where q swir is the reflectance at the shortwave infrared (SWIR) spectral band. The
capability of NDWI for estimating canopy water stress status that affects photosynthesis is limited by its sensitivity to the relatively small changes in relative
water content observed in natural vegetation and inability to discern changes in
canopy biomass from changes in canopy moisture status (Hunt and Rock 1989;
Gao 1996). Some studies, however, have shown that the NDWI is strongly correlated with leaf water content (equivalent water thickness) (Jackson et al. 2004)
and soil moisture (Fensholt and Sandholt 2003) over time. NDWI has been used to
derive a water scalar in a light use efficiency model (Xiao et al. 2005).
Satellite remote sensing has also been used to quantify LAI and fPAR (e.g.,
Myneni et al. 2002). These two variables characterize vegetation canopy functioning and energy absorption capacity (Myneni et al. 2002), and are key
parameters in most ecosystem productivity and biogeochemical models due to
their high correlation with GPP (Sellers et al. 1997).
The explanatory variables used in the data-driven approach include land cover,
EVI, LST, NDWI, fPAR, LAI, and photosynthetically active radiation (PAR), and
these variables can account for factors influencing NEE. The explanatory variables
used here are slightly different from those used previously (Xiao et al. 2008) in that
surface reflectance is not used here. All these variables can be obtained from
MODIS data products, which also avoid the complications and difficulties to
merge disparate data sources.
6.3 Methods
A data-driven approach (Xiao et al. 2008) is used to develop a predictive model for
NEE using flux observations from AmeriFlux and MODIS data streams. The
predictive model is then used to produce continuous NEE estimates for the conterminous U.S. over the period 2000–2009.
6.3.1 AmeriFlux Data
The AmeriFlux network consists of a number of active and inactive flux towers
across the U.S. (Fig. 6.1). The Level 4 NEE data were obtained for AmeriFlux
sites over the period 2000–2006 (Xiao et al. 2008). These sites are distributed
across the conterminous U.S. The Level 4 product consists of NEE data with four
different time steps, including half-hourly, daily, 8-day, and monthly. NEE was
calculated using the storage obtained from the discrete approach or using a vertical
CO 2 profile system, and was gap-filled using artificial neural network. The 8-day
NEE data (g C m
-2 day
-1 ) were used to match the compositing intervals of
MODIS data.
6 Assessing Net Ecosystem Exchange
153
capability of NDWI for estimating canopy water stress status that affects photosynthesis is limited by its sensitivity to the relatively small changes in relative
water content observed in natural vegetation and inability to discern changes in
canopy biomass from changes in canopy moisture status (Hunt and Rock 1989;
Gao 1996). Some studies, however, have shown that the NDWI is strongly correlated with leaf water content (equivalent water thickness) (Jackson et al. 2004)
and soil moisture (Fensholt and Sandholt 2003) over time. NDWI has been used to
derive a water scalar in a light use efficiency model (Xiao et al. 2005).
Satellite remote sensing has also been used to quantify LAI and fPAR (e.g.,
Myneni et al. 2002). These two variables characterize vegetation canopy functioning and energy absorption capacity (Myneni et al. 2002), and are key
parameters in most ecosystem productivity and biogeochemical models due to
their high correlation with GPP (Sellers et al. 1997).
The explanatory variables used in the data-driven approach include land cover,
EVI, LST, NDWI, fPAR, LAI, and photosynthetically active radiation (PAR), and
these variables can account for factors influencing NEE. The explanatory variables
used here are slightly different from those used previously (Xiao et al. 2008) in that
surface reflectance is not used here. All these variables can be obtained from
MODIS data products, which also avoid the complications and difficulties to
merge disparate data sources.
6.3 Methods
A data-driven approach (Xiao et al. 2008) is used to develop a predictive model for
NEE using flux observations from AmeriFlux and MODIS data streams. The
predictive model is then used to produce continuous NEE estimates for the conterminous U.S. over the period 2000–2009.
6.3.1 AmeriFlux Data
The AmeriFlux network consists of a number of active and inactive flux towers
across the U.S. (Fig. 6.1). The Level 4 NEE data were obtained for AmeriFlux
sites over the period 2000–2006 (Xiao et al. 2008). These sites are distributed
across the conterminous U.S. The Level 4 product consists of NEE data with four
different time steps, including half-hourly, daily, 8-day, and monthly. NEE was
calculated using the storage obtained from the discrete approach or using a vertical
CO 2 profile system, and was gap-filled using artificial neural network. The 8-day
NEE data (g C m
-2 day
-1 ) were used to match the compositing intervals of
MODIS data.
6 Assessing Net Ecosystem Exchange
153