Whitewash and bias, Painting over the historical record

[Stratos Ferric]

I answered in the post Chassisbot referred to in your top reply on your last post. Warming is currently going on. By how much in the 20th and and 21st centuries, I believe the biases in post #1 obscure the real picture. Until the ground station data is verified against a better standard, and the corrections applied consistently to the record, if at all possible, the record in my opinion, for the reasons given, remains in error.

And then we shall need to refine P. Brohan, J.J. Kennedy, I. Harris, S.F.B. Tett and P.D. Jones, Uncertainty estimates in regional and global observed temperature changes: a new dataset from 1850. J. Geophys. Res, 111, D12106, doi:10.1029/2005JD006548. (pdf 1.2Mb)

Unless my responses are being bumped off for being OT, which they are. The calibration is not in question, the effects of the change from historical housing of instrumentation on the data is.

BIases will obscure, but as the research in the paper you've posted reveals, even taking a 'worst case' view of the land surface data would not obscure the fact that we are unprecedentedly warm.

I've only skim read the paper, and the authors (see s 2.1.2 I think) do talk about cross-referencing data to check validity and reliability. They also mention measurement bias of the broad type you're alluding to - i.e. housing - but liit their comments to construction, which may be fair enough. Paint does not seem to be identified as a factor. It would be perfectly easy to lab test your theories.

[Chris Knight]

A great deal of work has been done to align satellite tropospheric measurements to radiosonde measurements. This means as far as possible satellite calculated temperatures match up to actual measurements taken with balloon thermometers. One thing that has become evident is that tropospheric temperatures do not seem to have risen as much as surface temperatures. A number of theories have been put forward for this including paint , heat island affects,lack of polar and sea measurements, but to me the most likely seems to be dust in the lowest ten meters of the atmosphere. Removal of forest and greater activity by humans has increased dust levels. What ever the reason for the discrepancy surface based temperature records whilst accurate are thought to be less accurate in portraying global warming when compared to satellite measurements.

BF you mention theories including paint - do you have any relevant references here?

[Stratos Ferric]

BF - do you think sea surface measurements are actually worth anything at all? I've become sceptical about their efficacy and suspect they are incredibly random. I read some articles on this but haven't the time to look them up right now.

I think this is one area where remote sensing is a huge leap forward. Historic point measurement is always going to cause problems; the SST varies significantly from year to year and place to place, and a low resolution data set may well be prone to significant fluctuation. That's why, if you look at Chris's paper, the margin of error for SSTs is huge.

[BrickFielder]

BF - do you think sea surface measurements are actually worth anything at all.

They seem to be the temperature with the most error in them and considering the amount of the surface which is sea then you would be right to worry. Scientists are however very good at picking up anomalies and adjusting records into line, a process which can also introduce error. I think my argument would be that radiosonde and tropospheric satellite data is likely to reflect the global warming situation best.

BF you mention theories including paint - do you have any relevant references here

No but I did read about it having some small effect somewhere. I think any effect is mostly likely swamped by changes in snow cover or increased desert.

[Chris Knight]

BIases will obscure, but as the research in the paper you've posted reveals, even taking a 'worst case' view of the land surface data would not obscure the fact that we are unprecedentedly warm.

I've only skim read the paper, and the authors (see s 2.1.2 I think) do talk about cross-referencing data to check validity and reliability. They also mention measurement bias of the broad type you're alluding to - i.e. housing - but liit their comments to construction, which may be fair enough. Paint does not seem to be identified as a factor. It would be perfectly easy to lab test your theories.

Unprecedently warm may be, but by what yardstick? - less than 0.02 degrees globally per year on averaged figures from all the participating land-based stations.

I agree that lab tests would be easy, but how many stations have records of when the old whitewash method (there are various recipes for whitewash too!) was abandoned in favour of paint (including type and brand), and when new housing (type and brand) replaced old, obsolete stuff. Did they run the old kit alongside the new kit in all weather conditions, from the coldest to the warmest extents of the range - not feasible, if no record spell of hot or cold weather is available - to make sure the new readings are consistent. To risk being tedious, a standard is required.

[Devonian]

I'm not sure that's really fair to Chris, nor that it stands up academically.

It's perfectly acceptable in academia to question things, even to expose flaws in an argument, without producing the definitive research paper of your own. The world wouldn't advance its knowledge much if we all had to wait to produce the Summa Theologica before uttering a word of critique.

