Sources and References

Primary textbook — Winter, J.D. (2010). Principles of Igneous and Metamorphic Petrology (2nd ed.). Prentice Hall. The standard upper-division petrology textbook, covering thermodynamics, phase diagrams, trace element and isotope geochemistry, and all major igneous-tectonic associations with exceptional clarity and depth.

Supplementary sources — USGS Volcano Hazards Program (volcanoes.usgs.gov); Smithsonian Institution Global Volcanism Program (volcano.si.edu); Frost, B.R. & Frost, C.D. (2019). Essentials of Igneous and Metamorphic Petrology (2nd ed.); Klein, C. & Philpotts, A. (2017). Earth Materials (2nd ed.).

Chapter 1: Igneous Rock Classification

Introduction to Igneous Systems

Igneous rocks form wherever molten rock — magma — cools and solidifies, whether at the surface as lava or at depth within the crust as an intrusion. They constitute the most volumetrically dominant rock type on Earth, making up the entire oceanic crust, the bulk of the continental crust in deep crustal levels, and virtually all of the planet’s mantle lithosphere. Understanding how to classify and interpret igneous rocks is thus the foundation upon which all subsequent discussion of magma generation, differentiation, and tectonic setting must rest. Classification is not merely a taxonomic exercise; it encodes information about temperature, pressure, volatile content, and source rock composition in ways that can be decoded through careful analysis.

The diversity of igneous rocks is enormous. From the black, glassy obsidian of a silicic volcanic flow to the pale-grey, coarse-grained granite of a mountain batholith; from the dark, olivine-rich dunite of the mantle to the frothy pumice of a Plinian eruption — these materials share the common origin of crystallizing from a melt but differ profoundly in mineralogy, texture, and chemistry. Modern classification schemes use both modal mineralogy (the actual proportions of minerals present) and whole-rock chemical composition as complementary frameworks, because texture and grain size often prevent direct modal counting in fine-grained or glassy volcanic rocks.

The TAS Diagram

The Total Alkalis vs. Silica (TAS) diagram is the primary geochemical classification scheme for volcanic rocks in which phenocryst mineralogy is obscured by a fine-grained or glassy groundmass. It plots the weight percent sum of Na₂O + K₂O (the “total alkalis”) on the vertical axis against the weight percent SiO₂ on the horizontal axis. The choice of these particular two variables is motivated by several considerations. Silica content is the single most important variable controlling the physical properties of magma — most notably viscosity, which in turn governs eruption style, eruption products, and hazard — and it broadly correlates with the degree of magmatic differentiation. The alkalis Na₂O and K₂O are highly incompatible elements that concentrate dramatically in residual melts during fractional crystallization and are also elevated in magmas derived from enriched or metasomatized mantle sources, making the alkali content an indicator of magma source and evolutionary pathway.

The subalkaline fields, occupying the lower portion of the diagram, encompass the most voluminous volcanic rocks on Earth. Moving from low to high SiO₂ along the subalkaline array, the rock names progress through basalt (SiO₂ 45–52 wt%, Na₂O + K₂O < 5 wt%), basaltic andesite (52–57 wt%), andesite (57–63 wt%), dacite (63–68 wt%), and rhyolite (>68 wt%). These terms are broadly equivalent to the compositional continuum from mafic to felsic composition, and they describe the dominant products of mid-ocean ridges, subduction zones, and silicic continental calderas respectively. The boundaries between fields are sharp lines in a diagram but represent a continuous and gradational spectrum of compositions in nature.

The alkaline fields occupy positions above the alkalic dividing line and encompass rocks with elevated alkali contents relative to their silica. On the low-silica, high-alkali side of the diagram, one finds tephrite/basanite (which contains modal feldspathoid minerals or olivine in excess of silica saturation), phonolite (the coarse-grained equivalent of which is called nepheline syenite), and foidite (rocks dominated by feldspathoid minerals such as leucite or nepheline). Moving toward higher silica within the alkalic fields, trachyte and trachydacite occupy intermediate positions, reflecting magmas with elevated total alkalis but sufficient silica to be quartz-undersaturated or just quartz-saturated. The trachyandesite (sometimes called mugearite or benmoreite in oceanic island settings) and trachybasalt fields connect the basaltic compositions to the trachytic end-member along evolutionary trends governed by fractional crystallization with alkali feldspar accumulation.

Why does the diagram use total alkalis rather than separating Na₂O from K₂O? The answer is largely pragmatic: the ratio Na₂O/K₂O varies considerably within compositionally similar rock suites as a function of source composition and partial melting degree, but the sum is more systematically correlated with the degree of silica saturation and the position within the alkaline versus subalkaline series. There are, however, separate classification schemes that distinguish the potassic and sodic series within the alkaline rocks — shoshonites (potassic trachybasalts) and absarokites for potassium-enriched arc-related lavas, for example, or sodic hawaiites and mugearites for the oceanic island alkali basalt differentiation series.

The QAPF Modal Diagram

Whereas the TAS diagram is designed for fine-grained volcanic rocks where minerals cannot be directly counted, the QAPF diagram applies to coarse-grained plutonic rocks (and, in principle, porphyritic volcanic rocks with abundant phenocrysts). QAPF stands for Quartz (Q), Alkali Feldspar (A), Plagioclase (P), and Feldspathoid (F), and the diagram is actually a double triangle sharing the AP base. The upper triangle has Q at its apex and A and P at its base corners; the lower triangle has F at its nadir and A and P again at its base. The two triangles are mutually exclusive because quartz and feldspathoids cannot coexist stably in the same rock — they react to form feldspar — so every plutonic rock plots in one triangle or the other.

In the upper (Q-A-P) triangle, moving from high-Q to the AP base, the rock names progress from quartzolite (>90% Q), through quartz-rich granitoid (Q = 60–90%), to granite (Q = 20–60%, A/(A+P) = 35–90%), granodiorite (Q = 20–60%, A/(A+P) = 10–35%), and tonalite (Q = 20–60%, A/(A+P) < 10%). At lower Q (0–20%), the corresponding rocks are quartz syenite (high A), quartz monzonite (moderate A and P), quartz monzodiorite/quartz monzogabbro (low A), and quartz diorite/quartz gabbro (essentially no A). Along the AP base where Q approaches zero, the rocks are syenite, monzonite, monzodiorite/monzogabbro, and diorite/gabbro respectively. In the lower (F-A-P) triangle, the rocks are undersaturated with respect to silica and contain feldspathoids: foid syenite (high A), foid monzosyenite, foid monzodiorite/foid monzogabbro, foid diorite/foid gabbro, and pure foidolite (>90% F) at the F apex.

The volcanic equivalents of the major plutonic rock types form paired names that reflect the same bulk composition expressed at different grain sizes and cooling histories. Granite and rhyolite are plutonic-volcanic pairs, as are diorite and andesite, gabbro and basalt, syenite and trachyte, and nepheline syenite and phonolite. This pairing is fundamental: the chemical classification by TAS and the modal classification by QAPF converge on the same underlying compositional continuum, viewed through different windows.

Ultramafic Rocks and the Colour Index

For rocks with less than 10% felsic minerals (feldspars plus feldspathoids plus quartz), the QAPF diagram becomes unhelpful because the felsic minerals are simply too scarce to define meaningful proportions. Such rocks are classified separately by their proportions of mafic minerals: olivine, pyroxene (orthopyroxene and clinopyroxene), hornblende, and phlogopite. The resulting classification scheme uses the colour index as a primary criterion.

Lherzolite deserves special emphasis because it represents the composition of the mantle from which basaltic magmas are generated. The four-phase assemblage of olivine + orthopyroxene + clinopyroxene + an aluminous phase (spinel or garnet depending on pressure) defines what geochemists call primitive or fertile mantle — a source rock that retains all the incompatible elements in roughly chondritic relative proportions. When this rock partially melts, the resulting liquid has basaltic composition, and the solid residue, progressively depleted in pyroxene and enriched in olivine and orthopyroxene, trends toward harzburgite and ultimately dunite.

The colour index also connects to the broader concept of the mafic–felsic continuum. Mafic (from magnesium and ferrum = iron) rocks are rich in Mg, Fe, Ca, and have relatively low SiO₂ (45–52 wt%), dark colours, and high densities (∼3.0 g/cm³). Felsic rocks (from feldspar and silica) are enriched in Si, Al, Na, and K, are light-coloured, and have lower densities (∼2.6–2.7 g/cm³). Intermediate rocks bridge these extremes. This continuum in chemical composition arises from the interplay of source composition, degree of partial melting, and subsequent fractional crystallization, and it is the backbone around which all subsequent discussions of magmatic differentiation are organized.

Chapter 2: Physical Properties of Magma

Viscosity and the Role of Silica Polymerization

Viscosity is among the most consequential physical properties of a magma because it governs whether that magma erupts effusively as a lava flow, explosively as a fragmented pyroclastic spray, or something in between. Viscosity is the resistance of a fluid to flow — the internal friction that one layer of fluid exerts on adjacent layers as they move at different velocities. For a Newtonian fluid, viscosity \( \eta \) (in Pascal-seconds) is the proportionality constant between the applied shear stress and the resulting strain rate. Magmas are generally treated as Newtonian at low crystal contents, with non-Newtonian (particularly Bingham plastic) behavior emerging as crystal fraction increases above approximately 40–50 vol%.

