332
height center estimates derived from short wavelength (X- or C-band) InSAR and a
ground DEM derived from longer wavelength InSAR (L- or P-bands, Neeff et al.
2005; Balzter et al. 2007a) or from another source such as LiDAR (e.g., Kellndorfer
et al. 2004; Andersen et al. 2008; Tighe 2012). Tighe et al. (2009) applied this
approach with the addition of correction factors for various forest ecotypes in US
and Canadian environments from semiarid to boreal. Integrating polarimetric
response with InSAR (i.e., PolInSAR; Cloude and Papathanassiou 1998) can
improve InSAR estimates of tree height (e.g., Balzter et al. 2007b). Tomography has
also shown promise in modeling the vertical distribution of canopy biomass, but
multiple acquisitions must be conducted within a short time to minimize temporal
decorrelation. The DEMs produced from InSAR can also be used to generate topographic indices for analysis of topographic complexity or roughness related to habitat diversity and biodiversity (Turner et al. 2003; Kuenzer et al. 2014). Fusion of
optical imagery and/or LiDAR with InSAR (e.g., as reviewed in Treuhaft et al.
2004) can also improve vertical canopy and topographic information.
Mapping of disturbance or environmental change can serve as an indicator of
potential impacts on habitat diversity and biodiversity. Many studies have been conducted in diverse applications that cannot be fully reviewed here. Most early applications were in mapping of deforestation, particularly in tropical regions where
deforestation had become a major issue (e.g., Rignot et al. 1997; van der Sanden and
Hoekman 1999). Temporal data have become widely used in land cover change
analysis using classification approaches (e.g., Thapa et al. 2013), analysis of backscatter change (e.g., Whittle et al. 2012; Mermoz and Le Toan 2016), and biophysical modeling for biomass loss (e.g., Mitchard et al. 2011). Other major radar
applications of environmental change with implications for biodiversity impacts are
burn and inundation detection and mapping. Early fire impact studies focused on
backscatter variations related to burn intensity classes (e.g., Kasischke et al. 1992).
More recent work has included temporal backscatter data in pre−/postburn analysis
(e.g., Tanase et al. 2015) and polarimetric analysis and decomposition in modeling
biomass changes due to fire (Martins et al. 2016). Radar is particularly useful in
detecting inundation under vegetated canopies due to specular reflection off the
water surface; in the case of inundated forests, penetration of the canopy by longer
wavelengths occurs with double bounce scattering off the water surface and tree
trunks (e.g., Kim et al. 2009) and phase differences between different polarizations
(e.g., Rignot et al. 1997).
Overall, use of radar for biodiversity and landscape diversity analysis, modeling,
mapping, and monitoring follows similar approaches to optical RS. Diversity of
land cover types or specific classes within a given ecotype may be directly mapped
using classification or modeling, while estimation of biophysical variables can serve
as indicators of spatial heterogeneity and potential habitat diversity or biodiversity.
The main contributions of radar are in its unique response to vegetation structure
that complements spectral reflectance characteristics of vegetation in the optical
regions. With the multitude of wavelengths, incidence angles, and polarizations
available, as well as the capability to acquire and process InSAR and polarimetric
data, much promise has been shown for mapping and monitoring land cover diversity- and biodiversity-related vegetation metrics. New satellite systems are being
A. Lausch et al.
height center estimates derived from short wavelength (X- or C-band) InSAR and a
ground DEM derived from longer wavelength InSAR (L- or P-bands, Neeff et al.
2005; Balzter et al. 2007a) or from another source such as LiDAR (e.g., Kellndorfer
et al. 2004; Andersen et al. 2008; Tighe 2012). Tighe et al. (2009) applied this
approach with the addition of correction factors for various forest ecotypes in US
and Canadian environments from semiarid to boreal. Integrating polarimetric
response with InSAR (i.e., PolInSAR; Cloude and Papathanassiou 1998) can
improve InSAR estimates of tree height (e.g., Balzter et al. 2007b). Tomography has
also shown promise in modeling the vertical distribution of canopy biomass, but
multiple acquisitions must be conducted within a short time to minimize temporal
decorrelation. The DEMs produced from InSAR can also be used to generate topographic indices for analysis of topographic complexity or roughness related to habitat diversity and biodiversity (Turner et al. 2003; Kuenzer et al. 2014). Fusion of
optical imagery and/or LiDAR with InSAR (e.g., as reviewed in Treuhaft et al.
2004) can also improve vertical canopy and topographic information.
Mapping of disturbance or environmental change can serve as an indicator of
potential impacts on habitat diversity and biodiversity. Many studies have been conducted in diverse applications that cannot be fully reviewed here. Most early applications were in mapping of deforestation, particularly in tropical regions where
deforestation had become a major issue (e.g., Rignot et al. 1997; van der Sanden and
Hoekman 1999). Temporal data have become widely used in land cover change
analysis using classification approaches (e.g., Thapa et al. 2013), analysis of backscatter change (e.g., Whittle et al. 2012; Mermoz and Le Toan 2016), and biophysical modeling for biomass loss (e.g., Mitchard et al. 2011). Other major radar
applications of environmental change with implications for biodiversity impacts are
burn and inundation detection and mapping. Early fire impact studies focused on
backscatter variations related to burn intensity classes (e.g., Kasischke et al. 1992).
More recent work has included temporal backscatter data in pre−/postburn analysis
(e.g., Tanase et al. 2015) and polarimetric analysis and decomposition in modeling
biomass changes due to fire (Martins et al. 2016). Radar is particularly useful in
detecting inundation under vegetated canopies due to specular reflection off the
water surface; in the case of inundated forests, penetration of the canopy by longer
wavelengths occurs with double bounce scattering off the water surface and tree
trunks (e.g., Kim et al. 2009) and phase differences between different polarizations
(e.g., Rignot et al. 1997).
Overall, use of radar for biodiversity and landscape diversity analysis, modeling,
mapping, and monitoring follows similar approaches to optical RS. Diversity of
land cover types or specific classes within a given ecotype may be directly mapped
using classification or modeling, while estimation of biophysical variables can serve
as indicators of spatial heterogeneity and potential habitat diversity or biodiversity.
The main contributions of radar are in its unique response to vegetation structure
that complements spectral reflectance characteristics of vegetation in the optical
regions. With the multitude of wavelengths, incidence angles, and polarizations
available, as well as the capability to acquire and process InSAR and polarimetric
data, much promise has been shown for mapping and monitoring land cover diversity- and biodiversity-related vegetation metrics. New satellite systems are being
A. Lausch et al.