Accuracy of climate station electronic sensors – not the best

Given all of the discussions recently on issues with the surface network, I thought it would be a good idea to present this excellent paper by Lin and Hubbard and what they discovered about the accuracy, calibration and maintenance of the different sensors used in the climatic networks of the USA. Pay particular attention to the errors cited for the ASOS and AWOS aviation networks, which are heavily used by GHCN.

Sensor and Electronic Biases/Errors in Air Temperature Measurements in Common Weather Station Networks*

X. LIN AND K. G. HUBBARD

School of Natural Resource Sciences, University of Nebraska at Lincoln, Lincoln, Nebraska

ABSTRACT

The biases of four commonly used air temperature sensors are examined and detailed. Each temperature transducer consists of three components: temperature sensing elements, signal conditioning circuitry, and corresponding analog-to-digital conversion devices or dataloggers. An error analysis of these components was performed to determine the major sources of error in common climate networks. It was found that, regardless of microclimate effects, sensor and electronic errors in air temperature measurements can be larger than those given in the sensor manufacturer’s specifications. The root-sum-of-squares (RSS) error for the HMP35C sensor with CR10X datalogger was above 0.2°C, and rapidly increases for both lower (<-20°C) and higher temperatures (>30°C). Likewise, the largest errors for the maximum–minimum temperature system (MMTS) were at low temperatures (<-40°C). The temperature linearization error in the HO-1088 hygrothermometer produced the largest errors when the temperature was lower than -20°C. For the temperature sensor in the U.S. Climate Reference Networks (USCRN), the error was found to be 0.2° to 0.33°C over the range -25° to 50°C. The results presented here are applicable when data from these sensors are applied to climate studies and should be considered in determining air temperature data continuity and climate data adjustment models.

Introduction

A primary goal of air temperature measurement with weather station networks is to provide temperature data of high quality and fidelity that can be widely used for atmospheric and related sciences. Air temperature measurement is a process in which an air temperature sensor measures an equilibrium temperature of the sensor’s physical body, which is optimally achieved through complete coupling between the atmosphere and air temperature sensor.

The process accomplished in the air temperature radiation shield is somewhat dynamic, mainly due to the heat convection and heat conduction of a small sensor mass. Many studies have demonstrated that to reach a higher measurement accuracy both good radiation shielding and ventilation are necessary for air temperature measurements (Fuchs and Tanner 1965; Tanner 1990; Quayle et al. 1991; Guttman and Baker 1996; Lin et al. 2001a,b; Hubbard et al. 2001; Hubbard and Lin 2002). Most of these studies are strongly associated with the study of air temperature bias or errors caused by microclimate effects (e.g., airflow speed in-side the radiation shields radiative properties of sensor surface and radiation shields, and effectiveness of the radiation shields). Essentially, these studies have assumed the equation governing the air temperature to be absolutely accurate, and the investigations have focused on the measurement accuracy and its dependence on how well the sensor is brought into equilibrium with the atmospheric temperature. Such findings are indeed very important for understanding air temperature measurement errors in climate monitoring, but it is well known that all microclimate-induced biases or errors also include the electronic biases or errors embedded in their temperature sensors and their corresponding data acquisition system components.

Three temperature sensors are commonly used in the weather station networks: A thermistor in the Cooperative Observing Program (COOP) that was formally recognized as a nationwide federally supported system in

1980; a platinum resistance thermometer (PRT) in the Automated Surface Observing System (ASOS), a network that focuses on aviation needs; and a thermistor in the Automated Weather Station (AWS) networks operated

by states for monitoring evaporation and surface climate data.

Each of these sensors has been used to observe climate data over at least a ten year period in the U.S. climate monitoring networks. The U.S. Climate Reference Network (USCRN) was established in 2001 and gradually and nationally deployed for monitoring long-term and high quality surface climate data. In the USCRN system, a PRT sensor was selected for the air temperature measurements. All sensing elements in these four climate monitoring networks are temperature sensitive resistors, and the temperature sensors are referred to as the maximum–minimum temperature system (MMTS), sensor: HMP35C, HO-1088, and USCRN PRT sensors, respectively, in the COOP, AWS, ASOS, and USCRN networks (see Table 1).