The structural explanation for why silica-rich magmas are so much more viscous than silica-poor ones lies in the polymerization of the silicate melt. In a silicate melt, silicon atoms occupy tetrahedral sites surrounded by four oxygen atoms: [SiO₄]⁴⁻ tetrahedra. At low silica contents (basaltic compositions), these tetrahedra are largely isolated or share only one or two corners, yielding a relatively depolymerized melt with abundant non-bridging oxygens (NBOs). As SiO₂ content increases, more tetrahedra link by sharing corners (bridging oxygens), forming chains, rings, sheets, and ultimately three-dimensional networks analogous to the structure of volcanic glass or quartz. This progressive polymerization dramatically increases the structural rigidity and therefore the viscosity of the melt.

The practical consequence of this Arrhenius relationship is enormous. A basaltic lava at 1200°C has a viscosity of approximately 10–100 Pa·s — comparable to thick syrup or engine oil — while a rhyolitic lava at 800°C may have a viscosity of 10⁸–10¹² Pa·s, comparable to cold pitch or glass. The difference of ten orders of magnitude in viscosity separates the effusive sheet flows of Hawaiian shield volcanoes from the slow, coulée-like extrusions of rhyolitic domes. Water (H₂O) dissolved in the melt acts as a powerful depolymerizer because H₂O molecules react with bridging oxygens to form hydroxyl groups, breaking Si–O–Si bonds and reducing the network connectivity. Even a modest addition of 1 wt% dissolved H₂O can reduce the viscosity of a rhyolitic melt by two to three orders of magnitude, bringing it closer to the range of andesitic compositions.

Density, Buoyancy, and Melt Ascent

The density of magma determines whether it is buoyant relative to its surroundings and therefore whether it will ascend through the lithosphere or pond at depth. Mafic magmas (basalts) have densities of approximately 2.6–2.9 g/cm³ at magmatic temperatures, while felsic magmas (rhyolites) are lighter at approximately 2.2–2.4 g/cm³. These values are substantially less than the density of the surrounding crystalline rock at the same pressure (olivine ∼3.3 g/cm³; plagioclase ∼2.7 g/cm³; lower crustal gabbros ∼3.0 g/cm³), so most magmas are positively buoyant relative to their host rocks and will ascend when connected conduits permit flow.

The major exceptions are important. Basaltic magma is actually denser than felsic lower crust (2.8–3.0 g/cm³ versus 2.7 g/cm³ for granodioritic lower crust), which means it is not universally buoyant everywhere in the crust. Basaltic magma ascending through felsic crust will, at some critical depth, pass through a density trap where the magma density equals or exceeds that of the surrounding rock. At this depth, upward buoyancy is zero or negative, and the magma may pond, forming a MASH zone (melting, assimilation, storage, and homogenization) if sustained heat input forces partial melting of the surrounding crust. Such ponding zones at the base of the crust are believed to play a crucial role in generating the silicic magmas characteristic of continental arcs.

Volatile Solubility and Henry’s Law

Volcanic gases — predominantly H₂O, CO₂, SO₂, HCl, and HF — are dissolved in magmas at depth and exsolve as pressure decreases during ascent, driving volcanic eruptions. The solubility of volatiles in silicate melts is a pressure-dependent phenomenon governed at low volatile contents by Henry’s Law:

The derivation of the square-root pressure dependence for H₂O solubility proceeds from equilibrium thermodynamics. The dissolution reaction is:

The equilibrium constant for this reaction is:

Assuming ideal mixing in the melt and ideal gas behavior, and noting that \( a_{\text{H}_2\text{O, gas}} \propto P_{\text{H}_2\text{O}} \), the concentration of dissolved hydroxyl groups (proportional to total dissolved water at moderate water contents) scales as \( [\text{OH}] \propto P_{\text{H}_2\text{O}}^{0.5} \). At higher water contents, molecular H₂O also becomes significant, and the total dissolved water shows a steeper, super-square-root pressure dependence. These solubility relations mean that at the relatively modest pressure of 200 MPa (approximately 8 km depth), a basaltic melt can dissolve about 5 wt% H₂O, and a rhyolitic melt about 8 wt% — far more than the actual water contents of most natural magmas (typically 1–6 wt% for arc magmas, 0.1–0.5 wt% for mid-ocean ridge basalts).

Vesiculation and Fragmentation

As a volatile-bearing magma ascends, decreasing pressure reduces the solubility of dissolved gases and causes them to exsolve as a separate vapor phase — a process called vesiculation. The resulting bubbles grow by diffusion of gas molecules from the melt toward bubble nuclei and by decompression expansion of the gas already in the bubble. The dynamics of vesiculation — specifically whether bubbles can grow and rise gently through the melt or become trapped in a highly viscous melt that fragments catastrophically — is the single most important factor separating effusive from explosive eruptions.

The vesicularity (volume fraction of bubbles) at which fragmentation occurs has been empirically and theoretically constrained at approximately 70–80 vol% for homogeneous foams, corresponding to the point where bubble walls become so thin that they rupture. However, fragmentation can occur at lower vesicularity in heterogeneous systems where large bubbles are surrounded by thin, rapidly extensional melt films. The physical contrast between the gentle degassing of Hawaiian basalt flows (low viscosity, bubbles rise freely) and the paroxysmal eruptions of Pinatubo or Taupo (high viscosity, bubbles trapped until catastrophic overpressure) is thus fundamentally a viscosity story rooted in SiO₂ content and dissolved water.

Chapter 3: Mantle Mineralogy and the Peridotite Source

Olivine, Pyroxene, and the Mantle Mineral Assemblage

The mantle that underlies the oceanic and continental crust and extends to the core-mantle boundary at 2900 km depth is composed predominantly of ultramafic rock — peridotite in the upper mantle, with progressively higher-pressure mineral assemblages at greater depths. Understanding the mineralogy of the mantle is essential to understanding where and how magmas are generated, because partial melting of specific mineral assemblages yields characteristic melt compositions.

Olivine (Mg₂SiO₄–Fe₂SiO₄ solid solution, forsterite–fayalite series) is the dominant mineral of the upper mantle, comprising approximately 60 vol% of fertile lherzolite. The Mg# (= Mg/(Mg+Fe) atomic ratio) of mantle olivine is approximately 0.89–0.91, reflecting the highly magnesian nature of the mantle. Olivine is an orthosilicate, meaning the [SiO₄] tetrahedra are isolated (not polymerized), making it one of the most refractory (high melting temperature) common silicate minerals. Olivine melts at approximately 1890°C at atmospheric pressure (for the pure forsterite end-member) and is the first mineral to crystallize from a cooling basaltic melt according to Bowen’s reaction series.

Pyroxenes in the mantle occur in two varieties. Orthopyroxene (enstatite–ferrosilite series, (Mg,Fe)₂Si₂O₆) forms the second most abundant mantle phase, comprising roughly 20–25 vol% of lherzolite, and provides the orthopyroxene component that makes the rock a harzburgite (olivine + opx) or lherzolite (olivine + opx + cpx). Clinopyroxene (diopside–hedenbergite–jadeite series, CaMgSi₂O₆ dominant in the mantle) comprises approximately 10–15 vol% of lherzolite. Clinopyroxene is the primary carrier of incompatible elements (REE, Sr, Ba in particular) in the mantle, and its progressive removal during melting is responsible for the progressive depletion of incompatible elements in mantle residues.

Aluminous Phases: Spinel vs. Garnet Stability

The fourth phase in fertile mantle peridotite is an aluminous mineral, and the particular aluminous phase present changes systematically with pressure, defining two distinct mantle mineralogical facies of profound importance for partial melting processes and the chemistry of erupted magmas.

Garnet lherzolite facies prevails at pressures above approximately 2.5 GPa (depths > 80 km), where the aluminous phase is pyrope-rich garnet ((Mg,Fe)₃Al₂Si₃O₁₂). Garnet is a nesosilicate with a dense cubic structure that becomes stable at high pressure because packing aluminum in the octahedral sites of garnet is energetically favored over the looser packing in spinel or plagioclase. The presence of garnet versus spinel has critical geochemical consequences: garnet strongly retains the heavy rare earth elements (HREE) and Y in its structure, while spinel does not. A melt generated at pressures above the spinel-garnet transition leaves residual garnet, which sequesters HREE and creates a melt with a distinctively high La/Yb ratio (LREE-enriched, HREE-depleted pattern). In contrast, melts generated in the spinel stability field have flatter REE patterns because HREE are not retained in any residual phase.

The spinel-to-garnet transition is not a sharp univariant boundary but a pressure-temperature region spanning roughly 2.0–2.8 GPa that depends on the Cr/(Cr+Al) ratio of the spinel, the Fe/Mg ratio of the bulk rock, and temperature. The reaction in a simplified system is approximately:

This reaction has a positive Clapeyron slope (the garnet field expands to lower pressures with increasing temperature), meaning that hot, asthenospheric upwellings will reach the spinel stability field somewhat deeper than a cold, depleted lithospheric root would.

Depleted vs. Primitive Mantle

Not all mantle rocks have the same composition. The concept of a primitive or fertile mantle composition — one that has never experienced a melting event and therefore retains all major and trace elements in approximately chondritic proportions — is an idealized end-member defined by mass balance calculations. The actual mantle sampled by xenoliths (pieces of mantle rock carried to the surface by rapidly ascending volcanics) spans a wide range of compositions reflecting their different melting histories.

