If you want one more inhomogeneity in station records, it seems that temperature measurements have been whitewashed – literally. The Stevenson screens used to be painted with whitewash which is a calcium carbonate with very specific infrared absorption properties. I gather that whitewash isn’t used much any more and so screens are routinely painted with modern white paint, which has very different infrared absorption properties. See here. Maybe Phil Jones should be collecting tree rings instead.
-
Tip Jar
(The Tip Jar is working again, via a temporary location) -
Pages
-
Categories
-
Articles
-
Blogroll
- Accuweather Blogs
- Andrew Revkin
- Anthony Watts
- Bishop Hill
- Bob Tisdale
- Dan Hughes
- David Stockwell
- Icecap
- Idsos
- James Annan
- Jeff Id
- Josh Halpern
- Judith Curry
- Keith Kloor
- Klimazweibel
- Lubos Motl
- Lucia's Blackboard
- Matt Briggs
- NASA GISS
- Nature Blogs
- RealClimate
- Roger Pielke Jr
- Roger Pielke Sr
- Roman M
- Science of Doom
- Tamino
- Warwick Hughes
- Watts Up With That
- William Connolley
- WordPress.com
- World Climate Report
-
Favorite posts
-
Links
-
Weblogs and resources
-
Archives
Climate Audit 
106 Comments
This is very telling, because if we are arguing about tenths of degrees then “measurement” error surely must play a significant role in the basic research. All suuc errors do not cancel each other out.
Eli or Gavin, I forget which, says that for every positive systematic bias he can think of a negative one. However, he never details them and I remain unconvinced. I think that you can only invoke the central limit theorem if there are a very large number of systematic errors that are both positive and negative. Otherwise it isn’t proper to consider that systematic errors cancel. The paint change probably has little or no effect on weather forecasting. There’s also the selection problem. If you only look for and correct errors in one direction, the remaining errors will all be in the other direction.
I can think of a negative bias.
Let’s say you move the weather station further north? “Current knowledge” would have us believe that there would be no impact upon the measured temperatures up to and including a move of 1,200 km. But it is possible that there might be cooler temperatures.
I’m not excited about differing IR properties of paints. Here‘s the IR spectrum of titanium dioxide, for example, which is the major pigment in metal-based white outdoor paint. It has a large absorption near 10 microns. Lime, CaCO3, has a large absorption near 6.5 microns due to the carbonate group.
I’d agree that some IR absorption will occur with any paint, but all paints will absorb some IR in some range or other, and convert that to heat. There may be a differential effect on the thermometer housing, but I’d be surprised if it were very large, from paint to paint — much smaller than the effects of nearby concrete, roads, and air-conditioner vents, for example.
The author at your linked site is probably also wrong in saying that non-absorbed IR goes straight to the wood. He’s assuming there that the paint is transparent to IR. But most of the impinging IR is most likely scattered back by the metal oxide particles of pigment, which are likely to be about the same size (a few microns) as the IR wavelength.
“Possible” and “might” are one thing. When you change the coating on the recording station with something having a different IR absorbing characteristic the results WILL change. There would really be no physical way around it not changing.
I think the asphalt parking lot and sidewalk are the most important features of the photo. When were they constructed and what homogeneity adjustment was made to compensate? It seems the data accumulators are happy adjusting their way to a uniform data set. That says that the apparent T rise has no meaning, even in relative terms.
Further research reveals the following:
I note that, contrary to the original author’s claims, the transmittivity of whitewash is generally higher than that of paint in the NIR/IR …
w.
Seems like the paint change would have much the largest impact in full summer sunlight, yet warming seems to be more in winter and night. This could be a big infra red herring
Any bias potential in the change from asbestos to “laminate”
http://en.wikipedia.org/wiki/Stevenson_screen
#7, Willis
You’re wrong. I think you’re reading your graph wrong. The units of the X axis on your graph is wavenumbers [1 / cm]
Thus, the shortest wavelengths are to the right on your graph [2000 wavenumbers = 5 microns], and the longer wavelengths are to the left [200 wavenumbers = 50 microns]
In the graph you posted, the transmittance of light is lower for CaCO3 between 5 and 10 microns, but lower for TiO2 above 10 microns.
Why such a fuss? Surely the boxes and measuring equipment are standardised? Paint 2 with whitewash, put them side by side and measure temps for a week. Then paint one with “new” paint, re-whitewash the other, and measure temps for a week. Any differences should be obvious.
It doesn’t sound like a hugely expensive task…
#9 The Wikipedia article suggests that the screens should be repainted every 2 years. This could lead to a heavy acculation of paint over 100 years so paint characteristics alone won’t tell the whole story. Anybody want to start taking paint cores?
Building a few screens and testing finishes seems like a good topic for a high school science project.
#8, Murray
And yet architects design homes with thermal mass to keep those homes warm throughout the night and during the winter.
The heat uptake will certainly be higher in summer sun than other times, but I suspect the benefits of being inside a slightly warm box are best appreciated when it’s cold out.
Re #7, Willis, the original link in Steve’s thread had emissivity vs wavelength and wavenumber. Normally, spectral transmittance is assocated with windows and gases. Spectral emissivity makes more sense for a paint specification.
