CONCLUSION:
Poor siting will bias the
temperature records
of a weather station.
It's disappointing the US
has such poor siting
at so many stations.
We were supposed to have
the best weather station
network in the world.
US surface stations are not
a large part of the global
average temperature.
That's why I'm
most worried about
all the surface
weather stations
outside the US,
and will seek some
information about them.
Most disappointing is that no
US government bureaucrats
ever bothered to check weather
station siting.
NOAA (US Commerce Department)
adjustments to raw data seem
to reduce the gap (different amounts
of warming) measured at high quality
siting stations versus low quality siting
stations -- I hope that was not just
from NOAA increasing the
warming rate at the high quality
stations, to match the higher
warming rate at low quality stations!
I say that because it's not common
in climate science that adjustments
reduce the global warming rate.
The US does have an advantage
compared with many other nations
-- we have a lot of weather stations.
A 5 degree latitude by 5 degree
longitude surface grid cell that
has many weather stations
within it won't be affected
as much by bad siting of
some of them.
A much bigger problem
is the surface grid cells
with no thermometers,
and those with missing data
from one or more weather
stations within the grid.
Over half the 2,592 global grid cells
include numbers filled in (guessed)
by government bureaucrats
-- and for those grid cells
that have no thermometers
at all, the average monthly
temperature for that grid cell
has to be a wild guess,
with no way to verify that guess.
DETAILS:
This important study
has not received
enough attention
in the past decade
since the initial
Watts, 2009 report
The mainstream media
is not interested.
Even skeptics may have
little interest because
potential US errors and
biases won't have much
of an effect on the global
average temperature.
Similar siting biases
have been identified,
casually, in other nations.
The US was supposed
to have the best
weather station network
-- look at the weather station
photographs I've included here.
The 2007 through 2009
Surface Stations project
revealed that many
of the US thermometer
weather station shelters,
used for calculating
temperature trends,
were located near
artificial heating sources.
The Surface Stations project,
led by the meteorologist and
blogger, Anthony Watts,
found that about 70%
of the weather stations
in the U.S. Historical
Climatology Network
are currently sited in locations
with artificial heating sources
less than 10 meters from the
thermometer, e.g., buildings,
concrete surfaces,
air conditioning units.
In 2007, Watts began wondering
how reliable weather station
records were.
He visited three of his local
weather stations, and was
shocked to discover improper
siting for two of them.
One thermometer
was surrounded by
heat-producing
radio electronics.
Another thermometer screen
(aka Stevenson screen = louvered wood box)
was only a few meters from
several different heat sources.
-- That screen was
beside an asphalt
parking lot, likely to
get very hot
on a sunny day.
-- It was also a few meters
from the exhaust fans
of air conditioning units
of a nearby building.
-- As Watts stood near
the thermometer, he could
feel warm exhaust air
from the nearby cell phone
tower equipment sheds
blowing past him !
Nobody ever checked
weather station siting.
Groups calculating
global warming trends
assumed the stations
were ok, without checking.
The U.S. weather stations
in the “Global Historical
Climatology Network”
were considered to be
the most reliable.
For the unadjusted dataset.
1890s-1930s: US warming
1930s-1970s: US cooling
1970s-2000s: US warming:
Mr. Watts focused on the
1,221 stations known as the
U.S. Historical Climatology
Network (or USHCN).
Watts inspected a large number
of stations, and recruited a team
of more than 650 volunteers
through his Surface Stations
website to inspect the rest.
By spring 2009, more than
860 stations had been inspected,
photographed and evaluated.
Thermometer stations
are supposed to be located
on a flat and horizontal grass
or low vegetation surface,
ideally more than 100 meters
from any artificial heat sources.
Watts and the Surface Stations
team found that only 1%
of the stations they surveyed
met these requirements !
The Surface Stations group
used the same rating system
that NOAA’s National Climatic
Data Center used when they were
setting up the
U.S. Climate Reference
Network (or USCRN)
-- "a high quality"
weather station network,
founded in 2003.
The National Climatic Data Center
maintains the Global Historical
Climatology Network and U.S.
Historical Climatology Network
datasets.
The preliminary Watts 2009 report
was challenged by NOAA’s National
Weather Service Forecast Office.
They did an independent
assessment of 276 stations
that the Surface Stations
team had rated.
