Open AccessUncertainty of Soil Water Characteristic Curve Measurements Using an Automated Evaporation Technique
Author(s) -
Breitmeyer Ronald J.,
Fissel Lauren
Publication year - 2017
Publication title -
vadose zone journal
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 1.036
H-Index - 81
ISSN - 1539-1663
DOI - 10.2136/vzj2017.07.0136
Subject(s) - replicate , soil water , environmental science , soil science , evaporation , measurement uncertainty , uncertainty analysis , mathematics , materials science , statistics , physics , meteorology
Core Ideas Replicate‐specimen measurement of SWCCs was performed with an automated technique. Inter‐specimen uncertainty for SWCC parameters was quantified. A comparison of automated evaporation technique vs the standard hanging column is presented. Soil water characteristic curve (SWCC) measurements are used to estimate hydrologic parameters for a variety of engineering and agricultural applications. The SWCC parameterization is often based on a single measurement or with little or no consideration and/or quantification of the uncertainty in collection of SWCC data. Uncertainty in the SWCC data stems from random variability in pore geometry associated with irregular granular media. Additionally, a substantial opportunity exists in SWCC measurements for systematic or experimental errors that can increase uncertainty. This study utilized an evaporation method for measuring 10 SWCCs for each of four different test soils. One low‐plasticity and three high‐plasticity silty sand soils were sampled from the Sagehen Creek Experimental Watershed in Sierra County, California. The uncertainty in the SWCC measurements was quantified by the distribution and variability of van Genuchten parameters evaluated via nonlinear regressions on each set of SWCC data. Despite a careful, replicated specimen preparation, variability in the measured SWCC data was indicated by variability of estimates of van Genuchten's fitting parameters α and n and residual volumetric water content θ r , suggesting that packing specimens to identical dry densities does not produce identical SWCC measurements for a single soil type. For the soils tested in this study, replicate specimen testing resulted in uncertainty in α between ×/1.2 and ×/1.5 the geometric mean estimate for α, between ±2 and ±6% of the mean estimate for n , and between ±2 and ±6% for θ r . Additional testing is required before applying these uncertainties generally to soils of different textural classifications.