Phil
Re#9
The original US Weather Bureau spec for “repainting” was due to the fact that the whitewash flakes off and washes off in rain easily. If USHCN CWO operators and CWO observers follwed the spec with lead, oil, or latex based paints, then yes tehre will be accummulation. Today I visited a USHCN site that was in pretty bad shape maintenance wise and had latex paint applied over the original whitewash and you can see it here
Also regarding paint pigments being opaque, the University of Arizona has a very convincing photgraphic demonstration of how at IR wavelengths, you can see right through paint layers:
A particularly convincing demonstration of the transparency of pigments to
the infrared is provided in Fig. 6. In this case, (Fig. 6a) an underdrawing is
made using charcoal on a white ground. A layer of paint (Fig. 6b) is then
applied to the drawing which is viewed with an infrared sensitive video camera.
The image, Fig. 6c, of the underdrawing can be seen clearly in the infrared
display.
The same technique is used to reveal historical paintings that have bene painted over on canvasses.
The original U of AZ article that shows the how pigments can be trasnparenet at IR wavelenghts can be found here
See the section “Pigment Response in the Infrared”
Re #15 Wow! Your photos are of this temperature site are classic. Located between pavement and wind-blocking vegetation, with a storage battery inside the enclosure, broken slats allowing additional radiation to enter and of course the paint question.
Have you surveyed other sites, and, if so, what did you observe?
The temperature plot for #15 is here , if it’s the same station.
Looks like +1C over about 30 years, perhaps as the trees grew or they paved more parking area.
I wonder if they cut down the oleander in the 1990s.
Re # 7 etc,
As a chemist can I venture that whitewash was originally calcium hydroxide (slaked lime), which over time can weather to make calcium carbonate. Binders and tints were commonly used also, so the light absorption spectrum of the whitewash might be more complex. Also, wet white paint has about 10% TiO2 by weight plus binders etc, so ditto comment.
Re positive temp bias,
Mercury thermometers commonly used in Australia had two iron pegs in the column, whhich the mercury pushed to max and min temps until reset by an observer with a magnet. Take the max temp. A transient hot event pushed up the peg. A transient cool event did not move it. Systematic one-way bias from hot to hotter.
Another source. If the observer used the magnet before the max and min temps had been reached for the observation period, obvious potential for severe error in both directions for max and min.
Another source known to miners. The vertical natural temperature gradient of near-surface earth is not constant. Some mines get very hot very fast as you mine down. Climate stations positioned above high gradient spots might give different results to ones not far distant over low gradient spots. After all, heat flow is the important property, not temperature as a proxy. I would surmise that NO GLOBAL CLIMATE MODEL CAN SUCCEED until the sea floor spatial temperature distribution is mapped over time.
RE:#17
David I have eyeballed many other sites, but never with such a critical eye as today. I am beginning a survey of every USHCN site in Northern California, and will continue to post them.
Note its not just the battery in the CRS, its the entire electronics package including the data logger, radio modem, and solar battery charger unit. If I remember my electricity laws correctly, devices that use electricity also create heat as resistance is encountered.
I’ll point out that this electronics package was setup by the NWS Sacramento office. Not the university.
Not to be outdone, the Grass Valley, CA Sewage plant hosts a weather station operated by the California Department of Water resources. You can see pictures of it here and here
My favorite parts of the pictures (besides the building, parked car, asphalt on both sides, and upslope heat off the hill, is the sewage digestor vat just downslope. Thank goodness this is not a USHCN station.
Oh but wait, there’s more! (photos below from from J. Goodridge, former California state climatologist)
From Quincy, CA at the CalTrans office, they measure temperature right next to a trash incinerator
And from Mt. Hamiltion, CA east of San Jose, where the Mt Hamilton astronomical observatory sits, they measure temperature on the roof of a building AND next to a chimney!
These sites are all rural?
Yes, they are.
Re #19
It is really interesting to see the data logger, modem etc. located inside the screen. I can see several problems with this approach. The data logger box will block air flow through the box which should cause higher measured temperature because surface temperature tends to be higher than the air temperature. Looking at the amount of electronics in the logger enclosure it seems obvious that the current drain probably is 100 … 500 mA minimum. With a 12V supply (the battery) this would translate into roughly 1 … 5W. The heat from the data logger as such should cause a temperature rise of tenths of degrees C especially on days with little wind or during nights. Should we call this effect EHI (Electronic Heat Island)? I am equally shocked to see the sceen used as storage for unnecessary bottles etc. The reason is again that extra material inside the screen blocks air flow and causes higher temperatures to be measured.
Jim, you say:
Yes, that’s wavenumber, sorry for not labeling the axis. My understanding is that the part of the IR spectrum of interest runs from a wavenumber of say 200 to 1500. Am I wrong? (Wouldn’t be the first time).
w.
#11,
It shouldn’t have been an expensive task, but there is no evidence that anyone has ever done it.
The reason for this is simple. The ground based network was never designed to find the extremely small changes that people are claiming to see.
