Indirect measurements are being used more commonly to constrain water flow models. These methods present many advantages for monitoring transient hydrologic processes, including the abilities to measure subsurface properties noninvasively, nondestructively, and rapidly. However, these indirect methods require more careful application than direct methods because they often have nonunique dependencies on properties of interest. Specifically, care must be taken in the both the interpretation of indirect measurements and in the design of monitoring networks that include these measurements. We present an analysis of the use of coupled water flow and electrical flow models to design optimal electrical resistance tomography (ERT) networks to monitor infiltration through a layered medium. We consider both surface and cross borehole ERT surveys. Results demonstrate that the design of surveys based on traditional interpretations of ERT survey sensitivity do not result in optimal survey designs. As a result, ERT surveys that are designed without the use of a coupled modeling approach introduce error and uncertainty in hydrologic parameter estimations. We propose that this coupled modeling approach can be made efficient enough to allow for real-time monitoring network optimization during infiltration.