Sunday, 1 May 2016

Christy and McNider: Time Series Construction of Summer Surface Temperatures for Alabama

John Christy and Richard McNider have a new paper in the AMS Journal of Applied Meteorology and Climatology called "Time Series Construction of Summer Surface Temperatures for Alabama, 1883–2014, and Comparisons with Tropospheric Temperature and Climate Model Simulations". Link: Christy and McNider (2016).

This post gives just few quick notes on the methodological aspects of the paper.
1. They select data with a weak climatic temperature trend.
2. They select data with a large cooling bias due to improvements in radiation protection of thermometers.
3. They developed a new homogenization method using an outdated design and did not test it.

Weak climatic trend

Christy and McNider wrote: "This is important because the tropospheric layer represents a region where responses to forcing (i.e., enhanced greenhouse concentrations) should be most easily detected relative to the natural background."

The trend in the troposphere should a few percent stronger than at the surface; mainly in the tropics. However, it is mainly interesting that they see a strong trend as a reason to prefer tropospheric temperatures, because when it comes to the surface they select the period and temperature with the smallest temperature trend: the daily maximum temperatures in summer.

The trend in winter due to global warming should be 1.5 times the trend in summer and the trend in the night time minimum temperatures is stronger than the trend in the day time maximum temperatures, as discussed here. Thus Christy and McNider select the data with the smallest trend for the surface. Using their reasoning for the tropospheric temperatures they should prefer night time winter temperatures.

(And their claim on the tropospheric temperatures is not right because whether a trend can be detected does not only depend on the signal, but also on the noise. The weather noise due to El Nino is much stronger in the troposphere and the instrumental uncertainties are also much larger. Thus the signal to noise ratio is smaller for the tropospheric temperatures, even if the signal were as long as the surface observations.

Furthermore, I am somewhat amused that there are still people interested in the question whether global warming can be detected.)

[UPDATE. Tamino shows that within the USA, Alabama happens to be the region with the least warming. The more so for the maximum temperature. The more so for the summer temperature.]

Cooling bias

Then they used data with a very large cooling bias due to improvements in the protection of the thermometer for (solar and infra-red) radiation. Early thermometers were not protected as well against solar radiation and typically record too high temperatures. Early thermometers also recorded too cool minimum temperatures; the thermometer should not see the cold sky, otherwise it radiates out to it and cools. The warming bias in the maximum temperature is larger than the cooling bias in the minimum temperature, thus the mean temperature still has some bias, but less than the maximum temperature.

Due to this reduction in the radiation error summer temperatures have a stronger cooling bias than winter temperatures.

The warming effect of early measurements on the annual means is probably about 0.2 to 0.3°C. In the maximum temperature is will be a lot higher and in the summer temperature it will again be a lot higher.

That is why most climatologists use the annual means. Homogenization can improve climate data, but it cannot remove all biases. Thus it is good to start with data that has least bias. Much better than starting with a highly biased dataset like Christy and McNider did.

Statistical homogenization removes biases by comparing a candidate station to its neighbour. The stations need to be close enough together so that the regional climate can be assumed to be similar in both stations. The difference between two stations is then weather noise and inhomogeneities (non-climatic changes due to changes in the way temperature was measured).

If you want to be able to see the inhomogeneities you thus need to have well correlated neighbors that have as little weather noise as possible. By using only the maximum temperature, rather than the mean temperature, you increase the weather noise. But using the monthly means in summer, rather than the annual means or at the very least the summer means, you increase the weather noise. By going back in time more than a century you increase the noise because we had less stations to compare with at the time.

They keyed part of the the data themselves mainly for the period before 1900 from the paper records. It sounds as if they performed no quality control of these values (to detect measurement errors). This will also increase the noise.

With such a low signal to noise ratio (inhomogeneities that are small relative to the weather noise in the difference time series), the estimated date of the breaks they still found will have a large uncertainty. It is thus a pity that they purposefully did not use information from station histories (metadata) to get the date of the breaks right.

Homogenization method

They developed their own homogenization method and only tested it on a noise signal with one break in the middle. Real series have multiple breaks; in the USA typically every 15 years. Furthermore also the reference series has breaks.

The method uses the detection equation from the Standard Normal Homogeneity Test (SNHT), but then starts using different significance levels. Furthermore for some reason it does not use the hierarchical splitting of SNHT to deal with multiple breaks, but it detects on a window, in which it is assumed there is only one break. However, if you select the window too long it will contain more than one break and if you select the window too short the method will have no detection power. You would thus theoretically expect the use of a window for detection to perform very badly and this is also what we found in a numerical validation study.

I see no real excuse not to use better homogenization methods (ACMANT, PRODIGE, HOMER, MASH, Craddock). These are build to take into account that also the reference station has breaks and that a series will have multiple breaks; no need for ad-hoc windows.

If you design your own homogenization method, it is good scientific practice to test it first, to study whether it does what you hope it does. There is, for example, the validation dataset of the COST Action HOME. Using that immediately allows you to compare your skill to the other methods. Given the outdated design principles, I am not hopeful the Christy and McNider homogenization method would score above average.


These are my first impressions on the homogenization method used. Unfortunately I do not have the time at the moment to comment on the non-methodological parts of the paper.

If there are no knowledgeable reviewers available in the USA, it would be nice if the AMS would ask European researchers, rather than some old professor who in the 1960s once removed an inhomogeneity from his dataset. Homogenization is a specialization, it is not trivial to make data better and it really would not hurt if the AMS would ask for expertise from Europe when American experts are busy.

Hitler is gone. The EGU general assembly has a session on homogenization, the AGU does not. The EMS has a session on homogenization, the AMS does not. EUMETNET organizes data management workshops, a large part of which is about homogenization; I do not know of an American equivalent. And we naturally have the Budapest seminars on homogenization and quality control. Not Budapest, Georgia, nor Budapest, Missouri, but Budapest, Hungary, Europe.

Related reading

Tamino: Cooling America. Alabama compared to the rest of contiguous USA.

HotWhopper discusses further aspects of this paper and some differences between the paper and the press release. Why nights can warm faster than days - Christy & McNider vs Davy 2016

Early global warming

Statistical homogenisation for dummies

Tuesday, 26 April 2016

Climate scientists are now grading climate journalism

Guest post by Daniel Nethery and Emmanuel Vincent Daniel Nethery is the associate editor and Emmanuel Vincent is the founder of Climate Feedback. Climate Feedback is launching a crowdfunding campaign today.

The internet represents an extraordinary opportunity for democracy. Never before has it been possible for people from all over the world to access the latest information and collectively seek solutions to the challenges which face our planet, and not a moment too soon: the year 2015 was the hottest in human history, and the Great Barrier Reef is suffering the consequences of warming oceans right now.

Yet despite the scientific consensus that global warming is real and primarily due to human activity, studies show that only about half the population in some countries with among the highest CO2 emissions per capita understand that human beings are the driving force of our changing climate. Even fewer people are aware of the scientific consensus on this question. We live in an information age, but the information isn’t getting through. How can this be?

While the internet puts information at our fingertips, it has also allowed misinformation to sow doubt and confusion in the minds of many of those whose opinions and votes will determine the future of the planet. And up to now scientists have been on the back foot in countering the spread of this misinformation and pointing the public to trustworthy sources of information on climate change.

Climate Feedback intends to change that. It brings together a global network of scientists who use a new web-annotation platform to provide feedback on climate change reporting. Their comments, which bring context and insights from the latest research, and point out factual and logical errors where they exist, remain layered over the target article in the public domain. You can read them for yourself, right in your browser. The scientists also provide a score on a five-point scale to let you know whether the article is consistent with the science. For the first time, Climate Feedback allows you to check whether you can trust the latest breaking story on climate change.

