FIRE and MUD Contents

Evolution of a Small Caldera Lake at Mount Pinatubo

By Nora R. Campita,1 Arturo S. Daag,1 Christopher G. Newhall,2 Gary L. Rowe,2 and Renato U. Solidum1

1Philippine Institute of Volcanology and Seismology.

2U.S. Geological Survey.


Collapse of Mount Pinatubo's edifice during its explosive eruptions on June 15, 1991, created a 2.5-kilometer-wide caldera. By early September 1991, a lake began to form on the caldera floor, mainly by spring discharge from the walls of the caldera augmented by rainfall and surface runoff. The lake became increasingly acidic with time, with pH changing from 6.0 in October 1991 to 1.9 in December 1992. Lake temperature over the same period remained about 38+-2°C, except near fumaroles or hot springs. The high initial pH value reflects water of meteoric origin dominating that from a deeper hydrothermal system. pH decreased as the lake absorbed acid magmatic gases, becoming an acid sulfate-chloride brine. Much of the acidification during the period February 1992 to December 1992 may have occurred between July and October, when a tuff cone and then a dome grew within the lake. A slight overall enrichment of Mg and Si suggests posteruption leaching of rock by hot waters in and beneath the lake.

Note to readers: Figures open in separate windows. To return to the text, close the figure's window or bring the text window to the front.


Mount Pinatubo's violent eruption on June 15, 1991, formed a nearly circular caldera about 2.5 km in diameter, and, beginning in September 1991, a lake began to form from spring discharge, supplemented by rainfall and surface runoff. Over time, such lakes can act as large condensers that absorb the heat and volatile emanations from vents on the caldera floor. In addition, such lakes can also serve as recycling centers for condensed volcanic fluids. Mass balance calculations for Poás Volcano, Costa Rica (Brantley and others, 1987, 1992; Rowe and others, 1992), and for El Chichón Volcano, Mexico (Casadevall and others, 1984), indicate that seepage of caldera (or crater) lake brine and subsequent recycling affects the lake chemistry.

For volcano monitoring purposes, increases in lake water SO4/Cl ratios prior to or during eruptive activity have been noted in the crater lakes of Zavaritsky Volcano, Russia (Menyailov, 1975), and Kusatsu-Shirane Volcano, Japan (Takano and Watanuki, 1990), as acid, volcanic gases have interacted with meteoric waters. In addition, the Mg/Cl ratio has been used along with pH variations at the crater lake of Mount Ruapehu, New Zealand, to monitor changes in gas flux from fumaroles and to document high-temperature water/rock interaction episodes (Giggenbach, 1974, 1983; Giggenbach and Glover, 1975). The eruption in the crater lake of Soufrière Volcano in Saint Vincent in 1971-72 revealed that the relative ease of leaching components of the lava was Na>Fe>Mg>K>Ca>>Si (Sigurdsson, 1977).


Figures 1A-D show development of Pinatubo's caldera lake. In August 1991, ash emission waned enough to permit views into the new caldera, and numerous springs could be seen discharging from the caldera walls. Spring discharge that collected in the vent was being heated rapidly and flashed to steam in frequent phreatic explosions. As a result, no lake was forming.

However, by early September, as the rainy season progressed and as rocks of the vent area cooled, surface runoff and discharge from these springs began to collect and a shallow lake covered much of the caldera floor. Runoff from a catchment area of about 5 km2 fed into a lake about 1/15 that size, so a meter of rain (average for the last half of August and early September at nearby Cubi Point and Dizon Mines; Rodolfo and others, this volume) could account for more than 15 m of water accumulation on the bowl-shaped caldera floor. Although our data are crude, we judge that the lake required both surface runoff and spring discharge to counter evaporation and to develop during this early period.

The lake rose slightly over the dry season from October 1991 through July 1992 (for general rainfall information, see fig. 3 of Umbal and Rodolfo, this volume)--adding evidence for the importance of spring discharge. Throughout this period, the surface area of the lake remained approximately constant, about 4x105 m2.

