Total Pageviews

Tuesday, October 9, 2018

HadCrut4 Global Temperature, 1850 – 2018. ... Absurd errors everywhere in Hadley Met Centre surface temperature "data"

HadCrut4 Global 
Temperature Data, 
from 1850 to 2018.

HadCRUT4 is a global 
surface temperature 
dataset.

Here's what 
I've heard
"HadCRUT4"
stands for:

Highly 
 Unconvincing 
  Data 
   Confirming
    Unfeasible 
     Theory
      4

The HadCRUT4 dataset 
is a joint production 
of the UK Met Office’s 
Hadley Centre and the 
Climatic Research Unit 
of the University 
of East Anglia.

Just like other 
surface data,
from the US,
it shows 
more warming 
than alternative 
measurement 
methodologies -- 
 Weather satellites 
and weather balloons
(which are both very similar
to each other, and completely ignored 
by the government bureaucrats
with science degrees).

Had CRUT4 provides 
temperature anomalies, 
versus a base period, 
for surface "grids"
(5 degree latitude 
by 5 degree longitude
surface areas) 
across the world,
averages for 
the two hemispheres,
and an average 
global temperature.

CRUTEM4 and HadSST3 
are the land and ocean 
components of HadCRUT4.

Dr John McLean 
has now shown
the surface data 
are haphazard, 
not worthy of 
real scientists ...
( as I have written 
so many times 
in this blog ),
... with freak data, 
empty grids (fields), 
Fahrenheit temps 
recorded as Celsius, 
mistakes in longitude 
and latitude, and
brutal "adjustments" ! 

Temperature data
used by climate models 
has more than 70 
different problems.  

The most basic quality control 
checks have not been done.


Some examples 
of OBVIOUS errors
coming in from the

national meteorological
services 
(I don't know how many of these local errors,
if any, are caught and fixed before compilation 
of the HADcrut4 average global temperature:

(1)
One town in Columbia 
spent three months 
in 1978, at an average 
daily temperature 
of over 80 degrees C.
which is unbelievably hot ! 
( Celsius, not Fahrenheit ! )

For April, June and July 
of 1978 Apto Uto 
(Colombia, ID:800890)  
had an average 
monthly temperature 
of  81.5° C., 83.4° C. 
and 83.4° C., respectively.

Obviously wrong !



(2)
One town in Romania 
stepped out from summer 
in 1953, straight into a month 
of Spring at minus -46° C. !

The monthly 
mean temperature 
in September 1953 
at Paltinis, Romania 
is reported 
as -46.4° C. 
(in other years 
the September average 
was about 11.5° C.).

Obviously wrong !



These are supposed to be
“average” temperatures 
for a full month !



(3)
St Kitts, a Caribbean island, 
was recorded at 0° C. 
for a whole month ... twice !

At Golden Rock Airport, 
on the island of St Kitts
in the Caribbean, 
mean monthly temperatures 
for December in 1981 and 1984 
are reported as 0.0° C. 

Obviously wrong !

But from 1971 to 1990 
the average
in all the other years 
was 26.0° C.



(4)
Boats whose location
was dry land (?) 
allegedly recorded 
ocean temperatures 
from as far as 
100 kilometers inland !

Obviously wrong !



Dr. McLean 
audited 
the HadCrut4 
global data 
from 1850 onwards 
for his PhD thesis, 
and then continued 
afterwards,
until his audit
was complete:

“I was aghast to find 
that nothing was done 
to remove absurd values
… the whole approach 
to the dataset’s 
create is careless 
and amateur, 
about the standard 
of a first-year 
university student.”
– John McLean

McLean's supervisor 
was Peter Ridd, 
famously fired 
for saying that 
“the science was not
being checked, 
tested or replicated”, 
and for suggesting 
we might not be able
to trust our institutions.



Data are 
incredibly sparse:


For two years the entire 
Southern Hemisphere 
temperature was estimated 
from one land-based site 
in Indonesia, 
and some ship data. 

We didn’t get 
50% global coverage 
until 1906. 

We didn’t consistently get 
50% Southern Hemisphere 
coverage until about 1950.

McLean’s findings: 
Almost no quality control 
on these data. 


Countries include 
“Venezuala”,
” Hawaai”, and 
the “Republic of K” 
( also known as 
South Korea ). 

One country is 
“Unknown” 
while other "countries" 
are not even countries, 
such as “Alaska”.


Yet government bureaucrats
“sell” their temperature trends
 as if they are accurate !



Climatic Research Unit data 
covers 10,295 stations, 
but 2693 of them,
( over 26% ) 
don’t meet the criteria 
for inclusion, described
in Jones et al 2012, 
the best description 
of what should and 
shouldn’t be included.

