Melting, Compaction and Reactive Flow: Controls on Melt Fraction and Composition Change in Crustal Mush Reservoirs

Changes in melt fraction and local bulk composition in high-crystallinity, crustal mush reservoirs are essential to produce the large volumes of low-crystallinity, silicic magma that are emplaced to form plutons, or erupted to surface. Heating (and cooling) is well understood and widely invoked in driving melt fraction change, but does not cause chemical differentiation because there is no separation of melt and crystals. Fractional crystallisation at high melt fraction is widely assumed to explain differentiation, but is inconsistent with the evidence that large-scale, long-term magma storage and evolution occurs in high-crystallinity mush reservoirs. Compaction has been suggested to explain melt fraction change and differentiation at low melt fraction, but compaction (and decompaction) causes simple unmixing (and mixing) of melt and solid crystals: to produce very refractory bulk composition by compaction, melt fraction must be driven down to very low values. Yet microstructural evidence demonstrating widespread compaction in crustal mush reservoirs at low melt fraction is lacking. Here we show that melt fraction change can be expressed in terms of heating/cooling and compaction, plus an additional term that we call ‘reactive flow’. Similarly, composition change can be expressed in terms of compaction and reactive flow. Reactive flow changes the local bulk composition, which causes ‘chemical’ melting (dissolution) and freezing (precipitation), distinct from ‘thermal’ melting/freezing caused by changes in enthalpy. We use numerical modelling to show that the contributions of compaction and reactive flow in a crustal magma reservoir are similar in magnitude. However, reactive flow opposes melt fraction and composition changes caused by compaction when compaction occurs in a temperature gradient that increases upwards at, for example, the base of a sill intrusion, or decompaction occurs in a temperature gradient that decreases upwards at, for example, the top of a sill intrusion. If compaction causes melt fraction decrease and creates a more refractory bulk composition, then reactive flow causes melt fraction increase and a more evolved bulk composition, and vice versa. Reactive flow means that very small melt fraction is not required to produce very refractory composition in a crustal magma reservoir, consistent with the relatively scarce microstructural evidence for widespread compaction. The apparent lack of compaction in crustal magma reservoirs, as com