abundance value in 2005 (Fig.8.). During the third food type and seasonal status were influenced to different degrees. In
stage of 2006–2008, the remaining shorebirds (e.g., the Green Sandpiper contrast, wetland area and connectivity was the most significant vari-
and Common Sandpiper) continued to decrease along with the birds able influencing the bird communities followed by shrub land and forest
using the near-water vegetation (e.g., the Pied Kingfisher, White- land. The indicator of MESH_W could significantly explain almost half of
breasted Water hen and Yellow Bittern). Simultaneously, birds bird communities' variation, and thus biodiversity, as shown in Fig. 10.
inhabiting grassland and forest land increased including several kinds The wetland bird groups (BEN, PIS, W2 and W3), which represent the
of thrushes and buntings. During the last stage, 23 bird species moved original birds, were significantly negatively affected by loss and frag-
out, and even the various Egrets, small Shorebirds, Swallows, Thrushes mentation of wetland and forest land. The birds that preferred O1, O2
and Buntings became extinct, and the bird species composition became were also affected by wetland patch size, as wetland was usually a
much more similar to that of a residential area or urban center. type of open area, and the boscage area was easily influenced by the
Furthermore, the MSA was calculated for the four different stages wetland. The MESH for shrub land (MESH_S) was tightly correlated
(see Appendix D for more detailed information).As shown in Fig. 9, with the open-area birds (O1, O2 and W1). With respect to seasonal sta-
the MSA values of the original species decreased from 8.28 to 6.36 to tus, all habitat types were contributors to the variation of bird species.
4.32 to 2.00 individuals per survey, respectively. Compared with the Overall, it might be concluded that a large patch area and well-
original/reference species, the bird biodiversity in the second, third connected wetlands were the most important habitats for bird biodiver-
and fourth stages were approximately 76.8%, 52.2% and 24.5% of the sity conservation in our study area.
baseline condition.
4. Discussion
3.4. Effects of habitat loss and fragmentation on bird diversity
4.1. How to measure the dynamic variations in bird diversity due to
Pearson correlation analysis was conducted between biodiversity in- urbanization
dexes and PLAND of different types. The richness, birds of BEN, M, R, F3,
W2, W3 and W are related in positive linear sense with PLAND of shrub Urban sprawl is currently a worldwide phenomenon, especially in
land and wetland but negative with PLAND of built-up area and developing countries, concentrating around urban centre and replacing
1570 X. Xu et al. / Science of the Total Environment 624 (2018) 1561–1576

2002-2003 2004-2005
90 100%
10 160 10
Bird species richness 80 140

Bird species richness

70 8 76.8% 8
120
60
6 100 6

MSA
50

MSA
40 80
4 4
30 60
20 2 40
2
10 20
0 0 0 0
#0050
#0171
#0276
#0348
#0365
#0492
#0542
#0552
#0707
#0821
#0882
#0931
#1145
#1211
#1315

#0050
#0171
#0276
#0348
#0365
#0492
#0542
#0552
#0707
#0821
#0882
#0931
#1145
#1211
#1315
Bird species richness MSA Bird species richness MSA

2006-2009 2010-2013
160 10 32 10
140 9 28 9
Bird species richness

Bird species richness

8 8
120 24
7 7
100 6 20 6
52.2%
MSA

MSA
80 5 16 5
60 4 4
12 24.5%
3 3
40 8
2 2
20 1 4 1
0 0 0 0
0050#
0171#
0276#
0348#
0365#
0492#
0542#
0552#
0707#
0821#
0882#
0931#
1145#
1211#
1315#

0050#
0171#
0276#
0348#
0365#
0492#
0542#
0552#
0707#
0821#
0882#
0931#
1145#
1211#
1315#
Bird species richness MSA Bird species richness MSA

Fig. 9. Value of the MSA indicator at different urbanization stages (the reference species were numbered from #0050 to #1322).

adjacent land cover types such as agriculture and more natural vegeta- Marzluff, 2017). Bird species are the most important indicators of eco-
tion (McGranahan and Satterthwaite, 2003; Pauchard et al., 2006) that system change. In undisturbed ecosystems, bird communities are
generally dramatically changed ecological conditions (Tomasevic and often dominated by urban-avoiders, such as aquatic species that

Table 3
The correlation between bird diversity and the proportion of landscape types.

Pearson Correlation PLAND

Sig. (2-tailed) Built-up area Bare ground Grassland Shrub land Forest Wetland Water

