Volcanic Events, pg. 2
Mount St. Helens History, pg. 3-15
Eyewitnesses, pg. 53-67
Absolute Times, pg. 81-82, 86
Activity Sequence, pg. 127-134
Gas Studies, pg. 190-191
Ash Clouds, pg. 323-333
Blast Dynamics, pg. 379-400
Rapid Deposition, pg. 466-478
Phreatic Explosions, pg. 509-511
New Lava Dome, pg. 540-544
Ash-Fall Deposits, pg. 568-584
Water Chemistries, pg. 659-664
River Water Quality, pg. 719-731
CHEMISTRY OF SURFACE WATERS
The composition of surface waters sampled in June reflects (1) alteration and dissolution of minerals and glass, including surface coatings, (2) leaching of elements from organic matter, particularly in Spirit Lake, and (3) precipitation of secondary products. Tephra and water-sample compositions, plotted in figure 386, suggest that potassium is relatively immobile, and that the Ca-Na ratio in surface water may reflect the composition of the materials in contact with the water. The relative mobility of elements in the tephra can be calculated from the leachate data presented in table 78 as the ratio ((concentration in leachate/concentration in unleached tephra) x 100) for each element. Those data show:
Cl> SO4> > B> Ca = Mn> Na> Mg = K = Sr> > Ba> > SiO2> Fe
Surface-water compositions display similar patterns, except Ca and Na change positions in the index. Silica, aluminum (not included in table 78), and iron are highly immobile; chloride and sulfate are highly mobile; and boron, calcium, and manganese are the most mobile of the remaining elements. Two inferences can be drawn from these results: (1) the highly mobile elements from the tephra existed as liquid or readily soluble coatings on grain surfaces, much as they did in encrustations around fumaroles, and (2) immobile elements, such as silica and aluminum, must dominate the residue formed by dissolution of the surface layers of glass, feldspar, and ferromagnesian minerals. The relative mobility of Ca and Mn suggests that these elements either were concentrated in soluble form at grain surfaces or occupied exchange sites. However, sodium is present in higher concentrations than calcium in most surface waters at Mount St. Helens, particularly in those on pyroclastic flows. The lower levels of calcium in surface waters may reflect selective removal to exchange sites on allophane (Wada, 1977), or high calcium levels in the leaching experiments may have resulted from solution of gypsum coatings on mineral and glass surfaces.
Concentrations of most trace metals in surface waters from Mount St. Helens are low compared to those in other thermal waters. (See Wollenberg and others, 1979.) Manganese concentrations are high in all samples, and dissolved iron levels were high in Spirit Lake in June 1980. The relatively low levels of iron in other surface waters and a reported decrease of iron concentrations in Spirit Lake since June (T. J. Casadevall, written commun., 1980) probably result from the precipitation of an iron oxide or sulfide phase. Oxidation and precipitation rates of manganese are retarded in neutral or acidic organic-rich water (Stumm and Morgan, 1970); the rates are slower than rates for iron in most waters, which may explain the high manganese concentrations.
The stoichiometry of weathering solutions may be used, in relatively simple systems, to suggest the nature and rates of weathering reactions (Stumm and Morgan, 1970). For instance, the weathering of andesine feldspar to kaolinite may be written thus (Stumm and Morgan, 1970): ------
The molar ratios of Ca:Na:H4SiO4 in weathering
solutions for this reaction should be 1:1:2, but the measured values for
the Spirit Lake inflow and the mean of the pyroclastic-flow ponds give
1:4.2:1.3 and 1:2.3:0.24, respectively. These discrepancies suggest that
congruent alteration of plagioclase does not control solution composition.
If the dissolution of the glassy fraction of the pumice is a controlling
factor, the Ca-Na ratio in solution should be near 0.34 (as reported for
separated glass samples, this volume, table 80, analyses 8 and 17), even
if silica and aluminum are conserved. Weathering of deposits rich in lithic
material or organic matter should result in higher Ca-Na ratios, but all
postulated weathering systems are complicated by the tendency of divalent
cations such as Ca+2 and Mg+2 to preferentially occupy
exchange sites in crystalline or amorphous material. Chemical data from
the leaching experiments, the Spirit Lake inflow, and the pyroclastic-flow
ponds give Ca-Na ratios, respectively, of 0.91, 0.24, and 0.43. Water from
the pyroclastic flows thus gives ratios that bracket the predicted values.
Stoichiometric data suggest that the composition of water in contact with
the pyroclastic material reflects the dissolution of glass rather than
the alteration of andesine to kaolinite, but laboratory experiments are
required to substantiate the conclusion.