The sub-oceanic mantle that produces MORB is typically depleted, having lost a small fraction of melt at some point in its history, perhaps during an earlier episode of oceanic crust formation. In contrast, the enriched mantle beneath ocean islands and continental intraplate settings may have been metasomatized — infiltrated by small-degree partial melts or fluids that added incompatible elements back to the peridotite — or may represent ancient recycled oceanic or continental crustal material that was returned to the mantle by subduction. These compositional heterogeneities in the mantle are detectable through radiogenic isotope ratios (Sr, Nd, Pb, Hf) and are the basis for the mantle isotopic component framework (HIMU, EM1, EM2, DMM, PREMA) discussed in the tectonic settings chapter.

Chapter 4: Partial Melting Theory

Why Rocks Melt Partially

When a rock in the mantle or lower crust is heated, decompressed, or fluxed with volatiles, it does not melt all at once. Instead, it undergoes partial melting: only a fraction of the rock transitions to liquid, while the remainder stays solid. This happens because the rock is a multi-component system with multiple mineral phases, each with a different melting temperature. When temperature rises to the point where the first melt appears (the solidus), the composition of that first melt is the eutectic-like composition that coexists with all solid phases. As more heat is added or pressure decreases further, the melt fraction grows, and the residual solid progressively loses the most easily-melted, incompatible-element-rich phases (pyroxene, then garnet or spinel) relative to the refractory olivine.

The degree of partial melting, denoted F (as a fraction between 0 and 1), fundamentally controls the composition of the magma generated and the volume of melt available for eruption. Small-degree melts (F ≈ 0.01–0.05) at deep levels produce strongly enriched, alkalic melts with high concentrations of incompatible elements — such as the nephelinites and basanites of continental rifts. Large-degree melts (F ≈ 0.20–0.30) produce tholeiitic basalts with relatively flat incompatible-element patterns, as generated beneath mid-ocean ridges. The key to understanding this is the partition coefficient and the trace element behavior captured by batch and fractional melting equations.

Batch Melting Equation: Derivation

The simplest melting model is batch melting (also called equilibrium melting), in which the melt remains in complete chemical equilibrium with the residual solid throughout the melting process and is only extracted at the end as a single batch.

Consider a trace element with initial concentration \( C_0 \) in the source rock. The system consists of a total mass normalized to 1, of which a fraction \( F \) is melt with concentration \( C_L \) and fraction \( (1-F) \) is solid residual with concentration \( C_S \). Mass balance requires:

The bulk partition coefficient \( D \) of the trace element between the solid assemblage and the melt is defined as:

This \( D \) is a weighted sum of the mineral-melt partition coefficients: \( D = \sum_i X_i K_d^i \), where \( X_i \) is the weight fraction of mineral \( i \) in the solid residue and \( K_d^i \) is the mineral-melt partition coefficient of that element for phase \( i \).

Substituting \( C_S = D \cdot C_L \) into the mass balance:

Fractional Melting Equation: Derivation

Fractional melting is the opposite endmember model: each infinitesimal increment of melt is immediately removed from contact with the residue as soon as it forms. No equilibrium is maintained between accumulated melt and solid; the melt that finally erupts represents the aggregate of all increments.

For an infinitesimal melt increment \( dF \), mass balance gives:

Expanding the left side and substituting \( C_L = C_S / D \):

Separating variables:

Integrating with the boundary condition \( C_S = C_0 \) at \( F = 0 \):

Rayleigh Fractionation and Fractional Crystallization

By exact analogy with fractional melting, Rayleigh fractionation describes the evolution of a trace element during fractional crystallization, where each infinitesimal crystal increment is immediately isolated from the remaining melt. Starting from a melt with initial concentration \( C_0 \) and extracting crystals with \( C_S = D \cdot C_L \):

Influence of Pressure, Temperature, and Water on Melting

Three independent mechanisms can drive partial melting in the mantle, corresponding to the three ways one can cross the solidus: by heating, by decompression, or by volatile addition. Each mechanism dominates in a different tectonic setting.

Melting by decompression operates at mid-ocean ridges and hotspot plumes. As mantle upwells beneath a spreading center or rises in a plume, it follows an approximately adiabatic path: temperature decreases by only about 0.3°C/km (the adiabatic gradient), far less than the steep pressure-dependence of the peridotite solidus (approximately 3°C/km in terms of the equivalent temperature change along a pressure gradient of 0.035 GPa/km). As a result, rising mantle approaches the solidus from below and eventually crosses it, initiating melting. The intersection depth is the onset of melting, and the column of mantle above this point that remains within the melting zone determines the total melt fraction generated.

Melting by volatile addition is the primary mechanism in subduction zone settings. Hydrous fluids and silicate melts released from the subducting slab migrate into the overlying mantle wedge, drastically lowering the solidus temperature. The effect of water on the peridotite solidus can be understood thermodynamically through the depression of melting point by dilution of the melt. In formal terms, adding water to a melt increases its entropy, stabilizing it relative to the solid phases and lowering the equilibrium solidus temperature. The solidus of water-saturated peridotite at 3 GPa is approximately 1000°C, compared to approximately 1500°C for the dry solidus at the same pressure — a depression of ∼500°C. At typical subduction zone conditions (pressures of 1–3 GPa, temperatures of 700–1000°C in the mantle wedge), the dry mantle is below its solidus, but even a small flux of aqueous fluid (perhaps 0.05–0.2 wt% H₂O added to the mantle) can depress the local solidus enough to initiate melting.

Melting by heat addition occurs when hot intrusions, mantle plumes, or radiogenic heat raises the local geothermal gradient above the solidus. This is the least common mechanism in the modern Earth but was likely significant in the Archean when mantle potential temperatures were higher by 100–200°C.

Chapter 5: Phase Diagrams and Crystallization

Binary Eutectic Systems: Forsterite–Fayalite

The olivine solid solution system, forsterite (Fo, Mg₂SiO₄) – fayalite (Fa, Fe₂SiO₄), is the simplest geologically relevant binary system. It exhibits complete solid solution at all compositions, meaning that at any bulk composition, the system consists of a single solid phase (olivine of variable Mg/Fe ratio) plus melt over a range of temperatures. There is no eutectic because both end-members are fully miscible in the solid state.

For the forsterite–fayalite system specifically, the phase diagram has a characteristic lens shape with a liquidus curve (above which only liquid is stable) and a solidus curve (below which only solid olivine is stable). Between these curves, both liquid and solid coexist. A cooling melt of intermediate composition (say, 50 mol% Fo) will reach the liquidus at about 1600°C and begin crystallizing forsterite-rich olivine (perhaps Fo₈₀). As cooling continues, the melt becomes progressively more fayalite-rich and the olivine also becomes more fayalite-rich, following the liquidus and solidus curves respectively. The solid and liquid compositions track continuous paths dictated by the equilibrium phase boundary. This behavior, in which both liquid and solid continuously change composition during crystallization, is called continuous reaction and contrasts with the discontinuous reactions seen in peritectic systems.

Diopside–Anorthite Eutectic and the Crystallization Sequence

The diopside (Di, CaMgSi₂O₆) – anorthite (An, CaAl₂Si₂O₈) binary system is a geological archetype for eutectic behavior: two end-members that are completely immiscible in the solid state but mutually soluble in the liquid state. The liquidus curves descend from the melting point of pure diopside (1392°C at 1 atm) and pure anorthite (1553°C) to meet at the eutectic point at approximately 1274°C and 42 mol% An. At the eutectic, three phases coexist at equilibrium: melt, pure diopside crystals, and pure anorthite crystals. This is an invariant point in the sense of Gibbs’ Phase Rule (\( F = C - P + 2 \); with C = 2, P = 3, F = 1 at constant pressure, so the eutectic is a fixed temperature at any given pressure).

The crystallization sequence depends on the bulk composition. For a melt of composition more Di-rich than the eutectic (say, 70 mol% Di, 30 mol% An): upon cooling to the liquidus at approximately 1360°C, pure diopside begins crystallizing. The melt composition evolves along the liquidus toward the eutectic as diopside is removed. When the melt reaches eutectic composition at 1274°C, anorthite also begins crystallizing simultaneously with diopside. Crystallization proceeds at constant temperature until all liquid is consumed, yielding an interlocking texture of diopside and anorthite in approximately eutectic proportions. For a melt of exactly eutectic composition, crystallization begins directly at the eutectic.

Peritectic Systems: Forsterite–Silica

The forsterite–quartz (SiO₂) binary system illustrates the peritectic relationship, which is fundamental to understanding Bowen’s discontinuous reaction series. Forsterite and quartz are not stably co-crystallizable: they react to form enstatite (MgSiO₃, also called pyroxene in this context):

The phase diagram has a peritectic point at approximately 1557°C and intermediate composition where this reaction occurs. A melt of intermediate composition cooling toward the peritectic will first crystallize forsterite; when the peritectic temperature is reached, the forsterite reacts with the silica in the remaining melt to form enstatite. Only after all forsterite has been consumed (if the bulk composition is more silica-rich than forsterite) does the crystallization proceed further. If cooling is rapid and the forsterite crystals are armored (coated with enstatite), the reaction may be incomplete, producing a disequilibrium assemblage of forsterite + enstatite + residual glass — a common occurrence in rapidly cooled basalts and komatiites.