Richness −0.833⁎⁎ 0.285 −0.757⁎ 0.713⁎ −0.260 0.720⁎ −0.118

Abundance −0.165 −0.546 −0.386 0.013 0.562 0.347 0.713⁎
BEN −0.880⁎⁎ 0.120 −0.834⁎⁎ 0.756⁎ 0.053 0.759⁎ 0.167
CAR −0.610 0.339 −0.508 0.588 −0.434 0.495 −0.379
PIS −0.873⁎⁎ 0.123 −0.880⁎⁎ 0.669⁎ −0.086 0.838⁎⁎ 0.109
FRU −0.670⁎ 0.152 −0.697⁎ 0.475 −0.332 0.709⁎ −0.106
INS −0.741⁎ 0.245 −0.685⁎ 0.571 −0.322 0.697⁎ −0.160
OMN −0.611 0.442 −0.455 0.632 −0.365 0.403 −0.318
M −0.823⁎⁎ 0.277 −0.729⁎ 0.706⁎ −0.216 0.698⁎ −0.121
R −0.789⁎ 0.277 −0.748⁎ 0.673⁎ −0.311 0.704⁎ −0.106
F1 0.001 −0.053 0.014 −0.011 −0.269 0.080 −0.233
F2 −0.352 0.647 −0.227 0.434 −0.739⁎ 0.177 −0.677⁎
F3 −0.822⁎⁎ 0.165 −0.780⁎ 0.688⁎ −0.165 0.748⁎ −0.016
F −0.556 0.292 −0.485 0.510 −0.450 0.486 −0.335
O1 −0.773⁎ 0.273 −0.696⁎ 0.675⁎ −0.231 0.652 −0.062
O2 −0.544 0.498 −0.371 0.612 −0.317 0.289 −0.353
O −0.767⁎ 0.369 −0.652 0.714⁎ −0.279 0.592 −0.161
W1 −0.661 0.360 −0.485 0.589 −0.172 0.468 −0.189
W2 −0.885⁎⁎ 0.131 −0.866⁎⁎ 0.739⁎ −0.024 0.796⁎ 0.127
W3 −0.891⁎⁎ 0.167 −0.897⁎⁎ 0.662 −0.113 0.853⁎⁎ 0.122
W −0.918⁎⁎ 0.181 −0.887⁎⁎ 0.741⁎ −0.076 0.829⁎⁎ 0.094
⁎ Correlation is significant at the 0.05 level (2-tailed).
⁎⁎ Correlation is significant at the 0.01 level (2-tailed).
 X. Xu et al. / Science of the Total Environment 624 (2018) 1561–1576 1571

Table 4 guilds like insectivores, granivores and omnivores (Conole and

Redundancy analysis using the bird diversity and metrics of habitat loss and Kirkpatrick, 2011; Reis et al., 2012).
fragmentation.
Species richness and abundance are the two main indicators used for
Significant Explained Proportion of total p-value measuring biodiversity, and they have been widely used to be correlat-
variables variance (%) explained variance (%) ed with urban sprawl (MacGregor-Fors, 2008; Yuan et al., 2014; Batáry
Food type MESH_W 0.459(0.870) 52.8 0.002 et al., 2017). However, in our study, we found that species richness
PLAND_W 0.117(0.870) 13.4 alone appeared to be an insufficient indicator of the real status of bird
PD_W 0.107(0.870) 12.3
biodiversity, as it mainly represents the number of species while ignor-
MESH_S 0.132(0.870) 15.1
PD_S 0.055(0.870) 0.06 ing the process of bird species replacement, especially the original spe-
Habitat MESH_W 0.427(0.924) 46.3 0.002 cies by urban adaptors. The indicator of MSA of the original species
type PLAND_W 0.129(0.924) 14.0 could provide supplemental information by using the original species
PD_W 0.119(0.924) 12.9 as a reference condition. Therefore, the combination of different indica-
MESH_S 0.107(0.924) 11.6
MPS_F 0.071(0.924) 0.08
tors (species richness, relative abundance and MSA) could account for
MPS_W 0.070(0.924) 0.08 the overall trend of the change in biodiversity, the process of species re-
Seasonal MESH_W 0.547(0.818) 66.9 0.002 placement and the dynamics of biodiversity loss under different levels
status MPS_F 0.099(0.818) 12.1 of urbanization.
PLAND_S 0.084(0.818) 10.3
LPI_F 0.043(0.818) 0.05
PD_S 0.045(0.818) 0.06 4.2. How wetland degradation influences bird biodiversity in urban area?

Wetlands in metropolitan areas are typically the most biologically

diverse ecosystems, but they face the greatest degree of degradation,
generally feed on fishes, plants or benthonic animals (Palomino and such as through fragmentation and invasion by non-native species as
Carrascal, 2006; Matsuba et al., 2016), but in urbanized areas, bird com- well as destruction (Pauchard et al., 2006; Quesnelle et al., 2013; Yuan
munities are often dominated by urban-tolerant species, such as forest et al., 2014). This study detected the dependence of bird biodiversity

Food type Habitat type

Seasonal status

Fig. 10. Redundancy analysis biplots showing correlation between habitat loss and fragmentation and bird diversity.
1572 X. Xu et al. / Science of the Total Environment 624 (2018) 1561–1576

