114
e.G. Collier
Whilst weather radar does offer a means of making wide area measurements over
specific important land and coastal areas, it clearly does not offer a practical method
of making such measurements over the ocean, or over very large remote river
catchments. The use of satellite data has consequently been investigated for this
purpose ever since meteorological satellites have been flown.
6.2.2 Use of visible and infrared satellite data
Initially only visible and infrared channels were available but over recent years there
has been increasing use made of passive (emitted from the Earth and clouds) microwave channels.
VisiblelInfra-red techniques derive qualitative or quantitative estimates of rainfall
from satellite imagery through indirect relationships between solar radiance reflected
by clouds (cloud brightness temperatures) and precipitation. A number of methods
have been developed and tested during the past fifteen years with a measured degree
of success.
There are two basic approaches, namely the 'life-history' and the 'cloud-indexing'
techniques. The first type makes use of data from geostationary satellites which
produce images usually every half hour. It has been mostly applied to convective
systems. The second type, also based on cloud classification, does not require a series
of consecutive observations of the same cloud system.
It must be noted, however, that up to now none of these techniques has been shown
to be 'transportable'. In other words, relationships derived for a given region and a
given time period may not be valid for a different location and/or season. Other
problems include difficulties in defining rain/no rain boundaries and inability to cope
with the rainfall pasterns at the meso or local scales. Scientists working in this field
are perfectly aware of these problems, and this is why it is current practice to speak
of the derivation of 'Precipitation Indices' rather than rain rates.
6.2.3 Use of passive microwave satellite data
VIS-IR measurements represent observations of the upper surface of clouds only. In
contrast it is often believed that microwave radiation is not affected by the presence
of clouds. This statement is not generally true. Its degree of validity varies with the
microwave frequency used as well as with the type of cloud being observed.
One major difference between infra-red and microwave radiation is the fact that
whilst the ocean surface emissivity is nearly equal to one in the infra-red, its value
(although variable) is much smaller in the microwave region (from 5 to 200 Ghz
here). Therefore, the background brightness temperature (TBB) of the ocean surface
appears much colder in the microwave. Over land the emissivity is close to one, but
varies greatly depending on the soil moisture.
As far as microwaves are concerned several different effects are associated with the
presence of clouds. They are highly frequency dependent. Absorption by cloud
droplets is roughly proportional to the square of the frequency. A 2 km thick (nonprecipitating) cumulus cloud transparent at 18 Ghz may be totally opaque (through
absorption) at 160 Ghz, the resulting effect being an increase of brightness tempera-
e.G. Collier
Whilst weather radar does offer a means of making wide area measurements over
specific important land and coastal areas, it clearly does not offer a practical method
of making such measurements over the ocean, or over very large remote river
catchments. The use of satellite data has consequently been investigated for this
purpose ever since meteorological satellites have been flown.
6.2.2 Use of visible and infrared satellite data
Initially only visible and infrared channels were available but over recent years there
has been increasing use made of passive (emitted from the Earth and clouds) microwave channels.
VisiblelInfra-red techniques derive qualitative or quantitative estimates of rainfall
from satellite imagery through indirect relationships between solar radiance reflected
by clouds (cloud brightness temperatures) and precipitation. A number of methods
have been developed and tested during the past fifteen years with a measured degree
of success.
There are two basic approaches, namely the 'life-history' and the 'cloud-indexing'
techniques. The first type makes use of data from geostationary satellites which
produce images usually every half hour. It has been mostly applied to convective
systems. The second type, also based on cloud classification, does not require a series
of consecutive observations of the same cloud system.
It must be noted, however, that up to now none of these techniques has been shown
to be 'transportable'. In other words, relationships derived for a given region and a
given time period may not be valid for a different location and/or season. Other
problems include difficulties in defining rain/no rain boundaries and inability to cope
with the rainfall pasterns at the meso or local scales. Scientists working in this field
are perfectly aware of these problems, and this is why it is current practice to speak
of the derivation of 'Precipitation Indices' rather than rain rates.
6.2.3 Use of passive microwave satellite data
VIS-IR measurements represent observations of the upper surface of clouds only. In
contrast it is often believed that microwave radiation is not affected by the presence
of clouds. This statement is not generally true. Its degree of validity varies with the
microwave frequency used as well as with the type of cloud being observed.
One major difference between infra-red and microwave radiation is the fact that
whilst the ocean surface emissivity is nearly equal to one in the infra-red, its value
(although variable) is much smaller in the microwave region (from 5 to 200 Ghz
here). Therefore, the background brightness temperature (TBB) of the ocean surface
appears much colder in the microwave. Over land the emissivity is close to one, but
varies greatly depending on the soil moisture.
As far as microwaves are concerned several different effects are associated with the
presence of clouds. They are highly frequency dependent. Absorption by cloud
droplets is roughly proportional to the square of the frequency. A 2 km thick (nonprecipitating) cumulus cloud transparent at 18 Ghz may be totally opaque (through
absorption) at 160 Ghz, the resulting effect being an increase of brightness tempera-