It is impossible to know 
exactly which sites 
are included in the final 
temperature analysis, 
and whether 
a site’s records 
have been adjusted. 

The sub-parts of the datasets 
contradict each other. 

The land dataset 
and the sea dataset 
should combine to
equal the global set, 
but they don’t 
always match. 

Which one is right?



In probably the worst 
systematic error, 
the past is rewritten, 
"cooled", in a
haphazard attempt 
to correct for site moves. 

While some corrections 
are necessary, 
these adjustments 
are brutally sweeping. 

As the site “ages” 
over time,
buildings and roads 
get built nearby, 
and sometimes 
air conditioners too, 
all of them
artificially warming 
the nearby 
weather station. 

So a replacement thermometer 
is placed in a nearby open location, 
surrounded by grass. 

Usually each separate 
national meteorology center 
compares both sites 
for a while, and figures out 
the temperature difference 
between them. 

Then they adjust the readings 
from the old locations 
down to match the new ones. 

But, when a thermometer
is relocated to a new site, 
the adjustments assume 
that the old station site 
was always being “heated” 
by nearby concrete and buildings. 

In reality, the artificial warming 
from buildings and concrete,
probably crept in slowly
with economic growth. 

"Correcting" (cooling) 
1880 to 1950 data, 
for example,
for the extra heat
of a nearby building
built in 1950, 
causes the old 
temperature records, 
from 1880 through 1949,
to be artificially cooled.

If you cool the past, 
and not the current year,
the trend of global warming 
increases,
even if all the adjustments
were "cooling adjustments" !

When the past is rewritten ,
to be colder than it really was, 
that makes global warming 
look faster than it really was. 

Thousands of 
men and women 
trudged through 
snow, rain and mud 
to take temperatures that 
a computer “corrected” 
a century later.



In Australia 
data adjustments 
increased 
the warming trend 
by as much as 40%.

Ken Stewart found 
some adjustments 
to old historic data 
in Australia 
reduced the 
earliest temperatures
by as much as 
-2 degrees C. ! 



Each national 
meteorological bureau 
supplies “pre-adjusted” 
national data. 

The Hadley Centre 
just accepts them
with no audits, 
and no checking.

This systematic error
creates an artificial
warming trend 
from incorrect
"adjustments".




In May 1861 
the global coverage, 
according to the 
grid-system method 
that HadCRUT4 uses, 
was only 12% !

That means in 1861,
there were data for 88%
of the Earth’s surface, 
yet iot was used
for a “global average”.  

“Until 1906 
global coverage 
was less than 50% 
and coverage 
didn’t hit 75% 
until 1956”    
 John McLean


Note that "coverage"
only means the grid
has at least 
one thermometer
-- it does not mean
daily readings 
were made 
every day of a month,
or even every day 
of a week.


If you look at 
how much daily 
temperature data 
are provided,
even today in 2018,
a majority of grids 
in a typical month 
have no thermometers,
(so obviously no data),
or the grid has 
incomplete data.


That means 
the monthly 
average 
temperature 
for most grids
requires 
wild guessing
by government 
bureaucrats,
whose wild guesses 
can never be verified,
or falsified !

There are large areas
in central Africa,
which NASA (U.S.) claims
are having record heat,
yet there are 
no thermometers there,
meaning NASA 
just wild guessed
the temperatures there,
and declared 
their guesses
to be a "record" !



In 1850 and 1851, 
official data for the 
Southern Hemisphere 
only included one 
sole thermometer 
in Indonesia, 
and some 
random ship data. 
that covered 
only about 15% 
of the oceans 
in the southern half 
of the globe, 
( and “covers” 
may mean 
as little as 
one measurement
in one month 
in a grid cell, 
though it is 
usually more. ) 

Sometimes 
there are national data 
that could have been used, 
but were not. 




On February 6,1851, 
Australian 
newspaper archives 
show temperatures 
in the shade 
hit 117 F. in Melbourne, 
(that’s 47 C.),. 
115 F. in Warnambool, 
and 114 F. in Geelong. 

Really hot !

The Australian 
Bureau of Meteorology
argued these
were not standard 
officially sited 
thermometers, 
and didn't use the data.

But surely those 
unofficial 
temperatures 
would have been
more useful 
than the data 
from the one 
"officlal" 
1851 thermometer 
located in Indonesia, 
5,000 to 10,000 kilometers 
away? 

While the Hadley dataset 
is not explicitly estimating 
the 1851 Melbourne temperature,
they are estimating 
the average temperature of
“the Southern Hemisphere” 
and “The Globe” 
( and Melbourne
is a part of that. )




This article was based 
on an excellent summary 
of the McLean report at,
www.JoNova.com

... because this cheapskate
did not buy the 135-page audit 
with more than 70 findings 
that was being sold for $8 
from Robert Boyle Publishing.