on habitat conditions, especially in wetlands, so wetland birds are the 100%, the species or habitat has a natural or low-impact status, but if the
dominant indicator species for biodiversity conservation, especially at indicator is 50%, the average abundance of the original species is 50% of
the beginning of human disturbance. Other studies have reached the the natural or low-impact state (UNSD, 2014). However, the question as
same conclusion and have noted the distinctiveness of original bird to what strategies should be adopted if the indicator is 80%, 50% or even
species as indicators because of their specific nesting and habitat re- lower must be further examined. Additionally, the MSA could be easily
quirements as well as the intensity of disturbance to wetlands by compared among different stages and even among different study sites,
human activities (Conole and Kirkpatrick, 2011). In this study, the wet- which might provide significant information to aid decision makers in
land birds, such as the benthic fauna eaters or fishing birds, rapidly choosing priority sites for restoration and conservation.
declined at the beginning of the urbanization process, and this phenom-
enon has been discussed by many other researchers, who suggest that 5. Conclusions
there are critical differences in bird composition between non-
urbanized and urbanized areas (Conole and Kirkpatrick, 2011; Reis Bird diversity is sensitive to environmental change, especially habi-
et al., 2012; Zhou et al., 2012). However, our study area, New Jiangwan tat change caused by urban growth. In this study, the influence of hab-
Town, was a relatively small and isolated area for wildlife, so broader itat loss and fragmentation caused by landscape change from being
scales should be considered in further studies, including the three undisturbed to human-dominated on bird diversity was examined.
main areas with high bird concentrations on the mouth of the Huangpu From our research, we concluded that 1) spatio-temporal variations in
River: Gongqing Forest Park, Paotaiwan Wetland Park and Binjiang habitat loss and fragmentation were triggered by urbanization; 2) the
Forest Park. number of species and relative abundance of various species groups var-
ied with different levels of urbanization; 3) when the landscape
changed from being undisturbed to human-dominated, the urban
4.3. Implications for bird conservation and habitat restoration avoiders moved out and the urban-adapted birds moved into the
study area; 4) the combination of species richness, relative abundance
Habitat loss and fragmentation have been widely discussed in land- and MSA accounted for the overall trend in the change in biodiversity,
scape and ecology studies, and the implementation of these research the process of species replacement (such as the change from the original
findings in landscape planning and habitat restoration would contribute species to the invasive species) and the dynamics of biodiversity loss at
to decreasing biodiversity loss and maintaining ecosystem function in in- different levels of urbanization; 5) the impacts of habitat loss and frag-
creasingly fragmented landscapes (Bregman et al., 2014). By considering mentation on bird species differed among bird communities, and the
different levels of urbanization, we attempted to propose different biodi- wetland percent cover and patch size were the most significant factors;
versity conservation strategies with this study. In the first two stages, 6) with the help of biodiversity accounting, a comprehensive overview
which were dominated by natural or semi-natural habitats, conserving of the biodiversity of the ecosystem in response to landscape change
the entire area was suggested, and ecological zoning tools could have could be presented, and the detailed species characteristics could be
been applied to prevent habitats from further loss and fragmentation useful for future habitat restoration under various degradation scenari-
(Löfvenhaft et al., 2004). In the third and fourth stages, temporal conser- os in an increasingly urban world.
vation could have been implemented (Kattwinkel et al., 2011) to protect
habitats, and temporal and transparent planning tools could have been Acknowledgments
used to increase political awareness to save as many biotopes as possible
(Bekessy et al., 2012). In the last stage, habitat reconstruction should This research was financially supported by the National Key Re-
have been adopted when the landscape was dominated by built-up area. search and Development Program of China (No. 2016YFC0502705),
Grassland in this study, artificial with intense maintenance, increas- the National Natural Science Foundation of China (No. 41501194), the
ing along the built-up areas, turned out to be negative correlation with Social Science Major Programme Foundation of China (14ZDB140),
bird species which implicated that the original and natural habitats with and the Excellent Scholars Support Plan on Science and Technology
self-maintenance, such as original wetlands, shrub lands could be alter- of Guizhou Province (KY[2014]252). We acknowledge the China Bird
natives to grasslands and lawns in the urbanization process from the Record Center, the first public and professional NGO for bird records,
perspectives of landscaping. The bird diversity respect to food type which provided a large amount of bird census data that were gathered.
and habitat requirement has statistic significant relations with the Thanks to the great efforts of volunteer citizen-scientists. Thanks to
PLAND of habitats which might also give us suggestions for bird conser- Yanxia Zhao for providing the open-source bird data. Finally, we are
vation from the landscape level. grateful to those who participated in the bird recording activities for
The MSA indicator might be used as a reference for determining the more than ten years and who largely contributed to this research
different levels of ecosystem degradation. For example, if the indicator is anonymously.

Appendix A. The landscape metrics measured in this paper were listed as below, which were calculated by the software Fragstats 4.2
(McGarigal K. 2015)

Abbr. Metric name Unit Application scale

PLAND Proportion of the land type % Class level

PD Patch density #/km2 Landscape/class level
LPI Largest patch index % Landscape/class level
ED Edge density m/ha Landscape level
LSI Landscape shape none Landscape level
SHDI Shannon's diversity index Landscape level
SHEI Shannon's evenness index Landscape level
MESH Effective mesh size ha Class level
MPS Mean patch size ha Class level
 X. Xu et al. / Science of the Total Environment 624 (2018) 1561–1576 1573

Appendix B. Bird species observed in and around Jiangwan Campus in 2002–2013, with respect to food type, seasonal status, and habitat type

