Remote Sensing of Leaf, Canopy, and Vegetation Water Contents for Satellite Environmental Data Records - Satellite-based Applications on Climate Change - page 338

areas (Burgan 1988 ). Burgan ( 1988 ) used AVHRR-NDVI data to estimate the
spatial extent of vegetation dryness (Burgan and Hartford 1993 ; Hardy and Burgan
1999 ). The estimated FMC data are used in fire potential models that include
weather, topography, vegetation type, and other variables (Keane et al. 2001 ,
2010 ; Dennison et al. 2008 ; Rollins et al. 2004 ; Rollins 2009 ; Chuvieco et al.
2010 ). However, NDVI is only indirectly correlated to FMC, so remotely sensed
CWC has the potential to be more accurate for estimating wildfire risk.
FMC depends on the accurate estimation of both CWC and dry matter ( C m ).
Becausethereisalargevariationin C m among different habitats, from thin mesic
leaves to thick succulent leaves in xeric habitats, C m is hard to predict. Further,
differences in leaf thickness, dry weight, and morphology cause large variation in
FMC for leaves at full turgor. LWC or CWC are linearly related to FMC when
vegetation types with similar C m are considered (Chuvieco et al. 2002 , 2003 ;
Zarco-Tejada et al. 2003 ;Makietal. 2004 ;Dennisonetal. 2005 ;Robertsetal.
2006 ; Dasgupta et al. 2007 ; Verbesselt et al. 2007 ). According to Shipley and Vu
( 2002 ), differences in C m among species accounts for 95% of total intraspecies
variation. The use of a constant C m throughout the year could be considered, but
an overall decrease is expected during the drought season (Garnier et al. 2001 ).
Remote sensing C m directly is difficult because absorption features associated
with dry matter are obscured by liquid water in the leaves (Fourty and Baret 1997 ).
With imaging spectrometers that have very high signal to noise, C m could be
estimated at about 1722-nm wavelength, because C m has an average absorption
coefficient somewhat greater than that of liquid water. On the other hand, there is
much more water than dry matter so that the 1722-nm absorption feature is not
plainly visible in green-leaf reflectance spectra. To uncover the spectrum of dry
matter, effects of liquid water “may be removed” by fitting the absorption
coefficients of water (Fig. 20.1 ) to the foliar reflectance spectrum (Gao and Goetz
1994 , 1995 ; Schlerf et al. 2010 ; Ramoelo et al. 2011 ; Wang et al. 2011a ). The
residuals between the foliar reflectance spectrum and the fitted equation show the
absorption features associated with dry matter, particularly at the 1,722-nm wave-
length (Gao and Goetz 1994 , 1995 ; Wang et al. 2011a ). The depth of the residual
absorption feature is directly related to C m (Wang et al. 2011a ).
Even for fresh green leaves, C m subtly influences the foliar reflectance spectrum,
so that narrow-band indices are useful for estimating FMC (Wang et al. 2011b , c ;
Romero et al. 2012 ). Wang et al. ( 2011b , c ) developed the Normalized Dry Matter
Index (NDMI) based on the absorption at 1722-nm wavelength. Because FMC is
the ratio of LWC/ C m (Table 20.1 ), Wang et al. ( 2011b ) hypothesized that the ratio
of a foliar water index and the NDMI should be related to FMC. The results are
promising (Fig. 20.5 ), but use of this index will require imaging spectrometers,
which probably will not be available for routine monitoring of FMC for another
decade.
FMC changes dramatically over a growing season, so data need to be acquired
much more frequently than may be possible with imaging spectrometers. Daily
weather sensors such as MODIS and VIIRS, however, do not have narrow bands
useful for the detection of dry matter. As discussed in the next section, inversion of