An example of Climate Feedback in action. Scientists’ comments and ratings appear as a layer over the article. Text annotated with Hypothesis is highlighted in yellow in the web browser and scientists’ comments appear in a sidebar next to the article. Illustration: Climate Feedback

Last year the scientists looked at some influential content. Take the Pope’s encyclical, for instance. The scientists gave those parts of the encyclical relating to climate science a stamp of approval. Other “feedbacks,” as we call them, have made a lasting impact. When the scientists found that an article in The Telegraph misrepresented recent research by claiming that the world faced an impending ice age, the newspaper issued a public correction and substantially modified the online text.

But there’s more work to be done. Toward the end of the year the scientists carried out a series of evaluations of some of Forbes magazine’s reporting on climate change. The results give an idea of the scale of the problem we’re tackling. Two of the magazine’s most popular articles for 2015, one of which attracted almost one million hits, turned out to be profoundly inaccurate and misleading. Both articles, reviewed by nine and twelve scientists, unanimously received the lowest possible scientific credibility rating. This rarely occurs, and just in case you’re wondering, yes, the scientists do score articles independently: ratings are only revealed once all scientists have completed their review.

We argue that scientists have a moral duty to speak up when they see misinformation masquerading as science. Up to now scientists have however had little choice but to engage in time-consuming op-ed exchanges, which result in one or two high-profile scientists arguing against the views of an individual who may have no commitment to scientific accuracy at all. Climate Feedback takes a different approach. Our collective reviews allow scientists from all over the world to provide feedback in a timely, effective manner. We then publish an accessible synthesis of their responses, and provide feedback to editors so that they can improve the accuracy of their reporting.

We’ve got proof of concept. Now we need to scale up, and for that we need the support of everyone who values accuracy in reporting on one of the most critical challenges facing our planet. Climate Feedback won’t reach its full potential until we start measuring the credibility of news outlets in a systematic way. We want to be in a position to carry out an analysis of any influential internet article on climate change. We want to develop a ‘Scientific Trust Tracker’ – an index of how credible major news sources are when it comes to climate change.

We’re all increasingly relying on the internet to get our news. But the internet has engendered a competitive media environment where in the race to attract the most hits, sensational headlines can trump sober facts. We’re building into the system a new incentive for journalists with integrity to get ahead. Some journalists are already coming to us, asking our network of scientists to look at their work. We want readers to know which sources they can trust. We want editors to think twice before they publish ideological rather than evidence-based reporting on global warming.

On Friday 22 May 2016, more than 170 countries signed the Paris climate agreement. But this unprecedented international treaty will lead to real action only if the leaders of those countries can garner popular support for the measures needed to curb greenhouse gas emissions. The fate of the Paris deal lies largely in the hands of voters in democratic countries, and we cannot expect democracies to produce good policy responses to challenges of climate change if voters have a confused understanding of reality.

Scientists from all over the world are standing up for better informed democracies. You can help them make their voices heard. We invite you to stand with us for a better internet. We invite you to stand with science.

Victor Venema: I am also part of the Climate Feedback community and have annotated several journalistic articles when they made claims about climate data quality. It is a very effective way to combat misinformation. Just click on the text and add a short comment; Climate Feedback will take care of spreading your contribution. If you are a publishing scientist, do consider to join.

* Photo at the top. Severe suburban flooding in New Orleans, USA. Aftermath of Hurricane Katrina. Photo by ark Moran, NOAA Corps, NMAO/AOC (CC BY 2.0)

Tuesday, 29 March 2016

Upcoming meetings on climate data quality

It looks like 2016 will be a year full of interesting conferences. I already pointed you to EGU, IMSC and EMS before. Here is an update with three upcoming European meetings, including two close deadlines.

The marine climate data community will hold its main workshop this July (18 to 22). The deadline has just been prolonged to the 6th of April, Wednesday next week.

The metrologists (no typo) organize a meeting on climate data, MMC2016. It will take place from 26 to 30 September and will be organized together with the WMO TECO conference.

And naturally we will have the European Meteorological Society meeting in Autumn. This year is an ECAC year (European Conference on Applied Climatology). The abstract submission deadline is in about three weeks, 21 Apr 2016, during EGU. So start writing soon. As always we will have a session on "Climate monitoring; data rescue, management, quality and homogenization" for the homogenization addicted readers of this blog.

If you know of more interesting conferences, do add them in the comments.


The Fourth International Workshop on the Advances in the Use of Historical Marine Climate Data (MARCDAT-IV) will be held at the National Oceanography Centre, Southampton, UK between the 18th and 22nd July 2016. The workshop will be arranged around the following themes:
  • Data homogenization (benchmarking, bias adjustments, step change analysis, metadata)
  • Quantification and estimation of uncertainty
  • Data management, recovery and reprocessing (digitisation efforts and reprocessing of previously digitised data)
  • Reconstructing past climates
  • Integrating In-situ / satellite data sources
  • Consistency of the climate across domain boundaries (land, ocean, surface, subsurface, atmosphere)
  • The role of ICOADS and applications of marine climate data
  • Review of the 10-year action plan
This looks like an invitation of people working on land data to also participate. I just asked some colleagues on the homogenization list and it looks like there are a decent number of weather stations near the coast. We could compare them to marine observations, I would especially be interested in comparing sea surface temperature observations.

This workshop is free to attend, but participants must register.

Key Dates:
Abstract submission deadline: 8th April 2016
Registration closes: 31st May 2016


International workshop on Metrology for Meteorology and Climate in conjunction with WMO TECO 2016 conference & Meteorological Technology World Expo 2016. It will be held in Madrid Spain from 26 to 30 September 2016.
During the last years an increasing collaboration has been established between the Metrology and Meteorology communities. EURAMET, the European association of metrology Institutes, is funding several projects aiming at delivering results of valuable impact for the meteorology and climatology science. The key aspect of such projects is the traceability of measurements and uncertainties of measured physical and chemical quantities describing the earth atmosphere. The MMC conference aims to give an opportunity to those two communities to present and discuss needs, methods, expertise and devices for cooperating in producing better data. The invitation is addressed to the metrology, meteorology and climate scientific communities and operators. Starting with the first MMC 2014 in Brdo, Slovenia, this time the MMC 2016 is organized Madrid, Spain, in conjunction with CIMO-TECO conference.
As far as I know the abstract deadline has not been determined yet, but write the date into your agenda, bookmark the homepage and contact the organizers to give you a notice once the deadline is known.


The conference theme of the Annual Meeting of the European Meteorological Society is: Where atmosphere, sea and land meet: bridging between sciences, applications and stakeholders. It will be held from 12 to 16 September 2016 in Trieste, Italy. The abstract submission deadline is 21 April 2016, during EGU2016.

MC1 Climate monitoring; data rescue, management, quality and homogenization
Convener: Manola Brunet-India
Co-Conveners: Hermann Mächel, Victor Venema, Ingeborg Auer, Dan Hollis 
Robust and reliable climatic studies, particularly those assessments dealing with climate variability and change, greatly depend on availability and accessibility to high-quality/high-resolution and long-term instrumental climate data. At present, a restricted availability and accessibility to long-term and high-quality climate records and datasets is still limiting our ability to better understand, detect, predict and respond to climate variability and change at lower spatial scales than global. In addition, the need for providing reliable, opportune and timely climate services deeply relies on the availability and accessibility to high-quality and high-resolution climate data, which also requires further research and innovative applications in the areas of data rescue techniques and procedures, data management systems, climate monitoring, climate time-series quality control and homogenisation.

