Unadjusted data of long period stations in GISS show a virtually flat century scale trend

Theo Goodwin says:
October 24, 2011 at 2:26 pm
That was a brutal drubbing of the one trick pony. Took him to school with that one you did. Wished I’d have written it!

@Smokey
So you basically accept the findings of climate research except for the consequences. So do you agree that 90% of the self-proclaimed “skeptics” are crazy for denying the earth is warming and that humans are causing (at least) much of it? What other scientific consensus’ do you deny? Evolution? The germ theory of disease? Do you think the evidence that smoking causes lung cancer is just a political ploy?
Climate-denial-gate has robbed the “skepticism” movement of whatever credibility it had.

DR_UK says:
October 24, 2011 at 1:22 pm
“But isn’t this a similar method of taking first differences that was discussed and criticised before? See Hu McCulloch’s 2010 post at Climate Audit http://climateaudit.org/2010/08/19/the-first-difference-method/.”
—
Thanks for the link. The idea is indeed similar. There is one difference, however: The CA post states: “Missing observations may simply be interpolated for the purposes of computing first differences (thereby splitting the 2 or more year observed difference into 2 or more equal interpolated differences).” In contrast to this approach, I did not fill the gaps with interpolated numbers.

George Turner says:
October 24, 2011 at 1:29 pm
If you follow some of the adjustments they make to the temperature record, we should worry less about the present getting warmer than the alarming rate at which the past keeps getting cooler. If present trends continue, millions of extra people in the 1940′s are going to freeze to death.
Thank you George!! This was the comment of the day for me. ( I hope my mum and dad aren’t affected because then I wouldn’t be here)

Brian says:
October 24, 2011 at 2:49 pm
“What other scientific consensus’ do you deny? Evolution? The germ theory of disease? Do you think the evidence that smoking causes lung cancer is just a political ploy?”
—
None of these, actually. You may also be surprised to learn that not all of us own a gun.
Anyone who likens the profundity of our understanding of global warming, and the quality of the evidence to those pertaining to any of the above theories only proves their ignorance.

Hansen’s own GISStemp data show that it is actually COOLER than his infamous “Scenario C” forecast, which was based on NO (i.e. ZERO, NONE, NOT ANY) man-made CO2 emissions after 2000!
http://www.real-science.com/doubt-temperatures-rising-fast-hansens-emissions
If Hansen paid any attention at all to HIS OWN DATA, as well as his own predictions, pronouncements, prognositcations and various other bloviations, he would have concluded years ago that CO2-based warming simply DOES NOT HAPPEN!

As Requested:
Off Averages & Anomalies Part 1A
In recent years a number of claims have been made about ‘problems’ with the surface temperature record: that it is faulty, biased, or even ‘being manipulated’. Many of the criticisms often seem to revolve around misunderstandings of how the calculations are done and thus exaggerated ideas of how vulnerable to error the analysis of the record is. In this series I intend to look at how the temperature records are built and why they are actually quite robust. In this first post (Part 1A) I am going to discuss the basic principles of how a reasonable surface temperature record should be assembled, Then in Part 1B I will look at how the major temperature products are built. Finally in Parts 2A and 2B I will then look at a number of the claims of ‘faults’ against this to see if they hold water or are exaggerated based on misconceptions.
How NOT to calculate the Surface Temperature
So, we have records from a whole bunch of meteorological stations from all around the world. They have measurements of daily maximum and minimum temperatures for various parts of the last century and beyond. And we want to know how much the world has warmed or not.
Sounds simple enough. Each day we add up all these station’s daily average temperatures together, divide by the number of stations and, voilá, we have the average temperature for the world that day. Then do that for the next day and the next and…. Now we know the world’s average temperature, each day, for all that measurement period. Then compare the first and last days and we know how much warming has happened – how big the ‘Temperature Anomaly’ is – between the two days. We are calculating the ‘Anomaly of the Averages’. Sounds fairly simple doesn’t it? What could go wrong?
Absolutely everything.
So what is wrong with the method I described above?
1. Every station may not have data for the entire period covered by the record. They have come and gone over the years for all sorts of reasons. Or a station may not have a continuous record. It may not be measured on weekends because there wasn’t the budget for someone to read the station then. Or it couldn’t be reached in the dead of winter.
Imagine we have 5 measuring stations, A to E that have the following temperatures on a Friday:
A = 15, B = 10, C = 5, D = 20 & E = 25
The average of these is (15+10+5+20+25)/5 = 15
Then on Saturday, the temperature at each station is 2 °C colder because a weather system is passing over. But nobody reads station C because it is high in the mountains and there is no budget for someone to go up there at the weekend. So the average we calculate from the data we have available on Saturday is:
(13+8+18+23)/4 = 15.5.
But if station C had been read as well it would have been:
(13+8+3+18+23)/5 = 13
This is what we should be calculating! So our missing reading has distorted the result.