Ternary Haplogranite System: Ab–Or–Qz

The haplogranite system consisting of albite (Ab, NaAlSi₃O₈) – orthoclase (Or, KAlSi₃O₈) – quartz (Qz, SiO₂) at approximately 2 kbar water-saturated conditions is the paramount phase diagram for understanding granite petrogenesis. The diagram contains a cotectic line (also called the cotectic curve or two-feldspar + quartz field boundary) along which melt coexists with two solid phases, and a thermal minimum (or minimum melt point) analogous to a ternary eutectic where the melt coexists with all three solid phases: albite, orthoclase, and quartz.

The effect of pressure on the minimum melt composition is instructive. As water pressure increases, the minimum migrates toward more Ab-rich compositions, slightly expanding the stability field of alkali feldspar. This means that granites crystallized at greater crustal depths (higher pressure) tend to contain more plagioclase-dominated feldspar assemblages relative to K-feldspar. Conversely, very-low-pressure granites (subvolcanic porphyries) may crystallize from melts closer to the minimum and will contain an approximately equal mix of alkali feldspar and quartz.

Bowen’s Reaction Series

N.L. Bowen’s landmark 1928 work The Evolution of the Igneous Rocks established the concept of an idealized crystallization sequence for a cooling basaltic melt that passes through a predictable series of mineral assemblages. Bowen divided this into two parallel sequences that merge at the end into the felsic minerals.

The discontinuous series on one side progresses: olivine → pyroxene → amphibole → biotite. Each transition involves a reaction between the earlier crystallized mineral and the evolving melt, where the mineral reacts with melt to produce the next phase. The olivine-to-pyroxene transition is peritectic in nature (as described above for forsterite + silica → enstatite), as is the pyroxene-to-amphibole transition. These reactions require time and close crystal-melt contact; if crystallization proceeds rapidly or crystals are physically removed from the melt, the reactions may be incomplete, preserving early phases in a more evolved melt.

The continuous series on the other side describes the continuous compositional evolution of plagioclase feldspar from Ca-rich anorthite (An₉₀) at high temperatures to Na-rich albite (An₁₀) at low temperatures, without discontinuities. This reflects the complete solid solution in the plagioclase system. The continuous series also describes the continuous evolution of olivine from Mg-rich to Fe-rich compositions as described earlier for the forsterite-fayalite system.

The two series converge in the late-crystallizing felsic minerals: potassium feldspar → muscovite → quartz (approximately), with the final residual liquid being of granitic composition. The series is thus a qualitative guide to the order of mineral precipitation from a basaltic melt undergoing equilibrium fractional crystallization at low pressure (crustal conditions). It should be emphasized that Bowen’s series is a simplification: real magmas involve dozens of components, pressures that change during ascent, volatile fluxes, and assimilation of wall rocks, all of which modify the crystallization sequence.

Chapter 6: Geochemistry — Trace Elements and Isotopes

Compatible and Incompatible Elements

The behavior of trace elements during melting and crystallization is governed by the partition coefficient between minerals and melt. Elements that are strongly partitioned into solid phases (high \( K_d \)) are termed compatible and are retained in residual minerals during melting, entering the melt only slowly. Elements that strongly prefer the melt (\( K_d \ll 1 \)) are termed incompatible and are rapidly concentrated in the liquid phase even at small melt fractions.

The High Field Strength Elements (HFSE: Nb, Ta, Zr, Hf, Ti) are highly charged, small ionic radius elements that are generally incompatible in mantle silicates but compatible in oxide and titanate phases. The Large Ion Lithophile Elements (LILE: K, Rb, Cs, Ba, Sr, Pb, Eu²⁺) are low-charge, large-radius elements that are mobile in aqueous fluids and are concentrated in the continental crust. The contrast between HFSE and LILE behavior in different tectonic settings — specifically the “arc signature” of LILE enrichment and HFSE depletion — is one of the most powerful geochemical discriminants available to petrologists.

REE Patterns and Normalization

The rare earth elements (REE: La, Ce, Pr, Nd, Sm, Eu, Gd, Tb, Dy, Ho, Er, Tm, Yb, Lu) are a coherent group of trivalent lanthanide elements that substitute for Ca in pyroxene, amphibole, and accessory minerals, with partition coefficients that vary systematically with ionic radius from the large LREE (La, Ce) to the small HREE (Yb, Lu). Because their partition coefficients are well-measured and span a wide range, REE patterns are among the most sensitive indicators of igneous processes available.

Normalization to primitive (or chondritic) mantle removes the saw-tooth pattern in absolute concentrations caused by the Oddo-Harkins rule (even atomic number elements are more abundant than adjacent odd ones) and expresses all REE concentrations relative to an undifferentiated, pre-melting mantle standard. After normalization, the slope of the REE pattern carries geochemical information:

A negative slope (LREE-enriched relative to HREE, high La/Yb ratio) indicates either (1) melting at pressures where garnet is a residual phase (garnet sequesters HREE), (2) a small degree of partial melting (incompatible LREE enter the melt preferentially), or (3) fractional crystallization that removed HREE-compatible minerals. Ocean Island Basalts (OIB) characteristically have such patterns.

A flat or weakly negative slope indicates melting at shallower levels where spinel (not garnet) is the residual phase, or large-degree melting that exhausted the garnet. Mid-Ocean Ridge Basalts (MORB) typically have flat or slightly depleted LREE patterns, consistent with melting in the spinel stability field from a depleted mantle source.

Spider Diagrams and Primitive Mantle Normalization

Spider diagrams (multi-element variation diagrams) extend the normalization concept to a broader suite of trace elements arranged along the horizontal axis in order of decreasing incompatibility (from the most incompatible on the left, such as Cs and Rb, to the least incompatible on the right, such as Yb and Lu). Elements are normalized to primitive mantle concentrations, so a flat horizontal line at a normalized value of 1 would indicate a rock with exactly primitive mantle composition.

When plotted on spider diagrams, different tectonic rock suites show characteristic patterns that serve as fingerprints. MORB shows a pattern depleted in the most incompatible LILE and LREE relative to the less incompatible elements — a pattern tilting upward to the right — consistent with derivation from a depleted mantle source. OIB patterns are typically elevated above MORB and show enrichment particularly in the most incompatible elements. Arc basalts display the distinctive arc signature: enrichment in LILE (Ba, Rb, Sr, K, Pb) accompanied by a sharp Nb-Ta trough (and often a Ti trough), reflecting the addition of slab-derived fluid (which efficiently mobilizes LILE but not HFSE) to the mantle wedge and the retention of Ti and Nb in residual or crystallized titanate phases.

Radiogenic Isotope Systems

While trace element patterns record the processes a magma has undergone (melting, crystallization, contamination), radiogenic isotope ratios record the time-integrated history of the source from which the magma was derived. The most widely used systems in igneous petrology are:

The Rb–Sr system: ⁸⁷Rb decays to ⁸⁷Sr with a half-life of 48.8 Ga. Rocks or reservoirs that have had high Rb/Sr ratios over geological time (such as the continental crust, which is enriched in Rb relative to Sr) will evolve to high ⁸⁷Sr/⁸⁶Sr ratios. Depleted mantle, which lost Rb preferentially during melting events, has low ⁸⁷Sr/⁸⁶Sr (∼0.7022). Continental crust has ⁸⁷Sr/⁸⁶Sr > 0.705, and can exceed 0.74 in old cratonic crust.

The Sm–Nd system: ¹⁴⁷Sm decays to ¹⁴³Nd with a half-life of 106 Ga. Because Sm and Nd are adjacent REE with similar partition coefficients (but Nd is slightly more incompatible than Sm), melting and crystallization events slowly fractionate Sm/Nd. The ɛNd notation expresses ¹⁴³Nd/¹⁴⁴Nd as a deviation in parts per ten thousand from the chondritic value at the time of measurement: ɛNd > 0 indicates a mantle source that has been depleted of LREE (like DMM), while ɛNd < 0 indicates a source enriched in incompatible elements or contaminated by old continental crust.

Chapter 7: Tectonic Settings of Magmatism

Mid-Ocean Ridges and MORB Petrogenesis

Mid-ocean ridges are the most extensive volcanic systems on Earth, producing approximately 21 km³ of basaltic magma per year and constructing the entire oceanic crust at a steady-state rate that balances subduction. The process is decompression melting of upwelling asthenospheric mantle, as described in Chapter 4. As the spreading plates diverge, mantle beneath the ridge axis flows upward to fill the void, and this ascending mantle crosses the dry peridotite solidus at depths of approximately 60–90 km beneath fast-spreading ridges and perhaps 30–60 km beneath slow-spreading ridges, initiating partial melting.

The pooling model for MORB petrogenesis argues that melts generated over a wide column of mantle (perhaps 50–100 km wide and up to 80 km tall) migrate laterally and vertically through the overlying lithosphere and pool beneath the ridge axis in a melt lens — a thin, typically less than 1 km thick magma chamber located within the lower crust at approximately 1.5–3 km depth. From this lens, the magma crystallizes, is periodically replenished by fresh primary melt from below, and erupts periodically to the seafloor. The shallow magma chamber is the locus of extensive fractional crystallization, predominantly of plagioclase + clinopyroxene ± olivine at low pressures (< 0.5 GPa), which drives the MORB composition toward the familiar tholeiitic trend: initially Mg-rich and then progressively Fe-enriched (the tholeiitic iron-enrichment trend) in contrast to the calc-alkaline trend of arc magmas.