In this session, we welcome contributions (oral and poster) in the following major topics:
  • Climate monitoring , including early warning systems and improvements in the quality of the observational meteorological networks
  • More efficient transfer of the data rescued into the digital format by means of improving the current state-of-the-art on image enhancement, image segmentation and post-correction techniques, innovating on adaptive Optical Character Recognition and Speech Recognition technologies and their application to transfer data, defining best practices about the operational context for digitisation, improving techniques for inventorying, organising, identifying and validating the data rescued, exploring crowd-sourcing approaches or engaging citizen scientist volunteers, conserving, imaging, inventorying and archiving historical documents containing weather records
  • Climate data and metadata processing, including climate data flow management systems, from improved database models to better data extraction, development of relational metadata databases and data exchange platforms and networks interoperability
  • Innovative, improved and extended climate data quality controls (QC), including both near real-time and time-series QCs: from gross-errors and tolerance checks to temporal and spatial coherence tests, statistical derivation and machine learning of QC rules, and extending tailored QC application to monthly, daily and sub-daily data and to all essential climate variables
  • Improvements to the current state-of-the-art of climate data homogeneity and homogenisation methods, including methods intercomparison and evaluation, along with other topics such as climate time-series inhomogeneities detection and correction techniques/algorithms (either absolute or relative approaches), using parallel measurements to study inhomogeneities and extending approaches to detect/adjust monthly and, especially, daily and sub-daily time-series and to homogenise all essential climate variables
  • Fostering evaluation of the uncertainty budget in reconstructed time-series, including the influence of the various data processes steps, and analytical work and numerical estimates using realistic benchmarking datasets
Next to this session, readers of this blog may also like: Climate change detection, assessment of trends, variability and extremes. And I personally like the session on Spatial Climatology, which has a lot to do with structure and variability.

Top photo by Martin Duggan, which has a CC BY 2.0 license.

Monday, 21 March 2016

Cooling moves of urban stations

It has been studied over and over again, in very many ways: in global temperature datasets urban stations have about the same temperature trend as surrounding rural stations.

There is also massive evidence that urban areas are typically warmer than their surroundings. For large urban areas the Urban Heat Island (UHI) effect can increase the temperature by several degrees Celsius.

A constant higher temperature due to the UHI does not influence temperature changes. However, when cities grow around a weather station, this produces an artificial warming trend.

Why don’t we see this in the urban stations of the global temperature collections? There are several reasons; the one I want to focus on in this post is that stations do not stay at the same place.

Urban stations are often relocated to better locations, more outside of town. It is common for urban stations to be moved to airports, especially when meteorological offices are moved to the airport to assist in airport safety. Also when meteorological offices can no longer pay the rent in the city center, they are forced to move out and take the station with them. When urban development makes the surrounding unsuited or when a volunteer observer retires, the station has to move, it makes sense to then search for a better location, which will likely be in a less urban area.

Relocations are nearly always the most frequent reason for inhomogeneities. For example, Manola Brunet and colleagues (2006) write about Spain:
“Changes in location and setting are the main cause of inhomogeneities (about 56% of stations). Station relocations have been common during the longest Spanish temperature records. Stations were moved from one place to another within the same city/town (i.e. from the city centre to outskirts in the distant past and, more recently, from outskirts to airfields and airports far away from urban influence) and from one setting (roofs) to another (courtyards).”
Since relocations of that kind are likely to result in a cooling, the Parallel Observations Science Team (ISTI-POST) wants to have a look at how large this effect is. As far as we know there is no overview study yet, but papers on the homogenization of a station network often report on adjustments made for specific inhomogeneities.

We, that is mainly Jenny Linden of Mainz University, had a look in the scientific literature. Let’s start in China were urbanization is strong and can be clearly seen in the raw data of many stations. They also have strong cooling relocations. The graph below from Wenhui Xu and colleagues (2013) shows the distribution of breaks that were detected (and corrected) with statistical homogenization for which the station history indicated that they were caused by relocations. Both the minimum and the maximum temperature cool by a few tenth of a degree Celsius due to the relocations.

The distribution of the breaks that were due to relocations for the maximum temperature (left) and minimum temperature (right). The red line is a Gaussian distribution for comparison.

Going more in detail, Zhongwei Yan (2010) and colleagues studied two relocations in Beijing. They found that the relocations cooled the observations by −0.81°C and −0.69°C. Yuan-Jian Yang and colleagues (2013) find a cooling relocation of 0.7°C in the data of Hefei. Clearly for single urban stations, relocations can have a large influence.

Fatemeh Rahimzadeh and Mojtaba Nassaji Zavareh (2014) homogenized the Iranian temperature observations and observed that relocations were frequent:
“The main non-climatic reasons for non-homogeneity of temperature series measured in Iran are relocation and changes in the measuring site, especially a move from town to higher elevations, due to urbanization and expansion of the city, construction of buildings beside the stations, and changes in vegetation.”
They show an example with 5 stations where one station (Khoramabad) has a relocation in 1980 and another station (Shahrekord) has two relocation in 1980 and 2002. These relocations have a strong cooling effect of 1 to 3 degrees Celsius.

Temperature in 5 stations in Iran, including their adjusted series.

The relocations do not always have a strong effect. Margarita Syrakova and Milena Stefanova (2009) do not find any influence of the inhomogeneities on the annual mean temperature averaged over Bulgaria. This while “Most of the inhomogeneities were caused by station relocations… As there were no changes of the type of thermometers, shelters and the calculation of the daily mean temperatures, the main reasons of inhomogeneities could be station relocations, changes of the environment or changes of the station type (class).

In Finland, Norway, Sweden and the UK the relocations produced a cooling bias of -0.11°C and relocations appear to be the most common cause of inhomogeneities (Tuomenvirta, 2001). The table below summarises the breaks that were found and what the reasons for them were if this was known from the station histories. They write:
“[Station histories suggest] that during the 1930s, 1940s and 1950s, there has been a tendency to move stations from closed areas in growing towns to more open sites, for example, to airports. This can be seen as a counter-action to increasing urbanization.”

Table with the average bias of inhomogeneities found in Finland, Sweden, Norway and the UK in winter (DJF), spring (MAM), summer (JJA) and autumn (SON) and in the yearly average. Changes in the surrounding, such as urbanization or micro-siting changes, made the temperatures higher. This was counteracted by more frequent cooling biases from changes in the thermometers and the screens used to protect the thermometers, by relocations and by changes in the formula used to compute the daily mean temperature.

Concluding, relocations are a frequent type of inhomogeneity. They produce a cooling bias. For urban stations the cooling can be very large. For the average over a region, the values are smaller, but especially because they are so common, they will have most likely a clear influence on global warming in raw temperature observations.

Future research

One problem with studying relocations is that they are frequently accompanied by other changes. Thus you can study them in two ways: study only relocations where you know that no other changes were made or study all historical relocations whether there was another change or not.

The first set-up allows us to characterize the relocations directly, to understand the physical consequences to move for example a station from the center of a city / village to the airport. In this way the differences are not subject to other changes specific to a network. So, the results can be easily compared between regions. The problem is that only a part of the parallel measurements available satisfy these strict conditions.

Conversely, for the second design (taking all historical relocations, also when they have another change) the characterization of the bias will be limited to the datasets studied and we will need a large sample to say something about the global climate record. But on the other hand, we can also analyze more data this way.

There are also two possible sources of information. The above studies relied on statistical homogenization comparing a candidate station to its neighbors. All you need to know for this is which inhomogeneities belong to a relocation. A more direct way to study these relocations is by using parallel measurements at both locations. This is especially helpful to study changes in the variability around the mean and in weather extremes. That is where the Parallel Observation Science Team (ISTI-POST) comes into play.

