A database of saturated hydraulic conductivity of fine-grained soils: probability density functions

Shuyin Feng, Paul J. Vardanega*

*Corresponding author for this work

Research output: Contribution to journalArticle (Academic Journal)peer-review

7 Citations (Scopus)
43 Downloads (Pure)


Saturated hydraulic conductivity is a key soil mechanics parameter which has widespread use in many geotechnical applications. In order to set up stochastic analyses, geotechnical modellers require databases to calibrate the parameter ranges and distributions employed. This letter uses a recently compiled database of saturated hydraulic conductivity measurements called FG/KSAT-1358 and reports on the fitting of various probability density functions to the data of void ratio, liquid limit, water content ratio and the negative natural logarithm of ksat. It is shown that the best fit distribution is the lognormal for void ratio, while the loglogistic distribution is most favoured for liquid limit and water content ratio, and the best fit distribution for -ln[ksat(m/s)] is the logistic function. The data of -ln[ksat(m/s)] is then subdivided according to liquid limit level, silt or clay classification, type of hydraulic conductivity test used and sample preparation/condition. When some subdivisions of the database are analysed, the best fit distribution is more variable with GEV and logistic being the most favoured for most of the studied subsets.
Original languageEnglish
Pages (from-to)255-261
Number of pages7
JournalGeorisk: Assessment and Management of Risk for Engineered Systems and Geohazards
Issue number4
Early online date15 Aug 2019
Publication statusPublished - 1 Nov 2019


  • Saturated Hydraulic Conductivity
  • Probability Density Functions
  • Database
  • Fine-grained soils
  • Akaike information criterion
  • corrected Akaike information criterion


Dive into the research topics of 'A database of saturated hydraulic conductivity of fine-grained soils: probability density functions'. Together they form a unique fingerprint.

Cite this