In July-August 1992, renewed eruptions formed a small tuff cone and then a dome in the center of the lake (fig. 1C; see also Daag, Dolan, and others, this volume). By late August, heavy rains had washed so much debris off the caldera walls that deltas extending from the eastern wall reached the still-growing dome, and the area of the lake was reduced to about 2.5x105 m2. A rain gauge installed on the north rim of the caldera recorded 280 mm of rainfall in September 1992, and by late September, deltaic sediment surrounded two-thirds of the dome. The dome stopped growing at the end of October.

Logistical difficulties and hazard precluded detailed bathymetry. However, a view in early August 1991 (fig. 1A) shows an inner crater whose floor was at least 100 m below what was soon to become lake level. Erosion of sediment from the caldera walls probably filled most if not all of the deep crater by October 1991 (end of the 1991 rainy season), so the average depth during the period October 1991 to July 1992 was perhaps in the order of 10 to 20 m. Shorelines in October 1991 and July 1992 photographs suggest that the lake rose several meters during the dry season, a period of relatively modest erosion, so the average lake depth probably increased slightly over that period.

Figure 1. Progressive development of the Pinatubo caldera lake. A, Pinatubo caldera on August 1, 1991, before the caldera lake began to form (photographed from the northwest by T. Casadevall, USGS). B, Pinatubo caldera lake as seen on September 10, 1991 (photographed from the northwest by T. Pierson, USGS). C, New lava dome that grew through the caldera lake, as seen on July 29, 1992 (photographed from the northeast by R. Arante, PHIVOLCS). D, Pinatubo caldera lake on November 22, 1992. Westward growth of deltas partially surrounded the dome and sharply reduced the surface area of the lake (photographed from the northeast by R. Arante, PHIVOLCS).

When a new dome began to grow through the lake in July and August 1992, occasional updoming of talus on lake sediment surrounding the dome suggested that the lake was only a few meters deep. In November 1992, various lengths of rope were dropped from a helicopter into what appeared to be the deepest part of the lake, immediately west of the dome. These lengths of rope were weighted by a small sandbag at one end and marked by color-coded floats at the other. Floats at the end of 5-, 10-, 15-, and 20-m ropes disappeared from sight, indicating greater depths; unfortunately, the soundings had to be curtailed when rotor wash splashed acidic water on the helicopter. From these admittedly sparse observations, we judge that the average depth of the caldera lake in 1991-92 was in the order of 10 m, shoaling faster by deposition and evaporation than it deepened by influx of water. Its volume increased to a maximum of about 10 7 m3 in October 1991 and had decreased to half of that volume or less by late 1992. (Editorial note: By 1994, the lake level was several tens of meters higher than in late 1992.)


Lake water samples from October 1991 to December 1992 were all collected and stored in polyethylene bottles. Temperature and pH were measured in situ (except in November 1991) by use of a portable digital pH-temperature meter. Samples were not filtered in the field but were passed through a 0.1-m millipore filter before analysis. The methods for chemical analyses of lake water samples are summarized in table 1. The results of those analyses are shown in table 2A.

Sample PCL-5 was taken from an area of bubbling at the lake shore and is thus a hybrid of hot spring discharge and the lake as a whole. Regrettably, no samples could be obtained of early springs that gushed from the caldera walls; we can only guess that they were some mixture of meteoric water and geothermal fluids described by Delfin and others (this volume).

Table 1. Summary of analytical methods used on caldera lake water of Mount Pinatubo.




By mannitol method (PHIVOLCS, PNOC).


By Mohr method (PHIVOLCS, PNOC).


By potentiometric method (F-electrode)


All cations are analyzed by atomic absorption spectrophotometry (PHIVOLCS, PNOC).


















By gravimetric and colorimetric method


Lake temperature and pH are shown in table 2A and in figure 2A. Temperature has remained more or less constant; pH has decreased sharply, from 6.0 in October 1991 to 1.9 in December 1992. Li, Na, K, Cl, SO4, Ca, Mg, and total dissolved solids (TDS) dropped from October 1991 to February 1992 but had increased again by December 1992 (table 2A and fig. 3A). F, Fe, and SiO2 generally increased from October 1991 straight through to December 1992. The Mg/Cl ratio (table 2B and fig. 3B) increased between October 1991 and February 1992 and remained more or less the same in December 1992.

