66
Chapter 8
satellite data. The main interest of the method is that information on land use
and synoptic data is not required. SEBAL works only during clear sky
conditions and with the assumption that the Meteosat image contains a
number of wet and dry pixels. A statistical relation between surface
temperature and surface albedo is used to identify the dry (high
and
albedo) and wet (low
and albedo) areas. Assuming that sensible heat
equals zero for the wet pixels and that evaporation is negligible for the dry
pixels, SEBAL computes the sensible H and latent LE heat fluxes for the
whole image. The remotely sensed derived input of SEBAL includes surface
temperature, albedo and NDVI. The method has been tested over Spain on
the basis of EFEDA data. The system proved capable to retrieve the
evaporative fraction
for soil and sparse vegetation. Hurk
et al. (1997) proposed a method to modify the soil moisture from the
comparison between the simulated and SEBAL (observed) evaporative
fractions. The minimization of forecast errors was achieved by adjusting the
soil reservoir. A preliminary study was conducted over the Iberian Peninsula
during a 7-day period in the summer of 1994. Figure 3 illustrates the results
of the assimilation cycle. In the control run, the initial soil moisture is taken
from the climatology and then evolves freely (no correction). In the
experimental run, the soil reservoir is corrected every day using the simple
assimilation procedure. Figure 3 shows that the assimilation method has a
positive impact on the mean bias and root mean square errors of screen level
temperature and humidity over Spain. Thus, these results suggest that
remotely sensed data could improve the estimation of surface fluxes and help
the assimilation of soil moisture in NWP models. However, some important
technical points remain to be solved before an operational implementation of
the method. For instance, the method has to be used regularly, in particular
in cloudy and moderate soil water conditions. The SEBAL algorithm could
also benefit from the use of vegetation classification and low level
atmospheric information (influence of wind speed, for instance).
Chapter 8
satellite data. The main interest of the method is that information on land use
and synoptic data is not required. SEBAL works only during clear sky
conditions and with the assumption that the Meteosat image contains a
number of wet and dry pixels. A statistical relation between surface
temperature and surface albedo is used to identify the dry (high
and
albedo) and wet (low
and albedo) areas. Assuming that sensible heat
equals zero for the wet pixels and that evaporation is negligible for the dry
pixels, SEBAL computes the sensible H and latent LE heat fluxes for the
whole image. The remotely sensed derived input of SEBAL includes surface
temperature, albedo and NDVI. The method has been tested over Spain on
the basis of EFEDA data. The system proved capable to retrieve the
evaporative fraction
for soil and sparse vegetation. Hurk
et al. (1997) proposed a method to modify the soil moisture from the
comparison between the simulated and SEBAL (observed) evaporative
fractions. The minimization of forecast errors was achieved by adjusting the
soil reservoir. A preliminary study was conducted over the Iberian Peninsula
during a 7-day period in the summer of 1994. Figure 3 illustrates the results
of the assimilation cycle. In the control run, the initial soil moisture is taken
from the climatology and then evolves freely (no correction). In the
experimental run, the soil reservoir is corrected every day using the simple
assimilation procedure. Figure 3 shows that the assimilation method has a
positive impact on the mean bias and root mean square errors of screen level
temperature and humidity over Spain. Thus, these results suggest that
remotely sensed data could improve the estimation of surface fluxes and help
the assimilation of soil moisture in NWP models. However, some important
technical points remain to be solved before an operational implementation of
the method. For instance, the method has to be used regularly, in particular
in cloudy and moderate soil water conditions. The SEBAL algorithm could
also benefit from the use of vegetation classification and low level
atmospheric information (influence of wind speed, for instance).