It is also possible to study specific relocations. The relocation of stations to airports was an important transition, especially around the 1940s. This temperature change is likely large and this transition quite frequent and well documented. One could focus on urban stations or on village stations, rather than studying all stations.

One could make a classification of the micro and macro siting before and after the relocation. For micro-siting the Michel Leroy (2010) classification could be interesting; as far as I know this classification has not been validated yet, we do not know how large the biases of the 5 categories are and how well-defined these biases are. Ian Stewart and Tim Oke (2012) have made a beautiful classification of the local climate zones of (urban) areas, which can also be used to classify the surrounding of stations.

Example of various combinations of building and land use of the local climate zones of Stewart and Oke.

There are many options and what we choose will also depend on what kind of data we can get. Currently our preference is to study parallel data with identical instrumentation at two locations, to understand the influence of the relocation itself as well as possible. In addition to study the influence on the mean, we are gathering data on the break sizes found by statistical homogenization for breaks due to relocations. The station histories (metadata) are crucial for this in order to clearly assign breakpoints to relocation activities. It will also be interesting to compare those two information sources where possible. This may become one study or two depending on how involved the analysis will become.

This POST study is coordinated by Alba Guilabert and Jenny Linden and Manuel Dienst are very active. Please contact one of us if you would like to be involved in a global study like this and tell us what kind of data you would have. Also if anyone knows of more studies reporting the size of inhomogeneities due to relocations, please let us know. I certainly have seen more such tables at conferences, but they may not have been published.

Related reading

Parallel Observations Science Team (POST) of the International Surface Temperature Initiative (ISTI).

The transition to automatic weather stations. We’d better study it now.

Changes in screen design leading to temperature trend biases.

Early global warming.

Why raw temperatures show too little global warming.


Brunet M., O. Saladie, P. Jones, J. Sigró, E. Aguilar, et al., 2006: The development of a new daily adjusted temperature dataset for Spain (SDATS) (1850–2003). International Journal of Climatology, 26, pp. 1777–1802, doi: 10.1002/joc.1338.
See also: a case-study/guidance on the development of long-term daily adjusted temperature datasets.

Leroy, M., 2010: Siting classifications for surface observing stations on land. In WMO Guide to Meteorological Instruments and Methods of Observation. "CIMO Guide", WMO-No. 8, Part I, Chapter 1, Annex 1B.

Rahimzadeh, F. and M.N. Zavareh, 2014: Effects of adjustment for non‐climatic discontinuities on determination of temperature trends and variability over Iran. International Journal of Climatology, 34, pp. 2079-2096, doi: 10.1002/joc.3823.

Stewart, I.D. and T.R. Oke, 2012: Local climate zones for urban temperature studies. Bulletin American Meteorological Society, 93, pp. 1879–1900, doi: 10.1175/BAMS-D-11-00019.1.
See also the World Urban Database.

Tuomenvirta, H., 2001: Homogeneity adjustments of temperature and precipitation series - Finnish and Nordic data. International Journal of Climatology, 21, pp. 495-506, doi: 10.1002/joc.616.

Xu, W., Q. Li, X.L. Wang, S. Yang, L. Cao, and Y. Feng, 2013: Homogenization of Chinese daily surface air temperatures and analysis of trends in the extreme temperature indices. Journal Geophysical Research Atmospheres, 118, doi: 10.1002/jgrd.50791.

Syrakova M. and Stefanova M., 2009: Homogenization of Bulgarian temperature series. International. Journal Climatology, 29, pp. 1835-1849, doi: 10.1002/jov.1829.

Yan ZW; Li Z; Xia JJ. 2014. Homogenisation of climate series: The basis for assessing climate changes. Science China: Earth Sciences, 57, pp 2891-2900, doi: 10.1007/s11430-014-4945-x.

* Photo at the top "High Above Sydney" by Taro Taylor used with a Attribution-NonCommercial 2.0 Generic (CC BY-NC 2.0) license.

Monday, 7 March 2016

Bernie Sanders is more electable

The bias in the American mass media are driving me crazy. So let me get onto my little soap box.

You can see the bias when they add the super-delegates to the pledged delegates to pretend that Clinton has a higher lead. If these super-delegates would actually vote against the wish of the primary voters, the disgust will be large enough to make sure that the Democrats would lose.

You can see the bias in that the media hardly speaks about money in politics. A main topic in both primaries.

What I want to talk about today is that the media generally assumes that Hillary Clinton is more electable in the general presidential election than Bernie Sanders. Often in the dismissive implicit way of the establishment. That can be a real opinion, but the evidence goes into another direction.

Money in politics

I am happy to admit that I am also biased. For me Bernie Sanders is clearly the best candidate, mainly because he wants to get money out of politics. Money in politics is bad for the public debate, for democracy, for the environment and for the economy.

Politicians who have to do what their donors tell them have no flexibility to compromise and get things done; they have to defend indefensible positions. This leads to a childish political debate. Nearly all Republican representatives in Washington reject basic scientific findings on climate change. This is childish. It is equivalent to putting your fingers in your ears and singing la, la, la, la. It matches the donor's preferences, but does not match their voter base. Half of the Republican voters support policies to reduce greenhouse gasses.

Wednesday, 24 February 2016

Australia flying blind into unchartered climatic conditions

The Australian Climate Council wrote a report explaining how the gutting of climate science at the main Australian research institute, CSIRO, will harm climate science in Australia: Flying Blind: Navigating Climate Change Without CSIRO. The Australian climate research program is the most important one in the Southern Hemisphere and seriously hurts humanities understanding of the climatic changes there, which are anyway already understudied because most research happens in the global north.

The report also shows that these planned cuts to CSIRO’s climate science division would breach Australia’s commitments under the Paris agreement.

While destroying research can be done quickly with such a decision, building up a research program takes decades. A next Australian government will not be able to get to the same level again by renewing funding. That will take a long, long time.

It is very ironic that this decision comes from a right wing government that does everything to increase the use of coal and thus make climate change worse. It comes from a group that often advocates an adaption-only strategy to climate change. If you want to adapt, you need to know what climatic changes you need to adapt to. For that you need to know what will happen at a local scale, often with respect to changes in extreme weather, often what will happen with precipitation and storms. It is cleat that climate change is real, that the global mean temperature will keep on increasing if humanity refuses to act, but all those "details" we need to adapt are far from clear. Australians needed that information, farmers, engineers, companies, governors and majors, will find that CSIRO will not be able to answer their telephone calls.

Wednesday, 17 February 2016

The global warming conspiracy would be huge

The concept of global warming was created by and for the Chinese in order to make US manufacturing non-competitive.
Republican front runner Donald Trump on Twitter

Snowing in Texas and Louisiana, record setting freezing temperatures throughout the country and beyond. Global warming is an expensive hoax!
Republican front runner Donald Trump on Twitter

How do you know the climate didn't actually cool?
Eric Worrall, the main contributor to WUWT

Why use discredited surface data which everyone knows is fraudulent?
"Scottish" "Sceptic"

I am working on a study to compare nationally homogenized temperature data with the temperatures in the large international collections (GHCN, CRUTEM, etc.). Looking for such national datasets, I found many graphs in the scientific literature showing national temperature increases, which I want to share with you.

Mitigation skeptics like to talk about "The Team", as if a small group of people would be "in charge". That makes their conspiracy theories a little less absurd, although even small conspiracies typically do not last for decades. The national temperature series show that hundreds of national weather services and numerous universities would also need to be in the conspiracy of science against mankind. To me that sounds unrealistic.

The mitigation skeptics have a rough time and nowadays more often claim that they do not dispute the greenhouse effect or the warming of the Earth at all, but only bla, bla, bla. Which is why I thought I would show that this post is not fighting strawmen by citing some of the main bloggers and political leaders of the mitigation skeptical movement at the top of this post.