ID Family Name Species Food Seasonal Habitat

#0014 Phasianidae Japanese quail Coturnix japonica OMN R O1

#0050 Phasianidae Common pheasant Phasianuscolchicus OMN R F3
#0084 Anatidae Mandarin duck Aix galericulata OMN M W1-F2
#0089 Anatidae Mallard Anasplatyrhynchos OMN M W1
#0090 Anatidae Spot-billed duck Anaspoecilorhyncha OMN M W1
#0096 Anatidae Philippine duck Anasluzonica OMN M W1
#0169 Coraciidae Dollarbird Eurystomusorientalis INS M/R F2
#0171 Alcedinidae Common kingfisher Alcedoatthis PIS R W3
#0180 Cerylidae Pied kingfisher Cerylerudis PIS R W3
#0193 Cuculidae Eurasian cuckoo Cuculuscanorus INS M F2-W3
#0195 Cuculidae Lesser cuckoo Cuculuspoliocephalus INS M F2
#0204 Centropodidae Lesser coucal Centropusbengalensis INS R F3-W3
#0221 Apodidae Fork-tailed swift Apuspacificus INS M O1
#0274 Columbidae Oriental turtle dove Streptopeliaorientalis FRU R O2
#0276 Columbidae Spotted dove Streptopeliachinensis FRU R O2
#0313 Rallidae White-breasted waterhen Amaurornisphoenicurus OMN R W3
#0323 Rallidae Common moorhen Gallinulachioropus OMN R W1-O1
#0324 Rallidae Common coot Fulicaatra OMN R W1
#0334 Scolopacidae Common snipe Gallinagogallinago BEN M W3
#0345 Scolopacidae Common greenshank Tringanebularia BEN M W2
#0348 Scolopacidae Green sandpiper Tringaochropus BEN M W2
#0349 Scolopacidae Wood sandpiper Tringaglareola BEN M W2
#0351 Scolopacidae Common sandpiper Actitishypoleucos BEN M W2
#0363 Scolopacidae Rufous-necked stint Calidrisruficollis BEN M W2
#0364 Scolopacidae Temminck's stint Calidristemminckii BEN M W2
#0365 Scolopacidae Long-toed stint Calidrissubminuta BEN M W2
#0380 Jacanidae Pheasant-tailed jacana Hydrophasianuschirurgus OMN M W3
#0388 Charadriidae Pacific golden plover Pluvialisfulva BEN M W2
#0392 Charadriidae Little ringed plover Charadriusdubius BEN M W2
#0393 Charadriidae Kentish plover Charadriusalexandrinus BEN M W2
#0492 Accipitridae Common buzzard Buteobuteo CAR M O1
#0508 Falconidae Common kestrel Falco tinnunculus CAR R O1
#0518 Podicipedidae Little grebe Tachybaptusruficollis PIS R W1
#0535 Ardeidae Little egret Egrettagarzetta PIS R W2-F2
#0539 Ardeidae Grey heron Ardeacinerea PIS R W2-F2
#0542 Ardeidae Large egret Casmerodiusalbus PIS R W2
#0543 Ardeidae Intermediate egret Mesophoyxintermedia PIS R W2
#0544 Ardeidae Cattle egret Bubulcus ibis PIS R W2
#0545 Ardeidae Chinese Pond-heron Ardeolabacchus PIS R W2
#0546 Ardeidae Striated heron Butoridesstriatus PIS R W2-W3
#0547 Ardeidae Black-crowned night-heron Nycticoraxnycticorax PIS R W2-F2
#0552 Ardeidae Yellow bittern Ixobrychussinensis PIS R W3
#0614 Laniidae Brown shrike Laniuscristatus CAR M F3
#0616 Laniidae Long-tailed shrike Laniusschach CAR R F3
#0630 Corvidae Azure-winged magpie Cyanopicacyana OMN R F2
#0636 Corvidae Black-billed magpie Pica pica OMN R F2
#0691 Muscicapidae White-throated rock thrush Monticolagularis OMN M F1
#0700 Muscicapidae Scaly thrush Zootheradauma OMN M F1
#0702 Muscicapidae Grey-backed thrush Turdushortulorum OMN M F1
#0707 Muscicapidae Eurasian blackbird Turdusmerula OMN R F2
#0712 Muscicapidae Eyebrowed thrush Turdusobscurus OMN M F1
#0714 Muscicapidae Pale thrush Turduspallidus OMN M F1
#0717 Muscicapidae Dusky thrush Turdusnaumanni OMN M O1
#0730 Muscicapidae Asian brown flycatcher Muscicapadauurica INS M F1
#0734 Muscicapidae Narcissus flycatcher Ficedulanarcissina INS M F1
#0735 Muscicapidae Mugimaki flycatcher Ficedulamugimaki INS M F2
#0745 Muscicapidae Blue-and-white flycatcher Cyanopticacyanomelana INS M F1
#0767 Muscicapidae Bluethroat Lusciniasvecica INS M F3
#0773 Muscicapidae Orange-flanked bush-robin Tarsigercyanurus INS M F1
#0787 Muscicapidae Daurian redstart Phoenicurusauroreus INS R F3
#0821 Sturnidae White-cheeked starling Sturnuscineraceus OMN M O2
#0823 Sturnidae Black-collared starling Sturnusnigricollis OMN R O2
#0829 Sturnidae Crested myna Acridotherescristatellus OMN R O1
#0850 Remizidae Chinese penduline tit Remizconsobrinus OMN M W3
#0862 Paridae Great tit Parus major OMN R F2
#0882 Hirundinidae Barn swallow Hirundorustica INS M O1
#0884 Hirundinidae Red-rumped swallow Hirundodaurica INS M O1
#0898 Pycnonotidae Light-vented bulbul Pycnonotussinensis FRU R F2
#0913 Cisticolidae Zitting cisticola Cisticolajuncidis FRU R O1
#0922 Cisticolidae Plain prinia Priniainornata FRU R W3
#0931 Sylviidae Manchurian bush warbler Cettiacanturians INS M F3
#0932 Sylviidae Japanese bush-warbler Cettiadiphone INS M F3

