How not to measure temperature

I’m surveying climate stations of record around California and documenting their condition as part of a larger project I’m doing. You’ll see more about it here in the near future.

Today I visited Marysville’s Fire Station, just off Hwy 70 at 9th and B Street, where they have the station of record for the city using the MMTS electronic sensor installed by the National Weather Service. The data from this station is part of the USHCN (US Historical Climatological Network) and is used in the computer modeling used to predict climate change.

The Marysville station is located behind the fire department building on a patio and is probably the worst site visited so far. In addition to the sensor being surrounded by asphalt and concrete, its also within 10 feet of buildings, and within 8 feet of a large metal cell tower that could be felt reflecting sunlight/heat. And worst of all, air conditioning units on the cell tower electronics buildings vent warm air within 10 feet of the sensor. Oh and lets not forget the portable BBQ the firefighters use a “couple times a week.” The area has been constantly added to, what was once a grass rear yard was turned to a parking lot, then more buildings added, then a cell tower with one, then two electronics buildings and the air conditioners…no report on how long the firefighters were BBQ’ing back there, when they figured out why I was asking all the questions they clammed up.

I can tell you with certainty, the temperature data from this station is useless. Look at the pictures to see why, and is it any wonder the trend for temperature is upward?

Above: Vehicles with hot radiators park within 6 feet of the temperature sensor!

Now compare Marysville to Orland, just 50 miles away, where there’s not been any significant change in the last 100 years at the measuring location. Its obvious that Marysville is measuring UHI (Urban Heat Island) effects.

So the question is, how does bad data like this slip into the NOAA and NASA GISS climate database?

0 0 votes
Article Rating
Newest Most Voted
Inline Feedbacks
View all comments
May 26, 2007 9:48 am

Anthony, nice work. The two graphs tell the whole story, and ther photos clinch it, Marysville should not be in the historical data base. Now, the question is across the county, how many other sites are in similar condition. I hope others will take up the challenge to visit all the sites, in all states, and do an assesment.

May 26, 2007 12:14 pm

Thank you for this research.
Great photos, I look forward to seeing more.

Steve McIntyre
May 26, 2007 1:22 pm

This is a very interesting post, although pictures are pretty sickening. Maybe the temperature gurus should spend a little time on quality control. I’ve cross-posted on this with a link here.
BTW GISS also has adjusted versions of both Marysville and Orland. It might be worthwhile for you to email GISS and ask them how they specifically made adjustments at this site. They probably won’t tell you, but you never know.

David Walton
May 26, 2007 1:31 pm

If problems similar to the Marysville Fire Station (or other factors) are typical of a significant number of sites used to collect surface temperature data supporting global warming then, indeed, the current cycle of global warming is — without question — completely anthropogenic.

May 26, 2007 1:38 pm

Anthony: You are on the front page above the fold at Climate Audit.

David Walton
May 26, 2007 2:04 pm

Senator James Inhofe addresses the U.S. Senate Committee on Public Works and Environment about Global Warming —
Inhofe Statement

May 26, 2007 10:31 pm

Dear Anthony, You ask, “So the question is, how does bad data like this slip into the NASA GISS model database?”
First Anthony, data like this did not “slip into the NASA GISS model database”, it has been driven in by the truckload.
GISS are not the worst offenders and their effort arrived late on the scene, after the GHCN data was assembled in the 90’s.
The best documentation of how global T databases were first assembled is by the Jones et al team in 1986, reprinted in 1991. Unfortunately these books are no longer in print but your DoE should rectify that. They had the Jones/CRU team contracted for years. My pages at;
try to describe the critical sequence.
See link to;
which has scanned copies of Fred Woods critique of Jones et al 1986 and the Wigley and Jones reply. I have a Table setting out replies to 9 points where W & J said Wood was in error. Wood’s paper is the only critique I am aware of in print.
Form your own view.
Also see my Table of Southern Hemisphere stations which shows 40% of Jones et al stations are cities plus 50,000 population, even using the out of date, wrong GHCN population numbers.
Scroll down that Table, tell me what Southern Hemisphere cities are NOT there.
So to recap Anthony, it is not that small places like you are documenting, “slipped” in, reality is that hordes of major cities were as I said, driven in by the truckload. It is just not the 40% either, remember that the urban data tends to be longer term, more complete, less fractured by gaps, easier to work with. Hence the resulting global trends tend to be dominated by the UHI affected metro data.
At this page below you can link to several pages where I discuss individual UHI affected cities and surrounding more rural sites.
At this page below you can link to various studies of 5 degree grid cells, aiming broadly to show how the more urban stations as used by Jones et al have UHI affected trends that are not supported by rural data.
Sorry to rave on.

May 27, 2007 1:26 pm

I have just added a map of CRU/Jones 1994 temperature stations in the San Francisco grid cell, over at ClimateAudit.

June 22, 2007 1:22 pm

Oh, this is just TOO delicions. Someone should forward this to the Washington Times. They love printing stuff like this.
The climate change reporter there is Joyce Price,

January 2, 2009 6:15 am

[…] questioned how NOAA climate monitoring stations were setup. It wasn’t until I stumbled on the Marysville California fire station and its thermometer that that I began to notice just how badly sited these stations are. When I started looking […]

January 3, 2009 4:11 pm

[…] and blogger, Anthony Watts, has been surveying official weather stations in the US since May 2007 and his findings often make incredible reading.  Indeed, many official weather stations are very […]

John Leake
January 5, 2009 10:11 am

Many years ago (early 1970’s) I set up a weather station for the US Forest Service on the Teton National Forest. The standards I was given from the US Weather Service were quite explicit that the Marysville weather station is not permitted. Things like no building or structure within 50 feet and no pavement, etc near the station. Using data from bad stations such as the one in Marysville is inexcusable and the weather station itself as a climate reporting station is at best questionable.
I find those who use the data without knowing what it is are at best guilty of bad data leading to bad conclusions. At worst they are guilty of fraudulent manipulation of the data to support their preconceived conclusion. The evidence suggests that some notable names in this area are apparently guilty of the later.

January 6, 2009 3:13 am

My first thought was that if any station had reported such startling data which contradicted the tenets of the High Faith of Global Warming, there would be an investigation.
My second thought was the quote attributed to Robert A. Heinlein:
“Never attribute to malice that which can be adequately explained by stupidity”

%d bloggers like this:
Verified by MonsterInsights