Joint reconstruction and multi-modality/multi-spectral imaging (or joint geophysical inversion) is of growing importance in a wide range of contemporary issues including cost-effective environmental and groundwater investigations, natural hazard monitoring, carbon dioxide sequestration and efficient prediction and extraction of fossil and renewable fuels. It is also emerging rapidly in biomedical and materials science imaging. It combines data acquired using different methods (or modalities) to provide more realistic images of the subject under investigation than achievable using an individual modality as now well-known in environmental and energy investigations. Combining observations of multiple physical phenomena on an object of investigation has potential for accurate predictions and hence risk reduction in decision making with data. In the environmental and energy industries, the challenge in this integrated imaging of the subsurface is how to combine large-volumes of correlated data from interrelated physical phenomena or disparate data from unrelated physical phenomena and taking into account the different support volumes of the data (due to the different spatial scales or foot-prints of measurement modalities). In this paper, I describe some important considerations for adequate sampling of subsurface targets and data homogenization (or pre-conditioning), which data sets and physical constraints are most important for the joint image reconstruction process to be successful, uncertainty analysis, and the recent advances in structure-coupled inverse modeling of spatio-temporal multiphysics observations in petroleum and environmental investigations.