We can’t just average stations together! If we do, every time a station from a warmer climate drops off the record, our average drops. Every time a station from a colder climate drops off, our average rises. And the reverse for adding stations. If stations report erratically then our record bounces erratically. We can’t have a consistent temperature record if our station list fluctuates and we are just averaging them. We need another answer!
2. Our temperature measurements aren’t from locations spaced evenly around the world. Much of the world isn’t covered at all – the 70% that is oceans. And even on land our stations are not evenly spread. How many stations are there in the roughly 1000 km between Maine and Washington DC, compared to the number in the roughly 4000 km between Perth & Darwin?
We need to allow for the fact that each station may represent the temperature of very different size regions. Just doing a simple average of all of them will mean that readings from areas with a higher station density will bias the result. Again, we can’t just average stations together!
We need to use what is called an Area Weighted Average. Do something like: take each station’s value, multiply it by the area it is covering, add all these together, and then divide by the total area. Now the world isn’t colder just because the New England states are having a bad winter!
3. And how good an indicator of its region is each station anyway? A station might be in a wind or rain shadow. It might be on a warm plain or higher in adjacent mountains, or in a deep valley that cools quicker as the Sun sets. It might get a lot more cloud cover at night or be prone to fogs that cause night-time insulation. So don’t we need a lot of stations to sample all these micro-climates to get a good reliable average? How small does each station’s ‘region’ need to be before its readings are a good indicator of that region? If we are averaging stations together we need a lot of stations!
4. Many sources of bias and errors can exist in the records. Were the samples always taken at the same time of day? If Daylight Savings Time was introduced, was the sampling time adjusted for this? Where log sheets for a station (in the good old days before new fangled electronic recording gizmos) written by someone with bad handwriting – is that a 7 or a 9? Did the measurement technology or their calibrations change? Has the station moved, or changed altitude? Are there local sources of biasing around the station? And do these biases cause one-off changes or a time-varying bias?
We can’t take the reading from a station at face-value. We need to check for problems. And if we find them we need to decide whether we can correct for the problem or need to throw that reading or maybe all that station’s data away. But each reading is a precious resource – we don’t have a time-machine to go back and take another reading. We shouldn’t reject it unless there is no alternative.
So, we have a Big Problem. If we just average the temperatures of stations together, even with the Area Weighting answer to problem #2, this doesn’t solve problems #1, #3 or #4. It seems we need a very large detailed network, which has existed for all of the history of the network, with no variations in stations, measurement instruments etc, and without any measurement problems or biases.
And we just don’t have that. Our station record is what it is. We don’t have that time machine. So do we give up? No!
How do stations’ climates change?
Let’s consider a few key questions. If we look at just one location over its entire measurement history, say down on the plains, what will the numbers look like? Seasons come and go; there are colder and warmer years. But what is the longer term average for this location? What is meant by ‘long term’? The World Meteorological Organisation (WMO) defines Climate as the average of Weather over a 30 year period. So if we look at a location averaged over something like a 30 year period and compare the same location averaged over a different 30 year period, the difference between the two is how much the average temperature for that location has changed. And what we find is that they don’t change by very much at all. Short term changes may be huge but the long term average is actually pretty stable.
And if we then look at a nearby location, say up in the mountains, we see the same thing: lots of variation but a fairly stable average with only a small long term change. But their averages are very different from each other. So although a station’s average change over time is quite small, an adjacent station can have a very different average even though its change is small as well. Something like this:
Comparing two adjacent stations
Next question: if each of our two stations averages only change by a small amount, how similar are the changes in their averages? This is not an idle question. It can be investigated, and the answer is: mostly by very little. Nearby locations will tend to have similar variations in their long term averages. If the plains warm long term by 0.5°C, it is likely that the nearby mountains will warm by say 0.4–0.6°C in the long term. Not by 1.5 or -1.5°C.
It is easy to see why this would tend to be the case. Adjacent stations will tend to have the same weather systems passing over them. So their individual weather patterns will tend to change in lockstep. And thus their long term averages will tend to be in lock-step as well. Santiago in Chile is down near sea level while the Andes right at its doorstep are huge mountains. But the same weather systems pass over both. The weather that Adelaide, Australia gets today, Melbourne will tend to get tomorrow.
Station Correlation Scatter Plots (HL87)Final question. If nearby locations have similar variations in their climate, irrespective of each station’s local climate, what do we mean by ‘nearby’? This too isn’t an idle question; it can be investigated, and the answer is many 100’s of kilometres at low latitudes, up to 1000 kilometres or more at high latitudes. In Climatology this is the concept of ‘Teleconnection’ – that the climates of different locations are correlated to each other over long distances.
Figure 3, from Hansen & Lebedeff 1987 (apologies for the poor quality, this is an older paper) plots the correlation coefficients versus separation for the annual mean temperature changes between randomly selected pairs of stations with at least 50 common years in their records. Each dot represents one station pair. They are plotted according to latitude zones: 64.2-90N, 44.4-64.2N, 23.6-44.4N, 0-23.6N, 0-23,6S, 23.6-44.4S, 44.4-64.2S.
