More bad siting in Australian weather stations

WUWT Commenter “Boy on a Bike” was inspired by Willis article on Darwin (See: The Smoking Gun At Darwin Zero) to have a look at stations in his part of the world, he didn’t have to look far. He’s found what we’ve been saying for years on WUWT. Just have a look in our Weather Stations Category. One notable example, Lampasas, TX

He writes:

UPDATE: The writer has misidentified the lighthouse shown in the photos below. it is actually Cape Leeuwin Lighthouse, not Cape Naturaliste. Not being familiar with Australian lighthouses, I did not catch this right away. But I have verified that it is Cape Leeuwin in Google Earth. The siting issue of the weather station at Cape Leeuwin remains a concern. I’ll add the temperature record at Cape Leeuwin as soon as I can locate it. UPDATE: Graphs have been posted in comments – Anthony

======

Australian weather records – how much can we trust them?

After reading a few articles on weather stations in the US and Australia over the last year or so, I decided to do a random check on one station in WA [Western Australia] to see what the records looked like.
I chose Cape Naturaliste, as I have visited it several times, and the lighthouse has been there for around a century. I guessed that the weather station would be at the lighthouse – and I was right.
One glance at the annual mean maximum temp from around 1900 to today would convince most people that we are all about to fry. However, I decided to have a look for photos of the site to check out the location of the weather station.
The weather station is located between the two buildings on the far right – and it is about a foot from an asphalt road. Note that the asphalt is a dark black colour – the colour of fresh tar.
Here’s a side on view.
If memory serves me correctly, when I visited this place back in the 1980s, this was a gravel road – not tarmac.
If the weather station has not been relocated since 1901, what impact would moving from a gravel road to tarmac have on temperature readings? Undoubtedly, it would skew them upwards. The key question is when the tarmac was laid – or relaid – and what impact this has had on the temperature record.

88 thoughts on “More bad siting in Australian weather stations

  1. Nice job, the answer is it has absolutely no impact. The IPCC has determined that only positive adjustments are required to temperature data as it is simply impossible that local additions of asphalt, air conditioner outlets would have any effect whatsoever on thermometers.

    In fact, I read somewhere that all thermometers are going to be replaced with clocks. The clocks will count the seconds since installation and this will be used as a true temperature measure. Nearby tree rings will be used for verification. Then scientists can get rid of all this denialist rubbish about ‘heat affecting temperature’ and such.

  2. @John Hooper If I understand it correctly, the Stevenson screen will block radiant heating. The asphalt actuayl accepts the radiant energy from the sun, and turns it into heat, with then heats the air. A Stevenson screen can not stop the effects of artificialy heated air. That is why there are rules for placement, and the point of the Surface Station audit (to determin the quality of the placements). See http://go2.wordpress.com/?id=725X1342&site=wattsupwiththat.wordpress.com&url=http%3A%2F%2Fwww.surfacestations.org%2F

  3. “What’s the point of the the Stevenson screen if not to counteract that?”

    To stop birds nesting inside.

  4. Are you sure were are looking at the correct photos here?

    From the linked page’s links I would say these photos are of Cape Leeuwin rather than Cape Naturaliste.

    I could be wrong of course ….

  5. The plot is too small to see the dates but it would seem that it departs from almost a flat line to rise up again in 1960… hide the decline trick temperature adjustment maybe?

  6. Lets find someone to go there and paint that tarmac in white! I bet we’ll see at the news that the Ice Age is comming, and that in 10 years that lighthouse will be half in ice.
    The alarmism does great tv ratings!! ;]

  7. This is an unfenced road and it looks nice in the pictures but is it inundated with tourists who would park along the road or would people visiting the nearby buildings tend to park along the road here?
    It seems to me that parking could be the big unknown here and that may also be influenced by nearby buildings that do not directly influence the temperature measurement but if one suddenly changed from an occasional storage centre to a works canteen, for example, it would influence vehicle behaviour.
    The idea of being away from roads and car parks is just as much about insulating the measurement from traffic as the road surface itself. and traffic would be the most difficult to compensate for.
    It could be that a scenarios such as I have described could be difficult to detect or identify.

  8. OT. The Cryoshpere Today says that they still are unsure of current ice imaging due to satellite problems. Does anyone have an accurate analysis of current arctic ice extent?

  9. If one is looking for GLOBAL averages does one need any adjustments at all? IF stations are subject to random change – eg moves up or down hill, or from fields to tarmac or vice versa, wouldn’t it be better to just use raw data? (My guess would be that any net bias would be over recording due UHI, more tarmac roads, more air con outlets, etc)
    If this is right, the use of adjustments is a cure worse than the disease ( which is error margins around the averages used ?
    My reason for asking is that I’m trying to persuade folk re AGW over hype and so far just get shouted down, mostly!