I think there are at least two questions here. The first is that this is just a point estimation from a sample. The sites in these photos are samples from a population of ideal temperature sites that would blanket the earth – oceans included. Even the popular media typically gives error ranges for poll results. What do you suppose is the standard error of the estimate from these temperature sites? Perhaps .1 or .2 C. If that is so then the news stories should say the twentieth century temperature went up .6 C plus or minus .2 C. After all we wouldn’t accept a headline that said that Hillary is more popular than Obama (39.6% to 39.5%)
The second issue is bias. There is a contributor who is in fact measuring this potential bias. I suggest that everyone wait until the results are in. Storing a paint can in the screen may or may not effect the site’s measurements. Most contributors to this site are quick to criticize the climate alarmists when they base their statements on theory rather than empirical observation – and rightly so. All the above discussion on site bias is theory – just speculation.
It could be that a temperature site is corrupted by a nearby parking lot. Yet in the real world a site that must be periodically serviced by a human will always be near at least a road. Similarly storing a paint can inside the screen box could be a tragic mistake but then again who knows? In the real world such things happen. I await experimental results.
If I have to wait too long I will go out to the garage and pull out my can of titanium oxide white paint. If I go to the hardware store and get some whitewash and a couple thermometers, this theoretical discussion is over.
The ground based temperature measurement is sounding more and more like pathological science, especially the accuracy claims.
I notice that visitors sometimes criticize this form of checking as merely “science fair” stuff. I agree that it’s merely science fair stuff. Lots of engineering details are not high-falutin science but it still needs to be done. When the mystification is removed from the multiproxy studies, none of them are out of the range of a science fair.
Why not not paint the things at all. Get back to nature. I’ve sat on white painted benches and unpainted benches, and it seems the unpainted ones will be cooler on the butt than the white one….. I said butt!
The potential bias from paint may be pretty small compared to the other big white elephant in the CRS I found yesterday.
Most, if not all, of the USHCN stations in California have been converted from manual observing stations (mercury/glass thermometers) to automated stations (Campbell Scientific CR10 dataloggers) that send the data back via VHF radio.
Only one problem:
see these three pictures of a USHCN station I visited yesterday, full report here:
Old manual station on the right, new automated
one on the left
Old manual station on the right, new automatedone on the left
Interior of automated station, note position of
thermometer/dp sensor
Catalog of electronics in enclosure
CR10 datalogger, DC voltage regulator/solar
battery charger, VHF radio modem estimated at max 5 watts ERP
I went looking for paint and found a toaster.
#24, Willis
I don’t know if, or what portion of, the 200-1500 wavenumber [~6.7 – 50 micron wavelength] range is most critical. I’m just saying that your curves don’t prove that the transmittance of CaCO3 is higher than TiO2 in the near-IR, as you stated – or tend to disprove the claimed possibility of measurable bias. As always, your curves looked great, by the way.
I do know the solar irradiance drops off considerably with wavelength, so we’d expect more watts of solar IR at lower wavelengths. The NIST site you suggested shows lower transmittance for CaCO3 between 1000-4000 wavenumbers [2.5 – 10 micron wavelength]. This suggests a TiO2-coated screen would allow more solar radiation in. A book I have on Optical Materials lists near-IR as being on the order of 1 micron. Wikipedia lists near to mid IR as being below 10 microns.
The 200-1500 wavenumber range you mention would seem to be about peak emission for a room temperature object, so if you’re concerned about thermal radiation from the screen to the thermometer, you may be right. An internally CaCO3 – coated screen would seem to more easily radiate energy to the thermometer. An externally CaCO3 coated screen might also transmit more radiated thermal energy from a nearby heat source like asphalt, concrete, or an incinerator to the underlying wood. Is radiation the primary mechanism of heat transfer in either case ??
I think the transmittance is not the whole story, however. What happens to the light that isn’t transmitted ? Is it absorbed or reflected ? The reflectance of IR light seems key to me. I don’t see a significant difference in effect between the heat being absorbed by the paint or transmitted to the wood underneath. Total absorption / transmittance will vary with coating thickness, but reflectance will not [since it’s a surface phenom…]. The better coating for a screen will be the one with the higher reflectance, and if there is a significant difference in reflectance the originator of this discussion may be on to something.
#23, Lars,
I think your average power estimates are way too high. This system should spend most of it’s life sleeping with a very low duty cycle. Most of the power will be dissipated by the transmitter after the sample is taken. Peak power may be in watts but average power is probably in milliwatts.
Anthony, next time you survey one of these sites can you try and find out the battery size and time between charges? This would allow us to bound the power dissipation. Battery voltage before and after replacement would allow a fairly accurate calculation. This information may be collected and transmitted to the basestation.
ed
Anyone know when the paint was switched? If the new paint absorbs more or less IR heat at it’s given environment? f the difference in absorbance is enough to change the temperature reading inside? If there was an impact on the temperature reading, has this not already been accounted for and corrected in the data?
I think it would be hard to conclude that the paint actually makes a difference if we don’t know the answers to these and other questions.
#32
Don’t forget that the battery is not 100% efficient – it also generates heat. The lead-acid battery in the picture is filled with water so it could also retain a small amount of heat absorbed during the day and release it at night.
RE32 Ed,
I think the solar cell dc power regulator used to keep the battery charged is the largest heat generator, any time power regulation is involved, heat is generated.