Notice how the correlation coefficients are highest for stations closer together and less so as they stretch farther apart. These relationships are most clearly defined at mid to high northern latitudes and mid southern latitudes – the regions of the Earth with higher proportions of land to ocean.
This makes intuitive sense since surface air temperatures of the oceanic regions are influenced also by water temperatures, ocean currents etc instead of just air masses passing over them, while land temperatures don’t have this other factor. So land temperatures would be expected to have better correlation since movement of weather systems over them is a stronger factor in their local weather.
This is direct observational evidence of Teleconnection. Not just climatological theory but observation.
A better answer
So what if we do the following? Rather than averaging all our stations together, instead we start out by looking at each station separately. We calculate its long term average over some suitable reference period. Then we recalculate every reading for that station as a difference from that reference period average. We are comparing every reading from that station against its own long term average. Instead of a series of temperatures for a station, we now have a series of ‘Temperature Anomalies’ for that station. And then we repeat this for each individual station, using the same reference period to produce the long term average for each separate station.
Then, and only then, do we start calculating the Area Weighted Average of these Anomalies. We are now calculating the ‘Area Average of the Anomalies’ rather than the ‘Anomaly of the Area Averages’ – now there’s a mouthful. Think about this. We are averaging the changes, not averaging the absolute temperatures.
Does this give us a better result? In our imaginary ideal world where we have lots of stations, always reporting all the time, no missing readings, etc., then these two methods will give the same result.
The difference arises when we work in an imperfect world. Here is an example (for simplicity I am only doing simple averages here rather than area weighted averages):
Let’s look at stations A to E. Let’s say their individual long term reference average temperatures are:
A = 15, B = 10, C = 5, D = 20 & E = 25
Then for one day’s data their individual readings are:
A = 15.8, B = 10.4, C = 5.7, D = 20.4 & E = 25.3
Using the simple Anomaly of Averages method from earlier we have:
(15.8+10.4+5.7+20.9+25.3)/5 – (15+10+5+20+25)/5 = 0.52
While using our Average of Anomalies method we get:
((15.8-15) + (10.4-10) + (5.7-5) + (20.4-20) + (25.3-25))/5 = 0.52
Exactly the same!
However, if we remove station C as in our earlier example, things look very different. Anomaly of Averages gives us:
(15.8+10.4+20.4+25.3)/4 – (15+10+5+20+25)/5 = 2.975 !!
While Average of Anomalies gives us:
((15.8-15) + (10.4-10) + (20.4-20) + (25.3-25))/4 = 0.475
Obviously both values don’t match what the correct value would be if station C were included, but the second method is much closer to the correct value. Bearing in mind that Teleconnection means that adjacent stations will have similar changes in anomaly anyway, this ‘Average of Anomalies’ method is much less sensitive to variations in station availability.
Now let’s consider how this approach could be used when looking at station histories over long periods of time. Consider 3 stations in ‘adjacent’ locations. A has readings from 1900 to 1960. B has reading from 1930 to 2000 and C has readings from 1970 to today. A overlaps with B, B overlaps with C. But C doesn’t overlap with A. If our reference period is say 1930 – 1960, we can use the readings from A & B. But C doesn’t have any readings from our reference period. So how can we splice together A, B, & C to give a continuous record for this location?
Doesn’t this mean we can’t use C since we can’t reference it to out 1930-1960 baseline? And if we use a more recent reference period we lose A. Do we have to ignore C’s readings entirely? Surely that means that as the years roll by and the old stations disappear, eventually we will have no continuity to our record at all? That’s not good enough.
However there is a way we can ‘splice’ them together.
A & B have a common period from 1930-1960. And B & C have a common period from 1970-2000. So if we take the average of B from 1930 to 1960 and compare it to the same average from A for the same period we know how much their averages differ. Similarly we can compare the average of B from 1930-1960 to the average for B from 1970-2000 to see how much B has changed over the intervening period. Then we can compare B vs C over the 1970-2000 period to relate them together. Knowing these three differences, we can build a chain of relationships that links C1970-2000 to B1970-2000 to B1930-1960 to A1930-1960
Something like this:
‘Chaining’ station histories together
If we have this sort of overlap we can ‘stitch together’ a time series stretching beyond more than one station’s data. We have the means to carry forward our data series beyond the life (and death) of any one station, as long as there is enough time overlap between them. But we can only do this if we are using our Average of Anomalies method. The Anomaly of Averages method doesn’t allow us to do this.
So where has this got us in looking at our problems? The Average of Anomalies approach directly addresses problem #1. Area Weighted Averaging addresses problem #2. Teleconnection and comparing a station to itself helps us hugely with problem #3 – if fog provides local insulation, it probably always had, so any changes are less related to the local conditions and more to underlying climate changes. Local station bias issues still need to be investigated but if they don’t change over time, then they don’t introduce ongoing problems. For example, if a station is too close to an artificial heat source, then this biases that station’s temperature. But if this heat source has been a constant bias over the life of the station, then it cancels out when calculate the anomaly for the station. So this method also helps us with (although doesn’t completely solve) problem #4. In contrast, using the Anomaly of Averages method, local station biases and erratic station availability will compound each other making things worse.
