Total Pageviews

Wednesday, October 24, 2018

Land surface temperature data sets are all DIFFERENT = more like junk science than real science !

from:
Journal  of  Geographical 
Research  Atmospheres 

"Land surface 
air temperature data 
are considerably different 
among 
BEST-LAND, 
CRU-TEM4v, 
NASA-GISS, and 
NOAA-NCEI"

by Yuhan Rao
Shunlin Liang
Department of 
Geographical Sciences, 
University of Maryland 

and Yunyue Yu
Center for Satellite 
Applications and Research, 
NESDIS, NOAA, 
College Park

Several groups produce gridded 
land surface air temperature (LSAT) 
data sets using station measurements 
and infilling to assess climate change. 

The Intergovernmental 
Panel on Climate Change's 
Fifth Assessment Report 
suggests that estimated 
global and hemispheric 
mean LSAT trends of 
different data sets 
are consistent. 

Little attention is paid 
to comparisons of 
local and regional 
temperatures.


A study of four 
LSAT data sets 
included: 
-- Berkley Earth 
Surface Temperature 
       (BEST‐LAND), 

-- Climate Research Unit 
Temperature version 4 
       (CRU‐TEM4v), 

-- US National Aeronautics 
and Space Administration 
Goddard Institute 
for Space Studies
       (NASA‐GISS), and 

-- US National Oceanic and 
Atmospheric Administration 
National Center for 
Environmental Information 
      (NOAA‐NCEI). 

The mean 
LSAT temperature
anomalies are 
remarkably different, 
because of the 
data coverage 
differences, 
with the magnitude 
nearly 0.4°C 
for the global and 
Northern Hemisphere, 
and 0.6°C for the 
Southern Hemisphere. 

On the regional scale, 
northern high latitudes, 
southern middle
‐to‐high latitudes, 
and the equator, 
show the 
largest differences, 
nearly 0.8°C. 

At the local scale, 
four data sets 
show significant 
variations 
over 
South America, 
Africa, 
Maritime Continent, 
central Australia, and 
Antarctica, 
which leads to 
remarkable differences 
in local trend analysis. 

The differences 
are associated 
with the availability 
of weather stations 
and the use of infilling 
( infilling is 
wild guessing 
numbers 
when there are 
no weather stations
in a surface area grid, 
and when there are 
missing data,
from one or more 
weather stations
within a grid.)

Data sets using only 
station observations 
have large uncertainties 
over station‐sparse areas. 

Data uncertainty, 
caused by limited and 
unevenly distributed 
station observations, 
must be reduced.


The uncertainty in 
instrumental 
temperature trends
“0.097–0.305 degrees C.
per decade, 
for recent decades 
(i.e., 1981–2017)” 
is as large, 
or larger than ,
the alleged overall 
warming trend itself 
for this period.


“for some areas, 
different data sets 
produce conflicting results 
of whether warming exists”, 
due especially to variations 
in the use of “infilling techniques”
— adding artificial temperatures 
to areas where there are 
no real-world measurements.

One data set trend can be 
“nearly 90%” different 
than another data set trend !

One of the things 
few people understand 
about measurement 
is resolution. 

There is no way to 
statistically improve
the resolution 
of a measurement. 

To get an 
accurate average 
global temperature, 
it is necessary to take 
equally spaced 
measurements, 
across the ENTIRE Globe, 
24/7,  365 days a year.

Nothing remotely close
to that is being done.

Instead, 
the number of 
weather stations 
being used
for the global average,
has declined a lot 
since the 1960s.

And the "wild guess"
infilling used for 
these "missing stations", 
and for vast areas of Earth 
with few stations, 
are without credibility. 

Since 1979 the surface 
temperature data 
can be compared with
both weather satellite and 
weather balloon data.
which are very similar
to each other, 
but not similar to 
surface measurements 
plus infilling 
( surface data show
MORE warming than the 
two other methodologies ). 

So, of course, the smarmy
pro-warming, anti-science
governmeent bureaucrats,
ONLY use the surface 
measurements / infilling !

Surprise, surprise !