360
G.A. Schultz
16.2.2 Remote Sensing Data, Types and Acquisition
In this chapter remote sensing data will be used for (a) the estimation of rainfall
runoff water parameters and (b) as basis for the computation of model input variables.
As discussed later (Sects. 16.2.3 and 16.2.4) the transformation of remote sensing data into rainfall and runoff requires knowledge of various catchment characteristics. Among these landuse and landcover play an important role since they
have - among other factors - an important influence on soil water storage capacity.
The relevant landuse and landcover information is obtained with the aid of remote
sensing data for the whole catchment area with a high resolution in space. Landuse
classification may use air photography or satellite imagery. Here particularly data
from the Landsat and SPOT satellites are of importance. Relevant techniques for
land cover and landuse classification are discussed in Chap. 7. In Sect. 16.2.3
relevant rainfall runoff model parameters will be estimated with the aid of Landsat
satellite imagery which forms the basis for landuse classification required for the
determination of soil water storage capacity.
Model input data for flood forecasting models may be derived from various remote sensing sources. Precipitation data with a high resolution in time and space
may be obtained from groundbased weather radar. Also satellite imagery may be
used for rainfall estimation and forecasting, e.g. data from geostationary satellites
like Meteosat, GOES, INSAT or GMS. These satellites have an adequate resolution in time (30 minutes) and a reasonable resolution in space (5 km x 5 km). If
the catchments considered are not too small, these data are valuable for runoff
estimation (see e.g. Chap. 19). Otherwise this information is not adequate and
other data (e.g. groundbased precipitation measurements, or - if available -
weather radar data) have to be used. Also snow data and derived from those
snowmelt runoff can be determined with the aid of satellite imagery. This topic is
dealt with in Chap. Ii.
For real-time flood forecasting purposes groundbased weather radar is highly
adequate, since the information has not only an adequate resolution in time and
space, but it is also available at one point (e.g. Met office with radar) for a large
region. For detailed discussion of the use of weather radar for precipitation estimation see Chap. 6. Satellite information like Meteosat data can also be acquired
in real-time and are thus available for real-time flood forecasts although their
accuracy is lower than that of radar.
Other remote sensing data, e.g. Landsat or SPOT can usually not be acquired in
real-time. This is, however, not necessary, since these data are usually needed only
for the determination of model parameters which do not change too much with
time.
16.2.3 Determination of Hydro-meteorological Information from Remote
Sensing Data
Model Parameters. In Sect. 16.2.4 the structure of a deterministic hydrological
model of the distributed system type will be presented, which transforms precipi-
G.A. Schultz
16.2.2 Remote Sensing Data, Types and Acquisition
In this chapter remote sensing data will be used for (a) the estimation of rainfall
runoff water parameters and (b) as basis for the computation of model input variables.
As discussed later (Sects. 16.2.3 and 16.2.4) the transformation of remote sensing data into rainfall and runoff requires knowledge of various catchment characteristics. Among these landuse and landcover play an important role since they
have - among other factors - an important influence on soil water storage capacity.
The relevant landuse and landcover information is obtained with the aid of remote
sensing data for the whole catchment area with a high resolution in space. Landuse
classification may use air photography or satellite imagery. Here particularly data
from the Landsat and SPOT satellites are of importance. Relevant techniques for
land cover and landuse classification are discussed in Chap. 7. In Sect. 16.2.3
relevant rainfall runoff model parameters will be estimated with the aid of Landsat
satellite imagery which forms the basis for landuse classification required for the
determination of soil water storage capacity.
Model input data for flood forecasting models may be derived from various remote sensing sources. Precipitation data with a high resolution in time and space
may be obtained from groundbased weather radar. Also satellite imagery may be
used for rainfall estimation and forecasting, e.g. data from geostationary satellites
like Meteosat, GOES, INSAT or GMS. These satellites have an adequate resolution in time (30 minutes) and a reasonable resolution in space (5 km x 5 km). If
the catchments considered are not too small, these data are valuable for runoff
estimation (see e.g. Chap. 19). Otherwise this information is not adequate and
other data (e.g. groundbased precipitation measurements, or - if available -
weather radar data) have to be used. Also snow data and derived from those
snowmelt runoff can be determined with the aid of satellite imagery. This topic is
dealt with in Chap. Ii.
For real-time flood forecasting purposes groundbased weather radar is highly
adequate, since the information has not only an adequate resolution in time and
space, but it is also available at one point (e.g. Met office with radar) for a large
region. For detailed discussion of the use of weather radar for precipitation estimation see Chap. 6. Satellite information like Meteosat data can also be acquired
in real-time and are thus available for real-time flood forecasts although their
accuracy is lower than that of radar.
Other remote sensing data, e.g. Landsat or SPOT can usually not be acquired in
real-time. This is, however, not necessary, since these data are usually needed only
for the determination of model parameters which do not change too much with
time.
16.2.3 Determination of Hydro-meteorological Information from Remote
Sensing Data
Model Parameters. In Sect. 16.2.4 the structure of a deterministic hydrological
model of the distributed system type will be presented, which transforms precipi-