An Introduction to Satellite-Based Applications and Research for Understanding Climate Change - Satellite-based Applications on Climate Change - page 8

Global climate change signals as small as a few percent per decade critically
depend on accurately calibrated level 1B (L1B) data and the derived Fundamental
Climate Data Records (FCDRs). Detecting small climate changes over decades is
a major challenge and also impacts the retrieval of geophysical parameters from
satellite observations. Without dependable FCDRs and their derivative Thematic
Climate Data Records (TCDRs), the trends calculated from the measurements
will be questioned. Cao et al. ( 2008 ) analyzed the consistency of calibrated
reflectance from the operational L1B data between AVHRR on NOAA-16 and
NOAA-17 and between NOAA-16/AVHRR and Aqua/MODIS, based on the
recent Simultaneous Nadir Overpass (SNO) observation time series. The SNO
approach advanced the science of satellite calibration to a higher level of accuracy
and reliability and now includes the intercalibration between polar and geosta-
tionary measurements. Even so, the measurement uncertainty is still too high
relative to the trends being monitored. As a consequence, a more stable calibra-
tion source has been sought: the Moon. The Moon is thought to be a reliable and
stable calibration reference for studying climate change from satellites (Cao et al.
2009 ). However, having a quality FCDR does not guarantee the same or equiva-
lent quality TCDR or derived trend.
To calculate confident trends, Zou et al. ( 2009 ) developed a calibrated data set
based on the SNO approach for the Microwave Sounding Units (MSU) on NOAA
satellites 10 through 14 over the period from 1987 to 2006. This intercalibrated data
set reduced intersatellite biases by an order of magnitude compared to prelaunch
calibration and resulted in a well-merged time series for the MSU channels 2, 3, and
4, which represent the deep layer temperature of the mid-troposphere (T2), tropo-
pause (T3), and the lower stratosphere (T4). From Zou et al.'s ( 2009 ) data set, the
trend patterns revealed the tropical mid-troposphere warmed at a rate of 0.28

0.19 K per decade, while the Arctic atmosphere warmed two to three times faster
than the global average. Even with this improved trend calculation, there is appre-
ciable regional variability not demonstrated in this single number.
Liu and Weng ( 2009 ) also reported findings about the warming trend in the
troposphere and the cooling trend in the stratosphere. However, Liu and Weng's
( 2009 ) analysis presents evidence that the lower stratosphere has warmed slightly
since 1996 and the warming trend in the lower stratosphere may be related to a
possible recovery of stratospheric ozone concentration. This points out that even
with highly calibrated data, the debate over climate trends will likely change from
data quality to one of improving our understanding of the dynamic effects. In this
regard, Qin et al. ( 2012 ) analyzed MSU brightness temperatures to estimate the
global climate trend in the troposphere and stratosphere using a new adaptive and
temporally local data analysis method - Ensemble Empirical Mode Decomposition
(EEMD). Using EEMD, a nonstationary time series is decomposed into a sequence
of amplitude-frequency-modulated oscillatory components and a time-varying
trend. The data from the NOAA-15 satellite over the time period from October
26, 1998 to August 7, 2010 shows that most trends derived from microwave
channels are nonlinear in the Northern Hemisphere with a few channel exceptions.
Although the decadal trend variation of the global average brightness temperature is
no more than 0.2 K, the regional decadal trend variation could be different by plus
or minus 3 K in the high latitudes and over high terrain.