I’ve recently updated the www.surfacestations.org website with the latest surveys and numbers. We have 534 stations surveyed. Here is where we stand now with USHCN station surveys:
click for a larger image
Climate Reference Network Rating Guide – adopted from NCDC Climate Reference Network Handbook, 2002, specifications for siting (section 2.2.1) of NOAA’s new Climate Reference Network:
Class 1 – Flat and horizontal ground surrounded by a clear surface with a slope below 1/3 (<19deg). Grass/low vegetation ground cover <10 centimeters high. Sensors located at least 100 meters from artificial heating or reflecting surfaces, such as buildings, concrete surfaces, and parking lots. Far from large bodies of water, except if it is representative of the area, and then located at least 100 meters away. No shading when the sun elevation >3 degrees.
Class 2 – Same as Class 1 with the following differences. Surrounding Vegetation <25 centimeters. No artificial heating sources within 30m. No shading for a sun elevation >5deg.
Class 3 (error ~1C) – Same as Class 2, except no artificial heating sources within 10 meters.
Class 4 (error >~= 2C) – Artificial heating sources <10 meters.
Class 5 (error >~= 5C) – Temperature sensor located next to/above an artificial heating source, such a building, roof top, parking lot, or concrete surface.”
During the next week, I plan to add a number of stations during my road trip, and Russ Steele is getting many also on his 3 month cross country road trip through the USA by mobile home.
If you are planning trips this summer, why not check out which stations have been surveyed here and see if any at the bottom of the list that have not been surveyed will be near your travels? We still have over 600 stations to go, and your help is needed!
Discover more from Watts Up With That?
Subscribe to get the latest posts sent to your email.
Well, using the minimum bias estimates:
CRN1 & 2: +0°C
CRN3: +1°C
CRN4: +2°C
CRN5: +5°C
The grand total is an average warm bias of 1.95°C per station.
I do not see any reference to altitude as a condition nor is there any condition for the sea breeze effect which can effect temperatures inland up to a couple of miles.
So wait a second here. I’m no rocket scientist but from where I sit are basically estimating the temperature increase and correcting for more than a 2 degree C difference on far more than half of their stations and then declaring an average temperature increase over the course of many decades of 0.7 degrees or so? Am I following the argument correctly?
evan be careful the estimate of error is must likely the max you would see
and not an average. So, for example you might see the 5C error only on TMIN
and you might see it only during certain seasons. also the error can be either a warming bias or a cooling bias
Evan & Steve: there must be a way to represent the sigma these kinds of errors induce in the ‘average’ temperature. I assume this would show how ridiculous sub-degree prognosis is from this database, especially with the added deviation from NASA’s estimated adjustments.
The grand total is an average warm bias of 1.95°C per station.
no.
those numbers represent a POSSIBLE MAXIMUM error.
i overed Anthony my help in getting that graph correct, but he doesn t sem to be interested.
REPLY: So that “sod” could get a personal handle on station bias, I offered to help him survey some stations, but he didn’t seem to be interested. Too busy blogging on Iraq I guess. There’s no graph here to get “correct”, since the scheme is from NOAA referenced above. Also, my policy is not to pay much attention to anonymous phantoms, such as yourself. If you want respect, put your name to your work.
Am I following the argument correctly?
Oh, yes.
As St. Mac would say, a mining outfit attempting to make such an assay would be up on charges.
However, it is important to point out that the warm bias (offset) is 1.95°C, not the delta (trend). Now most of those violations occurred after 1980 (exurban creep, MMTS conversion etc.), so a fair chunk of the delta is very likely to be a false result.
My guess is that about half or so of the increase in temperature measurements since 1980 is spurious, and that the actual world temperatures are roughly equivalent of the 1930s. I base this guess on:
a.) the site violation issues. We have seen the pics and we have also seen the jumps in the graphs as the microsite conditions near the stations has changed.
b.) such an adjustment would bring temperature correlations very close to the Pacific Decadal Oscillation and Atlantic Multidecadal Oscillation cycles.
those numbers represent a POSSIBLE MAXIMUM error.
Those are the “estimated” (NOAA’s phraseology) MINIMUM error. (All “plusses” are excluded from the calculation.
Furthermore, the errors associated with heat sink primarily manifest themselves at T-Max and T-Min (especially the latter). And how does the USHCN tally their temps? By averaging T-Max and T-Min.
those numbers represent a POSSIBLE MAXIMUM error.
Those are the “estimated” (NOAA’s phraseology) MINIMUM error. (All “plusses” are excluded from the calculation.)
