Global Precipitation Monitoring - Satellite-based Applications on Climate Change - page 80

over land and can only extend out a short distance over oceans. Satellites, therefore,
provide crucial information to fill in these huge data voids, especially over
unpopulated regions and oceans. By integrating all available satellite information
with surface measurements in a “seamless” manner, the best possible global
precipitation climatology can be assembled. This chapter will begin by briefly
describing the various satellite precipitation retrieval methods, present the current
state of combined satellite and surface rainfall techniques, show some examples of
such datasets to depict global precipitation patterns, and conclude by describing the
anticipated advances over the upcoming decade.
6.2 Satellite Precipitation Retrieval Methods
A number of different methods are used to retrieve rainfall from satellites and are
summarized in Table 6.1 . In general, the methods can be categorized into low Earth
orbiting (LEO) and geostationary Earth orbiting (GEO), as well as by their observ-
ing spectral ranges (visible, infrared, passive microwave, active microwave) or
“multispectral” (i.e., use of one or more of these individual spectrums). Some brief
background on the various retrieval techniques is described.
6.2.1 Visible and Infrared Methods
Visible (VIS) and infrared (IR) techniques were the first satellite methods to be
developed and are rather simple to apply (Lovejoy and Austin 1979 ). However,
these techniques typically show a relatively low degree of accuracy. On the other
hand, GEO weather satellite VIS and IR imagers uniquely provide the rapid
temporal update cycle (e.g., 30 min or less) needed to capture the growth and
decay of precipitating clouds.
A complete overview of the early work and physical premises of VIS and
thermal IR (10.5-12.5 μ m) techniques is provided by Barrett and Martin ( 1981 ),
while Kidder and Vonder Haar ( 1995 ) present some of the more recent results. The
rainfall retrieval in these wavelengths is based on the fact that bright (optically
thick) clouds are positively correlated with regions of convective rainfall (Woodley
and Sancho 1971 ). On the other hand, clouds with cold tops in the IR imagery
produce more rainfall than those with warmer tops (Scofield 1987 ). However, the
correspondence between cold tops and visible bright spots is far from perfect and is
not always well correlated with surface rainfall (especially in stratiform rainfall
regimes).
Various approaches have been developed to stress particular aspects of the
sensing of cloud physics properties to settle differences between VIS and IR
retrievals and measured rainfall. The methods are sometimes classified as cloud
indexing (e.g., Arkin and Meisner 1987 ), bi-spectral schemes (e.g., Lovejoy and
Austin 1979 ), life history (e.g., Griffith et al. 1978 ), and cloud model-based (e.g.,
Adler and Negri 1988 ). More recently, multispectral (i.e., VIS and IR combined)