Characterizing biodiversity in urban areas using remote sensing - Urban Remote Sensing: Monitoring, Synthesis and Modeling in the Urban Environment - page 274

20.1 Introduction

Previously, many studies describing biodiversity in urban
areas used the urban-rural gradient approach to study the ecol-
ogy of cities and towns around the world (McDonnell et al .,
1997). Although this approach has increased our knowledge of
ecosystem processes and the distribution of species in relation to
cities and urbanization, the use of remote sensor data in combi-
nation with previous gradient approaches has great potential for
broadening our understanding of urban patterns, moving from
a simplistic to a more complex, quantified and realistic view of
urban-rural gradients and allowing us to develop a comprehen-
sive terminology with respect to urban ecology (e.g., McDonnell
and Hahs, 2008; Pickett et al ., 2008). Moreover, combining field
studies and remote sensing within a broad hierarchal perspec-
tive, including landscape analyses, such methods could be used
to monitor long-term changes in urban landscapes. In the long
run this approach could provide decision makers and landscape
planners with important information about which urban green
areas to prioritize. Finally, in this emerging research field of urban
ecosystems and the relationships between urbanization and bio-
diversity, remote sensor data will make a crucial contribution
and has yet to meet its full potential.
The following sections discuss capturing biodiversity infor-
mation in urban areas and areas subject to urbanization using
remote sensor data sources; we consider both direct and indirect
approaches, the latter using spatial ecological modeling com-
bined with field surveys. We also discuss their applications for
monitoring, planning and management.

Urbanization, together with agriculture, is the most impor-
tant threat to biodiversity worldwide (Ricketts and Imhoff,
2003). Urbanization will probably soon exceed agriculture as
the dominant agent of loss, fragmentation and degradation of
habitats as a result of the increasingly urbanized human pop-
ulation (Marzluff and Ewing, 2001). Urbanization will have a
major effect on biodiversity, defined as the variability within
species, between species and of ecosystems, since it is predicted
that urban areas will increase by more than 600 000 km 2 in
developing countries and 500 000 km 2 in industrialized cities
between 2010 and 2030 (Angel, Sheppard and Civco, 2005).
People traditionally settle in areas with highly productive ecosys-
tems and abundant natural resources, therefore many cities
are located in areas important for biodiversity conservation, so
called ecological hotspots (Cincotta, Wisnewski and Engelman,
2000; Ricketts and Imhoff, 2003; Luck, 2007). Moreover, since
the urban footprint extends far beyond municipal boundaries,
urbanization may also reduce the diversity of native species at
regional and global scales (Grimm et al ., 2008). Accordingly,
cities can, if managed properly, play an increasingly impor-
tant role in sustaining the world's species (Rosenzweig, 2003;
Dearborn and Kark, 2010). Interestingly, Grimm et al . (2008)
considered cities to be microcosms of the kinds of changes that
are happening globally, making them informative test cases for
understanding socioecological system dynamics and responses
to change.
Urban green areas, which act as habitats, can thus be of direct
importance to the flora and fauna within cities but can also
be beneficial to humankind. Natural remnant habitats or green
areas in cities provide humans with a multitude of resources
and processes that are supplied by ecosystems. Examples of such
ecosystem services are rainwater drainage and noise reduction
(Bolund and Hunhammar, 1999). Urban green areas have also
been highlighted as being important for humanwell being (Grahn
and Stigsdotter, 2010). Moreover, Fuller et al . (2007) showed an
increase in psychological benefits with biodiversity in urban
green areas.
In the scientific world, urban fauna and flora, and their diver-
sity, have long been a neglected research area, while other more
natural environments have been prioritized. However, during
the last decade research into biodiversity in urban areas has
increased rapidly, as seen in an increasing number of publica-
tions (McKinney, 2006; Chace and Walsh, 2006; Marzluff, 2008;
McDonnell, Brueste and Hahs, 2009; Davies et al ., 2009). Despite
there being more studies on biodiversity in urban areas, remark-
ably few of them use remote sensing techniques such as satellite
images (Gottschalk, Huettmann and Ehlers, 2005; Bino et al .,
2008). This may be due to the fact that high spatial resolution
satellite images have only been widely available since 1999 with
the launch of IKONOS and NASA's Geocover dataset (Tucker,
Grant and Dykstra, 2004). Prior to this there were very few freely
available Landsat images globally (Bino et al ., 2008). This situa-
tion is now changing rapidly. New data capture techniques using
passive and active sensors, increased data availability and novel
image interpretation methods, together with ongoing develop-
ment of spatial ecological modeling approaches, mean that there
is a greatly increased potential for identification, mapping and
assessment of biodiversity components (e.g., Turner et al ., 2003;
Guisan and Thuiller, 2005; Gillespie et al ., 2008).

20.2 Remote sensing
methods in urban
biodiversity studies

The complexity of the biodiversity concept, as encapsulated in its
definition, implies that deriving information relating to the wide
array of biodiversity components, from genes to ecosystems, is
complicated. Even if field surveys are necessary and give the
most accurate information on vegetation, species, individuals
and genes, the combined picture of many existing field surveys
reveals a range of problems for comparative meta-analyses of, for
example, the effects of urbanization on biodiversity. The different
methods used may not be compatible with each other, there may
also be extremely uneven coverage with a strong bias towards
developed countries and existing knowledge centers, and, even in
relatively well-studied regions, data is often spatially biased (e.g.,
Pautasso andMcKinney, 2007). Furthermore, even if urban areas
are largely accessible for conducting field surveys of biodiversity
components, the increasing size of urban areas contributes to
making the annual costs of detailed field studies very high,
particularly since field surveys require skilled individuals and are
very time consuming.
In the light of this, remote sensor data is especially attrac-
tive as a cost-effective source of information on biodiversity,
since it offers a relatively inexpensive mean of providing com-
plete spatial coverage of environmental information over large
areas. Furthermore, remote sensing captures information in a
systematic manner, can be used to examine inaccessible areas,
and can be updated regularly (e.g., Luoto et al ., 2004; Duro
et al ., 2007; Levin et al ., 2007; Saatchi et al ., 2008). However,