Evaluation and optimisation of groundwater observation networks using the Kriging methodology Nicolaos Theodossiou * , Pericles Latinopoulos Division of Hydraulics and Environmental Engineering, Faculty of Civil Engineering, School of Technology, Aristotle University of Thessaloniki, 541 24 Thessaloniki, Greece Received 11 November 2004; received in revised form 8 February 2005; accepted 11 May 2005 Available online 26 July 2005 Abstract Groundwater simulation models have nowadays a decisive role in the development and application of rational water policies. Since the accuracy of the simulation depends strongly on the available data, the task of optimising the observation networks is of great importance. In this paper an application is presented aiming at the optimisation of groundwater level observation networks and the improvement of the quality rather than the quantity of the obtained data. This technique is based on the application of the Kriging methodology and the evaluation of its results in conjunction with the statistical analysis of the available groundwater level data. This procedure that involves different analysis methods of the available data, such as estimation of the interpolation error, data crossvalidation and time variation, is applied to a case study in order to demonstrate the potential of improvement of the quality of the observation network. Ó 2005 Elsevier Ltd. All rights reserved. Keywords: Groundwater resources management; Groundwater modelling; Observation networks; Kriging methodology 1. Introduction In order to simulate with adequate accuracy and be able to apply prediction and management scenarios to such a complex water resources system as a groundwater aquifer one must ensure the sufficiency of the collected input data. Groundwater level measurements have a decisive role in the formulation of the simulation model. Thus special attention is required not only during the actual measurements but also in the design of the groundwater level observation network. The extent of the network, its density or the frequency of the measurements must not be considered as constants, not even for a single aquifer. All these characteristics of the observation network depend on a number of factors such as the time and space variability of the hydraulic head, the hydrogeological parameter’s distribution, the velocity field over the aquifer etc. It should be noted that the extent and density of the network as well as the frequency of measurements affect directly time and money investment, an aspect that cannot be underestimated. In this paper an application of Kriging is presented aiming at the evaluation, organisation and optimisation of groundwater observation networks. The procedure proposes the evaluation of an existing observation network at a specific site through the geostatistical analysis of the groundwater level measurements. One of the basic tools used for the analysis of the water level measurements is the Kriging technique, whose basic principles are presented in the following paragraphs. * Corresponding author. Tel.: C302310995660; fax: C302310 995711. E-mail address: niktheod@civil.auth.gr (N. Theodossiou). 1364-8152/$ - see front matter Ó 2005 Elsevier Ltd. All rights reserved. doi:10.1016/j.envsoft.2005.05.001 www.elsevier.com/locate/envsoft Environmental Modelling & Software 21 (2006) 991e1000