18 Computation of HydroJogicaJ Data
411
Fig. 18.5. Location of the Tano river basin in Ghana, West Africa. MeteosatVIS Image
18.3.2 Assessment of the monthly area precipitation with the aid of multitemporal B2-Meteosat satellite imagery
Most of the existing cloud indexing techniques are based on a threshold value of a
certain cloud top temperature determined from the thermal infrared spectral band.
Griffith et al. (1978) states e.g. that precipitating cloud systems usually have a top
black body temperature of less than -20°C and a brightness of more than 80 DC
(Digital Counts) in the visible spectral band. Arkin (1979) gives a threshold value
for cloud top temperature of precipitating cloud systems with 235 K (-38°C). In
our case for the Tano river catchment sensitivity analyses showed that a threshold
value of -20°C provides optimal results. The technique presented here uses the IR
spectral band from the Meteosat in order to estimate the monthly area precipitation.
Since there is a trade-off between required model accuracy and amount and costs
of the data acquisition and processing, a compromise has been made in the technique presented here. The so-called B2 Meteosat data were used which contain
only every sixth pixel in a line and every sixth line in all three Meteosat spectral
channels. The time resolution was chosen to be 3 hours. This way the data costs
411
Fig. 18.5. Location of the Tano river basin in Ghana, West Africa. MeteosatVIS Image
18.3.2 Assessment of the monthly area precipitation with the aid of multitemporal B2-Meteosat satellite imagery
Most of the existing cloud indexing techniques are based on a threshold value of a
certain cloud top temperature determined from the thermal infrared spectral band.
Griffith et al. (1978) states e.g. that precipitating cloud systems usually have a top
black body temperature of less than -20°C and a brightness of more than 80 DC
(Digital Counts) in the visible spectral band. Arkin (1979) gives a threshold value
for cloud top temperature of precipitating cloud systems with 235 K (-38°C). In
our case for the Tano river catchment sensitivity analyses showed that a threshold
value of -20°C provides optimal results. The technique presented here uses the IR
spectral band from the Meteosat in order to estimate the monthly area precipitation.
Since there is a trade-off between required model accuracy and amount and costs
of the data acquisition and processing, a compromise has been made in the technique presented here. The so-called B2 Meteosat data were used which contain
only every sixth pixel in a line and every sixth line in all three Meteosat spectral
channels. The time resolution was chosen to be 3 hours. This way the data costs