  10. PS
    Annual mean maximum temperature?
    What have other charts shown?
    annual mean average temperature?
    Is this the “value added” data or raw data? If “value added” then we seem to have the fingerprint post 1960 growth correction here where each year more correction is added.

  11. John Hooper (09:32:26) :
    “What’s the point of the the Stevenson screen if not to counteract that?”

    The screen is to shield the instruments from direct exposure to sunlight, so they measure ambient air temperatures rather than direct radiant heating. It doesn’t counteract the ambient air getting additional heating from the micro UHI.

    I have a thermometer on my porch (not screened; I’m not doing science). When the sun reaches the point where it is shining directly on the thermometer, its reading jumps 20 degrees.

  12. With google earth it should be possible to locate those few, those valiant few, weather stations that have remained far from any contaminating artificial heat effects. Go beyond just checking whether roads, buildings, etc., have changed during the data record — because it is always possible that the way these roads were travelled on and buildings heated have changed. Then, once this selection has been made, check what those obviously raw and uncontaminated data series show about warming. Now, that would be interesting …

  13. Re: GP (09:50:44)

    I agree – the photos on the original page are listed as being of the Cape Leeuwin site. It’s not clear from either the radio amateur photos or from the “Lighthouses of Australia Inc” web site where the Stevenson screen is located at Cape Naturaliste.

    Hopefully someone will identify the location.

  14. The buildings around the station are not 109 years old, so it is reasonable to assume they were built after the station was in place. Also likely, the roadway was the width of buckboard and obviously gravel. This would have placed the roadway further away from station. And let’s not forget the concrete sidewalk leading up to the antenna as being an add-on sometime after 1900. It’s no surprise the temperature record is showing significant heating.

  15. From a quick Google Check that is Cape Leeuwin:

    recent picture of Cape Naturaliste:

    I don’t have as trained an eye as many here, so maybe the Cape Naturaliste station is visible here:

  16. Science. Doncha love it? Once question to the group. The “white paint” on the thermometer structures likely reflect most of the light in the visible spectrum, since I see them as white. BuT… wouldn’t a perfectly reflective and emissive surface be preferred. A coating that reflects all light including infrared? Watt is the absorption spectrum of the paint on the boxes. Watt is the emmissivity? …just to pile on.

  17. GP (09:50:44) :

    Are you sure were are looking at the correct photos here?

    From the linked page’s links I would say these photos are of Cape Leeuwin rather than Cape Naturaliste.

    I could be wrong of course ….

    GP is right. These are photos of Cape Leeuwin.

  18. Anthony writes:
    “Undoubtedly, it would skew them upwards.”

    Is there no doubt about that? It’s an honest question. I assume the argument is that it can’t skew the downwards, so there will be a skew upwards, however it may be to small to matter?
    Which brings me to my next question: Can’t this type of effect be easily measured in a controlled study? Sure, it wouldn’t be something done in an afternoon, but surely by empirical testing, some function of distance and surface area, etc, could be made, no?

  19. Hmmm … near antenna (appears to be a horizontal quad) is lashed-up to a pipe sticking out of the ground using rope …

    Far antenna support by an X-frame on placed on the ground.

    Appears to be a ‘field day’ setup (this may all be redundant info).
    .
    .

  20. For the station in Windsor Ontario, the last 11 years is all adjusted data but no unadjusted. Weird.

  21. Those picture are from the Cape Leeuwin lighthouse, not the Cape Naturaliste lighthouse. Both are in the Leeuwin Naturaliste National Park, but are some distance from each other.

    Is there a separate weather station for each lighthouse, did the naming convention change over time, is the record a local average, does the site reference (lat/long) match the photos, etc. Every station needs these questions answered by local examination.

    What was the “fireball” incident all about? Ball Lightning

  22. My question about all this is: do we have any examples of temperature graph where all the data is known, shows warming, and does not have any illogical corrections? I’m somewhat of a skeptic on AGW, so I enjoy seeing this nice work about the cases where there are artificial correctionsof various sorts. It seems a very reasonable request to the warming crowd to show us some “good” (i.e., verifiable) examples of their temperature assertions. If every rock turned up has a snake under it, at some point our skepticism approches certitude. Do we know of any such “good” cases?

  23. I’m new to this. Would someone please check my thinking? First, you have a temperature signal with errors. Second, you use this erroneous data in a nonlinear dynamic system to predict future events. But wouldn’t the derivatives in such a math model serve to MAGNIFY the error even more?!

  24. Call me old fashioned but.. I work in telecoms & when digital cameras became cheap, I got a box full and tasked our installation engineers to take before and after pictures. Handy for QA, familiarisation and occasional remote hands stuff.