I’m checking other nearby stations first to see if this placement is just a fluke, and then if its not, I’ll try to quantify the heat/power issue from the electronics.
I’m looking into getting a FLIR camera to get a picture of the inside and outside of the CRS to see just what is going on, I’m willing to bet an IR photo of a Stevenson Screen doesn’t exist yet.
Anthony,
I really admire your diligence here. Do you have a list of all the california sites. I think I could make some visits
for the project.
I might be able to help on test design ( did this once upon a time) and donate time and a few bucks.
The Bottom line: you want to IMPROVE the data collection/instrumentation methodology.this is not about “exposing a Bias” .
this is about improving the data. At least from my perspective it is about better data, better models, better predictions.
Anyone know how long it takes to warm up a lead/acid battery and for it to cool down? Is it possible that having an huge heat sink in the box may just elevate the minimium temperature a tad?
Another point, will the temperature response vary depending on the wind direction? The slats are designed to allow air to flow through the box, from the picture it appears that the movement of air will be hindered in one pair of directions, but unimpeaded in the the dimension, due to the size of the huge AC/DC converter.
Surely, all these boxes are supposed to be standardized?
37: That is a very good point. There’s a lot of mass in a lead-acid battery!
A few details:
CRS cotton region shelter, another name for a Stevenson screen.
Another kind of electronic replacement for Stevenson screens and liquid
in glass thermometers may be seen on the post next to the house at
At the USHCN site at ftp://ftp.ncdc.noaa.gov/pub/data/ushcn/
there is a station.history.Z file (which is now over eleven years
old) which lists some specific location information. The README.TXT
file describes the layout of the station history file.
Re: 36.37
Steve, thanks. My whole point here is to determine if the screens have contributed an error, and if so, how to quantify it. Once quantified, then a correction can be applied (hopefully). But I should warn everyone that from what I’ve seen looking at these stations through my career, standardization doesn’t seem to apply very well in placement, instruments, calibration, and maintenance. The intent of the network as designed by the USWB at its inception was to provide surface obs for forecasting, later for pilots. Climate was a secondary or tertiary consideration, and I think MarkW said it pretty well with “The ground based network was never designed to find the extremely small changes that people are claiming to see”.
I’m open to collaborative help. Thanks for the offer.
Doc, re37, I don’t think the battery goes up/down in charge that much during the day, except at night. I’m familiar with the solar panel chargers, and they will create heat during the day when regulating the solar panel power. At night the battery will act as a heat sink for sure….
So many exposure variables, and no clear way to quantify them all yet. One thing is clear though, greater entropy seems present that we thought possible. I’ll kno more once I complete site surveys. I have a complete USHCN California station list, and I’ll probably get them all done by the end of summer. My goal is a complete USHCN photographic and surey notes site database.
Anthony,
I might get up around Chester ( north of Quincy )in the next month or so, or at the latest in August. I think there is a
site there from the map ( http://www.ipm.ucdavis.edu/WEATHER/SITES/plumas.html)
it appears to be at the local airport.
Anthony Watts says:
“Doc, re37, I don’t think the battery goes up/down in charge that much during the day, except at night. I’m familiar with the solar panel chargers, and they will create heat during the day when regulating the solar panel power. At night the battery will act as a heat sink for sure….”
I was just wondering how how much energy it takes to heat it up and cool it down whenthe external temperature changes during 24 hours. During the night the battary will radiate heat, just because it contains 10’s of pounds of lead. I have no ide what the heat capacity of a standar lead acid battery is, but myguess is that it will be substancial. It will not have much effect on the maximium temperature, but should have a big impact on the minimium temperature.
huts when did they start putting such large heatsinks in the boxes and did the changeoverresullt in higher winter minimium temperatures? Do were have pairs of nearby boxes, say just across state lines, when one site was updated and the other not ? Can we get a list showing when thechange frommanual to electronic was made and look at the per/post temperature record.
Seems to me we need a few controls: properly built Stevenson enclosures erected in appropriate locations as close as possible to a randomly chosen selection of current stations (and another test set similarly erected near stations suspected of being ‘off’).
Measure temps for a period… and discover the size of the uncertainty. For this round, don’t even imagine “correcting” — most likely, the number of variables involved is higher than the forcings of a GCM 😉
Well, more info. I find:
SOURCE (big pdf)
Not a whole lot seems to be published about the IR reflectivity of whitewash.
w.
Academics don’t want to do the stuff they call “normal science”. That stuff doesn’t make it into Nature. Instead, it’s derided, much as people are doing here. There is a cost to investigating the quality of the world’s weather stations. Yet no one (until now?) is going to give you a grant to study THAT.
That’s the business of science. You want to leave that in the hands of the academics?
I work in an accredited (commercial) testing laboratory, working to published national standards. We’d never be allowed to get away with the shoddy data collection techniques that are being exposed here.
If there is a standard for these screens and for their re-painting, then these must be followed to the letter.
The only conclusion I can draw from this is that the methods used to collect surface temperature data are not fit for the climatological purposes for which they are being applied. Surely one of the things that all the climate change research money should be being spent on is a purpose-built network of temperature sensors that will produce reliable data.
