416
1. Papadakis and G.A. Schultz
(5x5 km resolution) passes a designated threshold for one day. Analogue gridded
fields are created using raingage data only. Observed daily precipitation is interpolated. The Forecast Component consists of hydrologic models and software to
produce flow and stage hydro graphs based on inputs supplied by the Preprocessor.
The User Interface displays satellite imagery data, gridded data, and time series
data; it also serves as the interface between the forecaster and the NFS software,
data base and models. The results gained by this technique were used for monitoring and forecasting purposes in the Nile river system.
18.5 Summary and Discussion
Since it is not possible to measure runoff directly with the aid of remote sensing
information an indirect approach is suggested which uses remote sensing data for
the estimation of rainfall and transform it into runoff with the aid of a rainfallrunoff-model. This way, a long time series of monthly river flows can be generated
which allows the estimation of the expected future performance of a planned water
project in terms of reliability indices.
The approach presented here shows that the use of multi-temporal satellite imagery allows the estimation of monthly area precipitation values essential for the
transformation into runoff values by means of a rainfall runoff model.
The application of the approach presented here, however, is limited. Catchment
characteristics, the rainfall type, the area extension of the catchment and the geographical transferability form the main constraints.
The relationship between rainfall and runoff depends mainly on the hydrological
behaviour of the catchment area. Therefore, the reconstruction of historical river
flows based on a rainfall runoff model calibrated with the aid of a short observed
flow record has to take into consideration possible changes in the catchment characteristics, such as residential areas, land use, etc. These changes could be quantified using satellite systems such as NOAA! A VHRR, Landsat or SPOT.
The use of spectral information obtained from the visible, the thermal infrared
and the water vapour spectral region allows estimation only of convective rainfall
events. This means that only regions with mainly convective rainfall activity can
be contemplated. Such regions are concentrated around the Equator.
There is a relationship between the spatial resolution of satellite data and the
extension of the catchment area under consideration. By using geostationary satellite systems only areas greater than 10 000 km 2 can be modelled adequately.
The relationship between spectral cloud information and rainfall is empirical and
is valid only for the area in question. Previous investigations (Wylie, 1979;
Griffith et al. 1981; Adler and Mack, 1984) have shown that the direct transfer to
other geographical regions is not possible without further analysis for the specific
area. However, the use of the water vapour channel seems promising for overcoming this problem.
Additional spectral information such as passive microwave data could contribute
to removing the limitations mentioned above, because the rainfall and passive
microwave information have a direct physical relationship to each other. The dis-
1. Papadakis and G.A. Schultz
(5x5 km resolution) passes a designated threshold for one day. Analogue gridded
fields are created using raingage data only. Observed daily precipitation is interpolated. The Forecast Component consists of hydrologic models and software to
produce flow and stage hydro graphs based on inputs supplied by the Preprocessor.
The User Interface displays satellite imagery data, gridded data, and time series
data; it also serves as the interface between the forecaster and the NFS software,
data base and models. The results gained by this technique were used for monitoring and forecasting purposes in the Nile river system.
18.5 Summary and Discussion
Since it is not possible to measure runoff directly with the aid of remote sensing
information an indirect approach is suggested which uses remote sensing data for
the estimation of rainfall and transform it into runoff with the aid of a rainfallrunoff-model. This way, a long time series of monthly river flows can be generated
which allows the estimation of the expected future performance of a planned water
project in terms of reliability indices.
The approach presented here shows that the use of multi-temporal satellite imagery allows the estimation of monthly area precipitation values essential for the
transformation into runoff values by means of a rainfall runoff model.
The application of the approach presented here, however, is limited. Catchment
characteristics, the rainfall type, the area extension of the catchment and the geographical transferability form the main constraints.
The relationship between rainfall and runoff depends mainly on the hydrological
behaviour of the catchment area. Therefore, the reconstruction of historical river
flows based on a rainfall runoff model calibrated with the aid of a short observed
flow record has to take into consideration possible changes in the catchment characteristics, such as residential areas, land use, etc. These changes could be quantified using satellite systems such as NOAA! A VHRR, Landsat or SPOT.
The use of spectral information obtained from the visible, the thermal infrared
and the water vapour spectral region allows estimation only of convective rainfall
events. This means that only regions with mainly convective rainfall activity can
be contemplated. Such regions are concentrated around the Equator.
There is a relationship between the spatial resolution of satellite data and the
extension of the catchment area under consideration. By using geostationary satellite systems only areas greater than 10 000 km 2 can be modelled adequately.
The relationship between spectral cloud information and rainfall is empirical and
is valid only for the area in question. Previous investigations (Wylie, 1979;
Griffith et al. 1981; Adler and Mack, 1984) have shown that the direct transfer to
other geographical regions is not possible without further analysis for the specific
area. However, the use of the water vapour channel seems promising for overcoming this problem.
Additional spectral information such as passive microwave data could contribute
to removing the limitations mentioned above, because the rainfall and passive
microwave information have a direct physical relationship to each other. The dis-