    Is it too much to expect that the agencies responsible for these stations might do similar, or ask site operators to send them pics? Even better, get out of the office and inspect them themselves.

  25. Is there no doubt about that? It’s an honest question. I assume the argument is that it can’t skew the downwards, so there will be a skew upwards, however it may be to small to matter?

    If a structure were built to block the sun, that would bias readings lower. The black road surface will store heat during the day and release it at night usually raising the low temp for the day.

    Which brings me to my next question: Can’t this type of effect be easily measured in a controlled study? Sure, it wouldn’t be something done in an afternoon, but surely by empirical testing, some function of distance and surface area, etc, could be made, no?

    It is not easy. It depends in large part on the amount of sunlight hitting the road surface and the blackness of the surface. Its pretty hard to get a numerical average for leaves and dirt on the road, precipitation, degree of cloud cover at that point, and the effects of parked and moving vehicles.

    Moreover, an induced error of sat 2F or 3F is mostly going to be unnoticeable for use by locals on a daily basis. Forex would you do anything differently if told the temp was 76F versus 74F?

  26. “”” crosspatch (09:29:45) :

    Well, there is an obvious step up in the data in the mid-1990s.

    John Hooper (09:32:26) :

    What’s the point of the the Stevenson screen if not to counteract that? “””

    Well I don’t see any obvious step up in the mid-1990s. I do see a short period when the data appears to drop down, and then “steps up” to about where it seemed to be heading anyway.

    I’ve never seen a real Stevenson Screen; but I presume that it has inside it a thermometer which is essentially immersed in the air inside the owl box, and attempts to measure the temperature of that air; or at least that part of the air that is in contact with the thermometer.

    When you read a thermometer (assuming it is accurately calibrated) the only thing you can be sure of is the temperature of the thermometer; which is what it is recording. If your intent is to determine the temperature of something else, other than the thermometer, well then you have a problem; and that is the big question in temperature measurments. Just what temperature is the sensor responding to, other than its own.

    But I’ll assume that a lot of smart people have given a lot of thought to the design of the setup in the owl box, and that the sensor reading is a respectable representation of the internal air temperature.

    Also I assume, that either the box is hermetically sealed, or that it is open to the outside atmosphere; hopefully the latter, so that the box is free to exchange air between the inside and the outside whenever thermal gradients or winds cause that to happen.

    The thermal radiation from the tar road is certainly going to warm that outside air near the box differently from what a stony road would have.

    The direct radiation from the road surface is hopefully excluded from the box by the design of the louvres which I seem to recall those contraptions have (maybe I should get my own owl box). I assume that the box is made of wood and painted with some highly reflecting Titanium oxide or somesuch paint, that rejects electromagnetic radiation but is amenable to heat conduction from contact with the air.

    I have given a lot of thought to the question of exactly how would I go about constructing a “weather proof” enclosure for an accurate temperature sensor, that would adapt quickly to the temperature of its immediate surroundings, and not respond to any EM radiation source including the sun. Something tells me that I should make it small; the smaller the better, so I would opt for a semiconductor bandgap temperature sensor, that is part of an integrated circuit that converts the analog temperature to a digital serial data stream that simply repeats, as fast as the A-D can recycle. Of course the update rate would be kept low; maybe once per second so that the chip can power down between readings to minimise self heating.

    You can make some bloody low power CMOS ICs these days that wouldn’t self cook by more than the inherent uncertainty of the temperature sensor.

    The real trick is to have the thermal environment around the IC such that it facilitates conduction and heat exchange with the air (or ground), while eliminating EM radiation, even direct incident sunlight.

    I can’t say I have a solution I’m really in love with yet.

  27. I prefer unadjusted raw temperatures. Not anomalies. Although not as insidious as “models”, still there is the subjective choice of base period. Preferably monthly mean max temps. for the entire period of record. Just my $.o2 worth.

  28. This is how science should work. Check the facts, show your working, allow debate about your assumptions and results. If you are wrong, it’s back to the drawing board.

  29. Frank Skog (11:15:29) :

    “Can’t this type of effect be easily measured in a controlled study?”

    The GCM modelers should really show off their prowess and understanding of radiative and convective heat transfer by creating a program to model the temperature in a Stevenson screen in various real life installations. Then this microclimate model could be verified by instrumenting a real life site and comparing the results.

    Of course, we have to remember this “historical” temperature measuring system was only originally intended for weather data. There is probably a good meterological reason the weather reports only report the temperature to the nearest degree – not tenths or hundredths – not that critical.