Anthony Watts, the weather presenter hosting Watts Up With That (WUWT), typically claims that only half of the warming is real, although he recently softened his stance for the USA and now only claims that a third is not real. If half of the warming in the global collections were not real, many scientists would have noticed that the global data does not fit to their local observations.

Plot idea: 97% of the world's scientists contrive an environmental crisis, but are exposed by a plucky band of billionaires & oil companies.
Scott Westerfeld

And do not forget all the other scientists studying other parts of the climate system, the upper air, ground temperatures, sea surface temperature, ocean heat content, precipitation, glaciers, ice sheets, lake temperatures, sea ice, lake and river freezing, snow, birds, plants, insects, agriculture. One really wonders with Eric Worrall how on Earth science knows the climate didn't actually cool.

Another reason to write this post is to ask for help. For this comparison study, I have datasets or first contacts for the countries below. If you know of more homogenized datasets, please, please let me know. Even if it is "only" a reference. Also if you have a dataset from one of the countries below: multiple datasets from one country are very much welcome.

Countries: Albania, Argentina, Armenia, Australia, Austria, Belgium, Benin, Bolivia, Bulgaria, Canada, Chile, China, Congo Brazzaville, Croatia, Czech Republic, Denmark, Ecuador, Estonia, Finland, France, Germany, Greece, Hungary, Iran, Israel, Italy, Latvia, Libya, Macedonia, Morocco, Netherlands, New Zealand, Norway, Peru, Philippines, Portugal, Romania, Russia, Serbia, Slovakia, Slovenia, South Africa, Spain, Sweden, Switzerland, Tanzania, Uganda, Ukraine, United Kingdom, United States of America.
Regions: Catalonia, Carpathian basin, Central England Temperature, Greater Alpine Region.

Alpine region

The temperature for the Greater Alpine Region from the HISTALP project (Ingeborg Auer and colleagues, 2007). The lower panel shows the temperature for four low altitude regions. The top panel their average (black) and the signal for the high altitude stations (grey). All series are smoothed over 10 years.


The increase in the annual temperatures and the decrease in annual precipitation in Armenia. From Levon Vardanyan and colleagues (2013), see also Artur Gevorgyan and colleagues (2016).


The temperature signal over Australia for the day-time maximum temperature (red), the mean temperature (green) and the night-time minimum temperature. Figure from Fawcett and colleagues (2012).


From Lucie Vincent of Environment Canada and colleagues (2012).

The Czech Republic

Changes in mean annual and seasonal temperature time series for the Czech Lands in the period 1800–2010. The part of series calculated from only two stations is expressed by a dashed line. Figure by Petr Stepanek of the Global Change Research Institute CAS, Brno, Czech Republic.


The temperature change in China over the last 106 years, the annual mean temperature and the seasonal temperatures from QingXiang Li and colleagues (2010).


The famous Central England Temperature series of the Hadley Centre.


The annual average temperature in Finland. National averages are more noisy than global averages. Thus to show the trend better the graph adds the decadal average temperature. From Mikkonen and colleagues (2015).


The temperature signal since 1900 in India according to Kothawale et al. (2010) of the Indian Institute of Tropical Meteorology (IITM), Pune.


The temperature series of Italy since 1800 according to Michele Brunetti and colleagues (2006).

Middle America and Northern South America

These graphs show the change in the number of warm days (maximum temperature) and warm night (minimum temperature) and the number of cold days and cold nights computed from daily data from several countries in Middle America and in the North of South America. Figures from Enric Aguilar and colleagues (2005).

The Netherlands

Annual mean temperatures of the actual observations at De Bilt (red), the De Bilt homogenised series (dark blue), the previous version of the Central Netherlands Temperature series (CNT2,7; light blue) and the current Central Netherlands Temperature series (CNT4,6; pink). Gerard van der Schrier and colleagues (2009) from the Dutch weather service, KNMI. De Bilt is a city in the middle of The Netherlands were the KNMI main office is. The Central Netherlands series is for a larger region in the middle of The Netherlands.

New Zealand

The famous New Zealand 7-stations series.


Observed annual mean temperature anomalies in the Philippines during the period 1951–2010 computed by Thelma A. Cincoa and colleagues (2014).


Temperature averaged over Russia from the annual climate report of ROSHYDROMET (2014). The top panel are the annual averages, the four lower panels the seasons (winter, spring, summer and autumn). No homogenization.

The variability in winter is very high. According to mitigation sceptic Anthony Watts this is due to Russian Steam Pipes:
I do know this: neither I nor NOAA has a good handle on the siting characteristics of Russian weather stations. I do know one thing though, the central heating schemes for many Russian cities puts a lot of waste heat into the air from un-insulated steam pipes.
Then it would be surprising that such large regions are affected in the same way and that the steam pipe years are also hot in the analysis of global weather prediction models and satellite temperature datasets.


Temperature trends computed by José Antonio Guijaro (2015) of the Spanish State Meteorological Agency (AEMET) for 12 river catchments within Spain. Homogenization with CLIMATOL.

The Spanish temperature dataset of the URV University in Tarragona. The panels on the left show the minimum temperature, the panels on the right the maximum temperature. The top panels show raw data before homogenization, the lower panels the homogenized data. The maximum temperature before 1910 had to be corrected strongly because of the use of a French screen before this time.

United States of America

The minimum and maximum temperature of the lower 48 states of the United States of America computed by NOAA. You can see it is an original American-made graph because it is in [[Fahrenheit]].


The temperature signal in Switzerland computed by Michael Begert and colleagues of the MeteoSchweiz. The top panel show original station time series, the lower panel shows them after removal of non-climatic changes.

Related reading

Climatologists have manipulated data to REDUCE global warming

Charges of conspiracy, collusion and connivance. What to do when confronted by conspiracy theories?

If you're thinking of creating a massive conspiracy, you may be better scaling back your plans, according to an Oxford University researcher.


Aguilar, E., et al., 2005: Changes in precipitation and temperature extremes in Central America and northern South America, 1961–2003. Journal Geophysical Research, 110, D23107, doi: 10.1029/2005JD006119.

Auer, I., Böhm, R., Jurkovic, A., Lipa, W., Orlik, A., Potzmann, R., Schöner, W., Ungersböck, M., Matulla, C., Briffa, K., Jones, P., Efthymiadis, D., Brunetti, M., Nanni, T., Maugeri, M., Mercalli, L., Mestre, O., Moisselin, J.-M., Begert, M., Müller-Westermeier, G., Kveton, V., Bochnicek, O., Stastny, P., Lapin, M., Szalai, S., Szentimrey, T., Cegnar, T., Dolinar, M., Gajic-Capka, M., Zaninovic, K., Majstorovic, Z. and Nieplova, E., 2007: HISTALP—historical instrumental climatological surface time series of the Greater Alpine Region. International Journal of Climatology, 27, pp. 17–46. doi: 10.1002/joc.1377.

Begert, M., Schlegel, T. and Kirchhofer, W., 2005: Homogeneous temperature and precipitation series of Switzerland from 1864 to 2000. International Journal of Climatology, 25, pp. 65–80. doi: 10.1002/joc.1118.

Brunetti, M., Maugeri, M., Monti, F. and Nanni, T., 2006: Temperature and precipitation variability in Italy in the last two centuries from homogenised instrumental time series. International Journal of Climatology, 26, pp. 345–381, doi: 10.1002/joc.1251.

Cincoa, Thelma A., Rosalina G. de Guzmana, Flaviana D. Hilarioa, David M. Wilson, 2014: Long-term trends and extremes in observed daily precipitation and near surface air temperature in the Philippines for the period 1951–2010. Atmospheric Research, 145–146, pp. 12–26, j.atmosres.2014.03.025.