So this looks like a better method.
Which is why all the surface temperature analyses use it!
The Average of Anomalies approach is used precisely because it avoids many of the problems and pitfalls.
In Part 1B I will look at how the main temperature records actually compile their trends.

Off Averages & Anomalies, Part 1B
In Part 1A we looked at how a reasonable temperature record needs to be compiled. If you haven’t already read part 1A, it might be worth reading it before 1B.
There are four major surface temperature analysis products produced at present: GISTemp from the Goddard Institute of Space Sciences (GISS); HadCRUT, a collaboration between the Hadley Research Center and the University of East Anglia Climate Research Unit (HadCRUT); The US National Oceanic And Atmospheric Administration’s (NOAA) National Climatic Data Center (NCDC); and the Japanese Meteorological Agency (JMA). Another major analysis effort is currently underway: the Berkeley Earth Surface Temperature Project (BEST), but as yet their results are preliminary.
GISTemp
We will look first specifically at the product from GISS, at how they do their Average of Anomalies, and their Area Weighting scheme. This product dates back to work undertaken at GISS in the early 1980s with the principle paper describing the method being Hansen & Lebedeff 1987 (HL87).
The following diagram illustrates the Average of Anomalies method used by HL87
Reference Station method for comparing stations series
This method is called the ‘Reference Station Method’. One station in the region to be analysed is chosen as station 1, the reference station. The next stations are 2, 3, 4, etc., to ‘N’. The average for each pair of stations (T1, T2), (T1, T3), etc. is calculated over the common reference period using the data series for each station T1(t), T2(t), etc., where “t” is the time of the temperature reading. So for each station their anomaly series is the individual readings – Tn(t) – minus the average value of Tn.
“δT” is the difference between their two averages. Simply calculating the two averages is sufficient to produce two series of anomalies, but GISTemp then shifts T2(t) down by δT, combines the values of T1(t) and T2(t) to produce a modified T1(t), and generates a new average for this (the diagram doesn’t show this, but the paper does describe it). Why are they doing this? Because this is where their Area Averaging scheme is included.
When combining T1(t) and T2(t) together, after adjusting for the difference in their averages, they still can’t just add them because that wouldn’t include any Area Weighting. Instead, each reading is multiplied by an Area Weighting factor based on the location of each station; these two values are then added together and divided by the combined area weighting for the two stations. So series T1(t) is now modified to be the area weighted average of series T1(t) and T2(t). Series T1(t) now needs to be averaged again since the values will have changed. Then they are ready to start incorporating data from station 3 etc. Finally, when all the stations have been combined together, the average is subtracted from the now heavily-modified T1(t) series, giving us a single series of Temperature Anomalies for the region being analysed.
So how are the Area Weighting values calculated? And how does GISTemp then average out larger regions or the entire globe?
They divide up the Earth into 80 equal area boxes – this means each box has sides of around 2500km. Then within each box they divide these up into 100 equal area smaller sub-boxes.
GISTemp Grids
They then calculate an anomaly for each sub-box using the method above. Which stations get included in this calculation? Every station within 1200 km of the centre of the sub-box. And the weighting for each station used simply diminishes in proportion to its distance from the centre of the sub-box. So a station 10km from the centre will have a weighting of 1190/1200 = 0.99167, while a station 1190 km from the centre will have a weighting of 10/1200 = 0.00833. In this way, stations closer to the centre have a much larger influence while those farther away an ever smaller influence. And this method can be used even if there are no stations directly in the sub-box, inferring its result from surrounding stations.
In the event that stations are extremely sparse and there were only 1 station within 1200 km, then that reading would be used for a sub-box. But as soon as you have even a handful of stations within range, their values will quickly start to balance out the result. And closer stations will tend to predominate. Then the sub-boxes are simply averaged together to produce an average for the larger box – we can do this without any further area averaging because we have already used area averaging within the sub-box and they are all of equal area. Then in turn the larger boxes can be averaged to produce results for latitude bands, hemispheres, or globally. Finally these results are then averaged over long time periods.
Remember our previous discussion of Teleconnection, and that long term climates are linked over significant distances. This is why this process can produce a meaningful result even when data is sparse. On the other hand, if we were trying to use this method to estimate daily temperatures in a sub-box, the results would be meaningless. The short term chaotic nature of daily weather would swamp any longer range relationships. But averaged out over longer time periods and larger areas, the noise starts to cancel out and underlying trends emerge. For this reason, the analysis used here will be inherently more accurate when looked at over larger times and distances. The monthly anomaly for one sub-box will be much less meaningful than the annual anomaly for the planet. And the 10-year average will be more meaningful again.