The contrast between slow-spreading ridges (< 40 mm/yr full rate, such as the Mid-Atlantic Ridge) and fast-spreading ridges (> 80 mm/yr, such as the East Pacific Rise) is profound. Fast-spreading ridges maintain a persistent, well-developed axial magma lens, a smoother topographic profile, and erupt compositionally homogeneous lavas over large areas because the lens efficiently pools and homogenizes inputs from a wide mantle column. Slow-spreading ridges have much lower magma supply rates, cooler thermal structures, and often lack a persistent melt lens; instead they have deep rift valleys flanked by fault scarps, irregular off-axis topography, and more compositionally variable lavas reflecting less efficient pooling and less extensive fractional crystallization. The extreme case is ultra-slow spreading ridges (< 20 mm/yr, such as the Gakkel Ridge in the Arctic Ocean), where magma supply may be so limited that large sections of the ridge expose mantle peridotite at the seafloor rather than basaltic crust.

Subduction Zones: Slab Dehydration and Arc Magma Genesis

Subduction zones are the sites of the most violent volcanism on Earth, the birthplaces of the continental crust, and the repositories for recycled oceanic materials that fundamentally heterogenize the mantle over geological time. The fundamental process driving arc magmatism — volatile fluxing of the mantle wedge — has already been introduced in the context of volatile-induced melting, but the details of slab dehydration and melt generation are complex and deserve full treatment.

As a subducting slab descends into the mantle, it carries with it the altered oceanic crust (basalt + gabbro + peridotite that have been hydrated and carbonated by seawater interaction at the ocean floor), sediments, and the hydrated harzburgitic mantle of the slab itself (the slab mantle or subducted lithospheric mantle). The hydrated minerals in these rocks are stable at crustal pressures but break down progressively with increasing pressure and temperature during subduction, releasing aqueous fluids (and at higher temperatures, silicate melts) that rise buoyantly into the overlying mantle wedge.

The fluid released from slab dehydration rises into the base of the mantle wedge and lowers the local solidus, initiating partial melting of the wedge peridotite. Because the fluid is aqueous (not silicic), it efficiently mobilizes LILE (which are fluid-mobile) but does not transport HFSE (which are fluid-immobile in aqueous systems). The melt produced in the fluxed wedge thus inherits this fluid signature: it is enriched in Ba, Rb, Sr, K, Pb relative to primitive mantle, and shows negative anomalies in Nb, Ta, and Ti — the classic arc signature seen on spider diagrams. The physical-chemical basis for the Nb-Ta depletion is debated: the most likely explanation is retention of Nb and Ta in residual rutile (TiO₂) in the subducted oceanic crust or in titanate phases crystallizing during fractional crystallization in the crust of the arc.

Calc-alkaline differentiation is the characteristic magmatic evolution trend in arcs: unlike the iron-enrichment trend of tholeiitic MORB differentiation, arc magma suites typically show little iron enrichment and instead evolve toward higher SiO₂ with approximately constant FeO/MgO ratios. This behavior is attributed to the elevated oxygen fugacity of arc magmas (inherited from the oxidized slab fluid), which causes early crystallization of magnetite (Fe₃O₄). Magnetite crystallization removes Fe from the melt, preventing iron enrichment. The elevated water content of arc magmas also suppresses plagioclase crystallization (water lowers the plagioclase stability field), allowing amphibole to crystallize instead, which has a lower SiO₂ content and higher Al₂O₃ and is therefore more effective at driving the melt toward silicic compositions through fractional crystallization.

Adakites represent a special case in which the subducted oceanic crust itself partially melts rather than merely dehydrating. This occurs when the slab is young and hot (< ∼25 Ma at the trench) or when subduction rates are slow, keeping the slab geotherm in the stability field of slab melting (∼800–900°C at 3–4 GPa). Slab melts react with the overlying mantle peridotite as they ascend, producing lavas with high Sr/Y and La/Yb ratios (reflecting residual garnet in the slab, which retains Y and HREE), high Sr concentrations, and moderate SiO₂. These adakitic signatures are also invoked for some Archean tonalite-trondhjemite-granodiorite (TTG) suites, suggesting that Archean subduction involved hotter slabs than modern subduction.

Ocean Islands: Plumes and Isotopic Heterogeneity

Ocean Island Basalts (OIB) are produced by mantle plumes — anomalously hot, buoyant columns of mantle material that rise from the deep mantle (possibly the core-mantle boundary region) and impinge on the base of the lithosphere, producing hotspots such as Hawaii, Iceland, Galápagos, and the Azores. OIB are geochemically distinct from MORB in being enriched in incompatible elements, having higher La/Yb ratios, and displaying a wide range of isotopic compositions that require multiple, chemically distinct mantle source components.

The Hawaiian shield-building sequence illustrates systematic chemical evolution from shield-building to post-erosional volcanism. The main shield phase produces predominantly tholeiitic basalt (high degree of melting from the hot plume core, relatively depleted in incompatible elements relative to later stages). As the lithosphere moves off the hotspot and the plume melts a progressively thicker, cooler lithosphere, the melt fraction decreases and the magmas become alkalic — transitioning to alkali basalt, then hawaiite, mugearite, benmoreite, and ultimately trachyte in the rejuvenated stage (Honolulu Volcanics, for example). The decreasing degree of melting and/or melting from greater depth (within the garnet stability field) explains both the increasing alkalinity and the increasing La/Yb ratio through the volcanic sequence.

Continental Settings: Flood Basalts, Rifts, and Granites

Large Igneous Provinces (LIPs) are the most voluminous volcanic events in Earth’s history, emplacing millions of cubic kilometers of basaltic magma in geologically brief intervals (< 1–5 Ma). The Deccan Traps of peninsular India (65–66 Ma) released approximately 5 × 10⁵ km³ of lava and represent one of the best-studied examples. The Columbia River Basalt Group (15–17 Ma) is a smaller but exceptionally well-documented LIP that erupted approximately 2 × 10⁵ km³ in the Pacific Northwest of North America through a series of deep crustal feeder dikes.

The origin of flood basalts is debated. The plume head model proposes that a newly ascending large mantle plume initially has a bulbous head that melts extensively as it impinges on the lithosphere, producing the rapid pulse of magmatism. This model is supported by the spatial and temporal association of many LIPs with the predicted arrival of deep mantle plumes (e.g., the Deccan Traps with the Réunion hotspot, the North Atlantic LIP with the Iceland plume). Alternative models involving edge-driven convection, delamination of dense lithospheric roots, or rifting-induced decompression melting have been invoked for LIPs that lack clear hotspot connections.

The Deccan Traps and their relationship to the Cretaceous-Paleogene (K-Pg) boundary is a topic of active debate. The main phase of Deccan eruption straddles the K-Pg boundary at 66.0 Ma, overlapping closely with the Chicxulub bolide impact. Estimates of SO₂ and CO₂ released by Deccan volcanism suggest significant environmental effects — potential acid rain events and short-term warming pulses from CO₂ — but most researchers consider the Chicxulub impact to have been the primary driver of the end-Cretaceous mass extinction, with Deccan volcanism potentially contributing as a stressor.

Continental rifts generate distinctive magmatic suites because rifting reduces lithospheric pressure (decompression melting) and often involves small-degree partial melts of enriched lithospheric or asthenospheric mantle at relatively low temperatures. The East African Rift System (EARS) is the premier modern example, extending from the Afar Triple Junction in Ethiopia through Kenya to Mozambique. The EARS produces a remarkable diversity of magma types, from tholeiitic basalts at the most extended portions to alkalic rocks in the less-extended rift flanks, and famously includes carbonatites — igneous rocks composed dominantly of carbonate minerals (calcite, dolomite) rather than silicate minerals — at Oldoinyo Lengai in Tanzania. Carbonatites require extremely small-degree partial melting (F < 1%) of a deeply carbonated, enriched mantle source and represent the most CO₂-rich, Ca-Na-K-enriched, silica-undersaturated melts in the igneous rock spectrum.

Granite petrogenesis in continental settings is classified by the scheme developed by Chappell and White (1974) and subsequently expanded:

Chapter 8: Volcanic Eruption Styles and Products

Eruption Style and the Viscosity Control

The explosivity of a volcanic eruption is ultimately controlled by the interplay between the rate of magma ascent (decompression rate), the viscosity of the melt, and the concentration and speciation of dissolved volatiles. When these factors conspire to prevent bubbles from migrating upward and escaping quietly from the melt, pressure builds until catastrophic fragmentation releases the stored energy in an explosive eruption. When viscosity is low or ascent rates slow enough to allow continuous degassing, effusive eruption dominates.

Strombolian eruptions are characteristic of low-viscosity, gas-rich mafic magmas in which the primary eruption mechanism is the rise and burst of large gas slugs through the magma column. In a slug flow regime, gas accumulates into large bubbles (10s of centimeters to meters in diameter) that rise faster than the surrounding melt and burst at the surface, throwing incandescent clots of molten lava (bombs and lapilli) 100–300 m into the air. The resulting deposits are coarse, poorly sorted, and mantled around the vent in a scoria cone. Individual Strombolian explosions last seconds to minutes and recur at intervals of minutes to hours. Stromboli volcano in Italy, the archetype, has been erupting continuously in this mode for at least 2000 years. The key control is the conduit geometry: slug flow requires that the conduit diameter be small enough relative to bubble size that slugs form rather than dispersing into fine foam.

