In my 07/10/2008 post, I described the average NCDC adjustments of the NMSU data set that I observed. Subsequently, I compared the WRCC data set for Tucumcari 4NE (http://www.wrcc.dri.edu/cgi-bin/cliMAIN.pl?nmtucu) with the NMSU data set. With the exception of a larger number of missing data points (19 vs. 7), The NMSU data matches the WRCC data much more closely than the USHCN data; especially during the interval 1915-2002 where that are no significant change in the monthly, seasonal or annual trend patterns.

I would describe the average of these WRCC adjustments of the NMSU data set as follows:
1) TMEAN – 1905-1914: increase ~0.86; s.d. ~.69°F
2) TMEAN – 1915-1948: increase ~0.07°F; S.D. ~0.19
3) TMEAN – 1949-2002: decrease ~0.03°F; S.D. ~0.06
4) The differences in the 1949-2002 interval are essentially within rounding limits (NMSU data is in tenths of °F; the WRCC data is in hundredths of °F).

Is it possible that the historical surface station data archived at the RCCs more closely represent the “raw” temperature data?

I would be pleased to send you a copy of my Excel workbook but I haven’t received your email yet. Also, I can’t find an email adress listed on your surfacesstations website. If I called th4e telephone number listed under contacts there, would I be able to obtain an email address?

REPLY: I’ve sent two emails to you, perhaps the MSN spam filter ate them.

email address is info [-at-] surfacestations dot org

Re: My earlier posts on NMSU v. NCDC Tucumcari data

I have now manually transferred the NMSU data for 1905-2002 to my Excel program to compare it to the USHCN data Although I didn’t expect to see any new trends, I was quite surprised to see significant step-wise differences between TMEAN, TMAX and TMIN in the two data sets in APR 1957 and APR 1981. Alan Cheetham also noted these differences in his 07/08 post.

I would describe the average of these NCDC adjustments of the NMSU data set as follows:
1) TMEAN – 1915-1956: increase ~0.10°F; 1957-1981: decrease ~1.27°F; and 1982-2002: increase ~0.05°F
2) TMAX – 1915-1956: decrease ~0.17°F; 1957-1981: decrease ~2.03°F; and 1982-2002: increase ~0.07°F
3) TMIN – 1915-1956: increase ~0.43°F; 1957-1981: decrease ~0.42°F; and 1982-2002: decrease ~0.01°F
I cannot think of any logical reasons for NCDC making adjustments of this nature to the NMSU data set.

Alan Cheetham, thanks for making me aware of your website. As regards your question about which of the two data sets, GHCN or GISS is correct, according to my analysis of the NMSU/AES data, it compares very closely (~0.05°C) with the right-hand column of your NOAA-GHCN data table. However, it doesn’t match the USHCN data set that I downloaded from http://cdiac.ornl.gov.

REPLY: I dropped you an email on this subject two days ago, would you kindly reply to info [-at-] surfacestations dot org please? I’d like to get a copy of the spreadsheet. – Anthony

I have not received your email message. I checked my SPAM filter to see it might have been blocked but found no record of that having occurred.

Re: My 04/07 13:21 post on NMSU v. NCDC Tucumcari data

I have now manually transferred the NMSU data for 1905-1949 to my Excel program to compare it to the USHCN data. I have not deciphered NCDC’s algorithm for adjusting the NMSU data but here are a few observations:

1. There are very few missing data in the NMSU data set; OCT and NOV in 1907 and AUG through DEC in 1920. It appears there may have been an instrument problem in 1920 that corrupted some data. It is not apparent to me how NCDC filled these missing data points in their data set.
2. In addition to the missing data points in the 1915-1949 data sets, there are a small number of data points where the differences between the NMSU and NCDC data exceed three STDEVs; the remining differences are less than 3°F.
3. NCDC adjustments to the 1915-1949 data sets a) increased TMAX for each month in the DJF and MAM seasons and decreased TMAX in the JJA and SON seasons, b) increased TMIN for each month in the DJF and decreased TMIN in MAM seasons; and decreased TMIN in the JJA in ~1/2 of the months and SON by ~1/3 of the months, c) increased TMEAN for each month in the DJF and JJA; decreased TMEAN in MAM seasons; and decreased TMEAN in SEP, OCT but increased it in NOV.
4. The details of the NCDC adjustments are beyond the scope of this post but suffice to say, they are relatively small but extensive. If I knew how to include Excel graphics in WordPress, I could illustrate the nature of the adjustments.
5. I don’t see any direct relationship between the NMSU and NCDC data sets.

Transferring data manually is a slow, boring process for someone with my limited typing skills. I’ll try to transfer more NMSU data when I have time but I doubt that I will see many new trends.

REPLY: Check your email for a message from me, Anthony

Whoops! It’s getting late in ME on the 4th … Excuse the typos, i.e., “downloaded” and “somewhere”.

Re: “UPDATE 7/1/08 One of our commenters “AnonyMoose” has brought this well done historical weather and climate report by NMSU superintendent Rex Kirksey to our attention:

http://tucumcarisc.nmsu.edu/documents/rr751.pdf”

I note that the TMAX, TMIN and TMEAN data contained in Table A2, A3 and A4 of the NMSU report, Weather Observations at the Agricultural Science Center at Tucumcari 1905–2002, are quite different from the TMAX, TMIN and TMEAN data that I downloded from http://cdiac.ornl.gov. (See my 07/02/08 post above.) The differences are quite significant.

