"Eric Swanson" wrote in message
...
In article ,

says...

On Thu, 20 Feb 2003 22:18:55 +0000, Daniel B. Wheeler wrote:

"Titan Point" wrote in
message
...
On Wed, 19 Feb 2003 10:31:25 +0000, Daniel B. Wheeler wrote:

[snip]
The debate is not about global warming but about anthropogenic
global
warming.
According to Warren Washington, that is not the debate. You are
welcome
to disagree, of course. But the increased temperature globally
within
the last 30 years strongly points to some particular effect which
has
taken place or has accumulated within that time frame.

Funnily, satellite records calibrated with randio-sonde measurements
have
found practically no increase in lower atmospheric temperature since
1979.
The increase of the last 30 years is nearly lost in the statistical
noise.
By no means is the range of temperature rise over the last century,
unusual or unprecedented.

Funny, that's incorrect.

S & C did not "calibrate" their analysis against the sonde data. They
took
some sonde data, ran it thru a simulation to create a synthetic MSU
signal
and ran that thru their algorithm. It's another simulation, which
those
who don't like simulations should find disgusting.

It was calibrated to a model of theoretical emissions using the known
radiosonde data. The problem is that you then add a dependency to the
radiosonde data which is hardly the best source of absolute calibration
standards.

"Several independent estimates of the surface to lower tropospheric
trends based on 20 radiosonde temperatures have yielded quite
different results. For example, since 1958 Brown et al. (2000) found a
warming of about 0.20°C/decade and Lazante et al. (2003)
obtained about 0.15 and 0.00°C/decade for adjusted and unadjusted data
(respectively), and Gaffen et al. (2000) found a warming of
0.08°C/decade. Radiosonde data indicate that the rates of surface,
tropospheric, and stratospheric temperatures have evolved in a
complex way over the past 40 years (Figs. 7 and 8). Gaffen et al. (2000)
showed that the tropical troposphere warmed relative to the
surface over 1960-1978, and thereafter cooled relative to the surface.
Others (Lanzante et al. ;Brown et al. 2000; Hegerl and
Wallace, 2002) noted similar multi-decadal changes in lapse rate. Over
the period 1959-1998, Angell (2000) and Lazante et al. (2002)
found no discrepancy between the overall warming rates at the surface
and in lower troposphere. These results illustrate that
lapse-rate changes may not easily be generalized to other periods. The
tropospheric/surface temperature trend difference varies from
one dataset to another, but most analyses show that there was a step
change around the 1976-77 time period, just prior to the start
of the satellite record. "


There are now three sets of results from the MSU data, 2 of which
produce
a warming trend similar to that seen in the surface record. The third
by
Spencer and Christy (to which you presumably refer) now also indicates
a
warming trend. You have consistently ignored the other two.

Also the problems with the whole MSU dataset.


BTW, the MSU does not measure "temperature". It measures microwave
intensity.

With the problems of dirunal drift, water vapor contamination, satellite
coverage, isolation of the signal and removal of stratosphere cooling,
and..


So whatever Washington thinks about the record of climate change in
the
last 30 years appears to be entirely removed from what was actually
recorded.

You ignore the effects of the Great Salinity anomalie, which may have
contributed
to the colder conditions of the late 1960's and 1970's.

That might explain the 'shift' in the radiosonde trends in 1976 noted
above?


The climate has warmed about 0.6C since 1900. There is no dispute
about
this from anyone. The question is whether the increase in carbon
dioxide - over and above that which would be expected to be
produced by
a warming climate - is significant to force the climate warmer in
a
significant AND biologically negative way.
The temperature has increased nearly a full degree Fahrenheit since
accurate temperatures have been kept, about 1970 I believe. While I
don't know about your figure of .6C since 1900, the increase of 1
degree
F. since 1970 is, I believe, well documented. And that figure was
assessed in 1990, after only 20 years of data. CO2 concentrations
were
likewise also observed to increase.

The satellite record shows a decadal increase of (Spencer and Cristy)
0.075C/decade or 0.225C over the last 30 years. It is also misleading
to
begin your measurements in a known cold period and then act surprised
when
you show a warming trend. That is called end-date distortion.

Funny, your own words refute your claim above of practically no
warming.