WIB, questions are fine - I've no problem at all with them. But this thread is more claim than a question. And it's a claim without, so far, evidence. I'm more than happy to look at any evidence that the state of a Stevenson Screen might effect temps in a way so far not allowed for- where is it? Where is the evidence that such things aren't already know/allowed for?

Edited October 14, 2007 by Devonian

[pottyprof]

Dev. You have some good contacts. Could you find out that answer please?

[Devonian]

Dev. You have some good contacts. Could you find out that answer please?

No contacts, I've just been to the Met O library many times. I'm a bit too busy atm to take the the amount of time off required to make a visit worthwhile, next month maybe.

I do though have the apparently unusual view that the people working at the Met.O/Hadley are the experts at monitoring weather and climate :lol:

Edit: I'm just reading the .pdf WIB linked to. Table 12 indicates to me that screen errors are of the order of .007C per year

Edited October 14, 2007 by Devonian

[pottyprof]

I do though have the apparently unusual view that the people working at the Met.O/Hadley are the experts at monitoring weather and climate :lol:

I would never have guessed...

Edit: I'm just reading the .pdf WIB linked to. Table 12 indicates to me that screen errors are of the order of .007C per year

But that's an average figure Dev.

I can't imagine that the difference between the screen coverings makes any difference to what temperature is recorded on a typical overcast day but is it not possible in direct sunlight?

I've asked a mate of mine to build me a couple of identical screens that will house temperature probes. One screen will be whitewashed and one will be glossed. I love practical experiments...

[Chris Knight]

WIB, questions are fine - I've no problem at all with them. But this thread is more claim than a question. And it's a claim without, so far, evidence. I'm more than happy to look at any evidence that the state of a Stevenson Screen might effect temps in a way so far not allowed for- where is it? Where is the evidence that such things aren't already know/allowed for?

I make several statements (claims?) in the original post and ask several questions. None of the statements or any of the questions have been dealt with comprehensively.

Shall we start point by point? (question)

A: Claim and question:

Measurement of temperatures since the beginning of the 20th century have been extensively standardised, to reduce bias due to inadequately housed instrumentation and recording methods. The first consideration is the box that houses the instruments.

1. The Stevenson Screen is a standard, so Stevenson Screens in the same location have remained unchanged in design over many years, right?

I answer that question myself:

Wrong, a Stevenson Screen is a general pattern for a ventilated wooden box, set at a standard height above the ground and protected from direct irradiation from the sun, in order to accurately measure the air temperature and humidity. A such it can house several other weather related instruments for ease of observation and recording from the same geographical location.

Nobody so far seems to have any argument with this. (claim)

I then expand on this:

Technologies change, even with regard to double louvred boxes. Different woods have different properties with regard to specific heat, response to humidity in regards to thermal and water absorbency and evaporation. At one time it was thought double layers of asbestos sheet would provide the best surface for the roof, with either a single slope or roof ridge. Later, asbestos containing materials were deemed hazardous and replaced with laminated wooden materials. Woods also age and rot and have to be replaced.

In the original design, the wood was designed to be covered in whitewash, a porous lime containing coating that needed to be replaced annually or as required which also had the added benefit of preventing timber rotting, The lime hardens, with the addition of various natural additives to form a white calcium carbonate based coat which is essentially opaque to infrared radiation, absorbing it with the typical spectrum of calcite, and re-emitting longer wave radiation.

I have had some corroboration on the rotting of timbers, and the earlier use of whitewash, at least until 1970 in various posts.

the evidence for the infrared behaviour of calcium carbonate can be found here:

http://64.233.183.104/search?q=cache:UylxP...;cd=1&gl=uk

The argument against this could be that the calcium carbonate still emits infrared radiation, so why does this make any difference?

I'll go on in my next post.

I would never have guessed...

But that's an average figure Dev. :lol:

I can't imagine that the difference between the screen coverings makes any difference to what temperature is recorded on a typical overcast day but is it not possible in direct sunlight?

I've asked a mate of mine to build me a couple of identical screens that will house temperature probes. One screen will be whitewashed and one will be glossed. I love practical experiments...

Smashing idea! Why not get one made out of 50mm expanded polystyrene too? I can't think of a better white insulator offhand, that would be easy to fashion using a craftknife and padsaw.