Vulcanian eruptions involve more viscous magmas (intermediate to silicic composition) in which the conduit is periodically sealed by a plug of degassed, crystallized lava or by a dome that grows over the vent. Gas pressure builds beneath the plug until the seal fails catastrophically, launching a discrete explosion that ejects the plug material as ballistics and fine ash, produces a convecting ash cloud, and may generate pyroclastic density currents if the eruption column collapses. Vulcanian explosions are typically discrete events lasting minutes, with quiet intervals in between while the plug reforms. The 1995–2010 eruptions of Soufrière Hills volcano, Montserrat, exemplify this style, with regular Vulcanian explosions punctuating ongoing dome growth.

Plinian Eruptions and Column Height

Sub-Plinian and Plinian eruptions represent the most powerful sustained explosive eruptions, characterized by stable, convecting eruption columns that can reach 20–45 km altitude (the stratosphere), depositing tephra over thousands of square kilometers. The eruption column is sustained by the thermal energy of hot pyroclasts continuously injected into the base of the column, which heat the entrained air, causing vigorous convection that drives the column upward.

The derivation can be traced through scaling analysis. The buoyancy flux is \( B = g \Delta \rho / \rho \cdot Q / \rho \approx g \alpha \Delta T Q / (\rho C_p) \) where \( \Delta T \) is the temperature excess of the eruption mixture. For a buoyant plume rising in an atmosphere with Brunt-Väisälä frequency \( N \), the maximum height scales as \( H \sim (B/N^3)^{1/4} \). Substituting \( B \propto Q \) gives \( H \propto Q^{1/4} \). This is one of the most important quantitative results in volcanology, allowing eruption rate to be estimated from column height observations even for ancient eruptions recorded in ash deposits.

Caldera-forming eruptions are the largest events, in which magma is withdrawn so rapidly from a large, shallow magma chamber that the overlying roof collapses into the evacuating chamber, producing a caldera — a roughly circular depression kilometers to tens of kilometers in diameter. The erupted products are sheets of ignimbrite (also called ash-flow tuff or welded tuff), deposited by the pyroclastic density current (PDC) — a ground-hugging mixture of hot gas and fragmented pyroclasts that can travel at 100–300 km/h and reach temperatures of 300–600°C. PDCs are among the most lethal volcanic phenomena, capable of overcoming topographic obstacles, traveling 100 km from source, and completely destroying everything in their path. The 74 ka eruption of Toba in Sumatra (∼2800 km³ DRE) and the 2 Ma Huckleberry Ridge eruption of Yellowstone (∼2500 km³) represent Holocene and Pleistocene caldera-forming events at this scale.

Pyroclastic Materials

The products of explosive volcanism are collectively called tephra — any airfall material deposited from an eruption column. Tephra is classified by grain size using a modification of the Wentworth scale:

Pumice is highly vesicular (60–90 vol% bubbles), light-colored (felsic composition) pyroclastic material whose density is low enough to float on water. Scoria is the mafic equivalent — less vesicular (20–60 vol% bubbles), dark, glassy, and denser than pumice. Both record the vesiculation state at fragmentation: pumice forms when a highly viscous silicic melt fragments while still highly vesicular, preserving the foam texture. Dense juvenile clasts are fragments of magma with little or no vesiculation, representing portions of the melt column that erupted from lower-vesicularity regions or where degassing was rapid enough to collapse bubbles before freezing. The full spectrum of juvenile clast density in a single eruption provides information about the vesiculation dynamics in the conduit.

Ignimbrite welding occurs when hot PDC deposits retain enough heat and volatile content for the glassy particles to deform and fuse together. The resulting texture is called eutaxitic: flattened pumice clasts (called fiamme, Italian for “flames”) compressed against a matrix of fused ash. Welded ignimbrites can have rock-like appearance and strength, be quarried as building stone (the ancient Romans used ignimbrite for construction), and preserve a record of emplacement temperature through the degree of welding and the nature of secondary crystallization (devitrification and vapor-phase mineralization).

Accretionary lapilli are concentric-layered, spheroidal ash aggregates 1–10 mm in diameter that form within volcanic clouds when moisture nucleates around fine ash particles and additional ash layers accrete, similar to the growth of a hailstone. Their presence in an ash fall deposit indicates that water was available (either ambient atmospheric moisture or phreatomagmatic steam), and they are often used as markers for water-magma interaction in the eruption.

Chapter 9: Volcanic Monitoring and Eruption Forecasting

Seismic Monitoring

Modern volcano monitoring integrates multiple independent geophysical and geochemical data streams, because no single parameter provides a reliable eruption forecast alone. The goal is to track the migration of magma and fluids through the volcanic edifice and plumbing system, which manifests as changes in ground deformation, seismicity, gas flux, and thermal emission.

Seismic monitoring is the cornerstone of volcano surveillance. Volcanic earthquakes are classified by their waveform characteristics, which reflect the physical source mechanism. Volcano-tectonic (VT) earthquakes have high-frequency (>5 Hz) impulsive waveforms and are caused by brittle fracture of rock — the same mechanism as ordinary tectonic earthquakes — but they occur in swarms that migrate toward the surface or toward the magma body, tracing the path of magma ascent or fluid migration. Long-period (LP) earthquakes have emergent, low-frequency (1–5 Hz) waveforms with strongly resonant codas. They are attributed to pressure oscillations within cracks, conduits, or reservoirs filled with fluid (either magmatic gas, hydrothermal fluid, or magma itself). The resonance frequency depends on the crack geometry and fluid properties: \( f = c_s / (2L) \) for a crack of length \( L \) filled with fluid with sound speed \( c_s \). LP swarms are often the most reliable precursor to impending eruptions, as they signal active movement of magmatic fluids.

Harmonic tremor is a continuous, nearly monochromatic ground vibration lasting minutes to hours or longer, with a dominant frequency and overtones (harmonics). It is associated with sustained fluid flow through cracks or conduits — either degassing of rising magma or transport of hydrothermal fluids — and at many volcanoes precedes and accompanies eruptions by hours to days. The amplitude and frequency of tremor can be tracked in real time on seismograms and provide a proxy for magma flux rate.

Ground Deformation

As magma intrudes into or accumulates within a volcanic edifice, it displaces the surrounding rock and causes measurable surface deformation. The pattern and magnitude of deformation encodes information about the depth, geometry, and volume change of the magmatic source.

GPS (Global Positioning System) and GNSS provide continuous three-dimensional displacement time series at individual stations, with typical precision of 1–5 mm horizontal and 5–10 mm vertical. Networks of campaign or continuous GNSS stations around volcanoes track the source inflation/deflation (Mogi source modeling treats the magma body as a pressurized sphere and can retrieve source depth and volume change from the surface displacement pattern), rift zone opening (dike intrusion episodes are recorded as divergence of stations on either side of the rift), and flank instability (slow lateral movement of unstable flanks, as observed on Kīlauea’s south flank at ∼6 cm/yr).

Gas Monitoring

Volcanic gases carry direct information about the degassing state of the underlying magma body and can change dramatically in the hours to days before an eruption.

DOAS (Differential Optical Absorption Spectroscopy) is the standard technique for measuring SO₂ flux from volcanic plumes. A UV spectrometer mounted on a vehicle or in a fixed scanning geometry measures the absorption spectrum of the plume against the sky background UV radiation. By fitting the measured spectrum to reference spectra of SO₂ (and other absorbers, hence “differential”), the column amount of SO₂ is determined. Multiplying by the plume speed (measured by anemometry, seismoacoustic methods, or multiple spectrometers) gives the mass flux in tonnes per day or kg/s. Because SO₂ is almost entirely magmatic in origin (crustal rocks contain negligible sulfur in comparison), SO₂ flux is a direct proxy for the amount of degassing magma, the magmatic flux rate, and the degree of open-system versus closed-system degassing.

Continuous SO₂ monitoring at active volcanoes has revealed important patterns. At Kīlauea, SO₂ flux typically runs 500–1500 t/d during normal lava lake activity at the summit, rising to > 10,000 t/d during the 2018 lower East Rift Zone (LERZ) eruption. At Popocatépetl in Mexico, SO₂ flux spikes precede dome growth and explosive events by hours to days. The ratio CO₂/SO₂ in volcanic gas is particularly informative: because CO₂ exsolves from magma at greater depths than SO₂ (due to its lower solubility), an increasing CO₂/SO₂ ratio often signals the ascent of fresh, deep magma into the system. Real-time Multi-GAS sensors (measuring CO₂, SO₂, H₂S, and H₂O simultaneously) deployed on crater rims or fumarole fields now provide continuous multi-species gas data that allows tracking of degassing path evolution.

Eruption Forecasting

Eruption forecasting integrates all monitoring data streams into probabilistic statements about the likelihood and timing of future eruptions. The field has moved away from deterministic prediction (stating exactly when an eruption will occur) toward probabilistic frameworks that acknowledge the fundamental uncertainties in the volcanic system.

