<p>Extensive efforts over decades have focused on quantifying chemical transport in subsurface geological formations, from microfluidic laboratory cells to aquifer field scales. Outcomes of these efforts have remained largely unsatisfactory, however, largely because domain heterogeneity (in terms of, e.g., porosity, hydraulic conductivity, geochemical properties) is present over multiple length scales, and “unresolved”, practically unmeasurable heterogeneities and preferential pathways arise at virtually every scale. While <em>spatial</em> averaging approaches are effective when considering overall fluid flow – wherein pressure propagation is essentially instantaneous and the system is “well mixed” – purely <em>spatial</em> averaging approaches are far less effective for chemical transport, essentially because well-mixed conditions do not prevail. We assert here that an explicit accounting of <em>temporal </em>information, under uncertainty, is an additional, but fundamental, component in an effective modeling formulation. As an outcome, we further assert that “upscaling” of chemical transport equations – in the sense of attempting to develop and apply chemical transport equations at large (length) scales, based on measurements and model parameter values obtained at significantly smaller length scales – is very much a holy grail. Rather, we maintain that it is necessary to formulate, calibrate and apply models using measurements at similar scales of interest, in both space and time.</p>