449
Remote Sensing and Modeling of Global Evapotranspiration
Cleugh et al. (2007) used the more theoretically based P–M (Equation 19.5)
(Monteith 1965) to estimate ET over Australia with MODIS data. On the basis of
Cleugh et al.’s (2007) model, Mu et al. (2007a) developed a remotely sensed ET model
(RS-ET) to obtain the first remotely sensed global terrestrial ET map, suggesting it
is applicable to operationally estimate global ET in near real time at satellite sensor
resolution. The RS-ET algorithm employs reanalysis surface meteorological data
from the Global Modeling and Assimilation Office (GMAO 2004; v. 4.0.0) in the
National Aeronautics and Space Administration (NASA), with MODIS land cover,
albedo, LAI, and enhanced vegetation index (EVI) as inputs for regional and global
ET mapping and monitoring. On the basis of the RS-ET model (Mu et al. 2007a),
Zhang et al. (2009) developed a model to estimate ET using remotely sensed NDVI
data; Yuan et al. (2010) modified the RS-ET model by adding the constraint of air
temperature to stomatal conductance and calculating the vegetation cover fraction
using LAI instead of EVI.
There are also other methods using remote sensing data to estimate global ET. For
example, Fisher et al. (2008) used the Priestly and Taylor (1972) method to estimate
global ET using AVHRR data; Jung et al. (2010) used a machine learning method to
upscale the FLUXNET tower data to calculate the global ET. A standard and nearreal-time remotely sensed terrestrial ET data product can provide critical information on the regional and global water cycle and resulting environment changes.
19.3 THEORETICAL BASIS OF THE GLOBAL
MODIS ET (RS-ET) ALGORITHM
The RS-ET algorithm uses the well-known P–M equation (Equation 19.5; Mu et al.
2007a). The RS-ET algorithm considers both the surface energy partitioning process and environmental constraints on ET and calculates ET as the sum of plant
transpiration and soil evaporation. The vegetation cover fraction is estimated with
MODIS EVI; the net solar radiation is partitioned into components of vegetation
and soil; stomatal conductance at the leaf level is controlled by VPD and minimum
air temperature; leaf conductance is then upscaled to the canopy level with MODIS
LAI; soil evaporation is estimated by the potential soil evaporation and reduced by
air humidity and vapor pressure deficit. The RS-ET algorithm has been validated
with measured vapor fluxes at eddy flux towers and tested at the global scale. In this
section, we detail how the RS-ET model estimates ET by using MODIS and daily
meteorological data.
19.3.1 MatheMatical deScRiPtion
In the RS-ET model, ET is the sum of water lost to the atmosphere from the soil surface through soil evaporation and from plant tissues via transpiration (Figure 19.2).
Plant transpiration and soil evaporation are separated by vegetation cover fraction
(F C ) derived from MODIS EVI. Plant transpiration is largely determined by stomatal
openness at the leaf level and canopy conductance at the canopy level, and therefore,
canopy conductance is critical to ET estimate. Evaporation from soil is constrained
by relative humidity and vapor pressure deficit.
Remote Sensing and Modeling of Global Evapotranspiration
Cleugh et al. (2007) used the more theoretically based P–M (Equation 19.5)
(Monteith 1965) to estimate ET over Australia with MODIS data. On the basis of
Cleugh et al.’s (2007) model, Mu et al. (2007a) developed a remotely sensed ET model
(RS-ET) to obtain the first remotely sensed global terrestrial ET map, suggesting it
is applicable to operationally estimate global ET in near real time at satellite sensor
resolution. The RS-ET algorithm employs reanalysis surface meteorological data
from the Global Modeling and Assimilation Office (GMAO 2004; v. 4.0.0) in the
National Aeronautics and Space Administration (NASA), with MODIS land cover,
albedo, LAI, and enhanced vegetation index (EVI) as inputs for regional and global
ET mapping and monitoring. On the basis of the RS-ET model (Mu et al. 2007a),
Zhang et al. (2009) developed a model to estimate ET using remotely sensed NDVI
data; Yuan et al. (2010) modified the RS-ET model by adding the constraint of air
temperature to stomatal conductance and calculating the vegetation cover fraction
using LAI instead of EVI.
There are also other methods using remote sensing data to estimate global ET. For
example, Fisher et al. (2008) used the Priestly and Taylor (1972) method to estimate
global ET using AVHRR data; Jung et al. (2010) used a machine learning method to
upscale the FLUXNET tower data to calculate the global ET. A standard and nearreal-time remotely sensed terrestrial ET data product can provide critical information on the regional and global water cycle and resulting environment changes.
19.3 THEORETICAL BASIS OF THE GLOBAL
MODIS ET (RS-ET) ALGORITHM
The RS-ET algorithm uses the well-known P–M equation (Equation 19.5; Mu et al.
2007a). The RS-ET algorithm considers both the surface energy partitioning process and environmental constraints on ET and calculates ET as the sum of plant
transpiration and soil evaporation. The vegetation cover fraction is estimated with
MODIS EVI; the net solar radiation is partitioned into components of vegetation
and soil; stomatal conductance at the leaf level is controlled by VPD and minimum
air temperature; leaf conductance is then upscaled to the canopy level with MODIS
LAI; soil evaporation is estimated by the potential soil evaporation and reduced by
air humidity and vapor pressure deficit. The RS-ET algorithm has been validated
with measured vapor fluxes at eddy flux towers and tested at the global scale. In this
section, we detail how the RS-ET model estimates ET by using MODIS and daily
meteorological data.
19.3.1 MatheMatical deScRiPtion
In the RS-ET model, ET is the sum of water lost to the atmosphere from the soil surface through soil evaporation and from plant tissues via transpiration (Figure 19.2).
Plant transpiration and soil evaporation are separated by vegetation cover fraction
(F C ) derived from MODIS EVI. Plant transpiration is largely determined by stomatal
openness at the leaf level and canopy conductance at the canopy level, and therefore,
canopy conductance is critical to ET estimate. Evaporation from soil is constrained
by relative humidity and vapor pressure deficit.