[Chris Knight]

No contacts, I've just been to the Met O library many times. I'm a bit too busy atm to take the the amount of time off required to make a visit worthwhile, next month maybe.

I do though have the apparently unusual view that the people working at the Met.O/Hadley are the experts at monitoring weather and climate :lol:

Edit: I'm just reading the .pdf WIB linked to. Table 12 indicates to me that screen errors are of the order of .007C per year

Table 12. Approximate typical contributions (°C) to standard errors of uncertainty

in CET.

CETmean CETmax CETmin

Calibration (Section 2.1: all timescales) 0.06 0.09 0.09

Precision and random screen errors (Sections 2.2 & 2.3)

(daily) 0.1 0.2 0.2

(monthly) 0.02 0.03 0.03

(annual) 0.007 0.008 0.009

That is for the CET, any idea about HADCRUT3?

[Stratos Ferric]

Table 12. Approximate typical contributions (°C) to standard errors of uncertainty

in CET.

CETmean CETmax CETmin

Calibration (Section 2.1: all timescales) 0.06 0.09 0.09

Precision and random screen errors (Sections 2.2 & 2.3)

(daily) 0.1 0.2 0.2

(monthly) 0.02 0.03 0.03

(annual) 0.007 0.008 0.009

That is for the CET, any idea about HADCRUT3?

So what we're saying is that for all intents and purposes the impact of screen housing on measured temperature is lost in the margins - given that the outturn is reported to 1 d.p.

[Stratos Ferric]

Unprecedently warm may be, but by what yardstick? - less than 0.02 degrees globally per year on averaged figures from all the participating land-based stations.

I agree that lab tests would be easy, but how many stations have records of when the old whitewash method (there are various recipes for whitewash too!) was abandoned in favour of paint (including type and brand), and when new housing (type and brand) replaced old, obsolete stuff. Did they run the old kit alongside the new kit in all weather conditions, from the coldest to the warmest extents of the range - not feasible, if no record spell of hot or cold weather is available - to make sure the new readings are consistent. To risk being tedious, a standard is required.

Chris,

there is an extent to which a standard is required, for certain purposes. The general rule in all measurement is that instrument calibration should be within the accuracy of the stated values quoted. For example, to legitimately cite temperatures to 0.1C it would be preferable to be able to measure to 0.01C, or otherwise quote a +/- value on the reading.

Where I'm struggling with this whole thread is what its relevance is. I suspect that your point about 0.02C per year is intended to suggest that the rate of warming is within the measurement variability caused by inconsistent screening of instruments. This would be true of the warming were not arithmeticaly increasing. The impact of your paint is a one time only impact, it does not increase temperatures (if it does at all) by increasing increments.

Assuming we all accept the Hadley data then what we're saying is that variability of measurement due to housing is much much smaller than the total measured warming around the world. In essence it's akin to quibbling whether or not a 400m running track is 1 cm more or less than 400m around when the current timing precision is at an interval if around 10cm. It's not relevant to anything.

Yes, standardisation is a good thing to have, but only if it adds value to the outcomes for which measurement is used. On the basis of what's presented so far I'm not seeing that it does.

[Chris Knight]

Chris,

there is an extent to which a standard is required, for certain purposes. The general rule in all measurement is that instrument calibration should be within the accuracy of the stated values quoted. For example, to legitimately cite temperatures to 0.1C it would be preferable to be able to measure to 0.01C, or otherwise quote a +/- value on the reading.

Where I'm struggling with this whole thread is what its relevance is. I suspect that your point about 0.02C per year is intended to suggest that the rate of warming is within the measurement variability caused by inconsistent screening of instruments. This would be true of the warming were not arithmeticaly increasing. The impact of your paint is a one time only impact, it does not increase temperatures (if it does at all) by increasing increments.

Assuming we all accept the Hadley data then what we're saying is that variability of measurement due to housing is much much smaller than the total measured warming around the world. In essence it's akin to quibbling whether or not a 400m running track is 1 cm more or less than 400m around when the current timing precision is at an interval if around 10cm. It's not relevant to anything.

Yes, standardisation is a good thing to have, but only if it adds value to the outcomes for which measurement is used. On the basis of what's presented so far I'm not seeing that it does.

