18 Computation of Hydrological Data
403
bands. These data are stored and are available for further analysis at any time. On
the basis of the capabilities of remote sensing briefly mentioned above and taking
into consideration the availability of long-term satellite data, a method is presented
here by which historical hydrological data, particularly rainfall and runoff values
can be generated on the basis of multi-spectral and multi-temporal satellite imagery for the period when satellite information is available.
18.2 General Approach
Planning and design of water management systems (e.g. multipurpose reservoirs)
with small risk of failure requires long-term series of hydrometeorological data. In
most countries of the developing world observed hydrological data contain either
many gaps or are not available at all. Thus hydrologists are urgently seeking new
ways of augmenting their conventional data supplies. Satellite remote sensing is
being explored as one possible answer to the data acquisition problem.
The general approach presented here is as follows (Papadakis 1994): At the beginning of the project planning after the feasibility study, if hydrological data are
not available, a hydrological network has to be installed which collects hydrometeorological data over usually two to three years. Thus on the basis of the collected
data and the information obtained from satellite imagery over the same period of
time, a mathematical model can be developed which connects the observed hydrometeorological data with data obtained from satellite imagery. The parameters
of the mathematical model can be calibrated on the basis of simultaneous satellite
data and ground truth. After the calibration it becomes possible to reconstruct
historical river flows with the aid of the mathematical model on the basis of the
satellite data alone for the period of time for which relevant satellite information
exists. This way the very short series of hydrological data (collected during the
planning period) can be extended considerably into the past, thus allowing an
estimate of the future performance of the water project. It is expected that the
technique presented here will become a valuable planning aid for design with
short-term series of hydrological data. Figure 18.1 gives an indication of the basic
principles of the method.
The approach presented here consists of three consecutive MODULS (Fig.
18.2).
IS.2.1 MODUL I: Satellite system, data processing
1. Satellite system
For the choice of a satellite system suitable for the problem on hand, the following
criteria should be considered:
Time resolution of the satellite system. On the one hand, for high accuracy a
good resolution in time is required, while on the other hand it may not be necessary to process too many data since, for most planning purposes, monthly data are
sufficient. Geostationary satellites such as Meteosat and GOES-W/E etc. have a
403
bands. These data are stored and are available for further analysis at any time. On
the basis of the capabilities of remote sensing briefly mentioned above and taking
into consideration the availability of long-term satellite data, a method is presented
here by which historical hydrological data, particularly rainfall and runoff values
can be generated on the basis of multi-spectral and multi-temporal satellite imagery for the period when satellite information is available.
18.2 General Approach
Planning and design of water management systems (e.g. multipurpose reservoirs)
with small risk of failure requires long-term series of hydrometeorological data. In
most countries of the developing world observed hydrological data contain either
many gaps or are not available at all. Thus hydrologists are urgently seeking new
ways of augmenting their conventional data supplies. Satellite remote sensing is
being explored as one possible answer to the data acquisition problem.
The general approach presented here is as follows (Papadakis 1994): At the beginning of the project planning after the feasibility study, if hydrological data are
not available, a hydrological network has to be installed which collects hydrometeorological data over usually two to three years. Thus on the basis of the collected
data and the information obtained from satellite imagery over the same period of
time, a mathematical model can be developed which connects the observed hydrometeorological data with data obtained from satellite imagery. The parameters
of the mathematical model can be calibrated on the basis of simultaneous satellite
data and ground truth. After the calibration it becomes possible to reconstruct
historical river flows with the aid of the mathematical model on the basis of the
satellite data alone for the period of time for which relevant satellite information
exists. This way the very short series of hydrological data (collected during the
planning period) can be extended considerably into the past, thus allowing an
estimate of the future performance of the water project. It is expected that the
technique presented here will become a valuable planning aid for design with
short-term series of hydrological data. Figure 18.1 gives an indication of the basic
principles of the method.
The approach presented here consists of three consecutive MODULS (Fig.
18.2).
IS.2.1 MODUL I: Satellite system, data processing
1. Satellite system
For the choice of a satellite system suitable for the problem on hand, the following
criteria should be considered:
Time resolution of the satellite system. On the one hand, for high accuracy a
good resolution in time is required, while on the other hand it may not be necessary to process too many data since, for most planning purposes, monthly data are
sufficient. Geostationary satellites such as Meteosat and GOES-W/E etc. have a