(continued on next page)

1574 X. Xu et al. / Science of the Total Environment 624 (2018) 1561–1576

(continued)

ID Family Name Species Food Seasonal Habitat

#0933 Sylviidae Brownish-flanked bush-war Cettiafortipes INS R F3

#0961 Sylviidae Oriental reed warbler Acrocephalusorientalis INS M W3
#0984 Sylviidae Yellow-rumped warbler Phylloscopusproregulus INS M F3
#0988 Sylviidae Yellow-browed warbler Phylloscopusinornatus INS M F1
#0993 Sylviidae Pale-legged warbler Phylloscopustenellipes INS M F1-F3
#0995 Sylviidae Eastern crowned warbler Phylloscopuscoronatus INS M F1
#1013 Sylviidae Masked laughingthrush Garrulaxperspicillatus INS R F3
#1017 Sylviidae Greater necklaced laughingthrush Garrulaxpectoralis INS R F1
#1036 Sylviidae Hwamei Garrulaxcanorus INS R F3
#1145 Paradoxornis Vinous-throated parrotbill Paradoxorniswebbianus INS R F3
#1172 Alaudidae Eurasian skylark Alaudaarvensis OMN M O1
#1174 Alaudidae Oriental skylark Alaudagulgula OMN M O1
#1198 Passeridae Eurasian tree sparrow Passer montanus OMN R F2
#1207 Passeridae White wagtail Motacilla alba OMN R O1-W2
#1211 Passeridae Yellow wagtail Motacillaflava OMN M O1-W2
#1212 Passeridae Grey wagtail Motacillacinerea OMN M O1-W2
#1218 Passeridae Orienfnl tree pipit Anthushodgsoni OMN M F2
#1239 Passeridae White-rumped munia Lonchurastriata OMN R F3
#1246 Fringillidae Grey-capped greenfinch Carduelissinica FRU R F2
#1293 Fringillidae Yellow-billed grosbeak Eophonamigratoria FRU M F2
#1307 Fringillidae Meadow bunting Emberizacioides FRU R F2
#1311 Fringillidae Tristram's bunting Emberizatristrami FRU M F3
#1314 Fringillidae Yellow-browed bunting Emberizachrysophrys FRU M F3
#1315 Fringillidae Rustic bunting Emberizarustica FRU M F3
#1316 Fringillidae Yellow-throated bunting Emberizaelegans FRU R F1
#1322 Fringillidae Black-faced bunting Emberizaspodocephala FRU M F3

Resident status: R = resident, M = migratory; Food Type: OMN = omnivores, INS = insectivores, FRU = frugivores, PIS = piscivores, BEN = benthivores, CAR = carnivores; Habitat type:
F1 = species that only use forested areas, F2 = forest species that also use open areas, F3 = species that use boscage areas, O1 = open area species, O2 = species that prefer open areas but
also use forested areas, W1 = swimming birds that use open water, W2 = waders, W3 = species that conceal themselves in marshes and aquatic areas with high grass.

Appendix C

Bird species moving into and out of the study area in different urbanization stages.
 X. Xu et al. / Science of the Total Environment 624 (2018) 1561–1576 1575

Appendix D

The change of bird species and the MSA in the 2nd stage (2004–2005), with the 1st stage (2002−2003) as the baseline.

Species ID (according to the bird list Appendix A) Richness MSA Using 2002–2003
species as baseline

Opening species 2002–2003 #0050#0089#0090#0096#0169#0171#0195#0204#0221#0274#0276#0313#0323#0334#0345 73 8.28 (100%)

Baseline #0348#0349#0351#0363#0364#0365#0380#388#0392#0393#0492#0508#0518#0535#0539
#0542#0543#0545#0546#0547#0552#0614#0616#0636#0700#0707#0714#0717#0773#0787
#0821#0823#0829#0850#0862#0882#0884#0898#0913#0922#0931#0932#0933#0961#1036
#1145#1172#1174#1198#1207#1211#1218#1239#1293#1314#1315#1316#1322
Species change Moving into (+) #0014#0193#0544#0730#0745#0995#1017#1212#1307 9
Moving out (−) #0050#0089#0090#0096#0169#0195#0221#0274#0345#0349#0363#0364#0365#0380#0388 35
#0393#0492#0508#0539#0542#0543#0614#0700#0773#0787#0821#0823#0850#0862#0932
#0961#1172#1174#1211#1315
Closing species 2004–2005 #0014#0171#0193#0204#0276#0313#0323#0334#0348#0351#0392#0518#0535#0544#0545 47 6.36 (76.8%)
#0546#0547#0552#0616#0636#0707#0714#0717#0730#0745#0829#0882#0884#0898#0913
#0922#0931#0933#0995#1017#1036#1145#1198#1207#1212#1218#1239#1293#1307#1314
#1316#1322

The change of bird species and the MSA in the 3rd stage (2006–2008).