It is a one time impact each time an old system box is painted instead of whitewashed or changed for a new one. The measured rate of global warming is of the order of fractions of degrees per century, for the 20th century, rising to up to 0.2 degrees per decade for the last 40 y, which is 0.02 deg per year, closing in on the figure that was quoted for the standard error for the CET alone in an earlier post!

I accept that the Hadley data may have been a serious source of error for a long time, merely due to a lack of standardization in the housings of instruments used for the time series of global land temperatures. with biasses downward for the period when traditional methods were used, say pre. 1970, and increasing biasses upwards since, due to boxes that are increasingly dry and warm because the paint technology now used ensures that they stay that way!

[Devonian]

It is a one time impact each time an old system box is painted instead of whitewashed or changed for a new one. The measured rate of global warming is of the order of fractions of degrees per century, for the 20th century, rising to up to 0.2 degrees per decade for the last 40 y, which is 0.02 deg per year, closing in on the figure that was quoted for the standard error for the CET alone in an earlier post!

I accept that the Hadley data may have been a serious source of error for a long time, merely due to a lack of standardization in the housings of instruments used for the time series of global land temperatures. with biasses downward for the period when traditional methods were used, say pre. 1970, and increasing biasses upwards since, due to boxes that are increasingly dry and warm because the paint technology now used ensures that they stay that way!

Again, Chris, you keep stating this paint business is a problem, you make various vague claims, yet you keep failing to put forward any concrete evidence.

So, please put a concrete figure on what you think are the errors involved. These will be the basis of a prediction based on your theory which could then be tested by us. Until you do people who are sceptical, open minded and interested in the science will not accept what you say.

But, I really suggest you ask the Met O what they think as a first step, I'm sure they have the answers and instead of a lengthy thread like this one you could just report what they say here.

As I've said, I'm as an owner as user of a standard Stevenson Screen. I built it myself in the late eighties, it's been moved but not been painted as, since I now have an AWS, it's rather neglected. The paint work is now in a poor state. But, do I think it's reading are effected significantly as a result? I honestly don't. Yes, on a hot still day then maybe they are a tad out, but today, on a cloudy October day? No. And how many hot still day do we get? No, for most of the year the readings will be as accurate as they ever have been - that's accurate to a few tenths of a degree. However, what did effect reading was moving the screen to a more exposed location. Now that DID have an effect, but one would expect changed locations to be different.

No, to convince those of us with field experience you've simply got to do better than keep stating you right. Get out there and get us some evidence.

Edited October 15, 2007 by Devonian

[Paul]

I've removed the past 4 posts from here, so a fresh slate now available for all posters to debate fairly.

[parmenides3]

Probably the best thing to do. Thank you, Paul.

:)P

[BrickFielder]

I did have to trawl around a bit to find some evidence to back this up and found that the facts seem to be that.

1) Whitewash reflects infra red light and modern paints do not.

Discussion about paints here

The instigator of these claims appears to be one Anthony Watts of Purdue’s meteorology department. To be fair he has set up some different screens and measured the different results with different paints. The results he obtained are not what you would expect. Firstly there is a difference but it is not uniform but at the maximum a difference of about 3degrees Farenheit was found.

The results discussion

mean daily max latex_temp - whitewash_temp= 4.1 F

mean daily min latex_temp - whitewash_temp= -0.8 F

mean daily latex_temp - whitewash_temp= 1.2 F

mean min latex_temp= 53.5 F

mean min whitewash_temp= 53.9 F

mean max latex_temp= 96.8 F

mean max whitewash_temp= 93.5 F

The spread for mean min/max temps for latex paint is 43.3 F; the spread for mean min/max temps for whitewash is 39.6 F.

Of course this is not peer reviewed and must be considered just conjecture however the discussion did point to a peer reviewed document about changes in thermometers used in surface temperature measurement.

This shows a difference of between 0.3 and 0.6C for LIG (liquid-in-glass ) used in the the Cotton Region Shelter (CRS) as opposed to th the thermistor sensor used in the MMTS or the platinum resistance thermometer (PRT) used in the ASOS.

Some Perspectives on Recent In Situ Air Temperature Observations

There are lots of interesting leads to follow in the discussions and there seem to be some valid reasons why surface temperature records could be questioned. This does not change the fact that Satellite measurements are also picking up warming, if to a lesser extent than surface records nor does it disprove AGW, its just another complex variable in a complex pattern.