The claims of Jones et al. and Hansen that they can determine global changes in the realms of tenths or hundredths of a degree just cannot be validated because of the precision and accuracy (or lack thereof) of the data on which their reconstructions are based.
Re #42,
steven,
At an airport you may not see anything resembling a Stevenson screen, but
something called an AWOS, or ASOS, unit. (automated weather observation
system, automated surface observation system). They may not all look
alike.
One photo is at http://www3.amss.nws.noaa.gov/amsstt.nsf/KRIC.jpg
and a couple are at
http://www.nws.noaa.gov/asos/index.html
You might be cautious about taking pictures of airport grounds in case
someone thinks you are casing the joint. 🙂
Re #32
Without beeing able to check the equipment in question it is difficult to say. If there is an interest in the topic I propose that we try to identify the different components and try to estimate the power requirements based on manufacturer specifications.
I start with the easy component:
– Vaisala combined humidity and temperature sensor. The power requirement is ca. 48mW calculated from assumed 12V feed. The information is taken from Vaisala Oy’s web page in Finland.
– The data logger at the top of the encloseure with wireing connectors is made by Campbell Scientific Inc. I don’t recognise the exact model. If somebody recognises the device please post information. Modern ones draw ca. 10mW at idle and 100mW when active. My guess is that the lower walue is probable.
– The device below the data logger is probably a battery charging controller handling input from the solar collector connected to the antenna. From the size of the solar collector I’d guess it is 5…10W, probably closer to 5W than 10W. Information about this device would be interesting to have. Small solar charging controllers often regulate charging by dumping excess power through a parallel load resistor. This device could in the worst case provide several W of heating. Assuming a 5W solar cell the average power consumption could at most be 1…2W I think in order to ensure that the battery stays reasonably charged. In order not to damage the battery the charge level should never go below 60%. An upper limit on the power drain assuming the battery discharges during the night would then be roughly 10W. Probably the power requirement is perhaps 25% of this to ensure that overcast days don’t cause problems –> 2.5W max total?
– The device unit behind the sensor end of the Vaisala probe is probably a radio modem of some kind. Info anyone?
– I’m not able to identify the device below the radio modem. My guess is that the one connected to the data logger through a narrow flat cable is a micro controller that reads serial data from the logger and transmits it over the modem.
– White unit at the bottom?
Any better ideas?
RE:49 Lars,
I’m going back to the site to get an electronics equipment inventory.
I’ll also get an IR contactless thermometer to see if I can get surface temps for the devices.
Bear in mind, we only have this one station so far. Its possible it is a fluke. I’m checking others on the Sac Valley floor this weekend, and I have a volunteer checking two others in the Sierra Nevada.
I’ll post new info as I have it.
Question: Does anybody have access to a FLIR digital camera? Getting one is not inexpensive. Typically $7000 – 10k+
#50. John Christy is from this area and recently re-collated temperature records from stations in this area showing a contrast between temperature histories in the valleys and higher up, which he attributed to irrigation effects.
#50, Anthony
Make sure you check temperature of the screen, ambient air, and electronics utilities box surface either before sunrise or at night – when one would expect the greatest confounding potential.
Re 37, 43
Respectfully,
The battery and other material inside the shelter won’t affect the temperature maxima and minima when considered as a “thermal mass” except to shift the peaks and valleys in time with respect to the exterior of the shelter. The battery will thermally act similar to way a capacitor works in an electrical circuit. The interior environment of the screen will hit the same highs and lows as the exterior it will just be later in the day. The time delay will be a function of many variables including but not limited to airflow through the screen and thermal resistivity of the materials involved.
I’m very interested in hearing the results of the “science fair” experiment.
RE:51 Steve,
I was in touch with Christy about a month ago, and I didn’t get the impression he did a hands on survey.
From his paper on California Central Valley Stations which you can access here
“We accessed from the
National Climatic Data Center (NCDC) daily maximum
(TMax) and minimum (TMin) temperature data
for all stations within the six counties of the central San
Joaquin Valley (Mariposa, Merced, Madera, Fresno,
Kings, and Tulare) from the valley eastward.”
Re: 53 I would agree if the “thermal mass” was passive, but its not, its active.
Energy is being added to the system every day because there is a solar photovoltaic panel driving the electronics, battery charge regulator, and the battery. During the day, energy is added, at night it is released. In both instances there is energy dissipated inside the shelter in the form of waste heat by the electronics.
If it were passive, just a mass, I agree it would be a time shift. In the case of an active energy addition, there will be an amplitude shift.
Re #53 and #55
For a passive system it won’t just be a delay, it will be smoothed as well provided the thermal mass of the battery is sufficient. The first thing you add to an electronic circuit to filter the high frequency noise is a capacitor. It has the potential to mask a transient low temperature minimum.
Having read the article about the Stephenson screens I decided to have a look at the Brogdale weather station. This is where the highest temperature in England ever, was recorded on the 10th of August 2003.
I have to say I was quite shocked to find it is a metal screen, white on the outside and black on the inside. I am no expert but I would have thought that metal would distort the results in an upward direction compared to painted wood.
It then begs the question how many more of these are being used in England or elsewhere. I don’t know of course how long ago this was changed to the present screen or if any work has been done on the differences to the traditional screens.