Table 2A. Result of chemical analyses of caldera lake water of Mount Pinatubo.

[nd, not determined; LANL, Los Alamos National Laboratory; PNOC/PV, Philippine National Oil Company/PHIVOLCS]









Date collected








Temp (°C)

40.0 nd 38.5 38.5 46.5 36.7 36.7

Analyzed by



4.79 5.21          


6.00 nd 2.74 2.78 2.73 1.9 1.9

Total acidity

80.00 110.00 230.00 235.00 230.00 320.00  

Mineral-acid acidity

0.00 30.00 100.00 80.00 90.00 80.00  

Concentration (ppm)





31.10 32.00 24.70 22.40 27.80 28.00 31.00




598.00 597.00 400.00 382.00 444.00 419.00 370.00


1,029.00 742.00 500.00 467.00 567.00 849.00 825.00




0.13 0.36 0.75 0.73 0.61 1.68  


<0.1 2.39 16.10 18.70 8.90 15.00 35 (as Fe3+)




80.00 68.00 45.00 43.00 50.00 58.00 54.00


0.72 0.70 0.50 0.40 0.50 0.70  


95.50 84.10 72.10 73.50 71.20 121.00 110.00




519.00 402.00 285.00 274.00 319.00 395.00 370.00










70.00 95.00 140.00 147.00 119.00 164.00 167.00


1,727.00 1,689.00 1,288.00 1,253.00 1,389.00 1,364.00 1,600.00


3.06           1.4

Total dissolved solids

4,109.00 3,510.00 3,237.00 3,130.00 3,462.00 4,185.00  

Charge balance (WATEQ4F)-3.1

-0.1 2.0 3.9 1.2 18.1 6.5  

Table 2B. Molar Mg/Cl ratios for caldera lake water of Mount Pinatubo.























Correlation spectrometer (COSPEC, Barringer Instruments) measurements of SO2 emission before the climactic eruption showed a dramatic increase, brief stoppage, and then an even greater increase, reaching more than 10,000 t/d on June 10 (Daag, Tubianosa, and others, this volume). SO2 emission during July and August (only 3 measurements before the caldera lake formed) were between 1,000 and 5,000 t/d. However, as soon as the lake formed, SO2 emission (regrettably, only one measurement) dropped to near the detection threshold for the measurements, about 20 t/d.

The next measurements of SO2 emission were in July-November 1992, during a period of increased seismicity and then growth of the dome through the caldera lake. Emissions were <300 t/d before the dome rose above lake level and generally were >300 t/d until the dome stopped growing in October.

These observations suggest that the lake absorbed most of the volcano's SO2 emission while it covered the vent. SO2 might also have been absorbed into a shallow hydrothermal system beneath the lake in 1992, but we cannot judge this from COSPEC measurements.

Figure 2. A, pH of Pinatubo caldera lake, October 1991­December 1992. B, Total acid (T.A.) and mineral-acid acidity (M.A.A.) of the lake during the same period.

Figure 3. A, Concentrations of species in the Pinatubo caldera lake, October 1991­December 1992. B, Mg/Cl ratio of Pinatubo caldera lake water for the same period.



The high initial pH of the lake (6.0) reflects a dominantly meteoric origin of the lake water, with much rainfall in the area from September to October 1991. The subsequent increase in acidity, accompanied by increasing total acidity and mineral-acid acidity (fig. 2B), occurred as the lake and its shallow hydrothermal system absorbed acid magmatic gases.


The initial enrichment of the caldera lake in Cl, SO4, and F can be attributed to addition of HCl, HF, and SO2 from fumaroles during early degassing of the 1991 magma. Solution of anhydrite from rocks within the caldera (mostly pre-1991 rock) could have added SO4, and the high concentration of B could have been from leaching of volcanic wall rocks (Ellis and Sewell, 1963). Mg, Ca, and K were probably leached from rock in and around the conduit and carried in residual hydrothermal fluids of Pinatubo.