[Chris Knight]

I did have to trawl around a bit to find some evidence to back this up and found that the facts seem to be that.

1) Whitewash reflects infra red light and modern paints do not.

Discussion about paints here

The instigator of these claims appears to be one Anthony Watts of Purdue’s meteorology department. To be fair he has set up some different screens and measured the different results with different paints. The results he obtained are not what you would expect. Firstly there is a difference but it is not uniform but at the maximum a difference of about 3degrees Farenheit was found.

The results discussion

Of course this is not peer reviewed and must be considered just conjecture however the discussion did point to a peer reviewed document about changes in thermometers used in surface temperature measurement.

This shows a difference of between 0.3 and 0.6C for LIG (liquid-in-glass ) used in the the Cotton Region Shelter (CRS) as opposed to th the thermistor sensor used in the MMTS or the platinum resistance thermometer (PRT) used in the ASOS.

Some Perspectives on Recent In Situ Air Temperature Observations

There are lots of interesting leads to follow in the discussions and there seem to be some valid reasons why surface temperature records could be questioned. This does not change the fact that Satellite measurements are also picking up warming, if to a lesser extent than surface records nor does it disprove AGW, its just another complex variable in a complex pattern.

As pointed out earlier, Satellite measurements by radiometer are indirect measurements that may only be calibrated by reference to actual temperature readings from ground stations that are under clear skies. Satellite radiometer response degrades over time due to the hostile environment of outer space, and the satellite network needs to share calibration data by repeated reference to actual ground station datum points. An old caveat is GIGO - Garbage In, Garbage Out.

[Paul]

Hi Chris, have you got any links which show that satellites are calibrated against ground stations as I have searched and can't find any. I know they are regularly compared against upper level temperatures but cannot find any evidence they are calibrated against anything on the ground?

[BrickFielder]

Satellite measurements by radiometer are indirect measurements that may only be calibrated by reference to actual temperature readings from ground stations that are under clear skies. Satellite radiometer response degrades over time due to the hostile environment of outer space, and the satellite network needs to share calibration data by repeated reference to actual ground station datum points.

This is partially true I think. Satellite measure microwave channels which correspond to levels in the atmosphere. The point about this is that Satellites do not measure surface temperature but give a temperature measure typically for the lower troposphere excluding the surface. When you say they are calibrated against ground stations I don't think you are correct what they are calibrated against is weather balloon temperature measurements (radiosonde). I think where you are correct is in saying that satellite data needs regular re-calibration and off course changes in radiosonde measurement methods may also affect the record.

Satellite measurements of the Earth’s microwave emissions

Here is a paper discussing the ins and outs and problems of calibrating the satellite measurements.

A Reanalysis of the MSU Channel 2 Tropospheric Temperature Record

The Report by the U.S. climate change science program in 2006 on the subject also makes interesting reaidng.

Temperature Trends in the Lower Atmosphere Steps for Understanding and reconciling differences 2006

[Chris Knight]

This is partially true I think. Satellite measure microwave channels which correspond to levels in the atmosphere. The point about this is that Satellites do not measure surface temperature but give a temperature measure typically for the lower troposphere excluding the surface. When you say they are calibrated against ground stations I don't think you are correct what they are calibrated against is weather balloon temperature measurements (radiosonde). I think where you are correct is in saying that satellite data needs regular re-calibration and off course changes in radiosonde measurement methods may also affect the record.

Satellite measurements of the Earth’s microwave emissions

re is a paper discussing the ins and outs and problems of calibrating the satellite measurements.

A Reanalysis of the MSU Channel 2 Tropospheric Temperature Record

The Report by the U.S. climate change science program in 2006 on the subject also makes interesting reaidng.

Temperature Trends in the Lower Atmosphere Steps for Understanding and reconciling differences 2006

Thanks for the useful references BF. Yes radiosondes and aircraft have been used for calibrating Tropospheric measurements, I have no doubt that the extreme care taken with the calibration of radiosonde sensors ensures an accurate temperature correlation with the satellite sensors, with the caveat that there are some uncertainties due to temporal and positional correlations of airborne and satellite data, but there are many ground based validation trials that have also been performed with various degrees of difference and other discrepancies that play a part in their repeated cross calibration.

