Taking into Account Data Accuracy for Interpretation of Slug Tests in Confined or Unconfined Aquifers

Abstract

Different methods may be used to interpret the data of slug tests performed in aquifers, which are the water column height, Z, and time, t. The data accuracy usually is not taken into account. However, all measured Z data contain a random error and may contain a systematic error. This paper is believed to be the first one to explain how to assess the random error and display it in plots and, then, how to extract the systematic error, using three diagnostic graphs (two semi-log plots and a derivative plot that unifies all theories, and yields user-independent results). The plot of logZ versus t with “error” bars has a distinctive look: all Z data have the same “error,” but the smallest logZ data are the most inaccurate. As a result, the error bar is small at early times (large Z), but it increases to become very large at late times (small Z), which may modify the interpretation of data. Finally, the paper quantifies the errors that are made when using the Hvorslev equation for curved plots without error bars. Most often, the curvature results from a systematic error on the assumed piezometric level. When this error is not acknowledged, the user is at liberty to interpret the data and extract a hydraulic conductivity, K, which fits some beliefs. This yields a large error on K, which is quantified using equations and graphs.