324
related to local energy, water, or nutrient conditions. Examples will be briefly
described of methods to disentangle the effects of water, energy, and nutrients on
plants in the context of vegetation.
Müller et al. (2014, 2016) applied principal component analysis (PCA) to extract
dominant LST patterns from time series (28 scenes covering 12 years) of ASTER
TIR images of the mesoscale Attert catchment in midwestern Luxembourg. The
PCA-component values for each pixel were related to land use/vegetation data and
to geological and soil texture data, indicating a strong information signal in the
temporal dynamics of LST data with regard to plant diversity.
Environmental disturbances have been investigated (e.g., by Duro et al. 2007)
making use of the negative relationship between vegetation density and
LST. Mildrexler et al. (2007) proposed a disturbance detection index based on this
principle that uses the 16-day MODIS Enhanced Vegetation Index (EVI) and 8-day
LST. They were able to successfully detect disturbance events such as wildfire,
irrigated vegetation, precipitation variability, and the recovery of disturbed landscapes at the continental scale.
Sun and Schulz (2015) could demonstrate that an integration of TIR data from
Landsat 5 and 8 was able to significantly enhance the classification results for different aggregation levels of land-use and land cover categories for a mesoscale
catchment in Luxembourg. This indicates the high potential of TIR data to support
more specific and selective plant species monitoring as relevant for biodiversity
research.
Environmental stress induced by long-term heat waves and/or a limited availability of water is likely to reduce stomata conductance, limit transpiration, and
thereby increase leaf surface temperature (Stoll and Jones 2007). The difference
between air temperature and leaf temperature combined with information on vegetation density can serve as an indicator of plant stress. Hoffmann et al. (2016) used
a spectral vegetation index and LST data from cameras mounted on UAVs to develop
a water deficit index (WDI). The WDI was highly correlated to eddy covariance
measurements of latent heat fluxes over a growing season, and that was used to map
spatially distributed water demands of various crops.
Environmental stress may also cause changes in leaves and the structure of
plants, dependent on their biophysiological characteristics. Buitrago et al. (2016)
found that two plant species [European beech, (Fagus sylvatica) and rhododendron
(Rhododendron catawbiense)], when exposed to either water or temperature stress,
experience significant changes in TIR radiance. The changes in TIR in response to
stress were similar within a species, regardless of the stress. However, changes in
TIR spectra differed between species, and these differences could be explained by
changes in the microstructure and biochemistry of leaves (e.g., cuticula).
Overall, the potential for exploiting LST information data in plant biodiversity
research is manifold. While LST is easily measured by thermometers at the point
scale, satellite RS TIR data are needed in order to derive LST routinely at high temporal and spatial resolutions over large spatial extents. However, the derivation of
LST from TIR data is a difficult task because such radiance measurements depend
A. Lausch et al.
related to local energy, water, or nutrient conditions. Examples will be briefly
described of methods to disentangle the effects of water, energy, and nutrients on
plants in the context of vegetation.
Müller et al. (2014, 2016) applied principal component analysis (PCA) to extract
dominant LST patterns from time series (28 scenes covering 12 years) of ASTER
TIR images of the mesoscale Attert catchment in midwestern Luxembourg. The
PCA-component values for each pixel were related to land use/vegetation data and
to geological and soil texture data, indicating a strong information signal in the
temporal dynamics of LST data with regard to plant diversity.
Environmental disturbances have been investigated (e.g., by Duro et al. 2007)
making use of the negative relationship between vegetation density and
LST. Mildrexler et al. (2007) proposed a disturbance detection index based on this
principle that uses the 16-day MODIS Enhanced Vegetation Index (EVI) and 8-day
LST. They were able to successfully detect disturbance events such as wildfire,
irrigated vegetation, precipitation variability, and the recovery of disturbed landscapes at the continental scale.
Sun and Schulz (2015) could demonstrate that an integration of TIR data from
Landsat 5 and 8 was able to significantly enhance the classification results for different aggregation levels of land-use and land cover categories for a mesoscale
catchment in Luxembourg. This indicates the high potential of TIR data to support
more specific and selective plant species monitoring as relevant for biodiversity
research.
Environmental stress induced by long-term heat waves and/or a limited availability of water is likely to reduce stomata conductance, limit transpiration, and
thereby increase leaf surface temperature (Stoll and Jones 2007). The difference
between air temperature and leaf temperature combined with information on vegetation density can serve as an indicator of plant stress. Hoffmann et al. (2016) used
a spectral vegetation index and LST data from cameras mounted on UAVs to develop
a water deficit index (WDI). The WDI was highly correlated to eddy covariance
measurements of latent heat fluxes over a growing season, and that was used to map
spatially distributed water demands of various crops.
Environmental stress may also cause changes in leaves and the structure of
plants, dependent on their biophysiological characteristics. Buitrago et al. (2016)
found that two plant species [European beech, (Fagus sylvatica) and rhododendron
(Rhododendron catawbiense)], when exposed to either water or temperature stress,
experience significant changes in TIR radiance. The changes in TIR in response to
stress were similar within a species, regardless of the stress. However, changes in
TIR spectra differed between species, and these differences could be explained by
changes in the microstructure and biochemistry of leaves (e.g., cuticula).
Overall, the potential for exploiting LST information data in plant biodiversity
research is manifold. While LST is easily measured by thermometers at the point
scale, satellite RS TIR data are needed in order to derive LST routinely at high temporal and spatial resolutions over large spatial extents. However, the derivation of
LST from TIR data is a difficult task because such radiance measurements depend
A. Lausch et al.