Estimating saturated soil hydraulic conductivity by the near steady-state phase of a Beerkan infiltration test

Abstract

Single-ring infiltration experiments carried out in the field, such as the Beerkan runs, allow easy and inexpensive characterization of soil hydraulic properties, and specifically saturated soil hydraulic conductivity, Ks, by maintaining the functional connection of the sampled soil volume with the surrounding soil. However, a single infiltration experiment is not enough to determine Ks. The simplest way to obtain the necessary additional data is based on the assessment of the soil texture and structure characteristics. In this investigation, a simplified method, named SSBI (Steady version of the Simplified method based on a Beerkan infiltration run), was developed to estimate Ks by only using the near steady-state phase of a Beerkan infiltration run and an estimate of the α* parameter. Testing the method against analytically generated infiltration data revealed low prediction errors of Ks (≤4.1%) for a wide range of soils and initial soil water conditions. A test with an extensive set of field data showed that the developed method yielded means and medians of Ks that were similar (i.e., differing by no more than a factor of two) to those obtained with a more data demanding procedure. Similar coefficients of variation, i.e. 126–141% or 62–67%, depending on the sampled soil from regions of Sicily and Burundi, respectively, were also obtained. Another field comparison of the SSBI method with the more classical single-ring pressure infiltrometer method yielded statistically equivalent Ks values (100–143mmh−1) but significantly shorter equilibration times in the former case (16min) than the latter one (31min). The developed method appears attractive due to the simplicity of both the experiment and the data analysis procedure. In addition, it allows quicker runs and makes use of smaller water volumes as compared with other, more popular, infiltrometer methods based on constant-head single ring experiments. Therefore, additional testing of the developed method is advisable.