Santa Rosa USHCN – click for a much larger image
Or maybe with the lack of grass, “goat” might be more appropriate.
Every once in awhile (like once a week) I happen upon a NOAA USHCN weather station that leaves me wondering – what were they thinking?
From this NOAA USHCN COOP weather station #298107, which was located here on 6-14-2007 , we get climate data. Unfortunately there appears to be a lack of attention to details. For example, ignoring the obvious other things, look at the green garden hose at the base of the MMTS mounting pole. That garden hose was apparently used to bury the cable in underground to the residence. Only one problem. If the end of the hose is not sealed, water will wick right down the cable during rain and eventually fill the buried hose. Eventually the water will seep into the cable it was supposed to protect, causing some reduced resistance, perhaps a short later. With thermistor systems like this, the sensor often has resistance into the kilohms…and moisture induced resistance changes are easily masked and often go unnoticed in data.
But all that and Dr. Seuss aside, here is what the data looks like before and after NASA GISS adjusts it. These are the USHCN “raw” and “homogenized” data plots from the GISTEMP website. The before and after is quite something to behold.
Click for original source graph from GISS
And here is the data after it has been “homogenized” using the GISTEMP algorithm, I changed the color hues to help differentiate them visually:
Click for original source graph from GISS
Note that other than my applying labels and doing a hue shift, these are the exact graphs presented on the NASA GISTEMP website.
They also present the data used to plot these graphs on each of the above linked pages. You can download it yourself by clicking on the “Download monthly data as text” links at the bottom of the pages linked by the graphs above. I can’t provide direct links here due to the links being dynamically generated and as temporary files at GISTEMP.
I downloaded and plotted both of these datasets myself, plus did polynomial fit trend lines. Note that missing data from 1920 has not been plotted on this graph, but simply skipped.
The visual result of the before and after adjustments by NASA GISS speaks for itself:
Click for larger image
What was down, is now up. Note the hinge point in the near present. The GISTEMP data stops at 2006, it will be interesting to see what this new location does for the data in 2007 and 2008.
Sponsored IT training links:
Successfully complete your certification with help of latest 70-271 questions, 642-072 simulations and 642-357 practice test.
Retired Engineer
I hope you are right about the current loop, but as I have a couple of patents involving constant current sources, I know that they too can be fickle, and should be calibrated on a consistent basis.
The most I’ve found online is a check box schematic (unofficial) that shows that a 1M resistor should read ~-50, a 10K resistor should read above 100, and the resistors from the check plug should read 77.1 plus or minus .3 degrees. I thought I had a line on the thermistor used, but I must be mistaken, it is < 10k at 21C.
I agree that it would be nice to see a schematic and parts list, including the wire gauge used in the installation. I think Anthony has an MMTS, but he’s using it and I think he would object if we took it apart 😉 Google my name if you want to discuss offline.
Looked at the GISS adjustments for Santa Rosa as well as for 4 other stations. It appears from the adjustments that I have looked at 2006 is a special year since they were all zero. Some have negative and some positive changes in the past so no overall bias towards heating or cooling unlike the Santa Rosa data which has the big heating trend changes.
See http://gallery.me.com/wally#100002 for plot
I must be due for a new computer monitor this one is getting a little snow in the picture.
Well I wouldn’t panic about the remoteness or lack thereof as far as wire resistance. If the gizmo does have something as cruded as a thermistor; only someone who is as dumb as a box of rocks would ever run just two wires to it.
If it is some sort of resistive sensor, or some other types it is going to have a four wire connection; a so-called Kelvin connection.
You run two wires out there to run some current through the sensor, and you run another pair of wires, all the way up into the barn owl box to the sensor (thermistor or whatever) itself to sense the Voltage right at the sensor.
The current bearing wires are called the “force” pair, and the Voltage sensing wires are called the “sense pair, and they carry essentially zero current.
The sense pair has to be controlled as to material and relative location of the two wires; A binary pair of wires with minimum insulation between them is required, to keep the two wires at the same temperature to eliminate thermal EMFs betwwen the wires due to any temperature difference. You can have a temperature gradient along the wires, but not between the pair. Then you want the loop area enclosed by the pair to be as close to zero as possible ehnce the thin insulation, so that the loop just not pick up induiuced EMFs from varying magnetic fields or AC current carrying wires.
