Changes in the observing conditions or in the environment of the meteorological stations introduce anomalous perturbations in the data series (inhomogeneities). Therefore, homogenization is of paramount importance to obtain reliable analysis of the variability of climatological series, and there exists an extensive literature about the different methodologies applied so far (comprehensive reviews can be seen in Peterson et al., 1998, Aguilar et al., 2003 and Beaulieu et al., 2008). This allows climatologists to choose their preferred methods for their research, often implementing those methods by themselves. But some of them have made their developments available to other users by preparing and documenting ready to use computer packages and distributing them in the Internet or by other means. This is very interesting for operational climatology units of many National Meteorological and Hydrological Services around the world, that sometimes can lack the needed expertise to make their own developments, or else can devote the required time to other priorities.

The European Science Foundation, by means of the COST Office, funded the Action ES0601 («HOME», Advances in Homogenisation Methods of Climate Series: An Integrated Approach) that enabled the inter-comparison of many participating methodologies. Its final conference was held in Budapest during the last week of October 2011, and the results were summarized by Venema et al. (2011). However, and thanks to the discussions held in the frame of this Action, many methods improved their algorithms, and further work is needed to update the information about their performances. Current efforts include the International Surface Temperature Initiative, the Spanish project MULTITEST (for monthly series) and the European project INDECIS (for daily series).

The following tables summarize the main characteristics of currently available homogenization packages. (The first version was compiled thanks to several developers that replied to a survey, complemented by the information gathered from Venema et al., 2011.) After the publication of the Guidelines on homogenization some updates have been incorporated from developers or other sources.)

Explanatory notes and other characteristics:

• The primary operation qualification is somehow controversial: While every piece of software is intended to perform a number of operations automatically, some people would qualify them as automatic when no subjective decision has to be made by the user, while others would reserve this adjective for the cases when no prompt is required to specify some parameters or answer some questions in order to allow the process to continue. In this more restrictive sense, the only absolutely automatic package is Climatol, where all needed parameters are specified when calling the homogenization function, while others only need to answer a limited number of questions (ACMANT, MASH, ...) or running a few scripts (USHCN). In these cases, a fully automatic operation can be achieved by constructing a master script to call the needed scripts and to provide the required answers to the prompts. On the other hand, RHtests works on single series, and it is the user who must supply the reference series in order to perform relative homogenization. Therefore, although quite automatic in the sense that only a limited human intervention is needed, it can be considered interactive because of the subjective important choices that must be taken.
• *Max. number of series = 9999 means that the limit is imposed by the available computer resources.
• GSIMCLI detection statistic is based on simulated probability distribution of the candidate series. Its manuals are available at https://gsimcli.readthedocs.org
• Packages provided as R source can be run on any platform where R can be installed (since R is an interpreted language). USHCN requires the compilation of the FORTRAN sources, following the directions of the package. (Compilation errors of 'unknown operator' arise if /usr/bin/sh is not a link to /usr/bin/bash.)
• AnClim/ProClimDB is a software suite that performs many kind of statistical analysis on climatological networks, but the full functionality is not in the freeware version. (It could not be tested in automatic mode because it is very Windows dependent).
• Climatol is highly missing data tolerant, and therefore can use the information of the short series of the network as well. Accompanying post-processing functions provide climatic summaries, OLS trends and grids of the homogenized series.
• HOMER relies on subjective judgment, and is therefore addressed to trained users. An accompanying program called SPLIDHOM allows the adjustment of daily series if the break dates are provided.

References:

Aguilar E, Auer I, Brunet M, Peterson TC, Wieringa J (2003): Guidelines on climate metadata and homogenization. WCDMP-No. 53, WMO-TD No. 1186. World Meteorological Organization, Geneve.

Beaulieu C, Seidou O, Ouarda TBMJ, Zhang X, Boulet G, Yagouti A (2008): Intercomparison of homogenization techniques for precipitation data. Water Resour. Res., 44, 20 pp.

Peterson TC, Easterling DR, Karl TR, Groisman P, Nicholls N, Plummer N, Torok S, Auer I, Böhm R, Gullett D, Vincent L, Heino R, Tuomenvirta H, Mestre O, Szentimrey T, Salinger J, Førland E, Hanssen-Bauer I, Alexandersson H, Jones P, Parker D (1998): Homogeneity Adjustments of 'In Situ' Atmospheric Climate Data: A Review. Int. J. Climatol., 18:1493-1518.

Venema V, Mestre O, Aguilar E, Auer I, Guijarro JA, Domonkos P, Vertacnik G, Szentimrey T, Stepanek P, Zahradnicek P, Viarre J, Müller-Westermeier G, Lakatos M, Williams CN, Menne M, Lindau R, Rasol D, Rustemeier E, Kolokythas K, Marinova T, Andresen L, Acquaotta F, Fratianni S, Cheval S, Klancar M, Brunetti M, Gruber C, Prohom Duran M, Likso T, Esteban P and Brandsma T (2012): Benchmarking homogenization algorithms for monthly data. Clim. of the Past, 8:89-115.