The basic specifications of each sensor system including operating temperature range, static accuracy, and display/output resolution can be found in operation manuals. However, these specifications do not allow a detailed evaluation, and some users even doubt the stated specifications and make their own calibrations before deploying sensors in the network. In fact, during the operation of either the MMTS sensor in the COOP or HO-1088 hygrothermometer in the ASOS, both field and laboratory calibrations were made by a simple comparison using one or two fixed precision resistors (National Weather Service 1983; ASOS Program Office 1992).

This type of calibration is only effective under the assumption of temporal nonvariant sensors with a pure linear relation of resistance versus temperature. For the HMP35C, some AWS networks may regularly calibrate the sensors in the laboratory, but these calibrations are static (e.g., calibration at room temperature for the data acquisition system).

It is not generally possible to detect and remove temperature-dependent bias and sensor nonlinearity with static calibration. In the USCRN, the PRT sensor was strictly calibrated from -50° to +50°C each year in the laboratory. However, this calibration does not include its corresponding datalogger. To accurately trace air temperature trends over the past decades or in the future in the COOP, AWS, ASOS, and USCRN and to reduce the influence of time-variant biases in air temperature data, a better understanding of electronic bias in air temperature measurements is necessary.

The objective of this paper is to carefully analyze the sensor and electronic biases/errors induced by the temperature sensing element, signal conditioning circuitry, and data acquisition system.

This implies that the MMTS temperature observations are unable to discriminate  ±0.25°C changes

in the lower temperature ranges (Fig. 5 and Table 2). The interchangeability of the MMTS thermistors is from

60.2°C from temperature -40° to +40°C and ±0.45°C elsewhere (Fig. 4). Two fixed resistors (R2 and R3) with

a 0.02% tolerance produced larger temperature errors of measurement in low temperatures, but the error

caused by the fixed resistor R19 in Fig. 1 can be ignored. Therefore, the RSS errors in the MMTS are from 0.31°

to 0.62°C from temperature -40°C to -50°C (Fig. 5).

The major errors in the HO-1088 (ASOS Temp/DP sensor) are interchangeability, linearization error, fixed resistor error, and self-heating error (Table 2 and Fig. 7). The linearization error in the HO-1088 is relatively serious because the analog signal (Fig. 3) is simply linearized from -50° to 50°C versus -2 to 2 V. The maximum magnitude of linearization error reached over 1°C (Fig. 7). There are four fixed precision resistors: R13, R14, R15, and R16 with a 0.1% tolerance. However, the error of temperature measurement caused by the R14, R15, and R16 can be eliminated by the adjustment of amplifier gain and offsets during onboard calibration operations in the HO-1088.

The error caused by the input fixed resistor R13 is illustrated in Fig. 7. Since this error was constantly varied from -0.2° to -0.3°C, it can be cancelled during the onboard calibration. It is obvious that a 5-mA current flowing through the PRT in the HO-1088 is not appropriate, especially because it has a small sensing element (20 mm in length and 2 mm in diameter). The self-heating factor for the PRT in the HO-1088 is 0.25°C mW21 at 1 m s21 airflow (Omega Engineering 1995), corresponding to the selfheating errors 0.5°C when the self-heating power is 2mW (Table 2 and Fig. 7). Compared to the linearization error and self-heating error, the interchangeability and LSB error in the HO-1088 sensor are relative small, ±0.1° and ±0.01°C, respectively (Table 2).