Whereas the Microwave radiometers can "see" through cloud (as opposed to satellite IR sensors), they cannot detect radiation from the surface through rain or other precipitation, and are subject to errors due to low and high level atmospheric aerosols, reflectivity of the ground, Elevation and gradient effects, seasonal variation of vegetation cover, the effects of limitations of the sensors due to pixellation saturation from high temperatures on land areas such as deserts, etc.

Over the oceans, there are additional errors due to emissivity of the surface water under different sea (and lake) conditions, the sampling variations of SST on ships, buoys, temporal and positional issues, and the limited global data available at any time.

The only gold standard for the question "what is the true ground temperature" are temperatures recorded over an area on the ground.

If these ground temperatures are correct, all well and good.

If there is an areal bias due to uncorrected land-based surface temperature errors, due to an as yet unknown bias, the algorithms that alter the sensor output on the satellite are altered so that the satellite data does not drift away from the land based (and sea based) data.

In effect their parameters are altered, on-the-fly, the satellites talking to each other, to some ground based computer systems, to radiosonde responders, and to some designated ocean buoys to fit the observed ground/sea based data.

The data they send back to receivers on earth are not raw data, it is basically the output of a computer model (a set of algorithms) designed to fit observed ground and ocean data.

I'd better stop and give some evidence of these concerns.

A fairly positive validation report from the source:

http://www.sciencedirect.com/science?_ob=A...17c61c37bd9d9ec

Remote Sensing of Environment

Volume 97, Issue 3, 15 August 2005, Pages 288-300

doi:10.1016/j.rse.2005.05.007

Copyright © 2005 Elsevier Inc. All rights reserved.

Ground measurements for the validation of land surface temperatures derived from AATSR and MODIS data

César Coll, , Vicente Caselles, Joan M. Galve, Enric Valor, Raquel Niclòs, Juan M. Sánchez and Raúl Rivas

Department of Thermodynamics, Faculty of Physics, University of Valencia, C/ Dr. Moliner 50, 46100 Burjassot, Spain

Received 30 November 2004; revised 3 May 2005; accepted 8 May 2005. Available online 7 July 2005.

Abstract

An experimental site was set up in a large, flat and homogeneous area of rice crops for the validation of satellite derived land surface temperature (LST). Experimental campaigns were held in the summers of 2002–2004, when rice crops show full vegetation cover. LSTs were measured radiometrically along transects covering an area of 1 km2. A total number of four thermal radiometers were used, which were calibrated and inter-compared through the campaigns. Radiometric temperatures were corrected for emissivity effects using field emissivity and downwelling sky radiance measurements. A database of ground-based LSTs corresponding to morning, cloud-free overpasses of Envisat/Advanced Along-Track Scanning Radiometer (AATSR) and Terra/Moderate Resolution Imaging Spectroradiometer (MODIS) is presented. Ground LSTs ranged from 25 to 32 °C, with uncertainties between ± 0.5 and ± 0.9 °C. The largest part of these uncertainties was due to the spatial variability of surface temperature. The database was used for the validation of LSTs derived from the operational AATSR and MODIS split-window algorithms, which are currently used to generate the LST product in the L2 level data. A quadratic, emissivity dependent split-window equation applicable to both AATSR and MODIS data was checked as well. Although the number of cases analyzed is limited (five concurrences for AATSR and eleven for MODIS), it can be concluded that the split-window algorithms work well, provided that the characteristics of the area are adequately prescribed, either through the classification of the land cover type and the vegetation cover, or with the surface emissivity. In this case, the AATSR LSTs yielded an average error or bias of − 0.9 °C (ground minus algorithm), with a standard deviation of 0.9 °C. The MODIS LST product agreed well with the ground LSTs, with differences comparable or smaller than the uncertainties of the ground measurements for most of the days (bias of + 0.1 °C and standard deviation of 0.6 °C, for cloud-free cases and viewing angles smaller than 60°). The quadratic split-window algorithm resulted in small average errors (+ 0.3 °C for AATSR and 0.0 °C for MODIS), with differences not exceeding ± 1.0 °C for most of the days (standard deviation of 0.9 °C for AATSR and 0.5 °C for MODIS).

Keywords: Land surface temperature; Split-window; Ground measurement; AATSR; MODIS

Corresponding author. Tel.: +34 963 543247; fax: +34 963 543385.