Furthermore, the errors associated with heat sink primarily manifest themselves at T-Max and T-Min (especially the latter). And how does the USHCN tally their temps? By averaging T-Max and T-Min.
and you might see it only during certain seasons. also the error can be either a warming bias or a cooling bias
There are possible cooling biases in the CRN2 category. But in the CRN3 through 5 cats, the bias parameters (waste heat and heat sink factors) are exclusively warming.
And as I said before, T-Max and T-Min (factored, of course, for TOBS) are the only relevant issues insofar as temperature measurement is concerned. Seasonal issues are relevant, though, and NOAA does not seasonally differentiate for their error estimates. (One would assume they were maximized in summer, minimized in winter, and the estimates are for the average.)
I want to re-emphasize that the word NOAA uses for these bias figures are “estimated”, not “maximum”.
CRN4 bias is 2 degrees C OR MORE. And CRN5 is 5 degrees C OR MORE.
And I note that the Rev is VERY conservative when rating the stations, overlooking many of the warming biases, always erring AGAINST warming. So the actual situation may be even worse than the Revs observations.
I do not see any reference to altitude as a condition nor is there any condition for the sea breeze effect which can effect temperatures inland up to a couple of miles.
“Far from large bodies of water, except if it is representative of the area” would seem to cover this contingency.
With only 13% of surveyed stations in in the CRN 1/2 categories, and assuming this remains representative of the whole set, we may be able to get an alternative gridded US average T sometime soon without having to fudge the numbers like NASA.
Will it match GISS?
Is it too late to worry about?
The problem is that most of those CRN1 & 2 stations are in the west coast, which is one area of the country that has seen real warming. They are not well distributed enough to create a reasonable grid.
This may change as the interior of the country is observed.
The new NOAA/CRN network (whenever they actually start it up) is supposed to address this problem.
I believe that Evan is correct in his numbers. I have failed in that I have been unable to travel here in Alabama and many of our stations particularly in the north of the state where I live haven’t been surveyed. Many of these are old stations and would represent a good study. I hope conditions here change in the near future as I have grand children that I \would like to get interested in doing this sort of work/research.
Theratings are shown at the top of the post and the network is in trouble.
13% of the surveyed stations have no bias = 0C
18% of the surveyed stations have minor bias = 2C
13% of the surveyed stations have major bias >= 5C
All of the bias descriptions given in class 3,4,and 5 are warming bias. By this survey it indicates thus far that 87% of the stations have a minimum of 1C bias and that the vast majority of the stations 67% have a bias greater than 2C. For these reasons I believe that Evan is correct in his reasoning.
Tom
sea breeze and altitude are natural conditions and not bias. The argument is that the rise in temperature is man made but if you are measuring the direct waste heat of man that isn’t an accurate measure of the regional or global temperature. That is the reasoning behind the class 3, 4, and 5 classifications, the nearness to direct man made waste heat or man made multipliers of natural heat (walls, asphalt or concrete or other unnatural surfaces nearby.
At least this is my interpretation of what I am reading, if I am incorrect please correct me.
Bill Derryberry
sod, you should note that even if the estimates are exaggerated by a factor of, say, three, they would still be highly significant.
I must correct myself,
13% no bias
18% minor bias 1C
56% moderate bias >= 2C
13% major bias >= 5C
my not reading my on notes.
The network is in trouble.
Bill Derryberry
Anthony – why not just plot graphs of the CRN1 & 2 stations and see how that compares to the official figures?
Evan says they are mainly on the West Coast, but looking at your map, they seem to be quite well distributed
– anyway plotting them would still give some interesting results
– you could also plot the CRN5 stations – and see how they compare to the good ones
Looking at the map, I see hardly any CRN2 stations at all. I think they are largely lost in the west coast sea of red.
The non-coastal south and Southwest (where actual cooling occurred seem under-represented.
Of course, if some sort of gridding procedure could be accomplished, that would give us a reasonable answer. (The CRN network is supposed to do this.)
Evan: “Far from large bodies of water, except if it is representative of the area” would seem to cover this contingency.
However, the Class 1 definition stipulates “at least 100 meters away”. Seabreezes can effect surface temperature up to a couple of miles inland depending on their strength that particular day. When the wind blows off shore, there is no seabreeze effect and the temperatures are very close to the inland temps. Without charting the diffference in number of days of each, your temp readings are not going to reflect a true read for the area. Also, is there a consideration for the temperature of the water over which the wind blows? Early spring Gulf water temps along the central west coast of Florida are generally in the upper 60’s so that seabreeze will effect the area much more at that time of year as opposed to summer when air temps are in the 90’s and water temps are in the high 80’s. It also works in reverse during winter. Gulf water temps are in the lower 60’s/upper 50’s. When a strong cold front blows down from the northwest over the water inland temps may hit the middle 30’s but the coast stays in the 40’s/50’s. If the cold front blows from the north or north east than the wind does not blow over the water and the coastal areas just as cold as inland (except for immediately next to the water).
