402
W.G.M. Bastiaanssen
The sensitivity of results of Atmospheric Models to the representation of land surface processes
was emphasized by Eagleson (1981). At that time, rather simple parameterizations of land
surface hydrology were applied, for instance that developed by Manabe (1969) and based on
earlier work by Thornthwaite and Mather (1955) and Budyko (1956). Land surface hydrology
was basically reduced to a soil water reservoir, assumed to be of equal size everywhere in the
world (i.e. water storage capacity was independent of soil type, depth, etc.). This land surface
parameterization scheme, however, was sufficient to show that soil water content can have a
significant effect on the allocation of net radiation into sensible, latent and soil heat flux and,
therefore, on atmospheric circulation processes (Mintz, 1982; Rowntree and Bolton, 1983).
More recent research with more advanced land surface hydrology parameterizations confirmed
that surface albedo, soil water content and surface roughness are very critical for adequate
simulations of the climate system (e.g. Sud et al., 1990; Milly and Dunne, 1994)
This paper describes some classes of remote sensing algorithms which cover different approaches
to estimate large scale energy balances from remote sensing measurements. The key-problem
associated with energy balance modelling on the basis of thermal infrared measurements is to
adequately determine the sensible heat flux from surface radiation temperatures. The aim of
this paper is to address the classical problem of inferring a vertical air temperature difference
from spectral radiances and shows how several classes of existing remote sensing algorithms deal
with this issue. Some new 'internal calibration' techniques to get meso-scale sensible heat fluxes
in their correct physical range are the cornerstones of the Surface Energy Balance Algorithm
for Land (SEBAL) and will be outlined (Bastiaanssen, 1995). Some examples of the mapping of
meso- scale sensible and latent heat fluxes during the EFEDA experiment conducted in Castilla
la Mancha, Spain (Bolle et al., 1993), will be described.
15.2 Land Surface Fluxes and their Relation to Remote
Sensing Estimates of Surface Characteristics
Radiometers measure the radiances in selected bands throughout the visible, infrared and microwave region at the top of the atmosphere and no possibility exists to measure the turbulent
fluxes directly from these platforms (Menenti, 1993). However, several land surface state variables and properties mentioned in Table 15.1 can be deduced from these spectral radiances
and are essentially the key parameters marked by Garrat (1993) for improving the description
of land surface and atmosphere boundary layer treatments.
Since only the parameters mentioned in Table 15.1 are measurable, regional latent and sensible
heat fluxes with current satellite systems need to be estimated indirectly. To obtain estimates
of meso- scale land surface evaporation, a remote sensing parameterization scheme has therefore
to be developed, i.e. a schematization of the sensible (H) and latent heat ()"E) fluxes based
on the parameters listed in Table 15.1. The conventional micro-meteorological framework for
turbulent fluxes needs to be briefly discussed prior to identify the data gaps between Table 15.1
and Hand )"E fluxes at composite land surfaces.
The relationship between mean turbulent momentum T, sensible H and latent heat )..E fluxes
are usually described by:
T = Pau~ (N m- 2 )
(15.1)
H = -PaCpT.u. (W m- 2 )
(15.2)
)"E = -)..qaq.U. (W m- 2 )
(15.3)
where T (N m- 2 ) is the momentum flux, Pa (kg m- 3 ) is the air density, u. (m S-l) is the friction
velocity, Cp (J kg- 1 K- 1 ) is the air specific heat at constant pressure, T. (K) is the temperature
W.G.M. Bastiaanssen
The sensitivity of results of Atmospheric Models to the representation of land surface processes
was emphasized by Eagleson (1981). At that time, rather simple parameterizations of land
surface hydrology were applied, for instance that developed by Manabe (1969) and based on
earlier work by Thornthwaite and Mather (1955) and Budyko (1956). Land surface hydrology
was basically reduced to a soil water reservoir, assumed to be of equal size everywhere in the
world (i.e. water storage capacity was independent of soil type, depth, etc.). This land surface
parameterization scheme, however, was sufficient to show that soil water content can have a
significant effect on the allocation of net radiation into sensible, latent and soil heat flux and,
therefore, on atmospheric circulation processes (Mintz, 1982; Rowntree and Bolton, 1983).
More recent research with more advanced land surface hydrology parameterizations confirmed
that surface albedo, soil water content and surface roughness are very critical for adequate
simulations of the climate system (e.g. Sud et al., 1990; Milly and Dunne, 1994)
This paper describes some classes of remote sensing algorithms which cover different approaches
to estimate large scale energy balances from remote sensing measurements. The key-problem
associated with energy balance modelling on the basis of thermal infrared measurements is to
adequately determine the sensible heat flux from surface radiation temperatures. The aim of
this paper is to address the classical problem of inferring a vertical air temperature difference
from spectral radiances and shows how several classes of existing remote sensing algorithms deal
with this issue. Some new 'internal calibration' techniques to get meso-scale sensible heat fluxes
in their correct physical range are the cornerstones of the Surface Energy Balance Algorithm
for Land (SEBAL) and will be outlined (Bastiaanssen, 1995). Some examples of the mapping of
meso- scale sensible and latent heat fluxes during the EFEDA experiment conducted in Castilla
la Mancha, Spain (Bolle et al., 1993), will be described.
15.2 Land Surface Fluxes and their Relation to Remote
Sensing Estimates of Surface Characteristics
Radiometers measure the radiances in selected bands throughout the visible, infrared and microwave region at the top of the atmosphere and no possibility exists to measure the turbulent
fluxes directly from these platforms (Menenti, 1993). However, several land surface state variables and properties mentioned in Table 15.1 can be deduced from these spectral radiances
and are essentially the key parameters marked by Garrat (1993) for improving the description
of land surface and atmosphere boundary layer treatments.
Since only the parameters mentioned in Table 15.1 are measurable, regional latent and sensible
heat fluxes with current satellite systems need to be estimated indirectly. To obtain estimates
of meso- scale land surface evaporation, a remote sensing parameterization scheme has therefore
to be developed, i.e. a schematization of the sensible (H) and latent heat ()"E) fluxes based
on the parameters listed in Table 15.1. The conventional micro-meteorological framework for
turbulent fluxes needs to be briefly discussed prior to identify the data gaps between Table 15.1
and Hand )"E fluxes at composite land surfaces.
The relationship between mean turbulent momentum T, sensible H and latent heat )..E fluxes
are usually described by:
T = Pau~ (N m- 2 )
(15.1)
H = -PaCpT.u. (W m- 2 )
(15.2)
)"E = -)..qaq.U. (W m- 2 )
(15.3)
where T (N m- 2 ) is the momentum flux, Pa (kg m- 3 ) is the air density, u. (m S-l) is the friction
velocity, Cp (J kg- 1 K- 1 ) is the air specific heat at constant pressure, T. (K) is the temperature