  30. durox (09:56:37) :

    Lets find someone to go there and paint that tarmac in white! I bet we’ll see at the news that the Ice Age is comming, and that in 10 years that lighthouse will be half in ice.
    The alarmism does great tv ratings!! ;]

    Reply:

    Maybe it is time for all of us to grab a bucket of paint ‘n brush and several rolls of aluminum foil ‘n duct tape then head for the stations near us. Paint all the tarmac/concrete white and duct tape aluminum foil over all the a/c exhausts. One year later analyse the temperature data to see if there is any effect…. /sarc

    or maybe we really should do it…HMMmmm

  31. It’s beginning to look like this all needs a do over with open science, no grants. The only possible hope to get to the bottom of all this is to do it all in the open…

    Personal computers are now capable enough and free software is easily available on free Linux. Start your Ubuntu’s and let’s have at it.

  32. re: Keith (10:01:51) :

    This is from a comment link on another WUWT post. A 6th grader and his did use the raw data to pick out the heat island effect.

    The video also shows how they got that data and why they chose it. I expect someone could do the same thing and get data for other areas.

  33. George E. Smith (11:32:47) :

    Sensor of choice = LM92

    Claimed accuracy 0.33ºC

    About the best reasonably priced sensor & you won’t get any improvement in accuracy without spending big bucks.

    Interface is I2C 2 wire.

    Otherwise needs power & Gnd.

    If I remember correctly, best accuracy is @ 5V

    DaveE.

  34. Anthony,

    Feel free to use this Cape Leeuwin graphic made with Hadcrut3 data. I had to infill 9 random months after 1960 before calculating the yearly mean. Used the mean of the monthly series to infill. Overall series trend of about 0.4C. If you crop off the cold first decade the trend is reduced to 0.12C.

  35. Talking of Australia, are you aware of Peter Spencer.
    I think he is worthy of a little “coverage”.

    http://agmates.ning.com/group/peterspencerhungerstrike?commentId=3535428%3AComment%3A37261&xg_source=msg_com_group

    Excerpt,
    ” High up on a 300m wind monitoring tower Farmer Peter Spencer “Saarahnlee” NSW Australia on the 1st December 2009 as a last resort commenced an indefinite hunger strike until the Australian government meets his demands”

    REPLY: Actually, no, I think he’s not worthy of any coverage here at all. His issue appears to be land use, and the website you give does a horrid job of explaining what “his demands” are. – Anthony

  36. [snip]

    Setting up pair of Stevenson screens next to an asphalt-challenged weather-station, zero them, then move one to an area surrounded by white blankets, and MEASURE the difference.

    Then submit it to peer-reviewed literature.

    And do the REALCLIMATE “press release on submission” trick.

  37. Fighting a running battle here because it bemuses me:

    http://blogs.theaustralian.news.com.au/jacktheinsider/index.php/theaustralian/comments/climategate_lame_by_any_other_name/desc#commentsmore

    Actually got to the stage where the blog writer (“Jack the Insider”) called me “feeble minded” and an “idiot”. Disappointing because I always used to enjoy his cynical musings on politics but he has got the proverbial bee in his bonnet on the ClimateGate issue. as we Aussies say… he’s going off like a frog in a sock!

    But on topic… Jack’s main gun (a Bsc in CSRIO called Marcus) claims the UHI is accommodated for in the data at GISS. How the heck would these guys know how to accommodate every change in the environment around every station in the US? and then by how much? It would be a tough job trying to keep one thermometer honest, let alone thousands. I smell poppycock, but would love some ammo to back it up.

    I’d post the Youtube vid… but I doubt it would get much traction.

  38. Bulldust says:

    as we Aussies say… he’s going off like a frog in a sock!

    How strange. I spent 40+ years in Australia and I have never heard that saying.

  39. The “snip” above was the suggestion that someone’s school science project should be to do the Stevenson screen benchmarking on asphalt vs white-blanket covered land.

  40. PS> I think I may have heard it in Queensland… but then I spent 6 years in Kalgoorlie as well, so I picked up a lot there. I was born in Australia many moons ago (think just after Geelong won their previous flag, in that town even, but not 9 months after :p ) and returned to Kalgoorlie in 1992, in September, for a Geelong-WC Eagles GF. Somehow Australia felt instantly like home (despite leaving when I was 2)… no way I can explain it. Love the place.

    Teaching at the School of Mines (WA) I ran into a chap called Peter Cook (from memory) who spoke wonderful Australian. I could understand every word … he spoke very clearly … but he used so many Aussie expressions I had no idea what he was saying 90% of the time.

    It took me longer to pick up the Aussie slang than it did to understand Afrikaans as a Dutch speaker (but that is another story).

  41. Thanks for the demotivator… I slipped in the first one (assuming Jack will allow it) for shits and giggles.

    Off to work for moi – off to see a major resource company present on a waste recycling solution. Win-win-win. That’s real environmentalism.

  42. Anthony,

    How about issuing a call for some serious whistle blowing from some of the other major climate research centers?