I manually transferred the NMSU data for 1905-1919 to my Excel program to compare it to the USHCN data. There were differences between every data point in the 15-year NMSU and USHCN data set. For example, the differences in TMEAN ranged from -4.12°F to+6.24°F and the average was -0.31°F; the differences in TMAX ranged from -3.14°F to+8.16°F and the average was 0.12°F; the differences in TMIN ranged from -2.89°F to+17.07°F and the average was +0.26°F.

I understand that NCDC adjusts surface station data but their algorithm is not apparent to me. Is there an explanation somewhare that would help me understand the differences between the NMSU and USHCN Tucumcari station data?

REPLY: This is odd, this station has a near perfect attendance record for daily data. I wonder if the differences are resulting from the total set of USHCN adjustments? Truly a puzzle worth exploring.

I mentioned in a post a while ago that I was puzzled by some step-wise changes that I observed in the difference betwen the average and median temperatures for some USHCN stations. I used the monthly data from http://cdiac.ornl.gov. I defined TMED=(TMAX+TMIN)/2 and delT=TAVG-TMED. I calculated the annual values as the 12-month arithmetic mean.

I was surprised to see delT remained quite constant for long intervals followed by abrupt step-wise changes. I was puzzled by the self-evident seasonality of the changes, i.e., a distinct DJF, MAM, JJA and SON pattern that suggested that the monthly data had been adjusted from time to time on a quarterly basis.

I decided to use the same methodology to examine the Tucumcari station data and found a similar pattern. I don’t know how to post an EXCEL plot on your website but I’ll try to briefly explain what I observed. First, there are two apparent step-wise changes in delT: 1914-15 and 1984-85. Second, in 1914-15, a step increase in annual delT=0.54 degF; DJF delT~1.3 degF; MAM delT~0.7 degF; JJA delT~0.2 degF and SON delT~0.4 degF. Third, in 1984-85, a step increase in annual delT=0.18 degF; DJF delT~0.5 degF; MAM delT~0.2 degF; JJA delT~no change and SON delT~0.05 degF. Fourth, I didn’t observe any significant change around 2000.

It appears to me that some algorithm was used to adjust the Tucumcari station data in 1914-15 and again in 1984-85 but it is not apparent to me if TAVG, TMAX, TMIN or all were adjusted.

Having said all of this, the step-wise changes Tucumcari station data are fewer and smaller than in other stations that I have examined.

Yes, that could be done. But if the information is counter to the agenda of the publisher it would never see print. You can give them information but you can’t make them print it.

There is enough around on the Internet here, at Climate Audit, Icecap, and other places for any interested journalist to find and write their own article. Basically what you are asking is for someone to “freelance” an article themselves and submit it to the paper for publication. In most cases that would be counter to the agenda of the publication and would fall on deaf ears.

Daniel, your entire discussion is transparent rhetoric. Calling the monitoring stations “crappy” was being quite gratuitous on my part. The temperature station siting problems often discussed by this blog only scratch the surface, if that alone isn’t enough. Come on MAN! Barbecues, inversion layers, concrete, wind blocks, air conditioners, jet exhaust, metal roofs, bodies of water, trash incinerators and etc. Just what a growing temperature monitoring network needs in the immediate vicinity! Add that to the multitude of other problems, the least of which is that these thermometers have a degree of accuracy of +/- 2 degrees.
Further, all of these stations have to be signed off by a NWS station manager. Their rationalization for the very poor siting is that they have to compromise with low paid or volunteer coops to get the job done. The result has been that the quality of the stations gets worse over time. And we are to use this network of temperature stations as the foundation of global warming evidence that, in the end, adds 180 billion dollars per year to the bill of the American energy consumer?

“Obviously, there is disagreement on how well that extrapolation occurs, but it’s disingenuous to suggest that there isn’t any valuable data contained whatsoever in these sites data records.”

That is just dabbling with the extreme. It’s not a question of finding ”any valuable data… whatsoever.” The quantity of the usable available data is tiny. Of the 500 + temperature stations surveyed so far, only 11% have sufficient siting quality. In addition, only about 4 % do not have the added problem of urbanization. That is about 20 out of 500. We can’t even find one station per gridcell to do an analysis with even questionable margin of error.

Could not this increased presence be contributing to the overall heat record due to the storage of heat during those periods when the old environment normally released heat and cooled? High humidity at night is like an additional blanket, holding in heat so that when the warming from the sun begins in the morning, the beginning start temperature is already higher than the historical record.

REPLY: Christy documented this problem in the Sacramento valley due to irrigation, see his paper in this Tucumcari post. Suburban sprawl and lawns should be easy to catch. Find a weather station that has been overwhelmed by lawn bearing suburbanization and look at the data.

Mind-blowing stuff on bacterial nucleation of ice:

http://pubs.acs.org/cen/science/86/8615sci1.html

This seems to put the old idea that forests nucleate their own rain on a fairly firm basis. Amazingly, you can look at this as an evolved dispersal mechanism for bacteria!

Lawton, Pielke Sr. and some more UAH folk have shown that deforestation reduces cloud cover, and also cover the reduction in transpiration issue:

http://www.sciencemag.org/cgi/content/abstract/294/5542/584
http://www.rainforest-alliance.org/neotropics/eco-exchange/2002/jan02-2.html

Hence there is direct local warming due to loss of transpiration and reduced cloud cover, which Roy Spencer thinks could be a warming cause.

This all makes the idea of cutting down rainforest to grow biofuels not only wrong in principle but potentially completely counter-productive, even leaving aside the short-term CO2 release from the trees themselves.