Species ID (according to the bird list Appendix A) Richness MSA Using 2002–2003
species as baseline

Opening species 2004–2005 #0014#0171#0193#0204#0276#0313#0323#0334#0348#0351#0392#0518#0535#0544#0545 47 6.36 (76.8%)

#0546#0547#0552#0616#0636#0707#0714#0717#0730#0745#0829#0882#0884#0898#0913
#0922#0931#0933#0995#1017#1036#1145#1198#1207#1212#1218#1239#1293#1307#1314
#1316#1322
Species change Movinginto (+) #0274#0349#0492#0508#0539#0614#0691#0702#0712#0734#0735#0773#0787#0821#0862 19
#0988#0993#1246#1311
Moving out (−) #0014#0193#0204#0313#0348#0351#0552#0717#0884#0933#1017#1307#1314#1316 14
Closing species 2006–2008 #0171#0274#0276#0323#0334#0349#0392#0492#0508#0518#0535#0539#0544#0545#0546 52 4.32 (52.2%)
#0547#0614#0616#0636#0691#0702#0707#0712#0714#0730#0734#0735#0745#0773#0787
#0821#0829#0862#0882#0898#0913#0922#0931#0988#0993#0995#1036#1145#1198#1207
#1212#1218#1239#1246#1293#1311#1322

The change of bird species and the MSA in the 4th stage (2009–2013).

Species ID (according to the bird list Appendix A) Richness MSA Using 2002–2003
species as baseline

Opening species 2006–2008 #0171#0274#0276#0323#0334#0349#0392#0492#0508#0518#0535#0539#0544#0545#0546 52 4.32 (52.2%)

#0547#0614#0616#0636#0691#0702#0707#0712#0714#0730#0734#0735#0745#0773#0787
#0821#0829#0862#0882#0898#0913#0922#0931#0988#0993#0995#1036#1145#1198#1207
#1212#1218#1239#1246#1293#1311#1322
Species change Moving into (+) #0103#0163#0313#0324#0630#0700#0717#0823#0984#1314#1316 11
Moving out (−) #0171#0274#0334#0349#0392#0544#0545#0546#0614#0691#0712#0730#0734#0745#0882 23
#0913#0922#0931#0993#0995#1036#1212#1246
Closing species 2009–2013 #0103#0163#0276#0313#0323#0324#0492#0508#0518#0520#0535#0539#0547#0616#0630 40 2.00 (24.5%)
#0636#0700#0702#0707#0714#0717#0735#0773#0787#0821#0823#0829#0862#0898#0984
#0988#1145#1198#1207#1218#1293#1311#1314#1316#1322

References Blandón, A.C., Perelman, S.B., Ramírez, M., López, A., Javier, O., Robbins, C.S., 2016. Tempo-
ral bird community dynamics are strongly affected by landscape fragmentation in a
Aronson, M.F.J., La Sorte, Frank A., Nilon, Charles H., Katti, Madhusudan, Goddard, Central American tropical forest region. Biodivers. Conserv. 25, 311–330.
Mark A., Lepczyk, Christopher A., Warren, Paige S., Nicholas, S.G. Williams, Braak, C.T., Smilauer, P., 2002. CANOCO reference manual and CanoDraw for Windows
Cilliers, Sarel, Clarkson, Bruce, Dobbs, Cynnamon, Dolan, Rebecca, Hedblom, user's guide: software for canonical community ordination (version 4.5). Microcom-
Marcus, Klotz, Stefan, Kooijmans, Jip Louwe, Kühn, Ingolf, MacGregor-Fors, Ian, puter Power Ithaca, NY.
McDonnell, Mark, Mörtberg, Ulla, Pyšek, Petr, Siebert, Stefan, Sushinsky, Bregman, T.P., Sekercioglu, C.H., Tobias, J.A., 2014. Global patterns and predictors of bird
Jessica, Werner, Peter, Winter, M., 2014. A global analysis of the impacts of ur- species responses to forest fragmentation: implications for ecosystem function and
banization on bird and plant diversity reveals key anthropogenic drivers. Proc. conservation. Biol. Conserv. 169, 372–383.
R. Soc. Lond. B Biol. Sci. 281. Cai, Y.T., Tang, S.M., Yuan, X., Wang, J.J., Ma, Z.J., 2011. Checklist and change of birds
Batáry, P., K., K., S.-R., M., C., D.E., 2017. Non-linearities in bird responses across urbaniza- in Shanghai. Journal of Fudan University(Natural Science) 50, 334–343 [In
tion gradients: a meta-analysis. Glosb. Chang. Biol. 1–9. Chinese].
Bekessy, S.A., White, M., Gordon, A., Moilanen, A., Mccarthy, M.A., Wintle, B.A., 2012. Carrete, M., Tella, J.L., Blanco, G., Bertellotti, M., 2009. Effects of habitat degradation on the
Transparent planning for biodiversity and development in the urban fringe. Landsc. abundance, richness and diversity of raptors across Neotropical biomes. Biol. Conserv.
Urban Plan. 108, 140–149. 142, 2002–2011.
1576 X. Xu et al. / Science of the Total Environment 624 (2018) 1561–1576