    The United States has some very protective whistle blower laws and they include whistle blowing for waste of federal funds, scientific fraud, violation of laws (such as FOIA), engaging in politics from one’s position as a federal employee, etc.

    There must be a few honest and brave souls inside of NOAA GISS and the others alphabet organizations involved. The withheld code and data exists on networked computers. They can’t hide it forever. Someone must have a change of heart and let the cats out of the bag so this can be cleared up before our elected officials spend our grandchildren’s futures to gild their own pockets.

    Can you issue an offer of safe harbor for these whistle blowers?

  43. If you like lighthouses:-

    Tim Blake-Crystal Machine:- Lighthouse

    nice job. Know anyone in Alice Springs?

  44. Keith (10:01:51)

    I have the same question. So now they have two people to shout at.

    Random variation may take care of station moves, but equipment changes may not be as random. Starting over with the raw data that would be a good beginning.

    There was a post in the last couple of days where the author chose to homogenize the data by lowering the old temperatures. I would tend to leave old temps alone, because being an old fogey, I tend to trust mercury thermometers more than the new stuff, and urban heat islands should not have been as prevalent. However, that just shows my own bias and personal biases are what got us where we are: junk.

  45. There are temperature records back to 1887 for Cape Naturaliste and Cape Leeuwin at http://www.waclimate.net

    There are also BoM official temperature data files on the page for download, and pop-up graphs based on the BoM data. I must say, based on the BoM records, Cape Leeuwin graphs as flat as a tack and Cape Naturaliste has increased but it doesn’t look anywhere near as dramatic as the BoM graph above.

  46. BOM manual for siting of weather stations states that the distance a Stevenson screen must be set up is to be at least 5 times the width of a nearby non-conforming surface from it. Eg if the road is 5 metres wide the equipment should be 25 metres from it. So that one is definitely upter.
    One of these days I’m going to check out the local station (Mackay).

  47. Bulldust (14:00:06) : But on topic… Jack’s main gun (a Bsc in CSRIO called Marcus) claims the UHI is accommodated for in the data at GISS. How the heck would these guys know how to accommodate every change in the environment around every station in the US? and then by how much? It would be a tough job trying to keep one thermometer honest, let alone thousands. I smell poppycock, but would love some ammo to back it up.

    OK, GIStemp does a rather poor job of UHI adjustment. It often ‘gets it wrong’ and makes the past record of a station cooler (it adjusts the past) thus making the slope more, not less, as a UHI adjustment ought to do.

    For example, Pisa, Italy, gets a 1.4 C adjustment the wrong way in an early data point.

    http://chiefio.wordpress.com/2009/08/30/gistemp-a-slice-of-pisa/

    This is done via “The Reference Station Method”. It “looks nearby” circling out up to 1000 km looking for a “rural” station, then uses that station to adjust for UHI. Small problems: The station may be in a far different climate zone and may have an upside down relationship. The “rural” station may not be very rural. They are often airports. Quantico Virginia “The crossroads of the Marines” and Diego Garcia (a major air base on an Island) are two examples..

    http://chiefio.wordpress.com/2009/08/23/gistemp-fixes-uhi-using-airports-as-rural/

    http://chiefio.wordpress.com/2009/08/26/agw-gistemp-measure-jet-age-airport-growth/

    http://chiefio.wordpress.com/2009/09/04/most-used-rural-airport-for-uhi-adj/

    And the percentage of airports in the “raw” cooked GHCN dataset regularly rises over time:

    http://chiefio.wordpress.com/2009/12/08/ncdc-ghcn-airports-by-year-by-latitude/

    There’s more, but I don’t want to load this up with links. If this isn’t enough, well, give me a holler over at my place and I’ll put together some more. The “Islands in the Sun” posting covers the Diego Garcia issue (if you look under AGW issues on the right of the blog).


  48. DaveE (13:38:48) :

    Sensor of choice = LM92

    Claimed accuracy 0.33ºC [this is at 30 deg. C. It shows +- 0.5 deg C over 10 to 50 C, -1 +1.5 accuracy window at -25 deg C]


    About the best reasonably priced sensor & you won’t get any improvement in accuracy without spending big bucks.

    Interface is I2C 2 wire.

    Otherwise needs power & Gnd.

    Hmmm … a higher-accuracy ‘fan controller’ chip via I2C bus … does not claim to be ‘instrumentation grade’.

    a) going to need some ‘line conditioning’ (for the digital data) if this thing is more than a few feet ‘down the line’.

    b) a micro/special control-box to ‘talk’ the I2C protocol and display the temp, also perhaps provide an interface to a PC (USB, serial, LAN, etc)

    c) Lightning/transient protection in the design (aside from what native protection the chip comes with); a must if you want it to survive in the real world.

    d) Not the most accurate temperature device available (accuracy limits ‘open up’ on then low end past 0 deg C, see pg 6 datasheet below) – see short treatise on RTDs, thermocouples and Thermistors for instance.

    e) LM92 Data sheet

    Disclaimer: we use the LM93 (a different, but similar function chip/fan control and supervision) in product, and it works as advert’d.
    .
    .
    .

