282
2006). In addition, savannas in South American and Central Asia and temperate
grasslands in the Western United States will also be heavily impacted. Thus, free,
open-access, global mapping satellite RS data sets are especially important for
grassland IAS detection and monitoring. These sensors lack the fine spatial resolution and full spectrum afforded by airborne imaging spectroscopy, which may be
necessary to separate native and non-native species of the same functional type.
However, multispectral imagery is often used for viewing widespread and abundant
invasives, which is key for monitoring overall ecosystem invasion onset and die-off,
but offers little help in terms of real-time or early IAS detection. In both cases, the
minimum percent cover required for mapping can vary across similar ecosystems
(Bradley 2014) and depends on sensor resolution and on how distinguishable the
invader is from the background. Even when a non-native grass is spectrally distinguishable, an acceptable detection rate is not always possible when patch sizes are
small relative to pixel resolution (Mladinich et al. 2006). Therefore, to ensure successful mapping, IAS targets must differ from the native community spectrally, phenologically, texture/morphologically, or architecturally (Bradley 2014). Analysis
considerations must include a careful evaluation of the relationship between vegetation characteristics and sensor resolutions, particularly in the spatial, spectral, and
temporal domains.
12.2.3 Aquatic Ecosystems
Although they cover a small portion of the Earth’s surface, aquatic ecosystems are
disproportionately important to global diversity. They are among the most diverse
and productive ecosystems on Earth and provide vital ecosystem services (Tabacchi
et al. 1998; Barbier et al. 2011). Aquatic ecosystems encompass multiple gradients,
such as water intermittency, microtopography, and salinity, leading to complex
environmental heterogeneity (Junk et al. 1989; Mitsch and Gosselink 2007). This
mosaic of diverse environmental conditions supports high biodiversity through multiple niches (Tockner et al. 2000; Ward et al. 2002).
Biodiversity losses in coastal and freshwater aquatic ecosystems are among the
highest in the world (Dudgeon et al. 2015; Waycott et al. 2009; Vörösmarty et al.
2010). At least 30%–50% of the world’s wetlands have been lost (Finlayson 2012;
Hu et al. 2017), and up to 35% of the extent of critical habitats like seagrasses and
mangroves have been destroyed just in the twentieth century (UNESCO 2018).
These ecosystems are among the most vulnerable to invasion because they are
highly connected, are used extensively by humans, and often are geographically
close to invasion foci such as ports or urban areas (Gherardi 2007; Williams and
Grosholz 2008).
Plants in aquatic ecosystems can be broadly classified into five functional types
or sets of species that occupy distinct spatial niches along the gradient from water
to land and often have similar characteristics. The five functional types considered
E. A. Bolch et al.
2006). In addition, savannas in South American and Central Asia and temperate
grasslands in the Western United States will also be heavily impacted. Thus, free,
open-access, global mapping satellite RS data sets are especially important for
grassland IAS detection and monitoring. These sensors lack the fine spatial resolution and full spectrum afforded by airborne imaging spectroscopy, which may be
necessary to separate native and non-native species of the same functional type.
However, multispectral imagery is often used for viewing widespread and abundant
invasives, which is key for monitoring overall ecosystem invasion onset and die-off,
but offers little help in terms of real-time or early IAS detection. In both cases, the
minimum percent cover required for mapping can vary across similar ecosystems
(Bradley 2014) and depends on sensor resolution and on how distinguishable the
invader is from the background. Even when a non-native grass is spectrally distinguishable, an acceptable detection rate is not always possible when patch sizes are
small relative to pixel resolution (Mladinich et al. 2006). Therefore, to ensure successful mapping, IAS targets must differ from the native community spectrally, phenologically, texture/morphologically, or architecturally (Bradley 2014). Analysis
considerations must include a careful evaluation of the relationship between vegetation characteristics and sensor resolutions, particularly in the spatial, spectral, and
temporal domains.
12.2.3 Aquatic Ecosystems
Although they cover a small portion of the Earth’s surface, aquatic ecosystems are
disproportionately important to global diversity. They are among the most diverse
and productive ecosystems on Earth and provide vital ecosystem services (Tabacchi
et al. 1998; Barbier et al. 2011). Aquatic ecosystems encompass multiple gradients,
such as water intermittency, microtopography, and salinity, leading to complex
environmental heterogeneity (Junk et al. 1989; Mitsch and Gosselink 2007). This
mosaic of diverse environmental conditions supports high biodiversity through multiple niches (Tockner et al. 2000; Ward et al. 2002).
Biodiversity losses in coastal and freshwater aquatic ecosystems are among the
highest in the world (Dudgeon et al. 2015; Waycott et al. 2009; Vörösmarty et al.
2010). At least 30%–50% of the world’s wetlands have been lost (Finlayson 2012;
Hu et al. 2017), and up to 35% of the extent of critical habitats like seagrasses and
mangroves have been destroyed just in the twentieth century (UNESCO 2018).
These ecosystems are among the most vulnerable to invasion because they are
highly connected, are used extensively by humans, and often are geographically
close to invasion foci such as ports or urban areas (Gherardi 2007; Williams and
Grosholz 2008).
Plants in aquatic ecosystems can be broadly classified into five functional types
or sets of species that occupy distinct spatial niches along the gradient from water
to land and often have similar characteristics. The five functional types considered
E. A. Bolch et al.