And why the range of 1200 km? This was determined in HL87 based on the correlation coefficients between stations shown in the earlier chart. The paper explains this choice:
“The 1200-km limit is the distance at which the average correlation coefficient of temperature variations falls to 0.5 at middle and high latitudes and 0.33 at low latitudes. Note that the global coverage defined in this way does not reach 50% until about 1900; the northern hemisphere obtains 50% coverage in about 1880 and the southern hemisphere in about 1940. Although the number of stations doubled in about 1950, this increased the area coverage by only about 10%, because the principal deficiency is ocean areas which remain uncovered even with the greater number of stations. For the same reason, the decrease in the number of stations in the early 1960s, (due to the shift from Smithsonian to Weather Bureau records), does not decrease the area coverage very much. If the 1200-km limit described above, which is somewhat arbitrary, is reduced to 800 km, the global area coverage by the stations in recent decades is reduced from about 80% to about 65%.”
Effect of station count on area coverage
It’s a trade-off between how much coverage we have of the land area and how good the correlation coefficient is. Note that the large increase in contributing station numbers in the 1950s and subsequent drop off in the mid-1970s does not have much of an impact on percentage station coverage – once you have enough stations, more doesn’t improve things much. And remember, this method only applies to calculating surface temperatures on land; ocean temperatures are calculated quite separately. When calculating the combined Land-Ocean temperature product, GISTemp uses land-based data in preference as long as there is a station within 100 km. Otherwise it uses ocean data. So in large land areas with sparse station coverage, it still calculates using the land-based method out to 1200 km. However, for an island in the middle of a large ocean, the land-based data from that island is only used out to 100 km. After that point, the ocean based data prevails. In this way data from small islands don’t influence the anomalies reported for large areas of ocean when ocean temperature data is available.
One aspect of this method is the order in which stations are merged together. This is done by ordering the list of stations used in calculating a sub-box by those that have the longest history of data first, with the stations with shorter histories last. So they are merging progressively shorter data series into a longer series. In principle, the method used to select the order in which the stations are being processed could have a small effect on the result. Selecting stations closer to the centre of the sub-box first is an alternative approach. HL87 considered this and found that the two techniques produced differences that were two orders of magnitude smaller than the observed temperature trends. And their chosen method was found to produce the lowest estimate of errors. They also looked at the 1200 km weighting radius and considered alternatives. Although this produced variations in temperature trends for smaller areas, it made no noticeable difference to zonal, hemispheric, or global trends.
The Others
The other temperature products use somewhat simpler methods.
HadCRUT (or really CRU, since the Hadley Centre contribution is Sea Surface Temperature data) calculate based on grid boxes that are xº by xº, with default value being 5º by 5º. At the equator, this means they are approximately 550 x 550 km, although much smaller at the polar regions. They then take a simple average all the anomalies for every station within that grid box. This is a much simpler area averaging scheme. Because they aren’t interpolating data like GISS, they are limited by the availability of stations as to how small their grid size can go, otherwise too many of their grids may have no station at all. And in grid boxes where there is no data available, they do not calculate anything. So they aren’t extrapolating / interpolating data. But equally, any large-scale temperature anomaly calculation such as for a hemisphere or the globe is effectively assuming that any uncalculated grid boxes are all actually at the calculated average temperature. However, to then build results for larger areas, they need to area weight the averages for the differing sizes of the grid boxes depending on the latitude they are at.
NCDC and JMA also use a 5º by 5º grid size and simply average anomalies for all stations within that grid then area weighted averaging is used to average grid boxes together. All three also combine land and sea anomaly data. Unlike HadCRU & NCDC, which use the same ocean data, JMA maintain their own separate database of SST data.
In this article and the Part 1A, I have tried to give you an overview of how and why surface temperatures are calculated, and why calculating anomalies then averaging them is far more robust than what might seem the more intuitive method of averaging then calculating the anomaly.
In parts 2A & 2B we will look at the implications of this for the various questions and criticisms that have been raised about the temperature record.

Of Averages & Anomalies Part 2B
In Part 1A and Part 1B we looked at how surface temperature trends are calculated, the importance of using Temperature Anomalies as your starting point before doing any averaging and why this can make our temperature record more robust.
In Part 2A and in this Part 2B we will look at a number of the claims made about ‘problems’ in the record, and how misperceptions about how the record is calculated can lead us to think that it is more fragile than it actually is. This should also be read in conjunction with earlier posts here at SkS on the evidence here, here & here that these ‘problems’ don’t have much impact. In this post I will focus on why they don’t have much impact.
If you hear a statement such as ‘They have dropped stations from cold locations so the result is now give a false warming bias’ and your first reaction is, yes, that would have that effect, then please, if you haven’t done so already, go and read Part 1A and Part 1B then come back here and continue.
Part 2A focused on issues of broader station location. Part 2B focuses on issues related to the immediate station locale.
Now to the issues. What are the possible problems?
Problems with ‘bad’ stations
One issue that has received considerable attention is the question of the ‘quality’ of surface observation stations, particularly in the US. How well do the stations in the observation network meet quality standards with respect to location and avoidance of local biasing issues, and how much might this impact on the accuracy of the temperature record.
