412
I. Papadakis and G.A. Schultz
were reduced significantly without much loss of accuracy as a sensitivity analysis
has shown.
The relationship between AFCC (Accumulated Fractional Cloud Cover) and
monthly area precipitation. Three-hourly Fractional Cloud Cover Indices (FCCI)
were determined as a function ofa chosen threshold cloud top temperature (-20°C
in this case). The AFCC which is of interest for the computation of the monthly
area precipitation is simply the sum of the three-hourly FCCI for the whole month.
Colour Plate 18.A shows the cloud development over the Tano River basin,
Ghana, West Africa by means of two successive Meteosat images of the IR spectral channel for a rainfall event in August 1984.
To estimate the monthly area precipitation, the empirical relationship of best fit
is determined using the least-squares method. Different relationships with numerical values depending upon the radiance and specific local climatological conditions, e.g. the threshold cloud top temperature (from the thermal infrared spectral
band) and the migration of the ITDZ (Inter-Tropical Discontinuity Zone), were
varied in order to fmd out an optimum agreement between observed monthly rainfall and spectral information from Meteosat. The AFCC and the monthly area
precipitation for the test catchment in Ghana are related by:
wet season
P = 10 1.4 . AFCC~13 s
(18.3)
dry season
P = 2.03 . AFCC\:2 s
(18.4)
where:
P:
Monthly area precipitation
AFCC: Accumulated Fractional Cloud Cover
Ts:
Threshold value of the cloud top temperature (here -20°C)
The equations described above and developed in the calibration procedure with the
aid of simultaneous satellite data (AFCC) and ground truth (observed precipitation) may be used for reconstruction of historical monthly area precipitation in the
catchment area on the basis of satellite data alone. As can be seen from Fig. 18.6
the agreement between observed and computed monthly area precipitation values
is reasonably good.
Moreover, according to a sensitivity analysis of the time resolution, two satellite
images per day taken at 20 and 23 GMT (Greenwich Meridian Time) are enough
to estimate monthly area precipitation for hydrological purposes in the Tano river
basin without significant loss in accuracy as compared to the use of all 48 images
per day.
I. Papadakis and G.A. Schultz
were reduced significantly without much loss of accuracy as a sensitivity analysis
has shown.
The relationship between AFCC (Accumulated Fractional Cloud Cover) and
monthly area precipitation. Three-hourly Fractional Cloud Cover Indices (FCCI)
were determined as a function ofa chosen threshold cloud top temperature (-20°C
in this case). The AFCC which is of interest for the computation of the monthly
area precipitation is simply the sum of the three-hourly FCCI for the whole month.
Colour Plate 18.A shows the cloud development over the Tano River basin,
Ghana, West Africa by means of two successive Meteosat images of the IR spectral channel for a rainfall event in August 1984.
To estimate the monthly area precipitation, the empirical relationship of best fit
is determined using the least-squares method. Different relationships with numerical values depending upon the radiance and specific local climatological conditions, e.g. the threshold cloud top temperature (from the thermal infrared spectral
band) and the migration of the ITDZ (Inter-Tropical Discontinuity Zone), were
varied in order to fmd out an optimum agreement between observed monthly rainfall and spectral information from Meteosat. The AFCC and the monthly area
precipitation for the test catchment in Ghana are related by:
wet season
P = 10 1.4 . AFCC~13 s
(18.3)
dry season
P = 2.03 . AFCC\:2 s
(18.4)
where:
P:
Monthly area precipitation
AFCC: Accumulated Fractional Cloud Cover
Ts:
Threshold value of the cloud top temperature (here -20°C)
The equations described above and developed in the calibration procedure with the
aid of simultaneous satellite data (AFCC) and ground truth (observed precipitation) may be used for reconstruction of historical monthly area precipitation in the
catchment area on the basis of satellite data alone. As can be seen from Fig. 18.6
the agreement between observed and computed monthly area precipitation values
is reasonably good.
Moreover, according to a sensitivity analysis of the time resolution, two satellite
images per day taken at 20 and 23 GMT (Greenwich Meridian Time) are enough
to estimate monthly area precipitation for hydrological purposes in the Tano river
basin without significant loss in accuracy as compared to the use of all 48 images
per day.