Fascinating discovery Chris, I was not aware metal was used.
Raises many issues as you say.
Warwick Hughes
#43 Doc,
The battery in question weighs ~13 lbs.
#57. Chris, can you supply a digirtal photo of the site?
#49 Lars,
I had missed the solar cell initially and thought this was a total loss system. I agree that the solar cell controller is the largest heat source. Hopefully it is colocated with the cell and not in the box. I still believe that the sensor, logger, controller, modem, and transmitter could live on a 100 mW average power budget.
In response to the large amount of interest I expect on this issue, and to accomodate the need for a repository of global surface station qualitative data, site surveys, and photographs. I will be setting up a website called http://www.surfacestations.org that will provide a downloadable site survey guide, and a way to upload the data and images.
I have acquired a database engine for this purpose, and I am in the process of implementing it.
for a model page see this http://www.intelliweather.com/ushcn_sites/CSUC_farm.htm
Re #60 Steve
I do have three photos of the site and the screen but my computer skill don’t seem to be good enough to post them, any tips would be welcome.
Since posting #57 I have found out more about the screens, they seem to be made of plastic and Aluminium they can be found at:
http://www.ukweathershop.co.uk/acatalog/Instrument_Screens.html
The following is the write up for the screens:
Manufactured and tested to UK Met Office standards this screen will meet the most demanding technical requirements. It is constructed from high performance plastics and polyester powder coated aluminium that will not degrade with time in the same way as more wooden types.
RE63 Chris use Flickr or Photobucket…you can upload pix and get a URL to post here, and the service is fre, just register.
Re #64
Thanks Anthony hopefully this will work
” Derek McQ says:
May 11th, 2007 at 10:39 am
Re 37, 43
Respectfully,
The battery and other material inside the shelter won’t affect the temperature maxima and minima when considered as a “thermal mass” except to shift the peaks and valleys in time with respect to the exterior of the shelter. ”
I suspect that a huge heat sink will not work like that, if you acknowlege that air temperature and dew-point are intertwined. You will get less dew inside the box as outside, as the transition from vapor to dew will not happen inside, i.e. Water vapor will fall out of the air outside the box, if it has an IR source inside.
This will mean that the box is dryer than the outside, first thing in the morning, and so will heat up more quickly. The temperature will fall more slowley at night, and the bottom of the “V” at the true minimum will become more “U” shaped. This means that the absolute low temperature will be higher.
I checked the website for this battery. It is indeed 13 pounds. More over, this is a new model, and has replaced the PS12180. That model was made from white/gray plastic, rather than black.
(spec for the one in the pic http://www.the-powersource.com/UBBat%20specs/UB12180.pdf)
Now I understand that the mmGW people believe that the rise in temperature they can measure over the last two decades is on par with an increase in energy of 4 w/m2. The spec sheet says that if the battery is discharged over the course of 20hr, at 10.8 volts, then it is rated at 9.9 W.
Here is a profile of the present daily temperature and humidity at the site.
http://www.wrh.noaa.gov/mesowest/getobext.php?wfo=sto&sid=CICC1&num=48
Thanks Chris,
Quite clean, this station, so I assume its new. The question is did this replace an existing shelter?
Puzzling that the UK Met office would diverge from the WMO spec on Stevenson Screens and approve this design (according to UKweathershop.com). Can you muck around and find out if they have any characteristics data or trial studies leading up to this approval?
Re #67
The site has been a weather station for many years as it was home to the National Fruit collection, part of the Ministry of Agriculture. As I was a Fruit Grower I visited the site many times, it was a conventional Stephenson screen. I have not been able as yet to find out when it was changed.
The manufacturer’s of the screen say that trials were done by the met office and that they are part holders of the patent, but I have not been able to track down the trial results.
Anthony,
Can you get a copy of the FMH-1B handbook? It is the Bible for federally certified weather observers. Many moons ago, the USAF and Navy used it to train thier observers. It stipulated all types of rules and procedures when taking weather observations. It devoted an entire chapter to taking accurate dry and wet bulb temperature readings.
I’ve noticed that now has an entire network of ametuer observers. Our local high school has its mini-weather station situated on the top of the school’s tar roof- situated next to two exhaust blowers coming up from the boiler room. I hope this crap doesn’t make its way into the GISS or HADCru databases.
Re #65
Chris,
Nice view of Brogdale – thanks.
In the second picture the screen is open, yet the third shows it padlocked. How did it come to be open in one and padlocked in the other?
I must say it looks like an excellent screen, I wish I had one.
Re #70
No mystery, a padlock one side just a clip the other
Well, I wouldn’t go around opening other people’s Stevenson Screens, but this is the home of self appointed auditing so your in good company 🙂
I don’t have the Federal Observers Handook, but I have this one, and I think it was more widely distribued to CWO volunteers.
Re #72
Peter Hearnden,
That honor appears to be shared with Risk Management Solutions, Ltd.
#69 JP
“Can you get a copy of the FMH-1B handbook?”
Is that FMH-1-2005 is available online here: http://www.ofcm.gov/fmh-1/pdf/FMH1.pdf?
I scanned through it and didn’t see much detail on equipment but maybe it isn’t the right document.
