Comment on ‘‘Determination of soil–water characteristic curve variables’’ by Qian Zhai and Harianto Rahardjo [Comput. Geotech. 42 (2012) 37–43] Zoheir Bellia ⇑ , Mohamed Rabhi, Mokhtar Dadouch, Moulay Smaîne Ghembaza Laboratoire de Génie Civil et Environnement, Université Djillali Liabès de Sidi Bel Abbès, Algeria article info Article history: Available online 7 April 2012 Keywords: Soil–water characteristic curve Air-entry suction Residual suction Parameters abstract The determination of soil–water characteristic curve variables, proposed by [1] provides more flexibil- ity compared to conventional methods such as [4]. However, if the method ‘‘A’’ described by the authors is used, the air-entry and residual suctions would be difficult to compute, because the param- eters m and n depend directly on the residual suction and vice versa, which leads to more unknowns than equations. It should be noted that the parameters of the original formula of [2] composed of the correction factor, are calculated in semi-log plot by using the neperian logarithm and not with the decimal logarithm. Crown Copyright Ó 2012 Published by Elsevier Ltd. All rights reserved. 1. Introduction The authors should be congratulated for their valuable contri- bution. The present discussion is mainly based on the method of determining the parameters m and n proposed by [1], as well as the calculation procedure of variables related to retention curve namely the air-entry and residual suctions. The fitting formulas proposed by the authors are those of the model of [2], with methods A and B defined by Eqs. (1) and (2) respectively. h ¼ 1  lnð1 þ w cr Þ lnð1 þ 10 6 cr Þ  ! 1 lnðe þð w a Þ n  ! m ð1Þ h ¼ 1 lnðe þð w a Þ n  ! m ð2Þ In [1] the parameters m and n of the method B, are calculated at the inflection point in semi logarithmic plot by using the decimal logarithm. m ¼ 3:67 ln h s h i  ð3Þ n ¼ 1:31 mþ1 h s m 1:616s 1 ð4Þ s 1 ¼j dh d logðwÞ j wa ð5Þ According to the authors, using the method B, the calculation of the parameters m and n of the original model [2], are calculated on an arithmetic plot, this remark is valid see Eqs. (6)–(8). But for the method A, it is totally different since the calculation was established according to a semi-log plot with neperian logarithm, see Eqs. (9)–(12). m ¼ 3:67 lnð h s h i Þ ð6Þ n ¼ 1:31 mþ1 h s m 3:72s 1 a ð7Þ s 1 ¼j dh dðwÞ j wa ð8Þ m ¼ 3:67 ln h s CðaÞ h i   ð9Þ n ¼ 1:31 mþ1 mCðaÞ 3:72s  ð10Þ s  ¼ s 1 h s  a 1:31 m ða þ c r Þ lnð1 þ 10 6 cr Þ ð11Þ s 1 ¼j dh d lnðwÞ j wa ð12Þ Therefore, the choice of logarithmic scales, neperian in [2] or deci- mal in [1,5], can significantly affect the calculation of calibration parameters in the case of a confusion between these last. In this section we discuss the procedure for calculation of air- entry and residual suctions. If the fitting is done by the method B, the parameters n and m are easily computable because they are independent of the residual suction. Now, using the method A, we will study the following cases: In the first, the parameters n and m would be calculated using a conventional method, i.e. 0266-352X/$ - see front matter Crown Copyright Ó 2012 Published by Elsevier Ltd. All rights reserved. http://dx.doi.org/10.1016/j.compgeo.2012.02.012 ⇑ Corresponding author. E-mail address: bellia_zoheir@yahoo.fr (Z. Bellia). Computers and Geotechnics 44 (2012) 20–21 Contents lists available at SciVerse ScienceDirect Computers and Geotechnics journal homepage: www.elsevier.com/locate/compgeo