Fawcett, R.J.B., B.C. Trewin, K. Braganza, R.J Smalley, B. Jovanovic and D.A. Jones, 2012: On the sensitivity of Australian temperature trends and variability to analysis methods and observation networks. CAWCR Technical Report No. 050.

Gevorgyan, A., H. Melkonyan, T. Aleksanyan, A. Iritsyan and Y. Khalatyan, 2016: An assessment of observed and projected temperature changes in Armenia. Arabian Journal of Geosciences, 9, pp. 1-9, DOI 10.1007/s12517-015-2167-y.

Guijaro, J.A., 2015: Temperature trends. AEMET Report.

Jain, Sharad K. and Vijay Kumar, 2012: Trend analysis of rainfall and temperature data for India. Current Science, 102.

Kothawale, D.R., A.A. Munot, K. Krishna Kumar, 2010: Surface air temperature variability over India during 1901–2007, and its association with ENSO. Climate Research, 42, pp. 89-104.

Li Q X, Dong W J, Li W, et al. Assessment of the uncertainties in temperature change in China during the last century. Chinese Sci Bull, 2010, 55: 1974−1982, doi: 10.1007/s11434-010-3209-1

Mikkonen, S., M. Laine, H. M. Mäkelä, H. Gregow, H. Tuomenvirta, M. Lahtinen, A. Laaksonen, 2015: Trends in the average temperature in Finland, 1847–2013. Stochastic Environmental Research and Risk Assessment, 29, Issue 6, pp 1521-1529, doi: 10.1007/s00477-014-0992-2.

Schrier, van der, G., A. van Ulden, and G. J. van Oldenborgh, 2011: The construction of a Central Netherlands temperature. Climate of the Past, 7, 527–542, doi: 10.5194/cp-7-527-2011

Ulden, van, Aad, Geert Jan van Oldenborgh, and Gerard van der Schrier, 2009: The Construction of a Central Netherlands Temperature. Scientific report, WR2009-03. See also Van der Schrier et al. (2011).

Vardanyan, L., H. Melkonyan, A. Hovsepyan, 2013: Current status and perspectives for development of climate services in Armenia. Report, ISBN 978-9939-69-050-6.

Vincent, L. A., X. L. Wang, E. J. Milewska, H. Wan, F. Yang, and V. Swail (2012), A second generation of homogenized Canadian monthly surface air temperature for climate trend analysis, J. Geophys. Res., 117, D18110, doi: 10.1029/2012JD017859.

Thursday, 11 February 2016

Early global warming

How much did the world warm during the transition to Stevenson screens around 1900?

Stevenson screen in Poland.

The main global temperature datasets show little or no warming in the land surface temperature and the sea surface temperature for the period between 1850 and 1920. I am wondering whether this is right or whether we do not correct the temperatures enough for the warm bias of screens that were used before the Stevenson screen was introduced. This transition mostly happened in this period.

This is gonna be a long story, but it is worth it. We start with the current estimates of warming in this period. There is not much data on how large the artificial cooling due to the introduction of Stevenson screens is, thus we need to understand why thermometers in Stevenson screens record lower temperatures than before to estimate how much warming this transition may have hidden. Then we compare this to the corrections NOAA makes for the introduction of the Stevenson screen. Also other changes in the climate system suggest there was warming in this period. It is naturally interesting to speculate what this stronger early warming may mean for the causes of global warming.

No global warming in main datasets

The figure below with the temperature estimates of the four main groups show no warming for the land temperature between 1850 and 1920. Only Berkeley and CRUTEM start in 1850, the other two later.

If you look at the land temperatures plotted by Berkeley Earth themselves there is actually a hint of warming. The composite figure below shows all four temperature estimates for their common area for the best comparison, while the Berkeley Earth figure is interpolated over the entire world and thus sees Arctic warming more, which was strong in this period, like it again was strong in recent times. Thus there was likely some warming in this period, mainly due to the warming Arctic.

The temperature changes of the land according to the last IPCC report. My box.

In the same period the sea surface temperature was even cooling a little according to HadSST3 shown below.

The sea surface temperature of the four main groups and night marine air temperature from the last IPCC report. I added the red box to mark the period of interest.

Also the large number of climate models runs produced by the Coupled Model Intercomparison Project (CIMP5), colloquial called IPCC models, do not show much warming in our period of interest.

CMIP5 climate model ensemble (yellow lines) and its mean (red line) plotted together with several instrumental temperature estimates (black lines). Figure from Jones et al. (2013) with our box added to emphasize the period.

Transition to Stevenson screens

In early times temperature observations were often made in unheated rooms or in window screens of such rooms facing poleward. These window screens protected the expensive thermometers against the weather and increasingly also against direct sun light, but a lot of sun could get onto the instrument or the sun could heat the wall beneath the thermometer and warm air would rise up.

A Wild screen (left) and a Stevenson screen in Basel, Switzerland.
When it was realised that these measurements have a bias, a period with much experimentation ensued. Scientists tried stands (free standing vertical boards with a little roof that often had to be rotated to avoid sun during sunrise and -fall), shelters of various sizes that were open to the poles and to the bottom, screens of various sizes, sometimes near the shade of a wall, but mostly in gardens and pagoda huts that could have been used for a tea party.

The more open a screen is, the better the ventilation, which likely motived earlier more open designs, but this also leads to radiation errors. In the end the Stevenson screen became the standard, which protects the instrument from radiation from all sides. It is made of white painted wood and has a measurement chamber mounted on a wood frame, it typically has a double board roof and double Louvred walls to all sides. Initially it sometimes did not have a bottom, but later had slanted boards at the bottom.

The first version [[Stevenson screen]] was crafted in 1864 in the UK, the final version designed in 1884. It is thought that most countries switched to Stevenson screens before 1920, but some countries were later. For example, Switzerland made the transition from Wild screens to Stevenson screens in the 1960s. The Belgium Station Uccle changed their half open shelter to a Stevenson screen in 1983. The rest of Belgium in the 1920s.

Open shelter (at the front) and two Stevenson screens (in the back) at the main office of the Belgium weather service in Uccle.

Radiation error

The schematic below shows the main factors influencing the radiation error. Solar radiation makes the observed maximum temperatures too warm. This can be direct radiation or radiation scattered via clouds or the (snow covered) ground. The sun can also heat the outside of a not perfectly white screen, which then warms the air flowing in. Similarly the sun can heat the ground, which then may radiate towards the thermometer and screen. However, the lack of radiation shielding also makes the minimum temperature too low when the thermometer radiates infrared radiation into the cold sky. This error is largest on dry cloudless nights and small when the sky radiates back to the thermometer, which happens when the sky is cloudy and the absolute humidity is high, which reduces the net infrared radiative cooling. The radiation error is largest when there is not much ventilation, which in most cases need wind. The direct radiation effects are smaller for smaller thermometers.

Schematic showing the various factors that can influence the radiation error of a temperature sensor.

From our understanding of the radiation error, we would thus expect the bias in the day-time maximum temperature to be large where the sun is strong, the wind is calm, the soil is dry and heats up fast. The minimum temperature at night has the largest cooling bias when the sky is cloudless and dry.

This means that we expect the radiation errors for the mean temperature to be largest in the tropics (strong sun and high humidity) and subtropics (sun, hot soil), while it is likely smallest in the mid and high latitudes (not much sun, low specific humidity), especially near the coast (wind). Continental climates are the question mark; they have dry soils and not much wind, but also not as much sun and low absolute humidity.

