402
I. Papadakis and G.A. Schultz
positive deviation in the estimated mean annual flow, all structures would be overestimated. Dam height would be higher than necessary, the reservoir would very
seldom be filled up to the required water level. The consequence would be that the
structures necessary for water power production such as the power station, conduits etc. would be much larger in size than required. The power production would
be below the planned target, and some machines would then remain idle or work
with low efficiency. For the planned irrigation schemes and water supply networks
there would be insufficient water to irrigate and supply the whole area and the
channels and other structures would be of disproportionate size.
Often the design of water management systems in many countries in the developing world suffers from inadequate or non-available sets of hydrometeorological
data. If hydrometeorological records are available, they are either too short or, if
they are sufficiently long, they include numerous gaps or discontinuities which
make these useless for further analysis (Gyau-Boakye, 1993; Gyau-Boakye and
Schultz 1994). The consequence is that very often water projects in these countries have to be designed under circumstances which are insufficiently representative for the project area. Instead of calculation of design data based on detailed
knowledge of the hydrological behaviour of the area under consideration, data are
prepared based on formulas, the validity of which has been proven for entirely
different areas.
Different mathematical-statistical and deterministic methods have been developed in the past which enable the designer to estimate missing data, generate synthetic sequences and derive estimates for design purposes (Gyau-Boakye, 1993).
The success of these methods, however, depends primarily on the length of the
available data considered for the identification of the system in question since
computation of systems reliability depends on long time series.
The remote sensing technology provides very useful methods of monitoring and
identifying many hydrological parameters and variables which are necessary for
the estimation of the project reliability. With the aid of remote sensing techniques,
relevant hydrological data can be obtained in a short time, at periodic intervals and
covering extended areas. Satellite imagery has been used in the fields of meteorology, watershed management, hydrologic modelling etc. Mathematical techniques
in a multidimensional feature space allow an accurate land use classification as
well as estimation of vegetation indices (Su, 1996). Changes in land use can be
quantified on the basis of multi-spectral and multi-temporal satellite imagery. Also
the identification of the potential location of dams in river valleys can be supported by remote sensing. A number of scientists have already used digital spectral
data from meteorological geostationary satellite systems to estimate precipitation
quantities with success (e.g Arkin 1979; Griffith 1987; Huygen, 1989; Stout et al.
1979; Sorooshian 1997). Furthermore, the estimation of runoff values for flood
forecasting, irrigation and other purposes may be based on remote sensing (Hardy
et al. 1989, Rott et aI., 1986; Tiwari et al. 1991; Kite, 1991; Koren et al. 1994,
1995).
For more than 20 years now, different satellite systems have monitored the globe
and will continue to do so, collecting a large amount of data in different spectral
I. Papadakis and G.A. Schultz
positive deviation in the estimated mean annual flow, all structures would be overestimated. Dam height would be higher than necessary, the reservoir would very
seldom be filled up to the required water level. The consequence would be that the
structures necessary for water power production such as the power station, conduits etc. would be much larger in size than required. The power production would
be below the planned target, and some machines would then remain idle or work
with low efficiency. For the planned irrigation schemes and water supply networks
there would be insufficient water to irrigate and supply the whole area and the
channels and other structures would be of disproportionate size.
Often the design of water management systems in many countries in the developing world suffers from inadequate or non-available sets of hydrometeorological
data. If hydrometeorological records are available, they are either too short or, if
they are sufficiently long, they include numerous gaps or discontinuities which
make these useless for further analysis (Gyau-Boakye, 1993; Gyau-Boakye and
Schultz 1994). The consequence is that very often water projects in these countries have to be designed under circumstances which are insufficiently representative for the project area. Instead of calculation of design data based on detailed
knowledge of the hydrological behaviour of the area under consideration, data are
prepared based on formulas, the validity of which has been proven for entirely
different areas.
Different mathematical-statistical and deterministic methods have been developed in the past which enable the designer to estimate missing data, generate synthetic sequences and derive estimates for design purposes (Gyau-Boakye, 1993).
The success of these methods, however, depends primarily on the length of the
available data considered for the identification of the system in question since
computation of systems reliability depends on long time series.
The remote sensing technology provides very useful methods of monitoring and
identifying many hydrological parameters and variables which are necessary for
the estimation of the project reliability. With the aid of remote sensing techniques,
relevant hydrological data can be obtained in a short time, at periodic intervals and
covering extended areas. Satellite imagery has been used in the fields of meteorology, watershed management, hydrologic modelling etc. Mathematical techniques
in a multidimensional feature space allow an accurate land use classification as
well as estimation of vegetation indices (Su, 1996). Changes in land use can be
quantified on the basis of multi-spectral and multi-temporal satellite imagery. Also
the identification of the potential location of dams in river valleys can be supported by remote sensing. A number of scientists have already used digital spectral
data from meteorological geostationary satellite systems to estimate precipitation
quantities with success (e.g Arkin 1979; Griffith 1987; Huygen, 1989; Stout et al.
1979; Sorooshian 1997). Furthermore, the estimation of runoff values for flood
forecasting, irrigation and other purposes may be based on remote sensing (Hardy
et al. 1989, Rott et aI., 1986; Tiwari et al. 1991; Kite, 1991; Koren et al. 1994,
1995).
For more than 20 years now, different satellite systems have monitored the globe
and will continue to do so, collecting a large amount of data in different spectral