  49. In my opinion the high trend at Cape Naturaliste is probably due to the proximity of the city of Busselton and what looks like holiday house type real estate development close to the Cape.

    Recently in an effort to determine the ‘pristine coastal site’ trend in Australia, I found that the following six stations have century long data and are nowhere near any urban development. These stations all have similar average trends (TMax + TMin / 2), the overall average being 0.59 degrees/century.

    CAPE LEEUWIN 084016 trend = 0.54 degrees/century
    CAPE MORETON LIGHTHOUSE 040043 trend = 0.74 degrees/century
    JERVIS BAY PT PERPENDICULAR LIGHTHOUSE 068151 trend = 0.39 degrees/century
    CAPE OTWAY LIGHTHOUSE 090015 trend = 0.58 degrees /century
    GABO ISLAND LIGHTHOUSE 084016 trend = 0.65 degrees /century
    WILSONS PROMONTORY LIGHTHOUSE 085096 trend = 0.65 degrees/century

    The data was taken from the BOM historical database, NOT the supposed high quality homogenised dataset.

  50. _Jim (21:46:39) :

    According to the LM93 datasheet, accuracy of the on-chip sensor is 3ºC

    I’ve not managed to find a sensor as cheap as the LM92 with similar accuracy (+/-1.5ºC) over its entire range.

    Buffer/isolators for I2C are cheap too.

    Look around for USB to I2C adaptors.

    443Mhz I2C & USB adaptors are also easy to come by.

    DaveE.

  51. E.M.Smith (21:31:28) :
    Thanks for the input, but it turns out Jack does not want my input anymore. I can no longer respond to his blog. I breached no rules, but clearly the temerity to oppose his opinion (wrong in fact) is enough to have a poster blocked.

    The running debate I had with him was regarding search hits for “cliamtegate” which he insists returns hits for the words “climate” and “gate”. I countered with the fact that the searched increased from hundreds of hits just after the CRU leak to thousands a few days later, and then millions and tens of millions in the next weeks.

    I must therefore assume, according to his logic, that there was suddenly a massive outpouring on the web regarding gates or climate, despite there being no interest in these subjects prior to mid November (at least on a search of climategate).

    Odd that he sees my debating this point with him as a bannable offense. The is after HE called me feeble minded and an idiot… in violation of his own posting rules. Hmmm funny how that works eh? Perhaps a letter to the senior editor of The Australian is warranted.

  52. PPS> Scratch that – he doesn’t allow the links in E.M. Smith’s post… submitting without (with credit of course).

  53. Exact location of cape naturaliste weather station can be obtained from BOM

    http://www.bom.gov.au

    Its no where near any ashphalt or buildings, there are no nearby settlemnets as its in the middle of a national park.

    The photos are of Cape leeuwin , the asphalt road has been there for at least the last 30 years. Please fix original article.


  54. DaveE (09:57:32) :

    According to the LM93 datasheet, accuracy of the on-chip sensor is 3ºC

    We do not have a tight requirement for temperature measurement using this IC. We *do* however need fan supervision/tach reading capability and PWM fan speed control, however, which this IC provides and that is the reason for choosing the LM93.

    Going a bit further in depth, we actually employ a thermistor as external air temperature sensor on account of the wider, and somewhat higher accuracy over that range and ‘read’ that using a 17 bit DVM chip ( an LTC2481 ).


    I’ve not managed to find a sensor as cheap as the LM92 with similar accuracy (+/-1.5ºC) over its entire range.

    A little limited as to accucacy at the low end don’t you think? And the datasheet shows no ‘performance’ below -25 degree C (again, this little chip is intended for fan control and related purposes; again, it is not instrumentation grade)

    It all depends on what you ‘need’ or perceive you need …


    Buffer/isolators for I2C are cheap too.

    Will they pass lightning tests (NOT just human body ESD)?


    Look around for USB to I2C adaptors.

    RIght. Like the Aardvark pods? That’s driving up costs rather substantially and OF COURSE this is not the *volume* (in quantity) answer (an OEM answer inloves the TI MSP uC line or a Microchip PIC).


    443Mhz I2C & USB adaptors are also easy to come by.

    DaveE.

    Are you proposing a one-of or something to be built in quantity? Does it have to work for more than six months of six years in the field?