Conclusions and discussion

This study provides a better understanding of temperature measurement errors caused by the sensor, analog signal conditioning, and data acquisition system. The MMTS sensor and the HO-1088 sensor use the ratiometric method to eliminate voltage reference errors. However, the RSS errors in the MMTS sensor can reach 0.3–0.6 under temperatures beyond -40° to +40°C. Only under yearly replacement of the MMTS thermistor with the calibrated MMTS readout can errors be constrained within ±0.2°C under the temperature range from -40° to +40°C. Because the MMTS is a calibration- free device (National Weather Service 1983), testing of one or a few fixed resistors for the MMTS is unable to guarantee the nonlinear temperature relations of the MMTS thermistor. For the HO-1088 sensor, the self-heating error is quite serious and can make temperature 0.5°C higher under 1 m/s airflow, which is slightly less than the actual normal ventilation rate in the ASOS shield (Lin et al. 2001a). The simple linear method for the PRT of the HO-1088 causes unacceptable errors that are more serious in the low temperature range. These findings are helpful for explaining the ASOS warm biases found by Kessler et al. (1993) in their climate data and Gall et al. (1992) in the climate data archives. For the dewpoint temperature measurements in the ASOS, such self-heating effects might be cancelled out by the chill mirror mechanism: heating or cooling the chill mirror body (conductively contains the dewpoint PRT inside) to reach an equilibrium thin dew layer–dewpoint temperature.

Thus, in this case, the selfheating error for dewpoint temperature measurements might not be as large as the air temperature after correct calibration adjustment. Likewise, the relative humidity data from the ASOS network, derived from air temperature and dewpoint temperature, is likely be contaminated by the biased air temperature.

Both resistance measurements in the HMP35C and USCRN PRT sensors are interrogated by the dataloggers.

The HMP35C is delivered from Campbell Scientific, Inc., with recommended measurement methods.

Even so, the HMP35C sensor in the AWS network can experience more than 0.28C errors in temperatures from

-30° to +30°C. Beyond this range, the RSS error increases from 0.4° to 1.0°C due to thermistor interchangeability, polynomial error, and CR10X datalogger inaccuracy. For the USCRN PRT sensor in the USCRN network, the RSS errors can reach 0.2°–0.34°C due to the inaccuracy of CR23X datalogger, which suggests that the configuration of USCRN PRT and measurement taken in the CR23X could be improved if higher accuracy is needed. Since the USCRN network is a new setup, the current configuration of the USCRN PRT temperature sensor could be reconstructed for better measurements.

This reconstruction should focus on the increase of signal sensitivity, the selection of fixed resistor(s) with smaller temperature coefficient of resistance, and the decrease of the self-heating power, so that it could be more compatible with the CR23X for longterm climate monitoring.

These findings are applicable\ to the future of temperature data generated from the USCRN network and possible modification of the PRT sensor for higher quality measurements in the reference climate network.

The complete Lin-Hubbard papert (PDF) is available here.

Advertisements

  Subscribe  
newest oldest most voted
Notify of
Jan Curtis

Great work Ken. I must read.

Lance Wallace

Anthony, the temperatures got screwed up here and make no sense. They appear to be thousands of degrees. The actual numbers might be -20 [superscript] degrees C but it gets changed to 2058 or some such thing.
REPLY: I noticed that as soon as I published. Some sort of translation error in ASCII character sets from the PDF to HTML. Fixed even before you wrote this. Refresh.- A

jack mosevich

This is a problem I have always wondered about. Have any calibration/accuracy studies been performed? This paper answers that question and I am humbled by how comlicated the answer to a simple question can be. I also wonder how accurate mercury/alcohol (older) thermometers fare?
Maybe the errors cancel and have a mean of zero. Do the GISS personel care I wonder?

Interesting stuff. In Brohan et al. 2006 (the HadCRUT3 and CRUTEM3 paper) the error associated with a single thermometer reading is estimated at 0.2C.

John Diffenthal

Garbage in – garbage out! It’s difficult to be amusing about this litany of poor design and configuration.
Has there ever been a systematic check of mercury thermometers against the data log?

“But that’s just a fact of life in the climate sciences. [No data for you!]”
Professor Phil Jones giving evidence at the Parliamentary inquiry into Climategate.

woodNfish

I’ve said all along that the error range of the equipment is large enough that there is no way to determine the small trends we were being told were occurring. All this furor over AGW for nothing, yet the juggernaut for regulating carbon continues unabated.

The tests need to be made in a humidity chamber so you can go from cold and dry to warm and humid quickly to see if condensation can be a factor.

