8 Evaporation
179
mates with measurements at individual stations within a homogeneous surface and
28 Wm- 2 when comparing spatial averages of both measurements and estimates.
Error propagation and Monte Carlo simulation studies for specific heat balance
models give significantly larger error bounds (e.g. Bastiaanssen et aI., 1994). The
precise reason for this discrepancy is unclear.
Errors due to spatial averaging were studied by Bresnahan et al. (1997) who
evaluated the error of estimate for decreasing spatial resolutions. The error in the
mean increased erratically as resolution coarsens due to the non-linearity of the
model. Absolute error increased asymptotically. These results seem to confirm that
spatial patterns in independent variables are important in determining the correct
averaging procedure and scale for a regional process model.
Airborne eddy correlation measurements have the potential of providing a better
standard for remote sensing of evaporation, although this author is not aware of
any successful study in this direction. These observations should also be useful to
understand how to select spatial samples of evaporation maps for comparison with
local measurements (see e.g. Schuepp et aI., 1989)
The issue of accuracy deserves further attention by recalling the results of Allen et
al. (1989) on the comparison of different equations to estimate potential evaporation
against accurate measurements with lysimeters. Deviations of monthly values from
measurements were 10 % to 15 % for the best formula and 30 % for the worst one.
This provides a measure of the challenge to meet when estimating instantaneous actual evaporation. On the other hand large changes in evaporation are strongly correlated with differences in land cover, so it is expected that useful estimates of actual
evaporation can be obtained with obserVations capturing such differences in land
cover.
8.7 Applications
A significant amount of evidence has been reviewed in the previous pages to
document the progress made in remote sensing of land evaporation. Not surprisingly the accuracy of estimates could be documented for relatively simple situations such as agricultural fields or small, adequately equipped watersheds. On the
other hand several applications to larger and more complex areas have been
documented in literature.
Seguin (1981 and 1989) described the use of thermal infrared observations to
assess and monitor agricultural production. The relevance of heat balance studies
for land degradation studies was mentioned by Myers et al. (1980). Sandholt and
Andersen (1993) used. AVHRR and meteorological data to estimate and map
evaporation in northern Sahel. Nemani et al. (1995) used spectral vegetation indices and surface temperature to assess the impact of land cover changes on climatic
processes through changes in net radiation and evaporation. Ottle and Vidal Madjar (1994) demonstrated that the performance of a hydrological model of the
Adour basin could be improved with remote sensing estimates of evaporation. The
independent estimates of groundwater losses by evaporation in the Libyan desert
179
mates with measurements at individual stations within a homogeneous surface and
28 Wm- 2 when comparing spatial averages of both measurements and estimates.
Error propagation and Monte Carlo simulation studies for specific heat balance
models give significantly larger error bounds (e.g. Bastiaanssen et aI., 1994). The
precise reason for this discrepancy is unclear.
Errors due to spatial averaging were studied by Bresnahan et al. (1997) who
evaluated the error of estimate for decreasing spatial resolutions. The error in the
mean increased erratically as resolution coarsens due to the non-linearity of the
model. Absolute error increased asymptotically. These results seem to confirm that
spatial patterns in independent variables are important in determining the correct
averaging procedure and scale for a regional process model.
Airborne eddy correlation measurements have the potential of providing a better
standard for remote sensing of evaporation, although this author is not aware of
any successful study in this direction. These observations should also be useful to
understand how to select spatial samples of evaporation maps for comparison with
local measurements (see e.g. Schuepp et aI., 1989)
The issue of accuracy deserves further attention by recalling the results of Allen et
al. (1989) on the comparison of different equations to estimate potential evaporation
against accurate measurements with lysimeters. Deviations of monthly values from
measurements were 10 % to 15 % for the best formula and 30 % for the worst one.
This provides a measure of the challenge to meet when estimating instantaneous actual evaporation. On the other hand large changes in evaporation are strongly correlated with differences in land cover, so it is expected that useful estimates of actual
evaporation can be obtained with obserVations capturing such differences in land
cover.
8.7 Applications
A significant amount of evidence has been reviewed in the previous pages to
document the progress made in remote sensing of land evaporation. Not surprisingly the accuracy of estimates could be documented for relatively simple situations such as agricultural fields or small, adequately equipped watersheds. On the
other hand several applications to larger and more complex areas have been
documented in literature.
Seguin (1981 and 1989) described the use of thermal infrared observations to
assess and monitor agricultural production. The relevance of heat balance studies
for land degradation studies was mentioned by Myers et al. (1980). Sandholt and
Andersen (1993) used. AVHRR and meteorological data to estimate and map
evaporation in northern Sahel. Nemani et al. (1995) used spectral vegetation indices and surface temperature to assess the impact of land cover changes on climatic
processes through changes in net radiation and evaporation. Ottle and Vidal Madjar (1994) demonstrated that the performance of a hydrological model of the
Adour basin could be improved with remote sensing estimates of evaporation. The
independent estimates of groundwater losses by evaporation in the Libyan desert