Chace, J.F., Walsh, J.J., 2006. Urban effects on native avifauna: a review. Landsc. Urban Ortega-Álvarez, R., MacGregor-Fors, I., 2011. Dusting-off the file: a review of knowledge
Plan. 74, 46–69. on urban ornithology in Latin America. Landsc. Urban Plan. 101, 1–10.
Ciach, M., Fröhlich, A., 2016. Habitat type, food resources, noise and light pollution explain Palomino, D., Carrascal, L.M., 2006. Urban influence on birds at a regional scale: a case
the species composition, abundance and stability of a winter bird assemblage in an study with the avifauna of northern Madrid province. Landsc. Urban Plan. 77,
urban environment. Urban Ecosystems 20 (3), 547–559. 276–290.
Cockell, C., Koeberl, C., Gilmour, I. (Eds.), 2006. Biological Processes Associated With Im- Pauchard, A., Aguayo, M., Peña, E., Urrutia, R., 2006. Multiple effects of urbanization on the
pact Events. Springer, Berlin. biodiversity of developing countries: the case of a fast-growing metropolitan area
Conole, L.E., Kirkpatrick, J.B., 2011. Functional and spatial differentiation of urban bird as- (Concepción, Chile). Biol. Conserv. 127, 272–281.
semblages at the landscape scale. Landsc. Urban Plan. 100, 11–23. Pettit, C.J., Raymond, C.M., Bryan, B.A., Lewis, H., 2011. Identifying strengths and weak-
Ditchkoff, S.S., Saalfeld, S.T., Gibson, C.J., 2006. Animal behavior in urban ecosystems: nesses of landscape visualisation for effective communication of future alternatives.
modifications due to human-induced stress. Urban Ecosys. 9, 5–12. Landsc. Urban Plan. 100, 231–241.
Edens, B., Hein, L., 2013. Towards a consistent approach for ecosystem accounting. Ecol. Pfisterer, A.B., Schmid, B., 2002. Diversity-dependent production can decrease the stability
Econ. 90, 41–52. of ecosystem functioning. Nature 416, 84–86.
Gaston, K.J., 2000. Global patterns in biodiversity. Nature 405, 220–227. Purvis, A., Hector, A., 2000. Getting the measure of biodiversity. Nature 405, 212–219.
González-Oreja, J.A., 2011. Birds of different biogeographic origins respond in contrasting Quesnelle, P.E., Fahrig, L., Lindsay, K.E., 2013. Effects of habitat loss, habitat configuration
ways to urbanization. Biol. Conserv. 144, 234–242. and matrix composition on declining wetland species. Biol. Conserv. 160, 200–208.
Gregory, R.D., Van Strien, A., Vorisek, P., Meyling, A.W.G., Noble, D.G., Foppen, R.P.B., Reis, E., López-Iborra, G.M., Pinheiro, R.T., 2012. Changes in bird species richness through
Gibbons, D.W., 2005. Developing indicators for European birds. Philosophical Trans- different levels of urbanization: implications for biodiversity conservation and garden
actions of the Royal Society of London B: Biological Sciences. 360, pp. 269–288. design in Central Brazil. Landsc. Urban Plan. 107, 31–42.
Gulezian, P.Z., Nyberg, D.W., 2010. Distribution of invasive plants in a spatially structured Remme, R.P., Hein, L., van Swaay, C.A.M., 2016. Exploring spatial indicators for biodiversi-
urban landscape. Landsc. Urban Plan. 95, 161–168. ty accounting. Ecol. Indic. 70, 232–248.
Haines-Young, R., Páramo, F., 2006. Land Accounts for Europe 1990–2000: Towards Inte- Robinson, G.R., Holt, R.D., Gaines, M.S., Hamburg, S.P., Johnson, M.L., Fitch, H.S., Martinko,
grated Land and Ecosystem Accounting. Office for Official Publ. of the European E.A., 1992. Diverse and contrasting effects of habitat fragmentation. Science 524–526.
Communities. Sandström, U.G., Angelstam, P., Mikusiński, G., 2006. Ecological diversity of birds in rela-
Hein, L., Obst, C., Edens, B., Remme, R.P., 2015. Progress and challenges in the develop- tion to the structure of urban green space. Landsc. Urban Plan. 77, 39–53.
ment of ecosystem accounting as a tool to analyse ecosystem capital. Curr. Opin. En- Schluter, D., Pennell, M.W., 2017. Speciation gradients and the distribution of biodiversity.
viron. Sustain. 14, 86–92. Nature 546, 48–55.
Jin, X.B., Zhou, B.C., Qin, X.K., Cui, Z.X., Xia, J.H., Si, Q., Liu, M.P., 2004. The biodiversity of the Schröter, M., Barton, D.N., Remme, R.P., Hein, L., 2014. Accounting for capacity and flow of
derelict Jiangwan airport in Shanghai. The 6th Conference Of National Biodiversity ecosystem services: a conceptual model and a case study for Telemark, Norway. Ecol.
Conservation and Sustainable Development Lijiang, Yunnan, China. Indic. 36, 539–551.
Jokimäki, Jukka, Kaisanlahti-Jokimäki, Marja-Liisa, Suhonen, Jukka, Clergeau, Philippe, Schröter, M., Remme, R.P., Sumarga, E., Barton, D.N., Hein, L., 2015. Lessons learned for
Pautasso, Marco, Fernández-Juricic, Esteban, 2011. Merging wildlife community ecol- spatial modelling of ecosystem services in support of ecosystem accounting. Ecosyst.