The upshot of investigations into this is that, at least in the US, a substantial proportion of stations have poor location quality ratings. However, analysis of the impact of the site quality problems by a number of independent analysts suggests that these problems have had almost no impact on the accuracy of the long term temperature record. How could this be? Surely that is the whole point of these quality rankings – poor quality sites can give bad results. So why wouldn’t they?
The definition of the best quality sites, Category 1 is as follows:
“Flat and horizontal ground surrounded by a clear surface with a slope below 1/3 (<19º). Grass/low vegetation ground cover 3 degrees.”
Down to Category 5:
“(error ≥ 5ºC) – Temperature sensor located next to/above an artificial heating source, such a building, roof top, parking lot, or concrete surface”
Lets consider a few of these factors. And remember, we are interested in factors that have an impact on long-term changes in the temperature readings at a site. If a factor results in a bias in the reading but this bias does not change over time, then it will not impact on the analysis since we are interested in changes – static biases get cancelled out in the analysis and have no long-term impact. Firstly, let’s look at the standard enclosure used for a meterological measurement station – the Stevenson Screen:
Stevenson Screen
The screen is designed to isolate the instruments inside from outside influences, particularly radiant effects from its surrounds and rain. It is usually made from a material such as wood or similar that is a fairly good insulator and isn’t going to change temperature too much because of radiant heating/cooling from its surroundings. The double-slatted design suppresses air movement from wind through the enclosure, minimising wind chill effects and restricting rain entry onto the instruments. The double-slatted design also means that any air rising from beneath the enclosure isn’t being preferentially drawn into or out of the box. And the design of the base allows air movement from below while shielding from radiation from below.
So what are the problems which can change the category of a temperature monitoring station to lower than 1?
Slope > 19º
A problem may arise if the station isn’t located on sufficiently flat ground. This can produce air movements that are caused by temperature, resulting in warmer air possibly moving towards the station. However, unless there have been really major earthworks around the site, this factor doesn’t change over time and is unlikely to have a long-term changing impact.
Grass/low vegetation ground cover >10 centimeters high
This can impact on air movements around the station. Also, if the vegetation changes substantially – low grass to shrubs and trees – then this could change water evaporation rates around the station and alter air temperatures. Major increases in vegetation might have a cooling effect on the station due to evaporative effects, while declines in vegetation back to Category 1 standards might have a warming impact. However, unless there is a regular and progressive change in the vegetation pattern around the station, this would not produce an ongoing change of any bias. If maintenance of vegetation around the station over its lifetime has been poor or erratic, then the bias may fluctuate up and down. This would create shorter term fluctuations in the bias but this would tend to cancel out in the longer term.
Shading when the sun elevation >3 degrees
If the degree to which the station and its surrounds are shaded over the course of the day changes, this can alter local heating. Primarily this is going to impact as a result of shading causing differing heating/cooling of the ground under/around the enclosure, resulting in changes in the temperature and flow rate of rising air up through the enclosure. Unless the cause of the shading varies over long, multi-year time frames such as trees growing or buildings rising, the shading effect is not a long-term changing biasing factor. Depending on the cause of the shading, this may cause changes in the bias over the course of a day and over the seasons, but as a multi-year bias, this would remain constant.
Not far enough from large bodies of water
This too is a static bias. The body of water would have a cooling effect due to evaporation that would vary with daily weather conditions and the seasons but would not be a multi-year biasing factor.
Static artificial heating sources
Essentially surfaces such a brick, concrete, bitumen, etc. that can act as local heat stores, greater than normal grass covered earth would be, that can then release heat either radiantly or by heating the surrounding air. These can be vertical structures, horizontal surfaces away from the enclosure, or a horizontal surface beneath the enclosure. The enclosures are designed to minimise radiant heat penetration into the enclosure from its surrounds, so the major impact of such static heating sources is going to be from heating surrounding air which may then pass through the enclosure. This will be worst when such a surface is very close to the enclosure, particularly beneath it, generating rising warmer air into the box.
Also an important factor will be the extent to which any such surfaces tend to form a partial ‘room’ around the enclosure, restricting horizontal air movement. Any such surface will tend to heat the air near/above it, causing that air to rise. More air is then drawn in to replace this, potentially flowing over or through the enclosure. If the distances involved and the geometry of the site result in this new air being warmer than the general surroundings, this could provide a warming bias for the site. Conversely if this replacement air is being drawn from a location that isn’t warmer then there may be no bias at all, possibly even a cooling effect. Ambient winds may also blow warmed air towards, or away from enclosure, depending on wind direction. And the effect of any such bias will vary over the course of the day and the seasons.
