Many empirical equations have been developed to predict the properties of saturation extracts, primarily electrical conductivity, from the properties of soil water extracts of 1 to 5 ratio. Soil water 1:5 extracts are more rapidly and reproducibly prepared than saturation extracts. However, the electrical conductivity of saturation extracts (ECse) is the benchmark to assess soil salinity. Analysis of the information provided by the electrical conductivity of soil 1:5 extracts (EC1:5) is a prerequisite to guide equation development and to use EC1:5 for soil salinity assessment. A total of 135 soil samples were taken from 39 sites at up to four different depths per site, down to a maximum depth of 95 cm in an irrigated agricultural area in SE Spain. Soil 1:5 extracts obtained from each sample were analysed for twelve chemical properties: concentration of sodium, ammonium, potassium, magnesium, calcium, chloride, nitrite, nitrate, sulphate, alkalinity and electrical conductivity. A principal component analysis (PCA) was applied to the correlation matrix of the log-transformed data set. The equilibrium status of 1:5 extracts with regard to gypsum was assessed with the chemical speciation programme SALSOLCHEMIS. The soil gypsum content was determined by using an adaptation of the classical method of total dissolution of soil gypsum, adequate to determine low gypsum contents, i.e. lower than 2%. Three principal components, accounting for 82% of the variance in the correlation matrix, were retained after eigenvector extraction. The first component accounted for 53% of the variance and was interpreted as representing the extract salinity due to gypsum dissolution. The second component accounted for 16% of the variance and was interpreted as representing the component of the extract salinity due to salts more soluble than gypsum, such as sodium and chloride salts. The information about soil salinity is split in two components in the 1:5 extracts, which are not easily resolved, whereas only one salinity component was obtained in the saturation extracts. The EC1:5 is a reliable property to estimate ECse, with a 95% confidence interval of ±1.2 dS m−1, only when it is lower than 1 dS m−1. When EC1:5 is higher than 2.4 dS m−1 it could also be reliably used to estimate soil salinity only if the 1:5 extract is gypsum-saturated, which is highly probable when the soil gypsum content is higher than 1.5%, or when the saturation extract is not gypsumsaturated, situation which happens when soil gypsum content is lower than 0.2%. When the EC1:5 is between 1 and 2.4 dS m−1, it provides poor information about the ECse, which could be between 4 and 14 dS m−1 unless the soil gypsum content is negligible, i.e. lower than 0.2%. The usefulness of the EC1:5 to estimate soil salinity can be extended when it is used in conjunction with an estimate of soil gypsum content. The development of more than one equation to estimate ECse from EC1:5 depending on soil gypsum content is recommended.

We thank the Conselleria de Medi Ambient from the Generalitat Valenciana for funding this research through project GV 0461/2006. We would also like to thank the two anonymous reviewers, and editor-in-chief for their fruitful comments and indications, which improved the article.

11 páginas, 7 figuras, 7 tablas.

Peer reviewed