Pops

It seems to me that you can have all the electronic gizmos and gadgets you like in order to accurately collect the data – even correlate it – but you still have the problem of human bias prior to the publication of the desired “end-product”, as some people refer to it these days. Until honest (new) hands are back at the tiller, can anyone trust any numbers published by anyone in the (climate) scientific world these days?

Philhippos

Two points:
1. I can still only see things like this in the conclusions: However, the RSS errors in the MMTS sensor can reach 0.3–0.6 under temperatures beyond 2408 to 1408C. Only under yearly replacement of the MMTS thermistor with the calibrated MMTS readout can errors be constrained within 60.28C under the temperature range from 2408 to 1408C.
2. I hoped to conatct via Tips & Notes to ask about possibility of checking out UK/European stations and where to find info but cannot get that link to work – only a blank page that takes no typing.

Also PRT (thermistors somewhat less so) are subject to stress changes. To detect them a number of cycles from cold to hot (to the hot limit) should be tried to see if there is a shift.

DR

We still keep a good ole mercury thermometer in our lab.

Gary Hladik

Uh-oh. I anticipate even more undocumented spaghetti “correction” code from GISS.

Henry chance

Blizzards in Spain. Do wind turbines attract snow? I wonder if the temps in Spain are high above freezing.

Also: .1% resistors are subject to drift over their life. It can be accelerated them by cooking them at the max inst temp for 1,000 hours or so. This is even more true for electronics.
I’ll come back and see what others say.
Let me leave you with this: electronic accuracy to 1% is not too tough. Initial calibration to .1% is not too tough. Holding .1% accuracy over life without frequent recalibration is very tough.
Then you have to go back and see how that relates to temp. BTW setting accuracy by manual trim pots is very bad. It should be done in software with values held in a local EEPROM. Second best is digital pots of sufficient resolution and stability.

Chris

I have wondered about the color stability of the plastic MMTS enclosures. Certain plastics such as PVC and perhaps others tend to yellow or gray slightly as they are exposed to weather and sunlight over several years. I guess this is a similar issue to weathering and paint issues on the Stevenson screens. This reflectance change might be enough to slightly increase heat absorption in the sensors and the resultant daytime readings. Don’t know if this has been considered a factor in long-term temperature trends. Don’t know if this is an issue with the materials used in the MMTS units.

As a layman who knows that when one is in doubt, try to find and work from first principles, I became a full-on sceptic when I enquired if there was a standard for climate thermometers, their siting, housing etc on a Guardian blog and got jumped all over as a “troll” and a “denier” by one of that paper’s more ardent Greenpeace activist/bloggers for my enquiry. In that short time my knowledge of climate has increased exponentially, but it now seems that my original question was very sensible, if a little innocent considering what we all know now about standards of data-selection and archiving in some quarters.

It appears that sensors in low temperature regions (Siberia, etc.) have the largest errors. Aren’t these regions showing the most warming?
“Do the GISS personel care I wonder?”
GISS has analyzed this problem and have come up with a correction factor:
Lower earlier temperatures and raise later ones.

Curiousgeorge

@ jack mosevich (09:11:08) :
Measurement system error is a very common concern throughout industry. Google “Gage Reproducibility and Repeatability”, or visit NIST ( http://www.nist.gov/index.html ) to learn more about it. In my experience in aerospace such errors are rarely evenly distributed. The distribution for a particular system tends to be skewed in one direction or the other, and knowledge of this behavior is usually incorporated by design engineers when determining tolerances.

Crikey !!!
it just keeps getting worse then we thought !!!

Max Anderson

Take a look at figure 3 in the paper. The signal conditioning circuit for ASOS stations uses variable resistors (potentiometers) to adjust gain and offset of the measurement. After working in the world of electronic measurements for 30+ years, I can tell you this is asking for trouble. Eventually these devices start to drift and/or become noisy which will cause more problems with the temperature data.
In today’s world of digital data processing there is seldom any excuse for using potentiometers in this sort of equipment.
REPLY: I’ve used hundreds of those devices, and yes, they are a source of drift – A

Tom Bakewell

Esteemed Antony, I still see text errors with the temp measurements as noted by Lance Wallace above.
What a great paper! Data is everything, and understanding what it is and how it is collected is my kind of science work.
Tom Bakewell