For one who wants us to believe that natural variability is the cause
of
all the changes in climate, you must also consider the fact that the
MSU
record begins near the peak of a sunspot cycle. Thus, if the solar
variability does have an impact, it should also be seen in the
satellite
record. Start-date distortion is just as much a problem as end-date
distortion.

Also, S & C's claims of little change in temperature were based upon
results
calculated up to 1997 and have since been changed as the result of the
discovery of several errors in the analysis, such as corrections for
orbital
decay and shift in time of day of Equitorial crossing.

"The existing dilemma can be traced to the pioneering work of Spencer
and Christy (1990). They compiled a record of
satellite-derived tropospheric temperatures and noted that the rate of
warming from the satellite record was negligible, especially
when compared to surface temperature increases. This curious result was
most prominent in the tropical and subtropical regions
(Fig.1). Spencer and Christy (hereafter referred to as the University of
Alabama at Huntsville Team --- UAH) took advantage of the
Microwave Sounding Unit (MSU) aboard NOAA polar orbiters. The MSU data
provided an all-weather integrated measure of temperature for
various atmospheric layers. There were four MSU channels, and their
weighting functions are depicted in Fig. 2 for channels 2 and 4,
for both the MSU and the Advanced MSU (AMSU) instrument. These two
channels have been used to depict both tropospheric and
stratospheric temperature trends. UAH derived a synthetic channel called
MSU2LT which was formed by differencing view-angles of MSU2
and AMSU5. For the lowest layers, the microwave data are affected by
changes in surface emissivity, but above the surface layer this
is not a factor. Hence some of the MSU2LT signal comes directly from
surface emissions: about 10% over ocean and 20% over land. Over
land the radiances are affected by changes in moisture, and over
mountains even more signal comes from the surface, so MSU2LT
values, although they are weighted for the lower half of the troposphere
are subject to increased noise over mountainous terrain,
including the Himalayas, Greenland and Antarctica. However, UAH also
compute a MSU2 temperature, which reflects the mid and upper
troposphere although some stratospheric emissions are included. "

"Trends in channel 4 have been produced and show the marked cooling in
the lowerstratosphere that have been linked to ozone
depletion and increases in greenhouse gases (IPCC, 2001). Such trends
influence MSU2, so examination of these other channels is also
needed to provide confidence that the ozone signal is being properly
simulated. It would also provide evidence as to whether the
volcanic signal (especially the aerosol signal of stratospheric warming)
is correctly included in models. These trends are quite
large relative to any observational uncertainty, but understanding the
causes of these trends is linked to understanding
tropospheric and surface trends. "

"Several other issues emerge, however, when assembling a homogenous
record of tropospheric temperatures. Each of the eleven
satellites used to measure temperature since 1979, has unique, local
sampling times that have changed over its lifetimes. This
introduces a diurnal temperature bias, even for temperatures well above
the surface layer, and substantial adjustments are required
to homogenize the data (Fig. 3). Moreover, each sounding unit on the
various satellites has some calibration uncertainties that had
to be assessed. The errors in the calibration of MSU have been addressed
through numerous analyses (Mo et al., 2001; Christy et al.,
2002; Wentz et al., 2001; Mears et al., 2002)."

"Wentz et al. (1998) found that in addition to biases related to
calibration and changes in diurnal sampling there were biases
introduced into the data due to decays in satellite orbit. Episodic
solar wind events reduce orbit altitude, and this significantly
affects the tropospheric temperature trends, especially those of the low
to mid troposphere. This correction was applied by the
Wentz team (hereafter referred to as Remote Sensing Systems --- RSS) in
their satellite derived temperature trends and also applied
by UAH in the latest versions (D and 5) of their data sets. Since 1995,
UAH have at different times calculated the 95% confidence
interval of the decadal trend of tropospheric temperature. These
estimates range from 0.03°C/decade to 0.06°C/decade, and the UAH
temperature trends during 1979-2001 for the two tropospheric layers
MSU2LT (surface to about 400hPa) and MSU2 (approximately Sfc to
100hPa) are 0.06 and 0.01°C/decade, respectively. In contrast, MSU2 as
processed by RSS finds a warming of approximately
0.10°C/decade. "