Short-term forecasting (hours to weeks) relies on detecting accelerating precursors — increasing seismicity rates (often described by the Failure Forecast Method using inverse rate plots of seismicity against time, projecting to a singularity that approximates failure time), increasing deformation rates, increasing gas flux, and changing gas ratios. Rapid acceleration of multiple parameters simultaneously provides the highest confidence in an imminent eruption. The challenge is distinguishing genuine precursors of eruptive activity from episodes of unrest that do not culminate in eruption — failed eruptions or aborted intrusions — which are in fact more common than eruptions at many restless calderas.

Long-term forecasting uses recurrence intervals from the geological record (eruption catalogues, dated tephra layers, historical records) combined with statistical models (Poisson process for memoryless volcanoes, or more sophisticated renewal processes for volcanoes with recharge dynamics) to estimate eruption probabilities over years to decades. The Volcanic Explosivity Index (VEI) scale, logarithmic from 0 to 8, provides a framework for comparing historical eruption sizes, with VEI 8 eruptions (> 1000 km³ tephra) occurring perhaps once per million years globally.

Chapter 10: Case Studies in Volcanology

Mount St. Helens, 1980

The May 18, 1980 eruption of Mount St. Helens in Washington State, USA, remains the most intensively studied volcanic event in history and fundamentally transformed understanding of volcanic hazards, particularly sector collapse and lateral blasts.

In the months prior to the eruption, an intense swarm of earthquakes and gradual but accelerating inflation of the north flank — a cryptodome of magma intruding horizontally into the edifice — caused the north face of the volcano to bulge outward at a rate of approximately 2 m/day. By mid-May, the north flank had displaced outward approximately 100 m. On May 18, a magnitude 5.1 earthquake triggered the sudden failure of the entire north flank: approximately 2.4 km³ of rock avalanched northward in the world’s largest historical landslide, instantaneously unroofing the pressurized cryptodome and triggering a lateral blast — a supersonic, ground-hugging surge of fragmented rock, gas, and magma that traveled at up to 480 km/h and devastated 600 km² of forest within seconds. The blast was directed at low angles (not vertically) and thus outran any conventional evacuation radius.

The depressurization of the unroofed magma chamber immediately triggered the Plinian eruption column, which reached 25 km altitude and deposited ash eastward across the United States. Simultaneously, pyroclastic density currents swept the newly formed amphitheater and adjacent valleys, and lahars (volcanic mudflows triggered by melting snow and ice and incorporation of debris) ran down river valleys. The eruption produced approximately 1 km³ of magma (dacite, ∼63 wt% SiO₂) and killed 57 people, most from the lateral blast.

The post-1980 phase of lava dome growth, punctuated by explosive dome-destruction events and renewed dome growth from 1980 to 1986 and again from 2004 to 2008, provided invaluable data on the relationship between dome extrusion rate, gas content, and explosive versus effusive behavior. The 2004–2008 spine-extrusion episode, in which large, nearly solid spines of degassed dacite were extruded at rates of up to 7 m/day with relatively few explosions, demonstrated that highly degassed, low-volatile melt can be erupted effusively even in a silicic, high-viscosity system.

Kīlauea and the 2018 Lower East Rift Zone Eruption

Kīlauea, on the Island of Hawai’i, is arguably the best-monitored and most continuously active volcano on Earth. Its activity is fundamentally driven by the Hawaiian mantle plume supplying basaltic magma to a shallow summit magma reservoir (∼2–3 km depth), from which magma can flow laterally through rifting zones extending to the east (East Rift Zone, ERZ) and southwest (Southwest Rift Zone, SWRZ). The 1983–2018 Pu’u ‘Ō’ō eruption on the middle ERZ was one of the longest-lived and most voluminous in Hawaiian history before it was eclipsed by the dramatic 2018 events.

In May 2018, the summit lava lake that had been active since 2008 within Halemaʻumaʻu Crater began dropping rapidly as magma drained eastward through the ERZ toward the lower ERZ. A series of at least 24 fissures opened in the residential Leilani Estates area (Lower East Rift Zone), erupting low-viscosity, high-temperature (1170°C) olivine tholeiite basalt (Fissure 8, later named Ahu’ailā’au) that produced the largest lava flow volumes of the eruption. The eruption was accompanied by SO₂ emissions of up to 50,000 t/d — the highest recorded at Kīlauea — reflecting the large magma discharge rate (estimated peak of 100–200 m³/s from Fissure 8). Lava entering the ocean created laze (lava haze, a caustic mixture of hydrochloric acid, steam, and fine glass particles from thermal shock) at the coastal entry points.

Lava tube formation during effusive Hawaiian eruptions is critically important for enabling lava to travel long distances while retaining heat. A lava tube is a natural conduit enclosed by a roof of solidified lava, insulating the flowing lava within from the atmosphere. The thermal insulation reduces heat loss so efficiently that lava can travel 50–100 km from the vent and still be erupted at the coast at temperatures > 1100°C. Lava tubes form when the surface of a lava flow solidifies while flow continues in the interior, or when roofed channels develop through thermal and mechanical processes. The 2018 LERZ lava flows traveled approximately 35 km from the main fissure zone to the ocean, enabled by an efficient tube system that developed within days of the eruption onset.

Pinatubo 1991

The June 1991 eruption of Mount Pinatubo in the Philippines is the largest eruption of the twentieth century and a landmark case study in successful eruption forecasting that saved tens of thousands of lives. After 600 years of dormancy, Pinatubo awakened in April 1991 with phreatic explosions. The response of the Philippine Institute of Volcanology and Seismology (PHIVOLCS) and the US Geological Survey, informed by rapidly escalating seismicity, SO₂ emissions (initially > 5000 t/d on June 5), and ground deformation, allowed for a timely evacuation of approximately 58,000 people from the immediate hazard zone.

The climactic eruption on June 15 generated a Plinian column reaching 34–40 km altitude, depositing dacitic (∼67 wt% SiO₂) tephra over > 5000 km² and producing numerous PDC that swept the volcano’s flanks. The estimated total erupted volume was approximately 5 km³ DRE (dense rock equivalent), making it a VEI 6 event. Pinatubo injected an estimated 20 million tonnes of SO₂ (10 Tg SO₂) into the stratosphere — the largest volcanic SO₂ injection of the twentieth century.

The success of the Pinatubo forecasting operation — based on a multi-parameter monitoring approach and the decisive communication of hazards to civil authorities — is now the model for international volcano crisis response. The eruption caused approximately 800 deaths (almost all from roof collapse under tephra or lahar-related causes, not from the eruption itself), a remarkably low toll for an eruption of this magnitude and for a population of several hundred thousand living within 30 km of the volcano.

Deccan Traps and the K-Pg Boundary

The Deccan Traps of peninsular India represent one of the largest continental flood basalt provinces in Earth’s history, covering approximately 500,000 km² with a preserved volume of approximately 5 × 10⁵ km³ and an estimated original volume of up to 1 × 10⁶ km³ before erosion. The eruptions occurred primarily between 66.5 and 65.5 Ma, straddling the Cretaceous-Paleogene (K-Pg) boundary at 66.04 Ma as determined by the iridium anomaly and shocked quartz layer that mark the Chicxulub bolide impact.

High-precision ⁴⁰Ar/³⁹Ar dating of Deccan lava flows shows that approximately 70–75% of the total erupted volume was emplaced within approximately 750,000 years, with the most intense phase (Wai Subgroup, ∼3 × 10⁵ km³) coinciding remarkably closely with the K-Pg boundary. This temporal coincidence has fueled debate about whether the Deccan volcanism contributed significantly to the end-Cretaceous mass extinction. Some models propose that large Deccan pulses could have released 0.5–1 × 10¹⁷ grams of CO₂ and up to 1 × 10¹⁶ grams of SO₂, potentially causing significant acid rain, ocean acidification, and greenhouse warming over thousands of years.

However, the geological and biological evidence overwhelmingly supports the Chicxulub impact as the primary driver of the mass extinction: the selectivity of extinctions, the global iridium anomaly precisely at the K-Pg boundary, the abundance of shocked quartz and impact spherules, and detailed stratigraphic records from marine sediments all point to an instantaneous global catastrophe. The current consensus is that Deccan volcanism was a contributing environmental stressor before and after the impact boundary, potentially weakening ecosystems prior to the impact and slowing biotic recovery afterward, but not the primary kill mechanism.

Chapter 11: Magmatic Ore Deposits

Introduction to Magmatic Ore Formation

Magmatic ore deposits are concentrations of economically valuable metals and minerals that form as direct products of magmatic processes — crystallization, immiscibility, and hydrothermal fluid exsolution — rather than by sedimentary or metamorphic mechanisms. They provide most of the world’s nickel, copper, platinum-group elements (PGE), chromium, and significant fractions of iron, titanium, and vanadium. Understanding their genesis requires integrating igneous petrology (crystallization sequences, phase equilibria) with economic geology (metal enrichment processes, exploration criteria).

The Sudbury Complex: Impact Melt and Ni-Cu-PGE Deposits

The Sudbury Igneous Complex (SIC) in Ontario, Canada, is unique among the world’s large igneous intrusions: it formed approximately 1850 Ma ago from the melting of the crust and upper mantle by a large (original impactor diameter ∼ 10–15 km) meteorite impact. The impact generated a superheated impact melt sheet of ∼12,000 km³, equivalent in volume to a small LIP, that differentiated in the crater floor as it cooled to produce the layered sequence of the SIC: a lower zone of norite (orthopyroxene gabbro), a transition zone of quartz gabbro, and an upper zone of granophyre (granitic texture).