Lance Wallace

Some errors corrected but there are still strange numbers like
3rd paragraph below Table 1: “2508C to 1508C”
5th para below Table 1: “unable to discriminate 60.258C…”
Conclusions line 5 “2408 to 1408C”
REPLY: It is strange that copy/paste from PDF does this in this case, but I think I’ve fixed them all. – A

It’s still wrong after refresh:
In the USCRN, the PRT sensor was strictly calibrated from 2508 to 1508C each year in the laboratory.
The linearization error in the HO-1088 is relatively serious because the analog signal (Fig. 3) is simply linearized from 2508 to 508C versus 22 to 2 V. [the ’22’ is wrong there as well]
Since this error was constantly varied from 20.28 to 20.38C, it can be cancelled
Even so, the HMP35C sensor in the AWS network can experience more than 0.28C errors in temperatures from 2308 to 1308C.
Beyond this range, the RSS error increases from 0.48 to 1.08C due to thermistor interchangeability,
The self-heating factor for the PRT in the HO-1088 is 0.258C mW21 at 1 m s21 airflow (Omega Engineering 1995),
or the USCRN PRT sensor in the USCRN network, the RSS errors can reach 0.28–0.348C due to
etc
REPLY: Yes strange that copy/paste from PDF results in this. But I think I’ve fixed them all now. -A

JonesII

Who cares! if afterwards any recorded temperatures will be “adjusted” to fulfill their global warming crazy dreams.

Mark Wagner

Maybe the errors cancel and have a mean of zero
unlikely given that:
some of the error is inherent in the component manufacturing. resistors of a given resistance, for example, are “mixed” with a certain ratio of the various “ingredients.” modern manufacturing processes ensure that each lot is virtually identical to the next, thus any fluctuation in purity would tend to always be off in the same direction. although components made by different mfgs could be different.
some of the error is due to degradation of the physical substance in the components, which would always be the same direction for a particular component type, although they would degrade at different rates due to differences in environment.
probably a few more I could think of

M. Simon (09:29:41) (09:38:41):
“…PRT (thermistors somewhat less so) are subject to stress changes. To detect them a number of cycles from cold to hot (to the hot limit) should be tried to see if there is a shift.”
That is certainly true of PRTs. The technical term is hysteresis. It means that if you cycle from a known baseline temperature, raise the temperature, then lower it back to the original temperature repeatedly, the PRT reading drifts around the baseline, never coming back to the exact same reading.
The effect isn’t great, but it’s definitely there. Hysteresis occurs in all thermocouples to varying degrees. That is why Mil-Spec government defense contracts require regular periodic calibration of thermocouples, RTDs, PRTs, etc.

John F. Hultquist

Anthony, FYI — I just visited the new (Sept. 09) weather station site in Cle Elum, WA (the one in the database now is defunct as it was when John S. did it). He alerted me to the altered-new Lat/Long and suggested I go have a look, which I did. Our families are both dealing with ill family members but we will get this done in a week or so. John H.
REPLY: Thank you for the effort. -A

JonesII

The gulf deep cold…so the gulf current would change in the near future:
http://weather.unisys.com/surface/sst_anom.gif

geo

When they were talking about outside -40C/+40C, I was thinking “that’s not good, but not terrible. . . problematic regularly in some parts of the southwest on the high side, and occassionally in parts of the upper midwest on the low side”.
Then they got to talking about outside -30C/+30C, and I just sighed. . . .

Steven Hill

If it’s low, throw it out. If it’s high, it’s climate change.