The NOAA assessment
confirmed the findings
of the Surface Stations
project
( see Menne et al., 2010 )
Many studies assessed the results
of the Surface Stations survey,
but there was no agreement
among the studies
Watts, 2009 (non peer-reviewed)
https://wattsupwiththat.files.wordpress.com/2009/05/surfacestationsreport_spring09.pdf
https://wattsupwiththat.files.wordpress.com/2009/05/surfacestationsreport_spring09.pdf
Martinez et al., 2012
abstract only:
https://www.sciencedirect.com/science/article/pii/S0022169412004696?via%3Dihub
abstract only:
https://www.sciencedirect.com/science/article/pii/S0022169412004696?via%3Dihub
Watts et al., 2011
see Fall et al., 2011, above
see Fall et al., 2011, above
Connolly & Connolly, 2014
http://oprj.net/oprj-archive/climate-science/11/oprj-article-climate-science-11.pdf
http://oprj.net/oprj-archive/climate-science/11/oprj-article-climate-science-11.pdf
The Watts, 2009 report
speculated about
a warming bias in
U.S. temperature trends,
but it did not attempt
to quantify
what the net bias was.
There is disagreement
over what effect
(if any) the poor station
siting problem has had
on estimates of U.S.
temperature trends.
Menne et al., 2010
suggested that the
National Climatic Data Center’s
adjustments accounted for
biases, which poor siting
may have introduced.
Muller et al., 2011
found linear trends
of the unadjusted
records of
the best stations
(rated 1, 2 or 3)
were comparable
to the worst stations
(rated 4 or 5).
Martinez et al., 2012
analyzed only
the state of Florida
stations, and used the
fully adjusted dataset.
The 22 Historical
Climatology Network
stations had different
trends for subsets
of the worst rated (4 & 5)
and best rated (1 & 2)
over the two periods
they considered
(1895-2009 and
1970-2009).
From this they concluded that
station quality does influence
temperature trends.
agreed with
Menne et al., 2010 that the
National Climatic Data Center’s
homogenization adjustments
reduced much (although not all)
of the difference between
the good quality and poor quality
subsets in the fully adjusted datasets.
"Homogenizing” station
records means
adjusting each record
to better match
those of its neighbors.
After homogenization,
all weather station records
will have more similar trends,
i.e., the station records
are fairly “homogeneous”.
Homogenization spreads the
siting biases into those stations
which had actually been okay
before homogenization.
In 2011, Watts published
the "final" results of the survey:
Out of the 1,221 stations in the
U.S. Historical Climatology Network,
they had ratings for 1,007 (82.7%).
Station Ratings, as of 2011:
Excellent quality - Rating 1 (1%)
Good quality------- Rating 2 (7%)
Intermediate -----– Rating 3 (21%)
Poor quality ------– Rating 4 (64%)
Bad quality -------– Rating 5 (6%)
Poor siting will bias the
temperature records
of a weather station.
Some researchers
claim these biases
have been removed
by a series of artificial
“homogenization”
adjustments.
Not true:
Those adjustments
just spread the biases
uniformly among many
stations, rather than
actually removing them.
Global and surface
weather station records
account for about 30%
of Earth's surface.
It's likely US temperatures
in the 1930s were warmer
than any decade since then.
even with CO2 concentrations
not far from pre-industrial levels.
That's why high quality siting
weather stations show,
but not the low quality siting
stations.
Records have been "adjusted"
over several recent decades,
between 1980 and 2018
to cool the hot 1930s,
so they are no longer the
hottest US decade.
This has led people to conclude
that recent U.S. temperatures
have become the hottest
on record (since 1880).
Changing numbers 50 years
later is junk science -- people
knew how to read a thermometer
in the 1930s.
Bad siting quality weather stations
make the 1975 to 2000 warming seem unusual.
make the 1975 to 2000 warming seem unusual.
Good siting quality weather stations
make the 1975 to 2000 warming seem natural
make the 1975 to 2000 warming seem natural
-- part of similar mild cyclical variations,
between periods of warming
and periods of cooling.
Glenns Ferry, Idaho
weather station.:
-- On a cold day, a nearby
power transformer
can be considerably
warmer than its
surroundings.
Perry, Oklahoma
fire department
weather station
-- The concrete
building wall
facing the
weather station
is warmer than
its surroundings.
Urbana, Ohio
weather station:
-- Located at a
wastewater
treatment plant.
The thermometer
(labelled “MMTS”
in the photo, after
Minimum-Maximum
Thermometer System)
is placed over
a concrete surface,
rather than on
a grass lawn, and
surrounded by
several buildings.
It is also beside
an air conditioning unit,
a refrigeration unit and the
plant effluent entrance.
It is also beside
an exhaust fan,
although it was said
to have been disabled
for 4 years before
the photograph.
Fallon, Nevada
weather station
station -- a Good
quality station
(Rating 2).
It's located on fairly
flat ground with natural
vegetation and is
more than 30 meters
from concrete, asphalt,
or any other artificial
heat source.
Napa State Hospital,
California
weather station
-- a Poor quality station
(Rating 4)
it is less than 10 meters from:
The hospital building;
An asphalt drive-way;
Exhaust from air-conditioning unit.
Santa_Rosa_Davis
weather_station_2
weather_station_2
-- a Bad quality station
(Rating 5).
It is on the roof of a building !