    Consideration of few more factors than just accuracy and dirt-cheap are required; what about sensor ‘drift’ over time? There might be better choices for a sensor than this considering other factors; the RTDs and Thermistors have come of age, some designs are offsetting as regards to cable length and ‘voltage’ reference (like the design we used with the LTC2481, the supply voltage is used as the ‘reference’ for both the IC and also used as excitation for the Thermistor with changes in one being the same ratio for the other and no effective change seen for Vcc changes).

    For anyone other than an OEM (original equipment manufacturer) I would look to pre-packaged solutions. If one wants a one-of solution AND accuracy but has budget limitations – check eBay for industrial solutions at a bargain price!
    .
    .

  55. _Jim (17:25:08) :

    We’re talking one-off for non-professional use here.

    The LM92 is ideal for this, it’s dirt cheap & ‘good enough’

    For total accuracy, you’re talking perhaps PT100 but it’s plain too expensive. I’ve used it extensively in a professional sense, but that’s not ‘home budget’

    Yes, thermistors are accurate but non linear and expensive to implement.

    DaveE.

  56. Keith (10:01:51) :

    If one is looking for GLOBAL averages does one need any adjustments at all? IF stations are subject to random change – eg moves up or down hill, or from fields to tarmac or vice versa, wouldn’t it be better to just use raw data? (My guess would be that any net bias would be over recording due UHI, more tarmac roads, more air con outlets, etc)
    If this is right, the use of adjustments is a cure worse than the disease ( which is error margins around the averages used ?
    My reason for asking is that I’m trying to persuade folk re AGW over hype and so far just get shouted down, mostly!

    The problem is that due to Anthony’s volunteer survey of the US sites, it’s glaringly apparent that a large majority of them have been out of spec in one or more ways for quite a number of years. Most of the out of specifications problems are direct or indirect heat sources. Direct heat coming from air conditioner exhausts or car engines, indirect heat coming from asphalt, concrete or reflections from shiny surfaces.

    In one case the instruments are in an irrigated lawn behind a hotel building with at least 100 air conditioners. The data from that station is total garbage. A recent superbad discovery is a station in the middle of a backyard garbage dump, surrounded by trees. It really doesn’t matter how its site problems would affect the readings, it’s so far from the ideal that it must be considered a junk station.

  57. dribble (07:44:24) :

    In my opinion the high trend at Cape Naturaliste is probably due to the proximity of the city of Busselton and what looks like holiday house type real estate development close to the Cape.

    Recently in an effort to determine the ‘pristine coastal site’ trend in Australia, I found that the following six stations have century long data and are nowhere near any urban development. These stations all have similar average trends (TMax + TMin / 2), the overall average being 0.59 degrees/century.

    CAPE LEEUWIN 084016 trend = 0.54 degrees/century

    The BoM official data (taken via Stevenson Screen recordings from 1910 to 2008) is interesting for Cape Leeuwin.

    The BoM annual average mean maxima are at:

    http://www.waclimate.net/max/cape-leeuwin.txt

    The BoM annual average mean minima are at:

    http://www.waclimate.net/min/cape-leeuwin.txt

    Cape Leeuwin
    The first 30 years of minima from 1910 to 1939 show an average of 13.1 degrees C, compared to an average 14.3 degrees C for the final 30 years from 1979 to 2008 – i.e. an increase of 1.2 degrees.

    The first 30 years of maxima from 1910 to 1939 show an average of 19.9 degrees C, compared to an average 20 degrees C for the final 30 years from 1979 to 2008 – i.e. an increase of .1 degrees.

    The average mean minimum from 1990 to 1999 was 14.38 degrees and from 2000 to 2008 it was 14.48 degrees. The average mean maximum from 1990 to 1999 was 20.1 degrees and from 2000 to 2008 it was 19.88 degrees.

    i.e. the average minimum at Cape Leeuwin has increased for 100 years and continues to rise, while the average maximum has remained steady and has dropped by .2 degrees over the past decade.

    The Cape Leeuwin recording station looks legit …

    http://www.waclimate.net/cape-leeuwin-station.html

    although it looks different from another BoM picture …

    http://www.bom.gov.au/climate/change/map/stations/009518.shtml

    Cape Naturaliste
    I doubt Cape Naturaliste is close enough to urban development to be influenced, the bitumen road being another question.

    The BoM annual average mean maxima are at:

    http://www.waclimate.net/max/cape-naturaliste.txt

    The BoM annual average mean minima are at:

    http://www.waclimate.net/min/cape-naturaliste.txt

    At Cape Naturaliste, the first 30 years of minima from 1910 to 1939 show an average of 11.6 degrees C, compared to an average 12.9 degrees C for the final 30 years from 1979 to 2008 – i.e. an increase of 1.3 degrees.

