have the same pigment content. Consequently, the method might be successful
when only one plant species homogeneously covers each pixel of the hyperspectral
image. For a homogeneous canopy such as heterogeneous grasslands (Zhang et al.,
2008; He and Mui, 2010), the method is not suitable because other factors such as
leaf age, light gradients, and leaf angle distribution would contribute to pigment
variation within such a canopy. Using species percentage cover data, the author
recently proposed an alternative method that could scale vegetation chlorophyll
content from the leaf to canopy or higher level for heterogeneous grasslands (Wong
and He, 2013). Understanding grassland chlorophyll properties are critical not only
at the leaf and canopy level but also at the landscape level because a study at this
scale could provide a better understanding of ecosystem nutritional and health
status. Further, landscape scale pigment estimation could provide information on
potential regional or global responses of vegetation to climate change.
The objective of this study was thus to bridge the gap in spatial scales through
estimating grassland chlorophyll content from leaf to landscape level. Specifically,
this study uses a spectral index to estimate grassland Chl a + b content at the leaf,
canopy, and landscape levels using ground and satellite remote sensing data. At the
leaf level, this study examined the relationship between leaf Chl a + b content data
and lab-derived hyperspectral reflectance data. The leaf-level Chl a + b content was
then scaled to the canopy level through a newly proposed simple method described in
the methodology section. The derived canopy-level Chl a + b was then correlated to
field hyperspectral data and space remote sensing data. The canopy-level Chl a + b
was further scaled to the landscape level to correlate with both field and satellite
remote sensing data.
7.2 STUDY AREA
This study was conducted at the Koffler Scientific Reserve at Jokers Hill, a 350-ha
field station owned by the University of Toronto and situated 50 km north of Toronto,
Ontario (44.03ʹ N, 79.29ʹ W; http://www.ksr.utoronto.ca). The study sites were
established in undisturbed old fields dominated by temperate tall grass prairies.
Some common species found in the study plots are common milkweed (Asclepias
syriaca), bird vetch (Vicia cracca L.), Canada goldenrod (Solidago), bog goldenrod
(Euthamia graminifolia (L.) Salisb.), Queen Anne’s lace (Daucus carota L.), rye
grass (Lolium perenne L.), red fescue (Festuca rubra L.), and Canada thistle (Cirsium
arvense).
7.3 DATA AND METHODS
7.3.1 Field Data Collection
Field data collection was conducted July 1 to 15, 2011, which is the maximum
growing season. This date collection period was chosen because chlorophyll is more
stable at the time window, while in the early or late growing season chlorophyll
128
ESTIMATING GRASSLAND CHLOROPHYLL CONTENT
when only one plant species homogeneously covers each pixel of the hyperspectral
image. For a homogeneous canopy such as heterogeneous grasslands (Zhang et al.,
2008; He and Mui, 2010), the method is not suitable because other factors such as
leaf age, light gradients, and leaf angle distribution would contribute to pigment
variation within such a canopy. Using species percentage cover data, the author
recently proposed an alternative method that could scale vegetation chlorophyll
content from the leaf to canopy or higher level for heterogeneous grasslands (Wong
and He, 2013). Understanding grassland chlorophyll properties are critical not only
at the leaf and canopy level but also at the landscape level because a study at this
scale could provide a better understanding of ecosystem nutritional and health
status. Further, landscape scale pigment estimation could provide information on
potential regional or global responses of vegetation to climate change.
The objective of this study was thus to bridge the gap in spatial scales through
estimating grassland chlorophyll content from leaf to landscape level. Specifically,
this study uses a spectral index to estimate grassland Chl a + b content at the leaf,
canopy, and landscape levels using ground and satellite remote sensing data. At the
leaf level, this study examined the relationship between leaf Chl a + b content data
and lab-derived hyperspectral reflectance data. The leaf-level Chl a + b content was
then scaled to the canopy level through a newly proposed simple method described in
the methodology section. The derived canopy-level Chl a + b was then correlated to
field hyperspectral data and space remote sensing data. The canopy-level Chl a + b
was further scaled to the landscape level to correlate with both field and satellite
remote sensing data.
7.2 STUDY AREA
This study was conducted at the Koffler Scientific Reserve at Jokers Hill, a 350-ha
field station owned by the University of Toronto and situated 50 km north of Toronto,
Ontario (44.03ʹ N, 79.29ʹ W; http://www.ksr.utoronto.ca). The study sites were
established in undisturbed old fields dominated by temperate tall grass prairies.
Some common species found in the study plots are common milkweed (Asclepias
syriaca), bird vetch (Vicia cracca L.), Canada goldenrod (Solidago), bog goldenrod
(Euthamia graminifolia (L.) Salisb.), Queen Anne’s lace (Daucus carota L.), rye
grass (Lolium perenne L.), red fescue (Festuca rubra L.), and Canada thistle (Cirsium
arvense).
7.3 DATA AND METHODS
7.3.1 Field Data Collection
Field data collection was conducted July 1 to 15, 2011, which is the maximum
growing season. This date collection period was chosen because chlorophyll is more
stable at the time window, while in the early or late growing season chlorophyll
128
ESTIMATING GRASSLAND CHLOROPHYLL CONTENT