Loes van Dam, University of Rhode Island


Complex tectonic plate motions and thermal heterogeneity in the Earth’s mantle give rise to three-dimensional, time-varying patterns in material transport towards the surface near mid-ocean ridges. The commonly accepted view of mid-ocean ridges is that they are passive features, where plate divergence results in vertical upwelling from deep inside the mantle. In this model, formation of the crust is uniform and symmetric about the ridge axis. Recent plate motion models have revealed that mid-ocean ridges are not stationary in time (Whittaker et al., 2015). Their migration may influence the pathlines of mantle rock motion towards the ridge axis. Both geophysical and geochemical observations further reveal that the mantle does not have a uniform temperature distribution. Spatial variations in mantle temperature can be either local (i.e., mantle plumes) or regional (i.e., basin-scale). Together, the driving force of the plate motions and the heterogeneous heat distribution in the mantle can impact melting, volcanism, and crustal formation at mid-ocean ridges.

The first chapter investigates the influence of mid-ocean ridge migration on isothermal upper mantle flow. We build a unique new laboratory apparatus and fluid reservoir that are ideally suited to studying this problem. Using this setup, experimental simulations of ridge migration are precisely controlled and repeatable. Through image processing and particle tracking, we find that the source region for new crust is constrained to much shallower depths than previously thought. Calculations of mantle melting indicate that melting may be suppressed at high migration rates and that melt production is not symmetry about the ridge axis. However, comparisons of natural ridge migration rates with published geochemical and geophysical data yield few meaningful correlations.

The second chapter assesses how buoyant thermal upwellings interact with the upper mantle flow field generated by a migrating ridge. We use a circular heater placed at the base of the fluid reservoir to simulate mantle plumes in a controlled way. Stationary ridges bisect the plume as it surfaces and plume material spreads laterally beneath the plate. Beneath a migrating ridge, the flow field can deform mantle plumes and thereby reduce their buoyancy. This results in enhanced loss of heat through diffusion. The deformation further inhibits the full ascent of a mantle plume, such that portions of the plume may rise towards the surface but not melt. This suppresses the total volcanic output that we expect from plumes. Hot portionts of unmelted plume material contribute to the global distribution of thermal heterogeneity in the upper mantle. The variability in volcanic output from deformation of rising mantle plumes has a natural counterpart. Large igneous provinces (LIPs) are thought to be the solidified lava outpourings resulting from melting of plume heads, which vary in volume by several orders of magnitude.

While controlled plumes yield valuable insight into the behavior of the largest buoyant upwellings in the Earth’s mantle, it is important to understand how thermal boundary layers naturally generate buoyant upwellings (active flow) that interact with (passive) plate-driven flow. We use a large rectangular heater at a scaled depth of 670 km to form a thermal boundary layer at the base of the mantle transition zone and allow the system to evolve naturally. We find two novel behaviors related to the formation and destruction of sheet-like upwellings. First, sheet-like upwellings collapse into plume-like conduits in the wake of a migrating ridge. Buoyant flow within the conduit stalls at depths of ~150 km, such that melt is generated but may be retained within the solid upper mantle. Depending on the melting model used, maximum melt fractions range from 0.1 – 0.23 with a resulting melt column of 30-75 km height. Ambient upper mantle material is entrained by the flow of such upwellings and may contribute in part to melting near the surface, especially if thermal diffusion has played a role during upwelling. This is dependent on the strength of the thermal boundary layer and the plate rates. Second, once vertical pathways for thermal boundary layer material have been established, a thermal upwelling will not form again directly beneath the ridge axis. Over the course of 36 Myr, we observe that material within the thermal boundary layer flows laterally towards the nearest upwelling, regardless of passive vertical flow beneath the ridge. This has implications for our understanding of the origins of thermal heterogeneity and the stability of thermal features over long timescales.