    The first 30 years of maxima from 1910 to 1939 show an average of 20.8 degrees C, compared to an average 21.4 degrees C for the final 30 years from 1979 to 2008 – i.e. an increase of .6 degrees.

    The average mean minimum from 1990 to 1999 was 12.94 degrees and from 2000 to 2008 it was 12.82 degrees. The average mean maximum from 1990 to 1999 was 21.4 degrees and from 2000 to 2008 it was 22.02 degrees.

    Cape Naturaliste is about 90 kilometres north of Cape Leeuwin and both are west coastal locations with most of their weather coming off the Indian Ocean and heading east.

    The Bom site descriptions are:
    Cape Naturaliste 109 m 0.5 km On hill at lighthouse
    Cape Leeuwin 13 m On the coast On coast

  58. My trend values for the ‘pristine coastal sites’ were calculated by
    1. Obtaining the historical monthly max/min data for all stations from the BOM (Price $40)
    2. Writing some software to amalgamate the provided annual max/min values along with the average for each station.
    3. The trend is then calculated for 1910-2008 (for these stations) with a simple linear regression function I obtained from the web.

    All of the sites mentioned appear to show a slight decline in temperature to 1950 followed by a steeper rise from 1950 – 2000 approx. What this means I dont know.

    I would have to disagree that Cape Naturaliste is not affected by urbanisation. I originally thought it was Busselton that is near the Cape but on closer inspection it is the town of Eagle Bay which I had thought was a holiday house encampment.

    Eagle Bay is quite close to the lighthouse area, and there has also been substantial clearing of the original bush in the same area. I would classify the Cape Naturaliste station as a definite candidate for urbanisation/land clearing effects.

  59. Why is this story still posted? It should have been removed when it was proven false. Is this site becoming a Gore presentation? Please move on.

  60. Don’t get me wrong. I think that this blog is the best source of info on the web but, when you’re wrong you’re wrong. Just admit it and move on.

  61. Chris Gillham,

    I have had a look at the data on the WAClimate site and it appears to be taken from the BOM High Quality dataset. I personally would not touch this dataset with a bargepole. Most of the ‘long term’ sites in this set are in fact splice-ups of data taken from a number of local stations to give the fraudulent appearance of a long term site. There are also ‘homogenisations’ to each stations record of unknown provenance (BOM say they provide information on this, but dont)

    Cape Naturaliste is however a genuine long term site. However the homogenizations, or adjustments, are extraordinary in my opinion.
    Compare for example the minimum temp graph for Cape Naturaliste in the Historical Database with the HQ dataset graph. The historical database graph has a trend of 0.5 degrees/century, the HQ version a trend of 2.0 degrees/century.
    For the maximum temps, the historical database has a trend of 1.84 degrees/century, the HQ dataset graph has a trend of 1.0 degrees/century.

  62. Unfortunately the GooglEarth street view does not identify the weather station at Cape Naturaliste. According to the Australian Bureau on Metereology the site is at:

    Station Details ID: 009519 Name: CAPE NATURALISTE
    Lat: -33.54 Lon: 115.02 Height: 109.0 m
    Source: http://www.bom.gov.au/products/IDW60701/IDW60701.94600.shtml

    Looking at GoogleEarth you see the lighthouse far north, three houses to the north, and the station would appear to be in a small area of bitumen not covered by street view. There is not much of a ckleared area as the bush comes right up to the bitumen. Street view pictures in the area give you a good idea of the vegetation in the area.

  63. Having said that the Long/Lat locations for the Aussie weather stations are very rough guides… the Leeuwin long lat coords puts it out to sea… so the previous post about the Naturaliste location is speculation.

  64. Almost tempted to go photograph a few of them … thanks for the link. I am off to Andorra next week, however, maybe I can snap one or two in the hills.

  65. Did you know that BoM measure the air pressure at Darwin on the first floor in an airconditioned office? They then apply a correction to it to get the “real” atmospheric pressure. Wouldn’t you think it would make more sense to measure the real atmospheric pressure instead? They reckon it is near enough to the real thing anyway. If that is their attitude to accurate data collection, is it any wonder there are so many question marks over the rest of their datasets?

  66. The mention in an earlier post about minimum and maximum and mean temperatures leads me to wonder whether mean or maximums are the right temperatures to use to monitor change.
    Obviously maximums occur during the heating phase of solar radiation incoming. Minimums occur during the cooling phase as thermal radiation is outgoing.
    Given the greenhouse effect is basically that of containing heat, any changes in it’s effectiveness should be directly indicated on it’s ability to maintain a certain degree of warmth.
    Apart from being a direct indicator, minimum temperatures are inherently likely to be subject to far less influence of outside factors that can cause both short and long term changes.
    Any thoughts?

Comments are closed.