Researchers from the C3 are the co-authors of an article that confirms the goodness of adjustments to the homogeneity of historical climate series Research

Enric Aguilar and Peter Domonkos are researchers at the Climate Change Centre C3 of the URV who have made a significant contribution to the article: Benchmarking homogenization algorithms for monthly data. The article was published on 10 January in the journal "Climate of the Past".
 

Homogenization improves the reliability of historical climate records. This is the conclusion of the study, whose main findings are published at www.clim-past.net/8/89/2012/cp-8-89-2012.pdf  in the open access journal "Climate of the Past", which is edited by the European Geosciences Union.

Climatic series are affected by various factors that can introduce artificial biases. These factors include: movements, changes in instrumentation, changes in the weather station, seasonal alterations, changes in the observation times, etc. All of these phenomena can obscure the true picture of the climate.

For this reason it is necessary correct the series in order to ensure that climate studies are accurate. This process is known as homogenization. For decades, climatologists have been aware of this problem and have applied different homogenization techniques that have been valid from a strictly statistical point of view.

In order to study the effect of these techniques on climate data, a series of tests have been carried out on datasets generated artificially using "surrogates". This allows the structure and properties of the original datasets to be reproduced with unprecedented accuracy. Non-climatic alterations were introduced to the original or homogenous benchmark, thus creating an inhomogenous benchmark.

The aim of the experiment was to apply different statistical techniques in order to homogenize the dataset; that is, to return the data to its original state free from non-climatic disturbances. Where this experiment differs from previous ones is that it has carried out a blind comparison: the group that generated the benchmark worked independently of the group that applied the methods for detecting and correcting inhomogeneities, thus increasing the validity of the results.
 
The results obtained indicate that after applying almost all of the methods analyzed, the series are more homogenous; that is, the changes applied have improved the researchers’ ability to correctly describe climatic changes. The experiment confirms that the trends observed in homogenized series are more reliable and accurate than trends that are observed in datasets that have not been adjusted; a finding that contradicts the claims of climate change sceptics.
 
The study contains contributions from 31 authors, including those who developed the simulated series and those who carried out the experiments on them. Among these are two researchers from the Climate Change Centre, C3 of the Department of Geographyat the Universitat Rovira i Virgili (Dr Enric Aguilar and Dr Peter Domonkos); a researcher at the State Meteorology Agency (Dr José Antonio Guijarro); a researcher from the  Meteorological Service of Catalonia (Dr Marc Prohom) and a researcher at the Andorran Centre for Snow and Mountain Studies (Pere Esteban).
 
This study was performed with support of the COST Action ES0601 – Advances in Homogenisation Methods of Climate Series: an Integrated Approach (HOME)www.homogenisation.org). HOME has also developed free software that allows users to apply both the techniques that have presented the best results and the innovations introduced on the basis of the experiments carried out. This software will soon be available to the international community.

The study could also not have been carried out without the administrative support of the URV Foundation and the collaboration of scientists from all 27 countries in the European Union, Andorra, the United States and Canada.