You can eliminate all kinds of pestilence such as plug contact resistances and other effects if you run a four wire system; but if there’s only twoi wires in that garden hose, then you are in a heap of trouble no matter if they are #2 AWG solid silver wires.
The art of Kelvin sensing of small Voltage signals is highly developed. We do it all the time in the semiconductor business for device testing; we even have Kelvin probes that can make two side by side electrically isolated contacts on a single silicon contact pad that may be no bigger than a few microns square.
Even so I wouldn’t waste such sophistication on a thermistor.
It turns out you can use four wire sensing in your hi fi setup, to eliminate the need for half inch diameter speaker cables from Monster Cable. A couple of zip cords, can do the job, if you run two pairs out to your speaker, and pour the juice down one pair, and connect the other pair to your amplifier’s output sense terminals. What, your KillerWatt ghetto blaster doesn’t have output sense terminals ?
I forgot to add, the two pairs of wires should be twisted pairs. That can alleviate the induced EMF problem since each twist will pick up basically the same EMF as the adjacent one but with opposite polarity. So that relieves the insulation problem a bit. The force pair should also be twisted to that it doesn’t radiate any glitches to the sense pair; for the same reason.
I hope Will Small is reading this blog – this piece of hard data goes directly to my first challenge for him.
I.e. First show that “AGW Climate Science” is – in fact – science.
Hi Will,
That would be specifically Points 1a – through to 1c,
1a. Are the Temperature data measurements credible, using calibrated instruments set to defined and commonly accepted industrial standards of operation.
1b. Are the Temperature Data storage methods credible, using techniques in accordance with defined and commonly accepted industrial standards for Data and Configuration Managment.
1c. Are the Temperature Data evaluation and interpretation methods and software, documented, transparent and freely available for the inspection of third parties.
from the http://wattsupwiththat.com/2008/11/26/atlantic-hurricane-season-sets-records/#comments post
The silver pipe with the cap on it is a well cap. Must be a newer well. The old ones were capped at ground level. And older than that, it was simply a hand-dug hole in the ground cribbed open if necessary. Most hand-dug wells that I have seen were 25 to 30 feet down and wide enough to fit at least two diggers into the hole.
Santa Rosa CA is 38.45N 122.73W
Ok so I have to ask…. Why is this station placed in this guy’s back yard? I mean seriously. Fast internet connection? Doesn’t really seem like it based on the suroundings aside from the ‘Dish network’ dish, but that’s not an i-net capable device.
Pamela, what are you talking about with respect to the well? I see a silver/gray item directly to the right of the sensor pole, but it’s standing on two legs as if it were a moon-lander. That’s not a well…. Sorry if my eyesight is doing me wrong here…
Finally, since we’re questioning the “science” here…. Did anyone look inside of the hose to see if it the end had been ‘stopped up’ with silicone or something similar? I mean… While it may be a little “redneck” it would keep the moisture out.
http://gallery.surfacestations.org/main.php?g2_view=core.DownloadItem&g2_itemId=1666 shows : 34.95 -104.68 7473100 for Santa Rosa
Is this an older location?
I just can’t buy that scientific equipment is placed in the shanty-town backyard with a garden hose. Surely even Al wouldn’t do that?
(note: big-bub: your coordinates don’t show me shanty-town either.)
Anthony: can you give us the coordinates of this location?
Apparently they silver rocket-ship looking item to the right of the MMTS is a non-recording rain gauges. Apologies for my ignorance.
See: http://gallery.surfacestations.org/main.php?g2_view=core.DownloadItem&g2_itemId=1666
for some nice equipment shots, and admission of MMTS “cabling problem”.
From Pamela Gray (18:20:18) :
Most hand-dug wells that I have seen were 25 to 30 feet down and wide enough to fit at least two diggers into the hole.
My dad and I ‘dug’ a well with one of those post hole diggers that you turn by hand. Add pipe sections as needed for depth. Gets a bit cumbersome lifting the dirt loads after the first 10 feet of extension… (Got 50 gpm from it!)
The boat and the wood pile will both get Very Hot in the summer sun (to the point where they will be hurtful to sit upon. I know this because I’ve sat upon such thing near Santa Rosa … During the winter, the same stuff will act to store solar heat to some extent and release it at night… Far more so than grass does.
Yes, please do explain more the difference between raw data and homogenized data. What is their explanation for this process and the resulting differences?
