Climate-Scale Oceanic Rainfall Based on Passive Microwave Radiometry - Satellite-based Applications on Climate Change - page 233

adjusting the difference in the T b for the pre- and post-boost T b s in the TMI channel
data (Shin and Chiu 2008 ) and by properly adjusting the T-R relation in the
algorithm by changing the earth incidence angle and redoing the radiative transfer
calculations (Chiu et al. 2010 ). During the transition period, another satellite, the
DMSP F13, is in stable operations. We use the METH product derived from the F13
as a calibration point and compare the differences for the pre- and post-boost
periods. The adjustment effectively eliminates the discontinuity introduced by the
TRMM boost.
15.5 Summary and Discussions
In this chapter, we discuss the theoretical bases of the METH technique, describe
the processing of the METH products, present the climatology of these parameters,
and discuss their relevance to climate studies. The uniqueness of this technique is
the determination of the background brightness temperature for the non-raining
portion and fitting the brightness temperature histogram to a mixed lognormal rain
rate distribution via a T-R relation derived from an atmospheric radiative transfer
model. The so-called beamfilling error is corrected using empirical data.
We briefly examine the characteristics of the rain rate parameters including the
unconditional rain rate, conditional rain rate, freezing level, and rain frequency.
These parameters are consistent with more recent and detail estimates, such as the
rain frequency computed from the CloudSat radar. Application to other microwave
sensors is rather straightforward and has been applied to TMI rather successfully.
The strength of this technique is well demonstrated in mitigating the discontinuity
of the TMI data record by simply changing the T-R relation in the algorithm.
We found no significant trend in the global (domain) average rainfall; however,
significant linear trends are detected in the equatorial belt 0-10ºN. Whether this
pattern is due to an intensification of the Hadley circulation or a shift of the rain
belts has yet to be determined. Two distinct modes of nonseasonal variations are
detected from an empirical orthogonal function analysis. The first mode is the well-
recognized ENSO mode, the associated time series of which show a correlation of
0.8 with a Southern Oscillation Index. The second mode is recognized as the ENSO
Modoki mode (or the central Pacific ENSO mode) and shows a correlation of 0.55
with an index of the ENSO Modoki. The ENSO Modoki mode leads the ENSO
mode by roughly 6 months.
This algorithm has been in operation for over 20 years and has served as an
important input to the Global Precipitation Climatology Project rain maps. With
improved understanding of the precipitation processes and the information col-
lected during major international missions, some of the crude physics and model
assumptions need to be revisited and improved so that uncertainties of climate-scale
rainfall can be better quantified and the data better utilized.