Sigró et al. (2024) wrote the following about time series homogenization:
“Several studies have compared these methods to assess their efficacy (Domonkos et al., 2021; Guijarro et al., 2023; Venema et al., 2012). The CLIMATOL software package (http://www.climatol.eu/), which was selected to test the homogeneity of these series, is the semiautomatic method with the best results in the detection and adjustment of inhomogeneities of monthly temperature and precipitation series.”
The text falsely suggests that Climatol was found to be the most accurate method by the referred method comparison tests. I have tried to achieve the publication of a corrigendum (from the main author) or a critical note (from the journal, proposing the text below) without any success.
Two of the referred studies are of the Spanish MULTITEST project (2015-2017), while the study of Venema et al. (2012) is about the tests of the COST ES0601 project (known as “HOME”, 2007-2011). In the era of HOME project the Climatol method had a markedly different content and the modest results of that are no longer relevant. By contrast, MULTITEST tested the modernized Climatol method.
The MULTITEST project has been the largest effort so far to quantify the accuracy of homogenization methods. Thousands of networks of synthetically generated monthly temperature and precipitation series were used in the project, which allowed to obtain statistically significant conclusions. For the large size of the test datasets, only automatic methods with default parameterization were tested. In the first part of the project, the automatic testing developed by Guijarro (2011) was applied, which provided fast results (Guijarro et al., 2023), but a disadvantage was that the test datasets were not saved and the homogenization accuracy for network mean characteristics were not checked. In the late phase of the project, 12 large-size temperature test datasets were saved, and for them the homogenization accuracy was checked both for individual time series and network mean series (Domonkos et al., 2021).
In the method comparison tests of MULTITEST the Climatol method was found to be the second best method in the homogenization of individual time series, while its results were more modest in the reconstruction of network mean trends. The method comparison test results are important characteristics of a homogenization method, but they are not the only important ones. Climatol is a free software which can be applied to a wide range of homogenization tasks. The method appropriately treats data gaps, and time series of highly varying observation periods can be homogenized together with relatively good results. Climatol can be used in fully automatic mode with a default parameterization, but various kinds of user interventions are allowed.
The homogenization of climatic time series is not a fully resolved scientific problem yet. Improvements in this area need the development, testing and use of various methods due to the complexity and high variety of practically occurring homogenization tasks. It is undoubtable that the Climatol method is a very valuable homogenization tool. However, overstatements and ambiguous statements should be avoided in scientific publications, since they may generate confusions.
References:
Domonkos, P., Guijarro, J.A., Venema, V., Brunet, M. and Sigró, J., 2021: Efficiency of time series homogenization: method comparison with 12 monthly temperature test datasets. J. Climate, 34, 2877-2891. https://doi.org/10.1175/JCLI-D-20-0611.1.
Guijarro, J.A., 2011: Influence of network density on homogenization performance. In Seventh Seminar for Homogenization and Qual-ity Control in Climatological Databases; Lakatos, M., Szentimrey T., Vincze, E. Eds.; WMO WCDMP-78; Geneva, Switzerland, 2011; pp. 11–18.
Guijarro, J.A., López, J.A., Aguilar, E., Domonkos, P., Venema, V.K.C., Sigró, J. and Brunet, M., 2023: Homogenization of monthly series of temperature and precipitation: Benchmarking results of the MULTITEST project. Int. J. Climatol. 43, 3994-4012. https://doi.org/10.1002/joc.8069.
Sigró, J., Cisneros, M., Pérez Luque, A.J., Pérez-Martínez, C. and Vegas‐Vilarrubia, T., 2024: Trends in temperature and precipitation at high and low elevations in the main mountain ranges of the Iberian Peninsula (1894–2020): The Sierra Nevada and the Pyrenees. Int. J. Climatol. 44, 2897-2920. https://doi.org/10.1002/joc.8487.
Venema, V., Mestre, O., Aguilar, E., et al., 2012: Benchmarking monthly homogenization algorithms. Climate of the Past, 8, 89-115. https://doi.org/10.5194/cp-8-89-2012.