By February 1992, the volcano had quieted and B, Cl, K, Li, Mg, Na, SO4, and TDS had decreased correspondingly. The Mg/Cl ratio increased (fig. 3B) as Cl decreased faster than Mg.

The sample of December 1992 reflected dome growth from July to October 1992, during which magmatic gases that were absorbed into the lake and the shallow hydrothermal system raised levels of Cl and F and further decreased pH. Increases in B, K, Mg, Ca, Na, SiO2, Li, and TDS concentrations (fig. 3A) and the Mg/Cl ratio (fig. 3B) can be explained mainly as interaction of the highly acidic lake water with volcanic materials (Giggenbach, 1974; 1983; Giggenbach and Glover, 1975). At least some of these species could have been derived from interaction of lake water with the high-temperature lava dome that eventually emerged above the caldera lake floor.

Mineralogic controls on lake-water composition were evaluated by use of the equilibrium thermodynamic model WATEQ4F (Ball and Nordstrom, 1991). October-November 1991 waters were approximately saturated with respect to gypsum, but increasing acidity and a concomitant increase in bisulfate concentration caused 1992 lake waters to be undersaturated with respect to gypsum. Lake water from December 1992 (sample PCL-6; USGS analysis, table 2A) was undersaturated with respect to gypsum and severely undersaturated with respect to primary and secondary aluminosilicate phases such as feldspars and clays and also with respect to the aluminum-sulfate-hydroxide mineral natroalunite. Natroalunite is found in sediments of the less acidic crater lakes (pH >2-3) such as Kusatsu-Shirane, Japan (Takano and Watanuki, 1990), whereas gypsum is characteristic of sediments of the more concentrated crater lakes (pH 0-1) such as Ruapehu (Giggenbach, 1974) and Poás (Brantley and others, 1987). December 1992 lake waters were in equilibrium with amorphous silica.

A lack of speciation data for reduced sulfur species precludes evaluation of the saturation state of native sulfur. However, disproportionation of SO2 degassed into the lake will result in the production of hydrogen sulfide and sulfuric acid. Subsequent oxidation of the H2S will cause native sulfur to precipitate in lake sediments and at the lake surface. The calculations suggest that with further degassing, pyroclastic material eroded into the lake will continue to dissolve, thereby increasing the concentrations of all rock-forming elements except silica. Lake sediments will consist mostly of pyroclastic material with small but significant quantities of amorphous silica and elemental sulfur. Future trends in lake water chemistry will be determined by variations in magmatic degassing, the degree of interaction between hot rock and lake water, and the amount of rainfall and spring water-derived recharge received by the lake.


The caldera lake of Mount Pinatubo reflects a combination of magmatic degassing into a hydrothermal system and into the lake itself, water-rock interaction in the hydrothermal system, rainfall and surface runoff, and evaporation.

The composition of the earliest sample suggests a predominance of meteoric water--from surface runoff and springs on the caldera walls, mixed with water from Pinatubo's preexisting hydrothermal system. Magmatic gases acidified the early lake, but, as time passed after the 1991 eruptions, the influence of magmatic gases declined. From the middle through late 1992, magmatic gases were again an important control on lake composition.

Evaporation and dilution by fresh rainwater are probably second-order controls on lake composition that would become increasingly important as volcanic output wanes, but our present data are insufficient to quantify their roles.


We thank the U.S. Navy, the U.S. Marine Corps, and Captain Agustin Consunji of Delta Aviation for help in obtaining the caldera lake water samples. We also thank Fraser Goff (Los Alamos National Laboratory), Aristeo Baltazar and personnel of the Philippine National Oil Corporation (PNOC) Geochemical Laboratory, and Ray van Hoven (USGS) for help with water analyses. Tom Casadevall (USGS) and David Sussman (UNOCAL) provided valuable reviews.


