285
in building species occupancy models based upon species ecological knowledge to
predict future invasions or direction of spread (Andrew and Ustin 2009; Rocchini
et al. 2015).
12.2.3.2 Emergent
Within the emergent functional type, the canopy is relatively uniform, composed
generally of only grasses, sedges, and reeds. These species are often mixed, and
patch sizes remain small even among species tending to grow as monocultures. The
canopy structure is typically erectrophilic, and spectral mixing with water is common, even with fine spatial scale imagery. In addition to spectral information and
temporal information, the texture of invasive and native species patches can be leveraged in mapping IAS and may be used to improve success. Samiappan et al.
(2017) used four methods to calculate texture indices as inputs into a SVM algorithm to map common reed (Phragmites australis). They took advantage of the 5 m
spatial resolution afforded by airborne (in this case UAS) imagery, though they
cautioned such an approach is unlikely to work if patches of IAS are smaller than a
few pixels or more mixed. However, texture has been shown to be advantageous
even with moderate spatial resolution imagery. For example, Arzandeh and Wang
(2003) successfully differentiated common reed and cattail (Typha angustifolia)
using Landsat TM by adding texture indices to increase pixel spectral information
content. For these reasons, hyperspectral aerial surveys have offered the best data
source for classifications for emergent communities. Using sensors such as CASI,
AVIRIS, and HyMap, many studies have mapped the emergent community, differentiating species within submerged and floating functional types (Hestir et al. 2008;
Jollineau and Howarth 2008; Hunter et al. 2010; Khanna et al. 2011; Hestir et al.
2012; Zhao et al. 2012). Occasionally, both spectrally rich and temporally strategic
data have been used together to map IAS (Laba et al. 2005; Hamada et al. 2007; Pu
et al. 2008).
12.2.3.3 Floating Macrophytes
Floating macrophytes have a simple canopy structure with vegetation growing close
to the water surface. They can spread over large areas and often grow as monocultures, so mapping them using RS has been relatively easy, except when two or more
floating species co-occur in a single ecosystem (Khanna et al. 2011; Cavalli et al.
2009). Floating macrophyte mats often appear very similar spectrally, for example,
water hyacinth, water primrose, and pennywort (Centella asiatica) (Khanna et al.
2011). Cavalli et al. (2009) separated three floating species with Landsat ETM+
data using spectral linear mixture modeling trained by high-quality spectral libraries
developed from field spectroscopy. However, without detailed spectral libraries
for a location, hyperspectral data are needed to differentiate between similar,
bright green uniform mats of floating species (Yang 2007; Khanna et al. 2011).
12 Remote Detection of Invasive Alien Species
in building species occupancy models based upon species ecological knowledge to
predict future invasions or direction of spread (Andrew and Ustin 2009; Rocchini
et al. 2015).
12.2.3.2 Emergent
Within the emergent functional type, the canopy is relatively uniform, composed
generally of only grasses, sedges, and reeds. These species are often mixed, and
patch sizes remain small even among species tending to grow as monocultures. The
canopy structure is typically erectrophilic, and spectral mixing with water is common, even with fine spatial scale imagery. In addition to spectral information and
temporal information, the texture of invasive and native species patches can be leveraged in mapping IAS and may be used to improve success. Samiappan et al.
(2017) used four methods to calculate texture indices as inputs into a SVM algorithm to map common reed (Phragmites australis). They took advantage of the 5 m
spatial resolution afforded by airborne (in this case UAS) imagery, though they
cautioned such an approach is unlikely to work if patches of IAS are smaller than a
few pixels or more mixed. However, texture has been shown to be advantageous
even with moderate spatial resolution imagery. For example, Arzandeh and Wang
(2003) successfully differentiated common reed and cattail (Typha angustifolia)
using Landsat TM by adding texture indices to increase pixel spectral information
content. For these reasons, hyperspectral aerial surveys have offered the best data
source for classifications for emergent communities. Using sensors such as CASI,
AVIRIS, and HyMap, many studies have mapped the emergent community, differentiating species within submerged and floating functional types (Hestir et al. 2008;
Jollineau and Howarth 2008; Hunter et al. 2010; Khanna et al. 2011; Hestir et al.
2012; Zhao et al. 2012). Occasionally, both spectrally rich and temporally strategic
data have been used together to map IAS (Laba et al. 2005; Hamada et al. 2007; Pu
et al. 2008).
12.2.3.3 Floating Macrophytes
Floating macrophytes have a simple canopy structure with vegetation growing close
to the water surface. They can spread over large areas and often grow as monocultures, so mapping them using RS has been relatively easy, except when two or more
floating species co-occur in a single ecosystem (Khanna et al. 2011; Cavalli et al.
2009). Floating macrophyte mats often appear very similar spectrally, for example,
water hyacinth, water primrose, and pennywort (Centella asiatica) (Khanna et al.
2011). Cavalli et al. (2009) separated three floating species with Landsat ETM+
data using spectral linear mixture modeling trained by high-quality spectral libraries
developed from field spectroscopy. However, without detailed spectral libraries
for a location, hyperspectral data are needed to differentiate between similar,
bright green uniform mats of floating species (Yang 2007; Khanna et al. 2011).
12 Remote Detection of Invasive Alien Species