However, since the main source of any such bias is the amount and layout of such surfaces and sunlight, these biases won’t change over multi-year time frames unless the area of the surfaces is changing. This could be due to construction, or changes in shading of these surfaces such as by trees growing or building construction nearby. And some of these shading changes could actual reduce the bias over time, resulting in a long-term cooling trend. Also to be considered is whether the site is included within a region that is or becomes urban, in which case the UHI adjustments mentioned previously may cancel out any bias completely. And we still have to allow for area weighting of data from such a site when averaged over the Earth’s land surface. And this doesn’t affect the oceans at all.
Dynamic artificial heating sources
These are similar to the static surfaces, but they are things that actively pump heated air into the environment. Things such as Air Conditioner condensers, Exhaust fans, Heater flues, Cooling towers, Vehicle exhausts, etc. As with the static sources, a key issue here is geometry. They are generating hot air which will tend to rise unless winds blow it towards the enclosure. Does any such device actively blow warm air towards the enclosure? Or does its operation tend to draw air in from elsewhere and over the enclosure? How distant is the device and what is the geometry?
Also how long does the device operate for; 24/7 or intermittently? A station may be next to a large car park, but unless there is continuous activity, even thousands of cars have no extra impact if they are all parked and empty. Does an Air-conditioner run 24/7 or just 9-5 weekdays? Is it a reverse cycle A/C unit also used for heating in winter or at night, in which case it will pump out colder air then that doesn’t rise? How much do these activities vary with the seasons? And ultimately do these activities grow in magnitude over multi-year time frames? Otherwise they again contribute to short-term intra-annual biasing but not multi-year effects. And they may be cancelled out anyway by UHI compensations.
Conclusions about ‘Bad’ Stations
The US network certainly isn’t as good as it should be. There are certainly factors operating there that influence short term daily and seasonal readings and these may have important implications for use in daily Meteorological forecasting which rely on absolute temperatures. However, for long term multi-year Climatological uses, it is perhaps easy to overestimate the impact of these problems.
It easy to understand how our subjective impressions, standing near a poor quality site, seeing an A/C roaring away or feeling the radiant heat from a concrete parking lot nearby, could lead us to think this is a big issue. But the combination of the screening properties of the enclosures, long-term averaging, anomaly-based averaging, and UHI compensation will certainly tend to remove many biases that do not have long term-trend changes. And area averaging over the Earth’s land surface combined with the fact that most of the Earth is water reduces any impact even further.
So it isn’t surprising that the long term temperature trend data doesn’t seem to be significantly affected by station quality issues. That is not to say that there may not be noticeable impacts on shorter term measures – local and seasonal trends and possibly daily temperature range (DTR) effects for example. But for the headline Global Temperature Anomaly, which is a main indicator of Climate Change, station quality issues appear to be a very minor issue, something that ‘all comes out in the wash’.
Station Homogenisation issues
Finally we come to ‘Station Homogenisation’ – the process of reviewing station data records looking for errors that are a result of how the measurement was taken, rather than what the temperature actually was.
A common misconception is that ‘the thermometer never lies’. That the raw data is the gold standard. As anyone who works in any field involving instrumentation knows, this isn’t true; there are always ‘issues’ that you have to monitor for. Any instrument, even a simple thermometer will have its own built in biases.
Sometimes there will be readings that are just plain whacky. And surrounding influences can have an impact. A thermometer out in the sunshine will have a different reading from one shaded by your hand for a few minutes. A caretaker who can work quickly taking the readings when the enclosure door is open will produce a different bias from one who works slowly, or reads the instruments in a different order. Bias and error is everywhere.
If readings at a station weren’t always taken at the same time of day, this can introduce biases. Changes in the instruments used can introduce a bias. Some readings can be just plain wrong. Imagine some scenarios:
The caretaker of a station may have had a ‘big night out’ and not read the thermometer very accurately. There is an error there but we probably can’t detect it.
The caretaker of a station may have a ‘big night out’ every Friday night. Now there might be a regular error in Saturday’s readings. With a pattern like this, we might be able to detect it with statistical analysis. We might be able to correct it but only if we are certain enough.
That caretaker might have had one ‘REALLY big night out’ and next morning broke the thermometer. He replaced it but did he record that fact in the station log? If he did, we know that a change of bias has been introduced between the two thermometers. Then we can compare readings from before and after and try to find the change. But only after we have years’ worth of data from both thermometers. And if he didn’t log it, then we only spot a problem if that station seems to have a strange change compared to nearby stations.
Over time the Stevenson Screen may have fallen into disrepair, resulting in a slow changing bias as outside influences start to penetrate. Then the site is updated with a new screen. Biases removed, although the new screen may have its own small bias. If we now about the change we can try to compensate for it. Eventually when we have enough data from before and after.
The caretaker at the station in Ushuaia right at the southern tip of Argentina records data through the early 1900’s. In Spanish, with poor handwriting – really hard to tell 7’s and 9’s apart. The log sheets are sent to Buenos Aires where the data from this and many other stations are collated and typed up onto summary sheets by a clerk with an old battered typewriter. Then they are filed away; 40 years later they are extracted, faded and old, photocopied on a poor quality early copier and mailed to the US for incorporation into climate databases. Where they must be copied into the database again by hand. How many errors have crept in during that process?