ogy with animal behavioral ecology for a better urban landscape planning. Landsc. Serv. 13, 64–69.
Urban Plan. 100 (4), 383–385. Scolozzi, R., Geneletti, D., 2012. A multi-scale qualitative approach to assess the impact of
Kattwinkel, M., Biedermann, R., Kleyer, M., 2011. Temporary conservation for urban bio- urbanization on natural habitats and their connectivity. Environ. Impact Assess. Rev.
diversity. Biol. Conserv. 144, 2335–2343. 36, 9–22.
Keinath, D.A., Doak, D.F., Hodges, K.E., Prugh, L.R., Fagan, W., Sekercioglu, C.H., Buchart, Shen, Z.Y., Hou, X.S., Li, W.K., Aini, G., Chen, L., Gong, Y.W., 2015. Impact of landscape pat-
S.H.M., Kauffman, M., 2017. A global analysis of traits predicting species sensitivity tern at multiple spatial scales on water quality: a case study in a typical urbanised
to habitat fragmentation. Glob. Ecol. Biogeogr. 26, 115–127. watershed in China. Ecol. Indic. 48, 417–427.
Löfvenhaft, K., Runborg, S., Sjögren-Gulve, P., 2004. Biotope patterns and amphibian dis- Sundell-Turner, N.M., Rodewald, A.D., 2008. A comparison of landscape metrics for con-
tribution as assessment tools in urban landscape planning. Landsc. Urban Plan. 68, servation planning. Landsc. Urban Plan. 86, 219–225.
403–427. Tomasevic, J.A., Marzluff, J.M., 2017. Cavity nesting birds along an urban-wildland gradi-
Luan, X., 2003. Studies on Avian Community of Shanghai and Planning of Conservation. ent: is human facilitation structuring the bird community? Urban Ecosys. 20,
[Thesis]. East China Normal University Shanghai, China. 435–448.
MacGregor-Fors, I., 2008. Relation between habitat attributes and bird richness in a west- UNSD, 2014. System of Environmental-Economic Accounting 2012—Experimental Eco-
ern Mexico suburb. Landsc. Urban Plan. 84, 92–98. system Accounting (In: New York).
Mackinnon, J., et al., 2000. A field guide to the birds of China. Hunan Educational Press, Vandewalle, M., Bello, F., Berg, M.P., Bolger, T., Dolédec, S., Dubs, F., Feld, C.K., Harrington,
Hunan, China. R., Harrison, P.A., Lavorel, S., Silva, P.M., Moretti, M., Niemela, J., Santos, P., Sattler, T.,
Marzluff, J., Bowman, R., Donnelly, R. (Eds.), 2012. Avian Ecology and Conservation in an Sousa, J.P., Sykes, M., Vanbergen, A.J., Woodcock, B.A., 2010. Functional traits as indi-
Urbanizing World. Springer Science & Business Media. cators of biodiversity response to land use changes across ecosystems and organisms.
Matsuba, M., Nishijima, S., Katoh, K., 2016. Effectiveness of corridor vegetation depends Biodivers. Conserv. 19, 2921–2947.
on urbanization tolerance of forest birds in central Tokyo, Japan. Urban For. Urban Whitworth, A., Villacampa, J., Serrano Rojas, S.J., Downie, R., MacLeod, R., 2017. Methods
Green. 18, 173–181. matter: different biodiversity survey methodologies identify contrasting biodiversity
McGarigal, K., Cushman, S.A., Neel, M.C., Ene, E., 2002. FRAGSTATS: Spatial Pattern Analy- patterns in a human modified rainforest — a case study with amphibians. Ecol. Indic.
sis Program for Categorical Maps. University of Massachusetts, Amherst. In. 72, 821–832.
McGranahan, G., Satterthwaite, D., 2003. Urban centers: an assessment of sustainability. Wood, J.M., Quinn, J.E., 2016. Local and landscape metrics identify opportunities for con-
Annu. Rev. Environ. Resour. 28, 243–274. serving cavity-nesting birds in a rapidly urbanizing ecoregion. J. Urban Econ. 1, 10.
McKinney, M.L., 2008. Effects of urbanization on species richness: a review of plants and Yeh, C.-T., Huang, S.-L., 2009. Investigating spatiotemporal patterns of landscape diversity
animals. Urban Ecosys. 11, 161–176. in response to urbanization. Landsc. Urban Plan. 93, 151–162.
McLellan, R., Iyengar, L., Jeffries, B., Oerlemans, N., 2014. Living Planet Report 2014: Spe- Yuan, Y.J., Zeng, G.M., Liang, J., Li, X.D., Li, Z.W., Zhang, C., Huang, L., Lai, X., Lu, L.H., Wu,
cies and Spaces, People and Places. World Wide Fund for Nature. H.P., 2014. Effects of landscape structure, habitat and human disturbance on birds:
Melles, S., Glenn, S., Martin, K., 2003. Urban bird diversity and landscape complexity: a case study in East Dongting Lake wetland. Ecol. Eng. 67, 67–75.
species–environment associations along a multiscale habitat gradient. Conserv. Zhou, D.Q., Fung, T., Chu, L.M., 2012. Avian community structure of urban parks in devel-
Ecol. 7. oped and new growth areas: a landscape-scale study in Southeast Asia. Landsc. Urban
Millennium Ecosystem Assessment, 2005a. Current State and Trends, Ecosystems and Plan. 108, 91–102.
Human Well-being. Island Press, Washington, DC.
Millennium Ecosystem Assessment, 2005b. Ecosystems and Human Well-Being: Biodi-
versity Synthesis. World Resources Institute, Washington, DC.