Parallel measurements

These theoretical expectations fit to the limited number of temperature differences found in the literature; see table below. For the mid-latitudes, David Parker (1994) found that the difference was less than 0.2°C, but his data mainly came from maritime climates in north-west Europe. Other differences found in the mid-latitudes are about 0.2°C (Kremsmünster, Austria; Adelaide, Australia; Basel, Switzerland). While in the sub-tropics we have one parallel measurement showing a difference of 0.35°C and the two tropical parallel measurements show have a difference of 0.4°C. We are missing information from continental climates.

Table with the differences found for various climates and early screen1. Temperature difference in Basel is about zero using 3 fixed hour measurements to compute mean temperature, which was the local standard, but about 0.25 when using minimum and maximum temperature as is used most for global studies.
Region Screen Temperature difference
North-West Europe Various; Parker (1994) < 0.2°C
Basel, Switzerland Wild screen; Auchmann & Brönnimann (2012) ˜0 (0.25)°C 1
Kremsmünster, Austria North-wall window screen; Böhm et al. (2010) 0.2°C
Adelaide, South Australia Glaisher stand; Nicholls et al. (1996) 0.2°C
Spain French screen; Brunet et al. (2011) 0.35 °C
Sri Lanka Tropical screen; in Parker (1994) 0.37°C
India Tropical screen; in Parker (1994) 0.42°C

Most of the measurements we have are in North West Europe and do not show much bias. However, theoretically we would not expect much radiation errors here. The small number of estimates showing large biases come from tropical and sub-tropical climates and may well be representative for large parts of the globe.

Information on continental climates is missing, while they also make up a large part of the Earth. The bias could be high here because of calm winds and dry soils, but the sun is on average not as strong and the humidity low.

Next to the climatic susceptibility to radiation errors also the designs of the screens used before the Stevenson screen could be important. In the numbers in the table we do not see much influence of the designs, but maybe we will see it when we get more data.

Global Historical Climate Network temperatures

The radiation error and thus the introduction of Stevenson screens affected the summer temperatures more than the winter temperatures. Thus it is interesting that the trend in winter is 3 times stronger in the (Northern Hemisphere, GHCNv3). In winter it is 1.2°C per century, in summer it is 0.4°C per century over the period 1881-1920; see figure below2.

Also without measurement errors, the trend in winter is expected to be larger than in summer because the enhanced greenhouse effect affects winter temperatures more. In the CMIP5 climate model average the winter trend is about 1.5 times the summer trend3, but not 3 times.

Temperature anomalies in winter and summer over land in NOAA’s GHCNv3. The light lines are the data, the thick striped lines the linear trend estimates.

The adjustments made by the pairwise homogenization algorithm of NOAA for the study period are small. The left panel of the figure below shows the original and adjusted temperature anomalies of GHCNv3. The right panel shows the difference, which shows that there are adjustments in the 1940s and around 1970. The official GHCN global average starts in 1880. Zeke Hausfather kindly provided me with his estimate starting in 1850. During our period of interest the adjustments are about 0.1°C; a large part of which was before 1880.

These adjustments are smaller than the jump expected due to the introduction of the Stevenson screens. However, they should also be smaller because many stations will have started as Stevenson screens. It is not known how large percentage this is, but the adjustments seem small and early.

Other climatic changes

So far for the temperature record. What do other datasets say about warming in our period?

Water freezing

Lake and river freeze and breakup times have been observed for a very long time. Lakes and rivers are warming at a surprisingly fast rate. They show a clear shortening of the freezing period between 1850 and 1920; the freezing started later and ice break-up started. The figure below shows that this was already going on in 1845.

Time series of freeze and breakup dates from selected Northern Hemisphere lakes and rivers (1846 to 1995). Data were smoothed with a 10-year moving average. Figure 1 from Magnuson et al. (2002).

Magnuson has updated his dataset regularly: when you take the current dataset and average over all rivers and lakes that have data over our period you get the clear signal shown below.

The average change in the freezing date in days and the ice break-up date (flipped) is shown as red dots and smoothed as a red line. The smoothed series for individual lakes and rivers freezing or breaking up is shown in the background as light grey lines.


Most of the glaciers for which we have data from this period show reductions in their lengths, which signals clear warming. Oerlemans (2005) used this information for a temperature reconstruction, which is tricky because glaciers respond slowly and are also influenced by precipitation changes.

Temperature estimate of Oerlemans (2005) from glacier data. (My red boxes.)


Temperature reconstructions from proxies show warming. For example the NTREND dataset based on tree proxies from the Northern Hemisphere as plotted below by Tamino.

Temperature reconstruction of the non-tropical Northern Hemisphere.

Paleo Model Intercomparison project

While the CMIP5 climate model runs did not show much warming in our period, the runs for the last millennium of the PMIP3 project do show some warming, although it strongly depends on the exact period; see below. The difference between CMIP5 and PMIP3 is likely that in the beginning of the 19th century there was much volcanic activity, which decreased the ocean temperature to below its equilibrium and it took some decades for it to return to its equilibrium. CMIP5 starts in 1850 and modelers try to start their models in equilibrium.

Simulated Northern Hemisphere mean temperature anomalies from PMIP3 for last millennium. CCSM4 shows the simulated Northern Hemisphere mean temperature anomalies (annual values in light gray, 30-yr Gaussian smoothed in black). For comparison various smoothed reconstructions (colored lines) are included which come from a variety of proxies, including tree ring width and density, boreholes, ice cores, speleothems, documentary evidence, and coral growth.

Sea surface temperature

Land surface warming is important for us, but does not change the global mean temperature that much. The Earth is a blue dot; 70% of our planet is ocean. Thus is we had a bias in the station data our period of 0.3°C, that would be a bias global temperature of 0.1°C. However, larger warming of land temperatures are difficult if the sea surface is not also warming and currently the data shows a slight cooling over our period. I have no expertise here, but wonder if such a large difference would be reasonable.

Thus maybe we overlooked a source of bias in the sea surface temperature as well. It was a period in which sailing ships were replaced by steamships, which was a large change. The sea surface temperature was measured by sampling a bucket of water and measuring its temperature. During the measurement, the water would evaporate and cool. On a steamship there is more wind than on a sailing ship and thus maybe more evaporation. The shipping routes have also changed.

I must mention that it is a small scandal how few scientists work on the sea surface temperature. It would be about a dozen and most of them only part-time. Not only is the ocean 2/3 of the Earth, the sea surface temperature is also often used to drive atmospheric climate models and to study climate modes. The group is small, while the detection of trend biases in sea surface temperature is much more difficult than in station data because they cannot detect unknown changes by comparing stations with each other. The maritime climate data community deserves more support. There are more scientists working on climate impacts for wine; this is absurd.

A French (Montsouri) screen and two Stevenson screens in Spain. The introduction of the Stevenson screen went fast in Spain and was hard to correct using statistical homogenization alone. Thus a modern replica of the original French screen build for an experiment, which was part of the SCREEN project.

Causes of global warming

Let's speculate a bit more and assume that the sea surface temperature increase was also larger than currently thought. Then it would be interesting to study why the models show less warming. An obvious candidate would be aerosols, small particles in the air, which have also increased with the burning of fossil fuels. Maybe models overestimate how much they cool the climate.

The figure from the last IPCC report below shows the various forcings of the climate system. These estimates suggest that the cooling of aerosols and the warming of greenhouse gases is similar in climate models until 1900. However, with less influence of aerosols, the warming would start earlier.

Stevens (2015) argues that we have overestimated the importance of aerosols. I do not find Stevens' arguments particularly convincing, but everyone in the field agrees that there are at least huge uncertainties. The CMIP5 figure gives the error bars at the right and it is within the confidence interval that there is effectively nearly no net influence of aerosols (ochre bar at the right).