REPLY: Yes strange that copy/paste from PDF results in this. But I think I’ve fixed them all now. -A
Still plenty need fixing – at least 19:
In the USCRN, the PRT sensor was strictly calibrated from 2508 to 1508C each year in the laboratory.
This implies that the MMTS temperature observations are unable to discriminate 6 [should be +/-]
0.258C changes in the lower temperature ranges (Fig. 5 and Table 2).
The interchangeability of the MMTS thermistors is from 60.28C from temperature 2408 to 1408C and 60.458C elsewhere (Fig. 4).
versus 22 [should be -2] to 2 V.
The maximum magnitude of linerization error reached over 18C (Fig. 7).
Since this error was constantly varied from 20.28 to 20.38C, it can be cancelled during the onboard calibration.
The self-heating factor for the PRT in the HO-1088 is 0.258C
mW21 at 1 m s21 airflow
corresponding to the selfheating errors 0.58C
Even so, the HMP35C sensor in the AWS network can experience more than 0.28C errors
the RSS error increases from 0.48
in the USCRN network, the RSS errors can reach 0.28

johnnythelowery

Dear Patchy:
Just to say thanks to everyone at Teri for the truck loads of cash over the years. I concluded my work and closed my office. In my grant app. i indicated i’d be studying the connection between CO2 and AGW. Well, I did and there isn’t one.
Cheers Mate.
Johnnny

A – don’t hate me for this, but…
>> 60.28C from temperature 2408 to 1408C and 60.458C elsewhere
there’s still one pesky gremlim wandering around just above the second table…

Brian D

Can this issue get any crazier? Can we get a true temp anymore?
Well it in the low 40’s and snow is melting away quite nicely. At least I think its in the low 40’s. Can’t be sure now. But the snow is melting and it feels warm out. I’ll go with that.

Fred Harwood

And did I read that PRTs are used to calibrate the temperature satellites?

John F. Hultquist

I wonder when the official temperature of the USA, like the time, will become the responsibility of the National Institute of Standards and Technology (NIST)? http://www.nist.gov/index.html
The issues raised by this paper are not going to get addressed in a large network of stations at airports, sewage treatment facilities, other NGO, and the backyards of hundreds of interested volunteers. However, I don’t believe this issue will have legs in the current cycle of politics and the need to find vast sums of money to support the ongoing spending.

Joe

Anthony,
We are putting all the weather data together to understand the worlds climate.
This is absolutely incorrect and will only produce confusing contridictions.
We have actually two weather systems.
The northern hemisphere has the most land mass and population density.
And the southern hemisphere with the most ocean mass and less animal and plant life.
These systems converge at the equator to roll back on themselves back to the poles. The bulge in our atmosphere at the equator is no coincident as heated air rises then moves back to the poles rotationally cooling down.
It is impossible for a hurricane to cross. If it did, it would loose all it’s energy very quickly as the rotational spin would be backwards to it and the energy the oceans are producing across the equator.

Annabelle

Thanks climategate2009 for that YouTube link of Jones giving evidence to the parliamentary enquiry.
It shows climate science in a terrible light – using Jones’s own words. Jones comes across as very scared and shifty – even now he is trying to obfuscate and not answer questions directly.

So, for both monitoring types over the range of temperatures found in nearly all habitable regions, -20degC to +50 degC, the bias is always positive. And as much as +1 degC bias at the higher temperatures.
Now, why am I surprised?
[not that I’m bothered about 1 degC either way – but that seems enough to have the world going crazy wasting a few more trillion on the carbon folly]

Walt The Physicist

When I tried to ascertain the accuracy of temperature measurement and asked Mr. Gavin (Real Climate), the answer was reference to GRL v.28,n13, pp.2621-2624 (2001). In this article it is shown that the accuracy of average monthly temperature is 0.2C/Square root of (60). The 0.2C was determined to be the accuracy of one temperature measurement with a thermometer. SQRT(60) comes from twice a day temperature recording – 60 data points. So, the accuracy is 0.0255C. This is an elegant solution to the accuracy problem – even with a crappy measuring device just make thousands of measurements and your accuracy will be fantastic. All that is ironically speaking, of course.

johnnythelowery

the……..’worse than we thought’ is the new baseline. When plotted, using this new baseline, the IPCC is building a ski jump!

Eddie

not to be a pain but i caught another translation error…
It is not generally possible to detect and remove temperature-dependent bias and sensor nonlinearity with static calibration. In the USCRN, the PRT sensor was strictly calibrated from 2508 to 1508C each year in the laboratory.
REPLY: Its the character formatting of the AMS document that is the pain…never had one do this before. I think I’ve caught them all now. -A

dearieme

OK, so these instruments should be error-analysed as we were taught in freshman physics. Why had it not been done before?

