294
control strategies. The acceleration in global change and biodiversity loss degrades
ecosystem resilience, threatening valuable ecosystem services. To preserve these
services will require global cooperation on IAS monitoring and control with RS is
a critical tool.
Each biome discussed in this chapter contains a unique complement of species.
As a result, a different method of RS and data fusion works best for each. However,
some methodologies can be valuable in all circumstances, such as increasing spectral information content. If only one IAS is of interest and it differs from its surroundings, multispectral data or use of photographs and texture analysis may be
enough to identify and map it. However, in most cases there are multiple IAS competing with one another and with native vegetation, with varying canopy complexity
and functional types. In such scenarios, difference in phenological characteristics
can be exploited for identification. For example, an IAS might be identified through
differing flowering times, flower colors, or earlier or later periods of senescence
relative to surrounding vegetation. This requires temporally dense data. In cases
where the invasion scenario is not simple or the data are not temporally sufficient,
fusion between RS and other data sources (e.g., habitat models, DEMs, climate
models) can be used to improve accuracy.
Data collection in the three domains of RS (spectral, spatial, and temporal) can
be optimized for a species based on the ecosystem type and image analysis approach.
For forests, lidar data are often a good addition to spectral information because they
can provide information on height and physical crown structure. For species below
the forest canopy, indirect methods such as models based on ecological knowledge
of the species may be necessary, or imagery may simply be collected during a leafoff period. IAS in grasslands often have similar spectral properties to natives,
requiring hyperspectral data, strategic image timing, or indirect modeling methods.
Aquatic ecosystems introduce many confounding factors due to presence of water
and its associated processes, necessitating high radiometric quality and good calibration. Because this biome is so complex, hyperspectral information and customized image timing are a must for differentiating IAS. Additionally, radiative transfer
modeling is often necessary to detect submerged and water column
IAS. Agroecosystems have minimal diversity, so fewer spectral data are required.
However, frequent assessment is necessary to allow a timely response to minimize
crop loss. RS detection of IAS in urban ecosystems requires varying methods and
unique adaptations because of the high potential for introductions and unusual landscape features, such as impervious surfaces.
These factors underscore the importance of mission design for two key data collection platforms. First, airborne platforms (piloted and unpiloted), which are vital
to rapid, local-scale assessments, must acquire data at key times relevant to IAS
phenology. As temperatures and biodiversity losses continue to increase, plant phenology is expected to continue to change (Primack et al. 2015; Wolf et al. 2017) and
airborne acquisition strategies must adjust accordingly. Second, satellite platforms
are critical to providing global-scale systematic monitoring of IAS. Current and
future missions must include high spectral resolution sensors with the capability to
create climate-relevant time series (a duration on the order of approximately a
E. A. Bolch et al.
control strategies. The acceleration in global change and biodiversity loss degrades
ecosystem resilience, threatening valuable ecosystem services. To preserve these
services will require global cooperation on IAS monitoring and control with RS is
a critical tool.
Each biome discussed in this chapter contains a unique complement of species.
As a result, a different method of RS and data fusion works best for each. However,
some methodologies can be valuable in all circumstances, such as increasing spectral information content. If only one IAS is of interest and it differs from its surroundings, multispectral data or use of photographs and texture analysis may be
enough to identify and map it. However, in most cases there are multiple IAS competing with one another and with native vegetation, with varying canopy complexity
and functional types. In such scenarios, difference in phenological characteristics
can be exploited for identification. For example, an IAS might be identified through
differing flowering times, flower colors, or earlier or later periods of senescence
relative to surrounding vegetation. This requires temporally dense data. In cases
where the invasion scenario is not simple or the data are not temporally sufficient,
fusion between RS and other data sources (e.g., habitat models, DEMs, climate
models) can be used to improve accuracy.
Data collection in the three domains of RS (spectral, spatial, and temporal) can
be optimized for a species based on the ecosystem type and image analysis approach.
For forests, lidar data are often a good addition to spectral information because they
can provide information on height and physical crown structure. For species below
the forest canopy, indirect methods such as models based on ecological knowledge
of the species may be necessary, or imagery may simply be collected during a leafoff period. IAS in grasslands often have similar spectral properties to natives,
requiring hyperspectral data, strategic image timing, or indirect modeling methods.
Aquatic ecosystems introduce many confounding factors due to presence of water
and its associated processes, necessitating high radiometric quality and good calibration. Because this biome is so complex, hyperspectral information and customized image timing are a must for differentiating IAS. Additionally, radiative transfer
modeling is often necessary to detect submerged and water column
IAS. Agroecosystems have minimal diversity, so fewer spectral data are required.
However, frequent assessment is necessary to allow a timely response to minimize
crop loss. RS detection of IAS in urban ecosystems requires varying methods and
unique adaptations because of the high potential for introductions and unusual landscape features, such as impervious surfaces.
These factors underscore the importance of mission design for two key data collection platforms. First, airborne platforms (piloted and unpiloted), which are vital
to rapid, local-scale assessments, must acquire data at key times relevant to IAS
phenology. As temperatures and biodiversity losses continue to increase, plant phenology is expected to continue to change (Primack et al. 2015; Wolf et al. 2017) and
airborne acquisition strategies must adjust accordingly. Second, satellite platforms
are critical to providing global-scale systematic monitoring of IAS. Current and
future missions must include high spectral resolution sensors with the capability to
create climate-relevant time series (a duration on the order of approximately a
E. A. Bolch et al.