Slightly less good agreement on the Great Lake, Utah: from http://www.met.utah.edu/research/saltlake/remote/vali_html/

Preliminary Validation of Satellite-derived LST

Although limited verification data exists, validation of satellite LST retrivals with in situ measurements shows general agreement to within 1.5 degrees Celsius. A crude validation conducted using bucket samples with a standard thermometer and a Raytec infrared sensor found satellite-derived LST measurements accurate to within 1.5 degrees Celsius on all but one occasion:(broken link to image)

Surveys of lake temperature conducted by the Utah Division of Wildlife Resources (DWR) GSL Ecosystem Project on 2 November 2004 under clear sky conditions also show that the satellite retrievals appear to be describing the magnitude of spatial LST variability relatively well. (broken link to image)

There are many sources of in situ data that may be used in the future for a complete validation study. The Hat and Gunnison Island weather stations both are equipped with lake temperature sensors, although the Hat Island station is above water level (the Gunnison station was resubmerged in September 2005). In addition, lake temperature data is also available from the USGS, Utah Geological Survey, and NOAA bouy data from the summer of 1991.

A diagram of the errors that may affect remote sensing of lake temperature data from another page on the same site:

The errors stated here are potentially amplified on larger off-land bodies of water.

2005 comprehensive paper on a new approach to calibration of remote sensing devices and discussion of biasses and sources of error involved. Available from https://www8.imperial.ac.uk/content/dav/ad/...%20(truths).pdf

truths.pdf

A powerpoint digest of the paper:

Teillet, P.M., G. Fedosejevs, R.P. Gauthier, N.T. O’Neill, K.J. Thome, SF. Biggar, H. Ripley and A. Meygret, A

generalized approach to the vicarious calibration of multiple earth observation sensors using hyperspectral data,

Remote Sensing of Environment, 77(3), 304-327,200la

covering the principles of remote sensor best practice, including ground measurements:

quasar.ppt

I hope these will be of use.

quasar.ppt

[The PIT]

Interesting BF downloaded the result files and may put them in Excel at some point. Didn't expect so much difference.

[Chris Knight]

'Ground frost' on the top of a Stevenson screen...isn't that part of the reason why the height is set at 4' or so.

Did the Antarctic lose most of its ice this year. Have you bothered to check any of the data.

Sorry for the delay in replying to this SF, a lot has happened in the interim.

"'Ground frost' on the top of a Stevenson screen...isn't that part of the reason why the height is set at 4' or so."

I have a car, a white Fiesta as it happens, and the windscreens are transparent. They get frosted before my lawn and my black tarmac drive get frosted, despite the 4' or so of height difference, when the air temperature is a few degrees above 0 deg C.

Forgive me if I am wrong, but wasn't the sea ice anomaly in the Antarctic in March 2007 just below the zero mark?

[Stratos Ferric]

Sorry for the delay in replying to this SF, a lot has happened in the interim.

"'Ground frost' on the top of a Stevenson screen...isn't that part of the reason why the height is set at 4' or so."

I have a car, a white Fiesta as it happens, and the windscreens are transparent. They get frosted before my lawn and my black tarmac drive get frosted, despite the 4' or so of height difference, when the air temperature is a few degrees above 0 deg C.

Forgive me if I am wrong, but wasn't the sea ice anomaly in the Antarctic in March 2007 just below the zero mark?

You're confusing different things there though Chris. Ordinarily, the first things to freeze will be the top metal surfaces of the car, simply because they are great conductors of heat and give up their energy most readily. Glass freezes fairly readily because it, too, is a reasonable conductor of energy. The ground freezes next, with variations depending on moisture and sand / air content. Air freezes last of all.

The fact that your car freezes is nothing to do with height above ground, and is unrelated to the vertical extent of a Stevenson Screen. It is quite possible for the temperature at the level of the grass be a 2-3 degrees or so below the temperature at the level of your windscreen, yet for your windscreen to be frozen. When I left home at 6am on Thursday my windscreen was starting to freeze, but the temperature outside was 3C. I suspect the temperature at ground level was a tad lower, but a SS reading WOULD have been 3C.

My point was that 'ground frost' on the top of a SS is a total and utter non sequiter. SSs are designed precisely to avoid the 'mismeasurement' of GF as if it is a general condition. The measurement of ground temperature is achieved by a grass thermometer. If SSs were metal you might have a point; the fact that they're not ought to suggest to you that the precise considerations of design and dimension are not chance happenings.