Ball, J.W., and Nordstrom, D.K., 1991, User's manual for WATEQ4F, with revised thermodynamic data base and test cases for calculating speciation of major, trace, and redox elements in natural waters: U.S. Geological Survey Open-File Report 91-183, 193 p., one diskette.

Brantley, S.L., Borgia, A., Rowe, G., Fernández, J.F., and Reynolds, J.R., 1987, Poás Volcano acts as a condenser for acid metal-rich brine: Nature, v. 330, p. 470-472.

Brantley, S.L., Rowe, G.L., Konikow, F., and Sanford, W.E., 1992, Toxic waters of Poas Volcano: National Geographic Research and Exploration, v. 8, no. 3, p. 328-337.

Casadevall, T.J., dela Cruz-Reyna, S., Rose, W.I., Jr., Bagley, S., Finnegan, D.L., and Zoller, W.H., 1984, Crater lake and post-eruption hydrothermal activity, El Chichón, Mexico: Journal of Volcanology and Geothermal Research, v. 23, p. 169-191.

Daag, A.S., Dolan, M.T., Laguerta, E.P., Meeker, G.P., Newhall, C.G., Pallister, J.S., and Solidum, R., this volume, Growth of a postclimactic lava dome at Mount Pinatubo, July-October 1992.

Daag, A.S., Tubianosa, B.S., Newhall, C.G., Tuñgol, N.M, Javier, D., Dolan, M.T., Delos Reyes, P.J., Arboleda, R.A., Martinez, M.L., and Regalado, M.T.M., this volume, Monitoring sulfur dioxide emission at Mount Pinatubo.

Ellis, A.J., and Sewell, J.R., 1963, Boron in waters and rocks of New Zealand hydrothermal areas: New Zealand Journal of Science, v. 16, p. 589-606.

Giggenbach, W., 1974, The chemistry of Crater Lake, Mt. Ruapehu (New Zealand) during and after the 1971 active period: New Zealand Journal of Science, v. 17, p. 33-45.

------1983, Chemical surveillance of active volcanoes in New Zealand, in Tazieff, H., and Sabroux, J.C., eds., Forecasting volcanic events: Amsterdam, Elsevier, p. 311-322.

Giggenbach, W., and Glover, R.B., 1975, The use of chemical indicators in the surveillance of volcanic activity affecting the crater lake on Mt. Ruapehu, New Zealand: Bulletin Volcanologique, v. 39, p. 70-81.

Menyailov, I.A., 1975, Prediction of eruptions using changes in composition of volcanic gases: Bulletin of Volcanology, v. 39, p. 112-125.

Rodolfo, K.S., Umbal, J.V., Alonso, R.A., Remotigue, C.T., Paladio-Melosantos, M.L., Salvador, J.H.G., Evangelista, D., and Miller, Y., this volume, Two years of lahars on the western flank of Mount Pinatubo: Initiation, flow processes, deposits, and attendant geomorphic and hydraulic changes.

Rowe, G.L., Jr., Brantley, S.L., Fernandez, M., Fernandez, J.F., Borgia, A., and Barquero, J., 1992, Fluid-volcano interaction in an active stratovolcano: The crater lake system of Poás Volcano, Costa Rica: Journal of Volcanology and Geothermal Research, v. 49, p. 23-51.

Sigurdsson, H., 1977, Chemistry of the crater lake during the 1971-1972 Soufrière eruption: Journal of Volcanology and Geothermal Research, v. 2, no. 2, p. 165-186.

Takano, B., and Watanuki, K., 1990, Monitoring of volcanic eruptions at Yugama crater lake by aqueous sulfur oxyanions: Journal of Volcanology and Geothermal Research, v. 40, p. 71-87.

Umbal, J.V., and Rodolfo, K.S., this volume, The 1991 lahars of southwestern Mount Pinatubo and evolution of the lahar-dammed Mapanuepe Lake.

FIRE and MUD Contents

PHIVOLCS | University of Washington Press | U.S.Geological Survey

This page is <>
Contact: Chris Newhall
Last updated 06.10.99