Just wondering if all this is considered.
well Evan, I would beg to differ. There isnt a qualitative study to back up the bias figures, so that is what we are undertaking. I take the ratings exactly as Dr. LeRoy proposed them, as estimates. Now in the first CRN study a CRN site ( class1)
was compared to a class 2 site ( ASOS) and the effects of site exposure were found to be on the order of .25C. This effect was modulated by winds and clouds. Simply, on some days you might see a .25C difference on cloudy or windy days you might see something different. So take a site. Assume a
FLAT temperature trend for 25 years. 0 trend. Now, introduce a .25C bias
sometime in its history. So, it goes from 0C to .25C, sometime in its history.
but only on certain days and during certain seasons. Now calculate the trend A bias will impact trend estimates when the trend is calculated over
the period when the bias was introduced. Afterwards, the bias does not impact the trend estimate. So, imagine a site that is class 1 in 1975
from 1975 to 1987 it shows no trend. In fact, its always the same temp 14C.
What’s the trend? ZERO. Now suppose in 1988 we introduce a BIAS of 1C.
And the site reads 15 C. and it reads 15C from 1988 to 2008. What the trend
from 1988 to today? ZERO. whats the trend from 1975 to 2008? about .04C per year. So Bias effects trend estimates when the period for which the trend is calculated extends over the change point caused by the bias. That is why it is important to document the time of the changes. Also, the effect of the bias on trend is a function of the length of the trend. Take a 100 year record with zero trend. Introduce a 5C bias step at year 50. Now calculate trends. Trends ending before the bias step will be zero. Trends starting after the bias step will be zero, trends that straddle the step will be effected
Finally, When I compared the class 5 sites to the Class 1 & 2 sites I found differences in trend. Thats what you look for. But you dont see anything close to 5C differences in absolute temps. Why? because the bias signal is not constant
Lets just take the effect of concrete. When a site is over concrete the daily TMAX may actually get depressed because the concrete is a great heak sink.
But when the sun goes down, then the TMIN spikes. Lets say it spikes by 5C
( you can see figures like this in vegas and Reno)
Now calculate the TMEAN for the day ( TMIN+TMAX)/2
So, TMIN has spiked by 5C over what it would normally be, a 5C error, but
TMAX is not effected or is actually a bit cooler in some instances.
So, the error in TMEAN will be 2.5C. And this only on sunny days where the concrete is exposed ( no shading, ect)
Same with AC units and other artifical sources of heat. Sometime they cause a bias, other times not. That is the reason why you should not site near them.
It makes the analysis of the data unfathomably complex.
Finding and quantifying microsite bias will be a very tough but very important piece of work.
I looks like there isn’t much left in Florida for me to offer to help with this summer. Or am I wrong? If there are any left, tell Tom (who I presume is the main surveyor in FL) to save some for the rest of us! 🙂
REPLY: maybe it is time to exapnd and get GISS stations that are not part of USHCN, there are a number of those in Florida, would you like to try those?
Ok, boarding the plane in 45 mins, using airport WiFi. Any new comments will not get approved until tonight, late.
timetochooseagain,
If by Tom you are referring to me, I have done no surveys. As you noticed there isn’t much left in Florida to survey but I would like to take up Anthony’s offer to look at GISS stations that are not part of USHCN. Anthony, were do I find the list and the locations? I found this site via Junkscience.com and the “How not to measure temperature” articles made so much sense even though I have no technical skill in that area. Over the past months I have gathered a general understanding of the processes and terminology by reading the articles and the corresponding comments section.
REPLY: See list of stations at http://www.surfacestations.org that are USHCN
for GISS see this list:
http://data.giss.nasa.gov/gistemp/station_data/
and click on Florida
locate accurate lat/lon and location description at NCDC MMS
http://mi3.ncdc.noaa.gov/mi3qry/login.cfm and use guest login.
Then enter station names from GISS
Slightly off topic but the BBC explain here what really happened with the Roger Harrabin email fiasco.
http://news.bbc.co.uk/player/nol/newsid_7350000/newsid_7355700/7355760.stm?bw=bb&mp=wm&news=1&ms3=6&ms_javascript=true&bbcws=2