Probabilistic potentiometric surface mapping

The paper provides a methodology to estimate the spatial variation of potentiometric surfaces, allowing for possible uncertainties such as instrument errors, human errors, and statistical errors. The spatial variation is expressed in terms of the mean estimation, the coefficient of variation of the estimation, and the variance of the data scatter. The methodology is illustrated by a case study. The paper provides the criteria to decide the best estimation model when a number of theoretically consistent models are available to estimate spatial variation. A procedure to detect anomalous data is discussed in the paper through an example. The paper also provides a procedure to estimate the confidence of the mean spatial variation estimation, and describes steps one should take to improve this confidence, if it is needed. A technique is also given to estimate the random-error component of the measurements.