Ed
I have made a second visit to the Chico State University Farm USHCN site and obtained additional detailed photos of the electronics to clearly show model numbers, etc. Additionally, another nearby feature is a large irrigation system to the soutwest.
http://www.intelliweather.com/ushcn_sites/CSUC_farm.htm
Re:#62
Anthony, thank you, that sounds like a great idea! On the site page, might I suggest adding a link to the site location in Google Earth (for convenience’s sake)?
Re:#63
For those interested in posting pictures, might I suggest you also consider using Panoramio.com? They allow you to upload up to a total of 2GB of .jpg photos (5MB max individual photo size IIRC), and although the photos are normally displayed in smaller sizes, it’s easy to have them display in full original size for downloading/detail viewing. The really nice feature is that they have a Google Earth-like interface that allows you to tag any/all photos with a geographical location, then (for example) browse by location.
Elsewhere in Steve’s threads I have been pushing the line that simple, well-thought modern experiments can do a lot to quantify errors. Re the screen with electronics inside, it is easy to place a mockup or actual one indide an insulated area like an old cool room, then run it a few days on and a few days off, with a mechanical temp recording device. If there is a measurable temp difference between electronics on and off, then there is an issue . If not, why try to calculate it from first principles when people would only argue about your assumptions and methodology?
Seems to me that there is a crying need for more micro-experiments. The UHI question could be answered by a network of identical stations filling a 5 deg grid cell containing a city. Same setup would allow then using interpolation mathematics to see how far apart they can be before they lose connectivity.
Some very important papers coming to light are micro experiments on (a) variation in temp at a site that is close to the ground or elevated and (b) variation in the shape of the temp record graph when readings are all taken at 3am or 6 am as opposed to max and min.
The old mercury-in-glass thermomenters that used to be used in Australia in the 1950s had metal pegs that the mercury pushed up to a maximum. They were reset with a magnet dragging them back to the mercury. As I’ve written before, a transient hot event would raise the peg, but a transient cool event would not lower it. So there is a simple example of instrument bias that makes hot hotter than it should be.
Finally, another simple micro experiment. Integrate a number of temp measurements over each day for a period of months as a heat flow proxy, then compare with max/min temp data. Be sure to correct for daylight saving time in observations.
#72 Peter H
Good to have you back. Where have you been all this time? How is the farming coming along? Expect your looking forward to 2007 being our new record highest CET year here in the UK?
KevinUK
Geoff, perhaps you’ve seen this, but the Barrow, AK UHI study seems well-thought out & implimented. This is on a town of 4600 in the Arctic on a level, windswept coastal site.
For some real whitewashing…http://www.sciam.com/print_version.cfm?articleID=0000A3DE-EF34-1CCF-B4A8809EC588EEDF and read about the discussion of coral. Sorry if already posted. Did not see it in this thread.
Re #80, Kevin,
I read CA every day. But, tbh, I don’t find I feel the need to post much anymore. Same old same old I guess.
Farming? Difficult. Weather? Nice try 🙂 but I don’t welcome warm records. Of course 2006 was, by some way, our warmest year in at least 300 years. It really would be something else if that was surpassed this year (though the 12 months to April end were another warm record).
That’s a shame for someone living in a post-glacial era.
The problem with Barrow is the vampires.
http://en.wikipedia.org/wiki/30_Days_of_Night
very nice study though.
Re: 83
Peter,
In spite of P D Jones “data” to the contrary, California’s Central Coast was cold enough in 2006 to push back the grape harvest by 3 weeks. This was also in spite of the media hype over 2 warm weeks in July, 2006.
Here in Chungnam Province, S Korea, we are still waiting for spring.
Re: 81
Beng,
Interesting study.
re 79:
Another useful experiment would be the comparison of north wall thermometers vs. stevensonhut thermometers.
This was the location of the dutch thermometer between 1734 and 1850 (north wall, outside central window, first floor above ground (= american second floor(!))
http://members.lycos.nl/ErrenWijlens/co2/zwanenburg2004/
re # 81,
Thank you beng. This is precisely the type of work that should be done before global assumptions are made. I had not seen the paper. It is important and illuminating in the context here.
#73
I remember an entire chapter devoted to procedures and Best Practices in taking, calculating both Wet bulb and dry bulb temps. When I was in the service, we used the TMQ-11, but when out in the field, we had to manually take these readings. The 2005 version of the FMH appears to reflect a more modern approach. For instance, in the past, the FMH stipulated that if the observer was making a manual temp/dewpoint reading, that he must take at least 3 dry and wet bulb readings within 10 minutes of observation time. If it was sunny, the readings should be done in the shade, and the observation point must be at least 10m from any buildings. Temp observations should not be taken from rooftops, over concrete roadways, etc…
I know there were also some “sanity checks” built into the process, such as rapid temp increases/decreases without any change in clouds, winds, precipitation, etc… Normally, the chief weather observer (a position that the USAF did away with some 25-30 years ago) would ensure that his team of weather observers made quality observations. Weather observing used to be position filled with professionals who took great pride in thier craft. Unfortunately, the field of professional weather observing went the way of the reel-to-reel and 8-track.