So, we can’t simply take the raw data at face value. It has noise in it. We need to analyse this data looking for problems and correcting them when we are confident enough of the correction. But also being careful that we don’t introduce errors through unjustified corrections. This requires care and judgment and it is sometimes a real detective story. And often corrections cannot be made until many, many years later because you need lots of data before you can spot changes in bias.
So this process of working through the data, trying to make it more accurate is ongoing.
But what of its impact on the temperature record? Again, if the biases at a station don’t change over time, they don’t affect our analysis. Individual errors matter but they will tend to be random, some higher, some lower so when we average over large areas and long time periods, they tend to cancel out. Again, it is problems that cause changing biases that matter. And analysis of changes due to Homogenisation in the record indicate that there are as many cooling changes as warming ones. Such as this from Brohan et al 2006:
Homogenisation Distribution
Conclusion
So, Part 1A looked at how we should calculate the temperature record and why the method used is very important to the result. And that this doesn’t necessarily match our intuitive idea of how it should be done; in this our intuition is often wrong. In Part 1B looked at how we DO calculate the temperature record, that is using the method outlined in part one and that the area weighting scheme used by one record is based on empirical evidence. In Part 2A we looked at some of the areas where the temperature record has been criticised with respect to its broader locale. And in this post we have explored issues related to the immediate surrounds of the station.
I think we have seen that there are many reasons why we tend to overestimate the effect of these problems. This conclusion is consistent with the evidence here, here & here from various analyses that show that these possible problems haven’t had any significant effect on the result.
My conclusion is that we can have a strong confidence in the results produced for the global temperature trend. Any problems will show up more in short-term patterns such as seasonal, monthly and daily trends. But the headline global numbers look pretty robust.
You will have to make up your own mind but I hope I have been able to give you some food for thought when you are thinking about this.

I just don’t get it with this site. On some posts (e.g. a recent one by Willis Eschenbach), there’s the argument that skeptics have never doubted that the world is warming. And then this article comes along to doubt that warming. Which is it? Maybe Anthony Watts can shed some light on this?

Have you heard of the Little Ice Age? Click that link and you will see two things. The first is the obligatory “I believe in global warming like you, so don’t cut off my funding!” message (see Climategate and it’s emails showing what is done to those who fail to support AGW). At the very least, this shows that you cannot say that the scientist here has an anti AGW bias. Second, there is this, yet another site (Wikipedia) that starts with the obligatory “we support AGW!” message, trying to say that the little ice age (according to the IPCC) was not worldwide, and then is followed by actual scientific evidence from , respectively, Europe, North America, Central America, Africa, Antarctica, Australia, New Zealand and Pacific Islands, and South America, which shows that the LIA was, indeed, worldwide, and that the IPCC was wrong or flat out lying. And if you look here, you see a 300 year record of 5 European cities showing about 150 years of cooling, followed by 150 years of warming. We also have records of the cold people wrote about then, such as the frost fairs held on the top of the ice of the frozen river Thames in England, and how George Washington dragged cannon across the frozen Hudson River. Conclusion, it was colder then. Then about 150 years ago, it started to warm up, and warmed up rapidly.
Thus, by the time we get to 1900 or so, it has warmed up from it’s formerly cold temperature. and since then the temperature has remained about the same. Thus we can clearly say “the world is warming”, since it is clearly warmer than the little ice age, and doubt modern, unprecedented warming since 1950, since we can see that for the last 100 years or a bit more, the temperature has not changed nearly as much as it did when the climate changed from the little ice age to the modern warm period.
To sum up:
There was a Little Ice Age, so called because it was cold.
It was consistently colder than now for quite some time, perhaps 300 years.
The Little Ice Age ended, and it warmed up rapidly.
For the last 100 years or more, in the USA at least, it has remained about the same, warmer than the Little Ice Age.
If global warming were true, it would not have remained consistently the same, it would have warmed rapidly after 1950, it did not.
This is according to the most reliable temperature measurement station we have, the ones that have remained constant for 100+ years.
Or in one sentence, it has definitely warmed up since the Little Ice Age, and has remained fairly warm for over 100 years, but has not warmed much more in that 100 years from what it was 100 years ago.
You just need a little perspective, compared to 200 years ago, the world has warmed, compared to 100 years ago, the world has not warmed.

Legatus says:
October 24, 2011 at 3:45 pm
Or in one sentence, it has definitely warmed up since the Little Ice Age, and has remained fairly warm for over 100 years, but has not warmed much more in that 100 years from what it was 100 years ago.
You just need a little perspective, compared to 200 years ago, the world has warmed, compared to 100 years ago, the world has not warmed.
It’s the last hundred years temperature data they’d been playing with as in New Zealand, the second mentions Salinger’s connection with CRU: http://climaterealists.com/index.php?id=6151
http://www.climateconversation.wordshine.co.nz/tag/nz-temperature-records/
It’s taken around forty years to put the record straight..