Equilibrium thermodynamics provides an additional method for examining
mineral stability and water chemistry. Figure 387 is a plot of stability
relations in the gibbsite-kaolinite-amorphous silica-smectite system. The
compositions of water from ponds on the pyroclastic flows, from the inflow
to Spirit Lake, and from Spirit Lake plot in the kaolinite stability field,
but the data suggest ponds in the debris avalanche are saturated with calcium
smectite. The stability field drawn for calcium smectite should be regarded
as approximate because of uncertainties in extrapolating thermodynamic
data to the composition of specific smectites. Metastable reaction products
may also persist at relatively low temperatures. We have sketched lines
on the stability diagram to show the progressive change in chemistry from
rainwater to the surface waters. If data for dissolved sodium and sodium
smectite are used (Tardy, 1971), the debris-avalanche pond waters still
plot in the smectite field, and the other waters remain in the kaolinite
field. Stability fields for chlorite, magnesium smectite, or other minerals
could be plotted by using additional axes. However, figure 387 is the simplest
diagram that includes the principal dissolved cations and silica.
Stability diagrams and data from the literature for soil-forming processes and hydrothermal alteration suggest that kaolinite (or halloysite) and smectite should be forming in the new deposits at Mount St. Helens. We have not positively identified neoformed smectite or kaolinite; the high levels of sulfur, iron, calcium, and sodium in encrustations at fumaroles and the presence of zeolites in the warm pyroclastic flows provide the strongest evidence for neoformed minerals. Mixed-layer chlorite-smectite and chlorite that are present in some thermal waters may be neoformed, in part.
However, the chemistry of surface waters, leaching experiments, and the appearance of glass and mineral surfaces under high magnification suggest that dissolution and alteration are occurring at present. The lack of evidence for crystalline aluminosilicates such as kaolinite suggests the presence of amorphous or nearly amorphous material, but we have not chemically analyzed this phase. Many investigators (see Ugolini and Zasoski, 1979) report rapid neofor-mation of minerals in ash-rich deposits. The absence of neoformed crystalline silicates in deposits from Mount St. Helens after 150 days of alteration may reflect either slow rates of dissolution or sluggish recrystallization of amorphous material, perhaps inhibited by the abundance of sulfate, chloride, and cations.
Despite the ambiguities in interpreting the leachate and surface-water chemistry in terms of mineral weathering, it is apparent that two interrelated proesses occur during the initial phases of alteration. Rapid surface reactions (fig. 388) and the dissolution of the finest particles apparently dominate the first seconds to hours (Busenberg and Clemency, 1976; Holdren and Berner, 1979) of rock-water contact. Then slower reaction rates occur as mineral-surface sites are filled by more stable ions, the pH of the soltion rises, and the kinetics of diffusion processes becomes important. In general, these initial, rapid reactions follow simple, first-order kinetics (White and Claasen, 1980), while long-term alteration is controlled by slower, diffusion-controlled reactions of the general form: Ci = kitn
where: Ci = dissolved activity of species i at time t ,We arbitrarily set n = 1/2 in plotting figure 388. Concentrations and times in figure 388 are approximate, but it is apparent that silica is released to solution at a rapid rate, chloride ion is not released after the initial stage of leaching, and the other constituents show initially rapid releases followed by slower rates. An extensive literature examines the causes of these slower rates (see, for instance, Petrovic, 1976; and Siever and Woodford, 1979); it is possible that control is exercised by surface reactions at lattice defects (see Berner, 1978; Berner and Holdren, 1979) or by the rate of diffusion through a patchy, amorphous aluminosilicate layer (Paces, 1973) that is slowly dissolving.
k = rate constant for species i,
t = reaction time, usually expressed in seconds,
n = exponent dependent on reaction, generally 0 < n < 1.
In closed systems, C, eventually approaches an upper value determined
by equilibrium (Helgeson, 1972), but in the open systems that characterize
most surface waters, "equilibrium" is a somewhat misleading concept, and
C is likely to reflect reaction kinetics and water contact time. More extensive
interpretation of reaction rates from surface-water chemistry requires
laboratory study of glass dissolution and a better understanding of contact
time, flow pathways, and other hydrologic factors.
FORMATION OF SECONDARY MINERALS
The amorphous alumina and silica-rich residue predicted by elemental mobility, reaction rates, and other factors noted above should crystallize to kaolinite or halloysite. In the presence of higher pH and a greater H4SiO4 concentration, neoformation may result in smectite. The pervasive smectite in the debris avalanche and in other deposits probably formed in the cone of Mount St. Helens over the past few hundred to few thousand years through hydrothermal action at relatively low temperatures. It may have been concentrated in fracture systems associated with domes, particularly the summit dome and Goat Rocks. Smectite, ML chlorite-smectite, and chlorite could be forming now in the debris avalanche, and perhaps in Spirit Lake as well. Sulfur, alunite, gypsum, and other products of acid-sulfate alteration have formed within 30 cm of the surface in the oxidized zone of some rootless fumaroles, and neoformed(?) zeolite minerals are present, without acid-sulfate species, at greater depths in these fumaroles. This mineral assemblage is similar to that described by Levering (1957) for the Valley of Ten Thousand Smokes in Alaska, and, on a minute scale, to the mineral assemblage at the active geothermal system at Steamboat Springs, Nev. (Schoen and others, 1974). There is no evidence that kaolinite-group minerals or other crystalline silicates had formed after about 150 days. Rodrique and others (1973) showed that kaolinite can be produced at ~200°C from solution in only a few weeks. However, the absence of kaolinite in the thermal ponds at Mount St. Helens (~25°C) is not surprising, as Hem and Lind (1974) have shown that crystallization of kaolinite takes several months to years at these temperatures. Furthermore, crystallization may be inhibited by the Si-Al ratio in solution by high-alkali concentrations (see Eberl and Hower, 1975), by the presence of certain forms of organic matter, or by high concentrations of sulfate and chloride ions. Subsurface temperatures at several sample sites in the pyroclastic flows remain above 200°C at present, so crystallization could be inhibited by one of the nonthermal factors.