When I was in the USAF, I was forced to move up to weather forecasting only after 2 years in the field. Most USAF weather observers are forced to go to forecasting school after they reach E-5.
To help Anthony Watts, I have put up a site survey for Lake Spaulding at NC Media Watch.
And, I have a rudimentary version of http://www.surfacestations.org up and running with two station reports.
I will be preparing a “how to” fill in the blanks step by step guide to enable anyone with reasonable observation and photography skills to go out and do site sureveys and send the results for posting here.
The goal here is to provide a database of quanitative and qualitative data though site surveys so that researchers can asses the value of the climatological and meteorological data gathered at the site.
While I plan to visit as many stations as I can, it is a monumental job, and I will need the help of hundreds, if not thousands of people to do surveys.
I will be installing a SQL based database engine to help with the task as well as to provide search features.
I welcome suggestions on all levels.
re #81,
How helpful! A carefully constructed study of a still-rural area (by population measures) — pop went from 300 to only 4600… and even with that they were able to demonstrate a several-degree UHI effect.
Searching for citations of the above study (didn’t find any among the Hockey Team, too bad…), I found
another study, of Oklahoma City, a more urban (and much warmer!) location.
Interestingly, the Oklahoma City study found a strong nocturnal UHI effect, somewhat offset by a very weak daytime UCI (Urban Cooling Island!) effect. (They interpret this as largely due to thermal lag: the rural area warming a bit quickly in the morning, and cooling far more quickly at night.) Overall however, they again found UHI of 1-2 degrees C.
Re: #50, #61
I don’t contest other estimates. The best thing is I think to do an inventory on site if it is possible. A list of devices to allow estimates of total power consumtion will be interesting to look at though. I hope Antony is correct about the power consumtion :). I think the most important device to locate and check is the solar cell battery charger/controller because depending on the design this could easily be the device generating most heat. Hopefully it is located close to the solar cell and not inside the enclosure. This is a very educational case! I look forward to get more information.
92: Anthony, are you interested in areas that are not in CA?
re 95 I will be makign it possible to post surveys for the entire GHCN and CWO network on the http://www.surfacestations.org website.
#79
re max-min mercury thermometers – didn’t you have the ones with U-formed capillary? Then the max temp iron bar was in the right arm and the min one in the left arm.
We still have one of those from 60’s in our 24 C incubation thermostatic chamber to check if the air-conditioning works properly.
Anthony Watts has a great post on the worst USHCN site he has found so far at Watt’s Up With That. You will not believe how bad, untill you see the cell tower airconditioner is blowing on the temperature sensor.
Omigosh.
Russ:
Great site, great story, great pictures, a great comparison site (matched roughly for mean temperature) and a
$64K question at the end. I will look forward to the full compilation of Anthony’s field visits.
#98, 99. That’s sickening. Compare that to Jones and Hansen’s claims that they’ve allowed for everything. I’m going to do up a post to draw attention to this.
This has 60 Minutes written all over it! Think of the video. Think of the effect of Leslie Stahl walking away from the current box with a portable highly calibrated temperature gauge at the ready and reading off the dropping temperature as she walks to an open grass covered area! I hope Anthony Watts is photogenic. The visibility could do a lot to shake loose the missing station location data.
Watt’s up asks “How does bad data like this slip into the NASA GISS model?”
Steve per DQA, NASA had some of the best written guidelines. Perhaps someone such as Anthony Watts should fill out a data quality complaint for this data site under the ACT. I would be glad to help. In fact, the details of meeting NASA’s guidelines should be a template where the changing data could be entered easily for each site.
From Wiki., the text of the DQA law
The guidelines under subsection (a) shall ‘€”
(1) apply to the sharing by Federal agencies of, and access to, information disseminated by Federal agencies; and
(2) require that each Federal agency to which the guidelines apply ‘€”
(A) issue guidelines ensuring and maximizing the quality, objectivity, utility and integrity of information (including statistical information) disseminated by the agency, by not later than 1 year after the date of issuance of the guidelines under subsection (a);
(B) establish administrative mechanisms allowing affected persons to seek and obtain correction of information maintained and disseminated by the agency that does not comply with the guidelines issued under subsection (a); and
(C) report periodically to the Director ‘€”
(i) the number and nature of complaints received by the agency regarding the accuracy of information disseminated by the agency; and
(ii) how such complaints were handled by the agency.
OMB, Guidelines for Ensuring and Maximizing the Quality, Objectivity, Utility, and Integrity of Information Disseminated by Federal Agencies, Final Guidelines (corrected), 67 Fed. Reg. 8452 (Feb. 22. 2002)
Perhaps this would interest 60 Minutes even more if NASA refuses to address such an obvious data problem. A federal agency refusing to obey federal law. Think of all the other potential faces in such a show.
Good idea, John. Make it NASA’s problem. Why don’t you post up a template and maybe Anthony or someone else will pursue? I wouldn’t mind seeing a template as well, since other cases are going to crop up/have already cropped up.
On second thoughts, Leslie Stahl carrying a portable reader will not work. She or any human would probably give off too much heat.. Best a series of boxes at regular intervals showing a trend as you move away from the hot spot!
I have just contacted NASA by email. I am trying to get their complete guidelines (if existant) for meeting the OMB’s DQA.