Towards optimal experimental design in hydrogeophysical studies

Hydrogeophysics attempts to combine multiple geophysical data sources to better constrain hydrological models of the subsurface. It is generally recognized that such data fusion attempts will be computationally demanding. Sophisticated (hydrogeophysical) inversion methods are one potential way to reduce the computational burden. Methods to reduce the number of measurements by avoiding redundant information (optimal experimental design) could also contribute to such a reduction. Such an optimisation of the experimental design requires that the quality of all possible designs can be first quantified and then maximized. In this contribution, we first review several methods from inverse problem conditioning to quantify the quality of the design. Then we apply these methods to a simple numerical example where borehole ground penetrating radar and electrical resistivity tomography are combined to measure soil water content variation with depth. Finally, we will discuss how time constraints (i.e. when monitoring transient processes) and financial constraints can be included in this framework. We hope that this contribution will encourage hydrogeophysicist to think about their own measurement design and how these can be improved for their specific case studies.