The economic significance of Sudbury lies in the Ni-Cu-PGE ore deposits concentrated in the contact sublayer and associated offset dikes at the base of the SIC. The ore minerals are primarily pyrrhotite, pentlandite, chalcopyrite, and cubanite forming massive sulfide bodies and disseminated sulfides in silicate rock. The PGE (Pd, Pt, Rh, Ru, Os, Ir) occur as minor sulfide, arsenide, and alloy phases associated with the main sulfides.

The genesis of magmatic sulfide ore involves sulfide liquid immiscibility: when the melt becomes saturated in sulfur, a separate immiscible sulfide liquid (primarily FeS-dominated) exsolves from the silicate melt. The sulfide liquid has an extraordinarily high capacity to scavenge chalcophile (sulfur-loving) metals — Ni, Cu, Co, and especially PGE — from the silicate melt. The efficiency of this scavenging is captured by the high partition coefficients of PGE between sulfide and silicate melts (\( K_d^{PGE}_{sulfide/silicate} \approx 10^4\text{–}10^6 \)).

The Bushveld Complex: Layered Intrusion and PGE Reefs

The Bushveld Igneous Complex of South Africa is the largest known layered mafic intrusion on Earth, with an areal extent of approximately 66,000 km² and a thickness up to 8 km. It intruded approximately 2060 Ma ago into the Kaapvaal Craton as a series of magma pulses that differentiated in place to produce an extraordinarily regular sequence of rhythmic layers alternating between olivine-rich, pyroxene-rich, and plagioclase-rich layers. The total volume of the Bushveld Complex is approximately 1 × 10⁶ km³.

Rhythmic layering in layered intrusions arises from cyclic changes in crystallization order caused by repeated injections of fresh primitive magma into a differentiating magma chamber. Each injection partially resets the system: the new hot, primitive magma mixes with or displaces the evolved resident magma, and upon cooling initiates a new crystallization sequence from chromite → olivine → pyroxene → plagioclase, producing a new unit in the layered sequence. The repetition of these cycles over millions of years produces the hundreds to thousands of individual layers visible in Bushveld outcrop.

The economic significance of the Bushveld is extraordinary: it hosts the Merensky Reef and the UG2 Chromitite layer, which together contain approximately 80% of the world’s known PGE reserves and the majority of its chromium resources.

The Bushveld Complex also produces significant quantities of vanadiferous magnetite (magnetite with V₂O₅ up to 1.5 wt%) from the upper magnetite layers, making South Africa the world’s largest vanadium producer. Vanadium is concentrated in the magnetite because of its high partition coefficient for magnetite relative to silicate melt and because Bushveld magnetite crystallized from a highly evolved, Fe-rich residual magma that had concentrated V over the full differentiation sequence.

Porphyry Copper-Molybdenum Deposits

Porphyry Cu-Mo deposits are the world’s most important source of copper and molybdenum, and they form at the magmatic-hydrothermal transition in subvolcanic to mid-crustal settings associated with intermediate to felsic intrusions in arc environments. Unlike the purely magmatic processes of Sudbury and Bushveld, porphyry deposits form through the exsolution and circulation of aqueous hydrothermal fluids from crystallizing magma bodies.

As a water-saturated, intermediate to felsic magma crystallizes in a porphyry system, the residual melt becomes progressively more water-rich (water is incompatible in common silicate minerals). When the melt reaches H₂O saturation — typically at melt fractions of 0.3–0.5 — a separate aqueous fluid phase exsolves. This fluid carries dissolved metals (Cu, Mo, Au) and ligands (Cl⁻, HS⁻, SO₄²⁻) that partition strongly from the melt into the fluid. The exsolved fluid is initially hot (> 700°C), single-phase (a supercritical or dense brine), and highly saline. As it ascends and cools, it may separate into a low-salinity vapor phase and a high-salinity brine, and eventually deposits Cu and Mo sulfide minerals (chalcopyrite CuFeS₂, molybdenite MoS₂, bornite Cu₅FeS₄) in a spatially zoned pattern of veins and disseminations centered on the intrusion.

The total Cu budget in a large porphyry system is remarkable: a world-class deposit such as El Teniente (Chile, the world’s largest underground mine) contains approximately 95 million tonnes of copper in an intrusive complex 10–15 km in diameter. Achieving this requires the exsolution of fluid from many km³ of crystallizing magma and efficient transport and deposition of Cu from the fluid. Isotopic and fluid inclusion studies consistently show that the ore-forming fluids are predominantly magmatic in origin (high-temperature fluid inclusions with magmatic δD and δ¹⁸O signatures), although meteoric water mixing may occur in the outer alteration zones.

Chapter 12: Summary of Igneous-Tectonic Associations

Synthesis

The diversity of igneous rocks on Earth — from the kilometre-wide lava flows of shield volcanoes to the granitic batholiths of mountain belts, from the ephemeral carbonatites of East African rifts to the million-tonne PGE ore deposits of ancient layered intrusions — can be understood as the expression of a limited number of fundamental processes operating in specific tectonic environments. The framework developed through these chapters allows these associations to be decoded systematically.

Mid-ocean ridges are the simplest case: adiabatic decompression of depleted asthenospheric mantle produces large-degree partial melts (F ≈ 10–25%) of tholeiitic basalt composition. Low-pressure fractional crystallization in the axial magma lens drives the melt composition through the iron-enrichment trend, producing differentiated basalts and, in exceptional cases, andesite and icelandite (iron-rich andesite). The geochemical signature is depletion of incompatible elements relative to primitive mantle, flat to slightly depleted REE patterns, and DMM-like isotopic compositions. MORB are the most voluminous magma type on Earth when measured by annual production rate.

Subduction zones invert this picture: volatile fluxing adds H₂O and LILE to the mantle wedge, lowering the solidus and triggering partial melting that produces water-rich, oxidized basalts. Fractional crystallization coupled with crustal assimilation and magma mixing drives these primary melts through a calc-alkaline differentiation trend, producing the andesite-dacite-rhyolite continuum that characterizes continental arcs. The REE signatures of arc magmas span a wide range depending on the depth and degree of melting, with adakitic signatures indicating garnet in the residue (slab melts or deep-crustal assimilation) and flat HREE patterns indicating spinel-stability melting of the wedge. The continental crust itself is broadly andesitic in composition, consistent with long-term subduction zone magmatic input, assimilation, and remelting.

Ocean islands represent deep mantle plume activity, sampling isotopically heterogeneous, enriched mantle sources that preserve recycled oceanic crust and sediment in proportions reflected by their EM1/EM2/HIMU isotopic fingerprints. Small-degree partial melts of deep, volatile-enriched sources produce alkalic, LREE-enriched melts; higher-degree melts from the hotter plume interior produce tholeiitic shield basalts. The Hawaiian shield-to-alkalic progression, the Galápagos archipelago showing a radial geochemical zonation from MORB-like to OIB-like compositions with distance from the hotspot center, and the Icelandic plume interacting with the Mid-Atlantic Ridge spreading center — all represent variations on the fundamental OIB theme.

Continental settings add the complexity of thick, compositionally diverse lithosphere. Rifts generate alkalic to tholeiitic basalt suites depending on extension rate and lithospheric thickness; granites form by partial melting of the lower or middle crust of diverse protolith compositions, generating I-, S-, and A-type variants. Flood basalts sample both plume head components and lithospheric mantle that has been metasomatized by previous subduction or mantle plume activity.

The magmatic ore deposits that form in these settings — Ni-Cu-PGE in large mafic systems, porphyry Cu-Mo at arc intrusions, Cr-PGE in layered intrusions, and carbonatite-hosted rare earth element deposits in rifts — represent the concentration of metals through the thermodynamic and kinetic processes of igneous differentiation: sulfide immiscibility, crystal fractionation, and hydrothermal fluid exsolution. Understanding the petrological and geochemical controls on these processes is thus not only intellectually satisfying but economically essential in a world of growing demand for the critical metals that fuel the energy transition.

The quantitative tools developed throughout these notes — the batch and fractional melting equations, the Rayleigh fractionation equation, the Henry’s Law solubility expressions, the column height–eruption rate scaling, and the R-factor framework for magmatic sulfide deposits — provide the means to move beyond qualitative description and constrain igneous processes with numbers. A student who can apply these equations to real geochemical datasets, interpret REE and spider diagrams in the context of tectonic setting, read a phase diagram to predict crystallization sequences, and assess volcanic hazard from monitoring data has achieved the core competencies of igneous petrology and volcanology at the undergraduate level.

The Earth continues to be reshaped by magmatic processes at rates that are geologically rapid and, for humans living near active volcanoes, urgently relevant. The 2018 Kīlauea eruption displaced thousands of residents and added new land to the island’s coast. The ongoing unrest at Campi Flegrei (Phlegraean Fields) in Italy — a large caldera system near Naples that has been inflating at an accelerating rate since 2005 — poses a potential hazard to one of the most densely populated volcanic regions in the world. Iceland sits atop the most active segment of the Mid-Atlantic Ridge, and its eruptions have repeatedly disrupted European air traffic and, in the case of the 1783–1784 Laki fissure eruption, may have contributed to climate anomalies that worsened famines across the Northern Hemisphere. The intellectual framework built in these chapters is thus not only a foundation for further academic study but a set of tools for understanding — and helping to mitigate — one of the most fundamental forces shaping our planet.