I’d like to refer this website to associates at another where I participate. I want to be sure I have it right:
The new station is put on line in 2007, but local records go back for decades. In order to claim continuity, the old records are modified using the temps from the new station and an algorithm. Is that about it?
Thanks
REPLY: No, in simple terms the GISS homogenity adjustment uses other station data within a radius of 1200 km of the station as a way of weighting all stations. So stations nearby like Lubbock, Amarillo, Phoenix, Albuquerque that have UHI issues then affect the data at rural stations. – Anthony
George E. Smith
Yep, there are many ways to do it, a few good and many bad. Unfortunately, there is very little information to go on. The one thing that I have been able to find is an unofficial check box that shows a 3 pin connector, and a test lpug that shows a resistor across two pins. I must say that it is not likely any kind of 4 wire measurement unless they are tying the low sense probe to ground, or connecting the source and sense leads in the back of the cable connector shell, and that doesn’t make a lot of sense either.
Oops! The link didn’t show. Google “MMTS check box” and you’ll find a NOAA page on it.
Answering my own question (from data on GISS homepage):
Is there a link to the stations which are ultimately used?
Re: Stations used. Nevermind – and sorry for the annoying questions.
Station list:
http://data.giss.nasa.gov/gistemp/station_data/station_list.txt
BBQ and a burn barrel (for burning trash?), if my eyes do not deceive me. Could it be that garbage is being burned 10 feet away?
Thermistors and wires.
In an earlier thread, Anthony mentioned a two wire cable. That eliminates any sort of ‘kelvin’ connection. (full disclosure – I used to work for Teradyne, high end automated test equipment outfit) Anthony also mentioned cable length problems in siting. I have nothing else to go on. 4-20 ma current loop is an industry standard, so I assumed that would be used as the MMTS is too old for any kind of digital on top of power type of measurement. It would make sense. Which may be why it isn’t a good assumption.
A thermistor works fine for a fish tank heater, cheap and reasonably stable. But totally worthless for scientific endeavors. The test notes other have uncovered cause me to believe that this system is really just a thermistor (those test resistor values would not work in a current loop environment).
So, after nearly 40 years in the mesaurement business, I have to conclude that the MMTS probably doesn’t suffer too much from ‘adjustment’. It’s not that good to begin with.
George E. Smith and Retired Engineer:
I hope you are still monitoring this.
If you look at the check box web page, you will see “Display(back)”. If you click on the “back” link you will see a 3 pin connector with only 2 pins loaded. I think that shoots the 4 wire theory. In fact, if you look at the “Cable” photo, it appears to be a sheilded twisted pair, but they are not even grounding the shield.
That leaves the current loop idea. Now, the check box schematic indicates that a 10K ohm resistor should read > 100 degrees. Since that is best case, let’s use that. Unless ohms law has completely left my feeble brain, I(R)=V, so for a 20ma current loop: .02(10000)=200V source needed. For a 4ma current loop .004(10000)=40V source needed. The 4ma loop is more reasonable, but still well above ttl or cmos levels. Looking at the display box, I don’t think that the transformer needed would fit, nor some kind of stepping power supply. My guess is that you only have the 5~12 volt range to power the electronics, and it looks like that is being supplied by an external power pack. At 12 volts, your best case current loop would be 1.2 ma, and what would good would that do?
My guess is that they are simply counting on the thermistor resistance to to swamp out any cable/connector resistance.
I put 15 years into test and measurement, but now I’m a database administrator, and have had the requisite frontal lobotomy. I would most humbly appreciate someone pointing where I am wrong. I love learning new things, or at least relearning something that I lost.
I work in the environmental department at a chemical plant. If I did anything like this, my plant would be fined a six or seven digit amount, the plant manager would be hauled off in handcuffs, and he would rat me out so that he would have a punching bag in prison (his words, not mine).
And as for Bill P’s quote, it makes perfect sense. However, it is confusing as it essentially says “we change the data in the urban stations to match the rural stations”. Why have the urban stations at all?
This is bordering on outright fraud.
While the adjustments may have an effect on US temperature curves, the overall GISTEMP global temperature anomaly is not heavily affected by adjustments. See the results of running GISTEMP with and without periurban adjustments here.
Perhaps we should turn this into a geocache type hunt.
Start a site like geocaching.com with the GPS location of stations so we can post a pic of the site. May even get a lil competitive to post the first of the site or post the most.
Then we can really start picking these sites apart.