Optimization of ERT Surveys for Monitoring Transient Hydrological Events Using Perturbation Sensitivity and Genetic Algorithms

available for completion of a survey is based on the physical process to be monitored. That is, all measurements A simple yet powerful algorithm is presented for the optimal allocacomprising a survey must be made rapidly compared with tion of electrical resistivity tomography (ERT) electrodes to maximize measurement quality. The algorithm makes use of a definition of the the rate at which the electrical conductivity changes in sensitivity of an ERT array to a series of subsurface perturbations. the subsurface. The time window divided by the time An objective function that maximizes the average sensitivity of a required per measurement defines the maximum number survey comprised of a large number of arrays is defined. A simple of arrays in a survey, A. Furthermore, the rate of change genetic algorithm is used to find the optimal ERT survey if there is of the electrical conductivity, EC, may change with time, a limited time allowed for survey. We further show that this approach leading to tightening or relaxation of the time frame, and allows for user definition of the sensitivity distribution within the a change in the number of arrays comprising a survey. targeted area. Results show clear improvement in the sensitivity distriElectrical resistivity tomography (note that for the purbution. The total sensitivity of the optimized survey compared with poses of this paper only surface electrodes are considtypically used surveys composed of one array type. This improved ered) has long been seen as a promising noninvasive, sensitivity will allow for more accurate monitoring of static and transient vadose zone processes. Furthermore, the algorithm presented nondestructive method (Edlefsen and Anderson, 1941). may be fast enough to allow for real-time optimization during timeRecent advances in ERT instrumentation and inversion lapse surveys. methods have increased the use of ERT for hydrologic investigations (Barker and Moore, 1998). These improvements include the use of multicore cable, addressed elecM subsurface hydrologic processes, and trodes, improved control and recording, and increased particularly those that occur within the vadose measurement accuracy, allowing for the use of very small zone, is difficult and expensive. Many monitoring methcurrents (i.e., tens to hundreds of milliamps). ods involve drilling for sampling or for access, which can Despite the increased use of electrical geophysical disturb the process under investigation and increase the methods, including ERT, applications are largely limited cost of the monitoring. Alternatives typically include burto monitoring static or very slowly changing conditions. ied instrumentation (e.g., time domain reflectometry, Electrical resistivity tomography has been applied to minthermocouples, or tensiometers). However, these point eral exploration (e.g., Griffiths and Barker, 1993), geomeasurements provide limited spatial resolution because logic mapping (e.g., Griffiths and Barker, 1993; Storz et they require a separate probe for each measurement al., 2000), groundwater table location (e.g., Yadav et al., point. In contrast, nondestructive, noninvasive geophysi1997), and groundwater contamination mapping (e.g., cal methods may offer high spatial and temporal resoluBuselli and Lu, 2001). Barker and Moore (1998) showed tion monitoring of shallow subsurface hydrological prothat ERT could be used to monitor transient processes cesses. in the shallow subsurface. However, their examples are Monitoring of hydrological problems differs in many limited to processes that occur slowly, on the order of ways from surveying for geological purposes. The most hours. An alternative to time-lapse monitoring involves important factor is the need to consider the measurement coupling the inversion of geophysical data with the physitime in hydrologic monitoring. Other differences include cal description of the monitored process (i.e., Richards’ the spatial scales of investigation (primarily that shallower equation in the case of infiltration), using the approach targets are of greater interest in most hydrologic investiof Seppanen et al. (2001). gations) and the required spatial resolution of the image As discussed above, the successful application of ERT (higher resolution is needed for hydrologic characterfor hydrological monitoring needs to address the requireization). ment for rapid measurement, accuracy, and high resoluAlthough the time required for a single ERT measuretion. Additional problems include the separation of the ment depends on many parameters, including the properresistivity to its components (primarily water content, ties of the subsurface at the time of measurement, it can chemical properties of the water, and electrochemical be treated as a constant (e.g., ≈15 s). The time window properties of the porous or fractured media). This can be particularly difficult under unsaturated conditions, or A. Furman and T.P.A. Ferré, Hydrology and Water Resources, Univ. when high solute concentrations are involved. of Arizona, Tucson, AZ 85721; A.W. Warrick, Soil, Water, and EnviIncreased accuracy can be achieved through noise reronmental Sciences, Univ. of Arizona, Tucson, AZ 85721; A. Furman, duction (Ritz et al., 1999), improved inversion algorithms, currently Inst. Soil, Water, and Environ. Studies, ARO, Volcani Ctr., assimilation of data obtained through other means (Yeh P.O. Box 6, Bet Dagan 50250, Israel. Received 12 May 2003. Original Research Paper. *Corresponding author ([email protected]). et al., 2002), and optimized selection of arrays used to Published in Vadose Zone Journal 3:1230–1239 (2004). © Soil Science Society of America Abbreviations: EC, electrical conductivity; ERT, electrical resistivity tomography; GA, genetic algorithms. 677 S. Segoe Rd., Madison, WI 53711 USA