Sparse data from older Mount St. Helens deposits and studies of alteration
in other volcanic deposits permit us to speculate about mineral alteration
that will occur in the new deposits. Halloysite noted by Mullineaux and
Crandell (1962) in a mudflow dating to 2,000 yr B.P. may reflect acid-sulfate
alteration in the cone, crystallization from allophane in the moist, acidic-weathering
environment near Mount St. Helens, or possibly both. Because Ugolini and
Zasoski (1979) report that halloysite forms rapidly from volcanic ash in
moist environments, and there is no evidence that the halloysite was a
hydrothermal mineral, we assume that the older halloysite is a product
of soil-forming processes. We infer that halloysite is likely to form in
the weathering zone of the 1980 deposits within a few hundred to a few
thousand years. Other likely minerals include kaolinite, cristobalite or
opal, and possibly vermiculite, in addition to smectite. These minerals
can be expected to form in leached zones beneath organic matter that is
already beginning to accumulate in some of the deposits; deeper, less weathered
portions of the deposits are likely to continue to reflect the mineralogy
of the parent material or its alteration products (ML chlorite-smectite
and chlorite). Ferromagnesian minerals will probably weather most rapidly,
followed by glass or plagioclase. (See Jackson, 1964.) Many of the highly
colored encrustations present in the pyroclastic flows are quite soluble
and thus ephemeral, with the exceptions of hematite and zeolites at depth,
which may provide lasting mineralogic evidence of fumarolic activity. (See
Lovering, 1957; Narebski and Paulo, 1973.) We would expect the most rapid
mineralogic changes in the cooling portions of the pyroclastic flows, where
glass devitrification will result in the formation of a feldspar and cristobalite
or tridymite mixture perhaps accompanied by halloysite or kaolinite. Finally,
neutral or slightly alkaline pH levels in the debris-avalanche, blast,
and mudflow deposits could promote the formation of additional smectite
and zeolites. However, as silica and cations are leached by the abundant
precipitation of the area, clay mineralogy of the deposits will probably
come to resemble that of the Silver Lake mudflow reported by Mullineaux
and Crandell (1962).
The mineralogy of encrustations at fumaroles, chemistry of surface water, and solution features observed by SEM study suggest that the deposits emplaced in 1980 are undergoing alteration as they cool and react with surface waters. The chemistry of waters warmed by the deposits and the results of ash-leaching experiments indicate that water reacts quite rapidly with glass, ferromagnesian minerals, and plagioclase, leaving amorphous residues rich in silicon, aluminum, and iron. With the exception of minerals at fumaroles and possibly smectite, ML chlorite-smectite, and chlorite at some warm ponds, neoformed minerals could not be detected in deposits after 150 days of alteration. Crystallization of the amorphous residue apparently proceeds slowly, even in deposits at 200° C, and may be inhibited by an excess of one or more cations, improper Si-Al ratios, abundant sulfate or chloride, or other factors.
Kaolinite formation in the new deposits is predicted by stability
relations calculated for most surface waters on those deposits, while smectite
stability is suggested for thermal ponds in the debris-avalanche deposit.
The prevalence of trioctahedral smectite in the < 2-m fraction of material
that composed the old cone suggests that alteration occurred there at neutral
to slightly alkaline pH and at temperatures of a few hundred degrees or
less, perhaps localized around and in the summit and Goat Rocks domes.
Opal, kaolinite, and halloysite were not found in any of the new deposits,
including the debris avalanche, so areas of acid-sulfate alteration in
the old cone must have had limited extent if they were present at all.
However, as silica and cations are leached from the newly formed deposits
during soil-forming processes, and as the pyroclastic flows undergo thermal
alteration, halloysite or kaolinite will probably form. The presence of
halloysite in an approximately 2,000-yr-old mudflow, noted by Mullineaux
and Crandell (1962), suggests that this mineral will probably form in the
new deposits within the next few hundred to few thousand years.
|START & IMAGES||