"The difficulty of adequately resolving the temperature trend issue is
attested to by the number of revisions to the original data
set of UAH and the magnitude of the corrections to the original data
that are required (Fig. 4). UAH have just issued version 5 of
their MSU temperature record (5 revisions over a 13-year period). RSS,
in work submitted for publication, has just released version
1. The rate of warming estimated by from these two science teams is
significantly different, despite the dedication of each team to
produce the most accurate temperature time series possible. There are at
least two differences in the techniques applied to generate
the respective time series. First, to correct diurnal drift errors, UAH
rely on adjustments derived from satellite measurements
themselves made at the appropriate diurnal time slots while RSS employs
information from a high-resolution climate model simulation
(Fig.3) to apply appropriate corrections. Secondly, the instrument
calibration adjustments depend upon temperature differences
observed by two satellites simultaneously. RSS use all periods of
simultaneous observations while UAH set thresholds to eliminate
some overlaps e.g., minimum of one- year overlap and a minimum level of
error reduction during the overlap. The resulting
calibration adjustments for each instrument are quite similar between
the two techniques except for one satellite, NOAA-9 which had
short overlaps with other satellites (Fig. 5). The difference in the
adjustment for NOAA 9 accounts for about 65% of the total
difference of the two MSU2 trends. Currently, RSS and UAH are sharing
data and computational algorithms to help explain the
difference. "

"UAH, in searching for independent methods to assess error statistics,
compared their satellite record with radiosonde (weather
balloon) instrumental data and with radiosonde- guided datasets such as
the global analyses produced by the National Centers for
Environmental Prediction. UAH contend that the similarity of temperature
trends between the radiosonde-based data and the
satellite-derived temperatures from UAH is important corroborative
evidence to help bolster the confidence in the tropospheric
trends produced by their team. In these UAH comparisons, controls were
established to eliminate stations with inhomogeneities.
However, in large compilations of radiosonde data, complications arise,
because the weather balloon data are also subject to time-
dependent biases because of numerous changes in instrumentation, site
location, proprietary calibrations and adjustments, and
ground-station processing methods. These changes are known to have
introduced significant time dependent biases in the temperature
record, most clearly visible in the stratosphere, and corrections are
not free of error. For example, Free et al. (2000) reported on
the results of several different research teams who attempted to adjust
the weather balloon data for time-dependent biases.
Inter-comparison of the various adjustments applied by the different
teams showed considerable disagreement among the teams related
to both the timing and magnitude of adjustments required (Fig.6) during
both the satellite and pre-satellite era. [Note: If two
teams identify a discontinuity at a station separated by as much as five
years within any pentad, this would be considered
agreement, at least in terms of calculating multi-decadal trends (Fig.
6)] Despite these problems, the trends from UAH compare
favorably with the radiosonde temperature trends. The greatest agreement
is for the lowest layer temperatures where radiosonde
inhomogeneities are smallest. Other investigators (Wentz et al., 2002;
Santer 2002) do not consider the UAH record to be completely
independent of the radiosonde record, particularly with regards to
decadal trends. Many of the radiosondes used by UAH are in the
temperate Northern Hemisphere, where the RSS and UAH results are quite
similar, although UAH's comparisons with the trends in the
tropics also show exceptional agreement."


According to measurements taken over the last century from properly
maintained, rural temperature stations, the warmest decade of the
20th
Century was the 1930s (for example 1934 saw fully one half of the
contiguous US in extreme drought, the "Dustbowl" years)

Your use of data for the U.S. ignores the fact that the "Dust Bowl"
was in large
part the result of cropping practices in the dry lands of the Great
Plains.
When drought hit, these poor soil conservation practices exacerbated
the problems.
Farms were abandoned and the bare soil without plant cover is much
hotter than
the same area would have been with plants/crops.

Not to mention that it confuses regional climate with global climate.


The very faint rise since 1970 is neither unprecendented nor
surprising.

The rise in carbon dioxide during the 20th Century appears to be a
confounding factor rather than cause of 20th Century warmth.

Most of the increase in atmospheric CO2 happened after WW II.
The time lags in the result make detecting the anthropogenic warming
very
difficult. Only during the last decade or so has the magnitude of the
AGW
approached and possibly exceeded the range of natural variability.

And the influence of a 'compound interest' factor such as the increase
in CO2 accumulates like interest in a fixed rate savings account.
Exponentially. This means that the start of the response is slow, but
build rapidly, and there is, as yet, no end in sight.