As a couple of people pointed out, error biases all tend to drift in the same direction. Was there any work done based on drift over time? not just the sensors themselves either. The wires they are attached to will oxidize over time and their resistance goes up. Small vibrations transmitted through the earth from a nearby highway produce tiny stress on the wires that also drives resistance up over time. If there is a capacitor in the circuit, its value will also change as it degrades over time.
I know a lot of people don’t believe the vibration thing. Might I remind you that those are human induced vibrations. They are 300 times as powerful as naturaly induced vibrations. This has been known for several decades due to research done by some boys on a beach who published considerable work on good vibrations.

Urederra

[quote]Henry chance (09:37:23) :
Blizzards in Spain. Do wind turbines attract snow? I wonder if the temps in Spain are high above freezing.[/quote]
There were below freezing here at night but not right now, but just for 1 degree. more freezing is coming tonight. I am living where the first wind turbines were placed in Spain. And, well, they are working OK right now. they were placed in a good windy spot.

James Chamberlain

accuracy smaccuracy. as long as it’s hot.

Steve in SC

Couple of things.
1) Thermistors are the absolute worst choice you could make for a temperature sensor for anything other than VERY VERY short term measurements. They (all models) have proven to be very drifty over time and temperature excursions. Nonlinear to boot.
2) The best choice would be RTDs. (you can tell these folks have not dealt with instrumentation very much at all just by their terminology) The second best choice would be thermocouples. Various types (E, K, J, etc)
3) The standard error for RTDs is +/- 1 deg F and for thermocouples is +/- 2 deg F. You can do better than that by on spot calibration of the “SYSTEM” which includes sensor, electronics, readout device, and any wire leads or batteries involved. You can generally expect 50% to 75% better accuracy with calibration. If you really want fine accuracy to .1 C or better, plan on calibrating for every reading.
4) The temperature sensor of choice in the pharmaceutical industry is the RTD. Calibrations are normally at 90 day intervals with some non critical monitoring temps out at 180 or yearly. None, I repeat NONE go beyond that for any reason. If anybody dies because of bad drugs caused by sloppy manufacture, heads roll, money is lost, the ax falls, and you know the rest.
5) While it is possible to resolve down to 0.01 deg C or F, getting inaccuracies down to that level is extremely expensive. (note the proper use of terminology)
6) It looks like (if the table is to be believed) that the ASOS would be the best setup of the bunch.
7) For those wondering the accuracy statement of a good calibrated lab grade mercury thermometer is in the neighborhood of +/- 1/2 degree.
The International Practical Temperature Scale of 1991(IPTS91) would be better than referring to Omega Engineering who are nothing but marketers and resellers. I can probably fish out some good references on how you actually do this stuff if anybody really needs it.

Douglas Hoyt

Instead of using PRT sensors which tend to drift, I would suggest that using sonic temperature measurements would give better results. A pdf file describing the theory is at http://www.wmo.int/pages/prog/www/IMOP/publications/IOM-82-TECO_2005/Posters/P3(09)_Germany_4_Lanzinger.pdf
They use a 20 cm baseline and get 0.3 C accuracy or better. If one built and instrument in the shape of a cross with the sound source in the center and 4 sensors on the outside, one could measure temperature and wind velocity very accurately. With a 5 meter baseline, a 0.01 C accuracy or better should be obtainable. The major error source then will probably be relative humidity and pressure. In order to get accurate temperatures below -25 C, they need to protect the electronics more, perhaps by burying it.
The people making temperature measurements might learn something if they talked to people specializing in metrology.

Jerry

I guess this has been my whole beef with the AGW theory, being an Engineer who uses pressure and temperature gauges in a closed system every day, the accuracy for which is purported never made any sense to me. The ranges for what the temperature of the earth (If there really is such a thing) could be exceeds the purported warming for the last century. No one will ever be able to convince me that measuring temperature is in any way a good proxie for what happened in the past or where we will go in the future.
As Pete in Oh Brother Where Art Thou Said ” That Don’t Make No Sense”