There is direct cooling of aerosols due to scattering of solar radiation. This is indicated in red as "Aer-Rad int." This is uncertain because we do not have good estimates on the amount and size of the aerosols. Even larger uncertainties are in how aerosols influence the radiative properties of clouds, marked in ochre as "Aer-Cld int."

Some of the warming in our period was also due to less natural volcanic aerosols at the end. Their influence on climate is also uncertain because of lack of observations on the size of the eruptions and the spatial pattern of the aerosols.

Forcing estimate for the IPPC AR5 report.

The article mentioned in the beginning (Jones et al. 2013) showing the CMIP5 global climate model ensemble temperatures for all forcings, which did not show much warming in our period, also gives results for model runs that only include greenhouse gases, which shows a warming of about 0.2°C; see below. If we interpret this difference as the influence of aerosols, (there is also a natural part) then aerosols would be responsible for 0.2°C cooling in our period in the current model runs. In the limit of the confidence interval were aerosols do not have a net influence, an additional warming of 0.2°C could thus be explained by aerosols.

CMIP5 climate model ensemble (yellow lines) and its mean (red line) plotted together with several instrumental temperature estimates (black lines). Figure from Jones et al. (2013) with our box added to estimate the temperature increase.

Conclusion on early global warming

Several lines of evidence suggest that the Earth’s surface actually was warming during this period. Every line of evidence by itself is currently not compelling, but the [[consilience]] of evidence at least makes a good case for further research and especially to revisit the warming bias of early instrumental observations.

To make a good case, one would have to make sure that all datasets cover the same regions/locations. With the modest warming during this period, the analysis should be very careful. It would also need an expert for each of the different measurement types to understand the uncertainties in their trends. Anyone interested in make a real publishable study out of this please contact me.

Austrian Hann screen (a large screen build close to a northern wall) and a Stevenson screen in Graz, Austria.

Collaboration on studying the bias

To study the transition to Stevenson screens, we are collecting data from parallel measurements of early instrumentation with Stevenson screens.

We have located the data for the first seven sources listed below.

Australia, Adelaide, Glaisher stand
Austria, Kremsmünster, North Wall
Austria, Hann screen in Vienna and Graz
Spain, SCREEN project, Montsouris (French) screen in Murcia and La Coruña
Switzerland, Wild screen in Basel and Zurich
Northern Ireland, North wall in Armagh
Norway, North wall

Most are historical datasets, but there are also two modern experiments with historical screens (Spain and Kremsmünster). Such experiments with replicas is something I hope will be done more in future. It could also be an interesting project for an enthusiastic weather observer with an interest in history.

From the literature we know of a number of further parallel measurements all over the world; listed below. If you have contacts to people who may know where these datasets are, please let us know.

Belgium, Uccle, open screen
Denmark, Bovbjerg Fyr, Skjoldnñs, Keldsnor, Rudkùbing, Spodsbjerg Fyr, Gedser Fyr, North wall.
France, Paris, Montsouris (French) screen
Germany, Hohenpeissenberg, North wall
Germany, Berlin, Montsouris screen
Iceland, 8 stations, North wall
Northern Ireland, a thermograph in North wall screen in Valentia
Norway, Fredriksberg observatory, Glomfjord, Dombas, North wall
Samoa, tropic screen
South Africa, Window screen, French and Stevenson screens
Sweden, Karlstadt, Free standing shelter
Sweden, Stockholm Observatory
UK, Strathfield Turgiss, Lawson stand
UK, Greenwich, London, Glaisher stand
UK, Croydon, Glaisher stand
UK, London, Glaisher stand

To get a good estimate of the bias we need many parallel measurements, from as many early screens as possible and from many different climatic regions, especially continental, tropical and sub-tropical climates. Measurements made outside of Europe are lacking most and would be extremely valuable.

If you know of any further parallel measurements, please get in touch. It does not have to be a dataset, also a literature reference is a great hint and a starting point for a search. If your twitter followers or facebook friends may have parallel datasets please post this post on POST.

Related reading

Parallel Observations Science Team (POST) of the International Surface Temperature Initiative (ISTI).

The transition to automatic weather stations. We’d better study it now.

Why raw temperatures show too little global warming.

Changes in screen design leading to temperature trend biases.


1) The difference in Basel is nearly zero if you use the local way to compute the mean temperature from fixed hour measurements, but it is about 0.25°C if you use the maximum and minimum temperature, which is mostly used in climatology.

2) Note that GHCNv3 only homogenizes the annual means, that is, every month gets the same corrections. Thus the difference in trends between summer and winter shown in the figure is like it is in the raw data.

3) The winter trend is 1.5 times the summer trend in the mean temperature of the CMIP5 ensemble for the Northern Hemisphere (ocean and land). The factor three we found in for GHCN was only for land. Thus a more careful analysis may find somewhat different values.


Auchmann, R. and S. Brönnimann, 2012: A physics-based correction model for homogenizing sub-daily temperature series. Journal Geophysical Research Atmospheres., 117, art. no. D17119, doi: 10.1029/2012JD018067.

Bjorn Stevens, 2015: Rethinking the Lower Bound on Aerosol Radiative Forcing. Journal of Climate, 28, pp. 4794–4819, doi: 10.1175/JCLI-D-14-00656.1.

Böhm, R., P.D. Jones, J. Hiebl, D. Frank, et al., 2010: The early instrumental warm-bias: a solution for long central European temperature series 1760–2007. Climatic Change, 101, pp. 41–67, doi: 10.1007/s10584-009-9649-4.

Brunet, M., J. Asin, J. Sigró, M. Bañón, F. García, E. Aguilar, J. Esteban Palenzuela, T.C. Peterson, P. Jones, 2011: The minimization of the screen bias from ancient Western Mediterranean air temperature records: an exploratory statistical analysis. International Journal Climatololgy, 31, 1879–1895, doi: 10.1002/joc.2192.

Jones, G. S., P. A. Stott, and N. Christidis, 2013: Attribution of observed historical near‒surface temperature variations to anthropogenic and natural causes using CMIP5 simulations. Journal Geophysical Research Atmospheres, 118, 4001–4024, doi: 10.1002/jgrd.50239.

Magnuson, John J., Dale M. Robertson, Barbara J. Benson, Randolf H. Wynne, David M. Livingstone, Tadashi Arai, Raymond A. Assel, Roger B. Barry, Virginia Card, Esko Kuusisto, Nick G. Granin, Terry D. Prowse, Kenton M. Stewart, and Valery S. Vuglinski, 2000: Historical trends in lake and river ice cover in the Northern Hemisphere. Science, 289, pp. 1743-1746, doi: 10.1126/science.289.5485.1743

Nicholls, N., R. Tapp, K. Burrows, and D. Richards, 1996: Historical thermometer exposures in Australia. International Journal of Climatology, 16, pp. 705-710, doi: 10.1002/(SICI)1097-0088(199606)16:6<705::AID-JOC30>3.0.CO;2-S.

Oerlemans, J., 2005: Extracting a Climate Signal from 169 Glacier Records. Science, 308, no. 5722, pp. 675-677, doi: 10.1126/science.1107046.

Parker, D.E., 1994: Effects of changing exposure of thermometers at land stations. International Journal Climatology, 14, pp. 1–31, doi: 10.1002/joc.3370140102.

Photo at the top a Stevenson screen of the amateur weather station near Czarny Dunajec, Poland. Photographer: Arnold Jakubczyk.
Photos of Wild screen and Stevenson screen in Basel by Paul Della Marta.
Photo of open shelter in Belgium by Belgium weather service.
Photo of French screen in Spain courtesy of SCREEN project.
Photo of Hann screen and Stevenson screen in Graz courtesy of the University of Graz.