ERUPT: A GRAPHICAL SIMULATION OF VOLCANIC ERUPTIONS FOR THE PERSONAL COMPUTER
(Copyright © 1991 UC)

(LA-UR-91-3205)

Kenneth H. Wohletz

Earth and Environmental Science Division

Los Alamos National Laboratory

Los Alamos, NM 87545

(see the ERUPT Page)

ABSTRACT

ERUPT is a user interactive computer program designed to simulate a wide range of volcanic activity and display it two-dimensionally. Written to take advantage of Microsoft Windows^TM 3.X, 95, and NT operating systems with multimedia, a DOS version is available requiring only EGA/VGA screen capability. The program involves numerical solution of basic physical laws of motion to reproduce a variety of eruption phenomena including: (1) Vulcanian/Plinian pyroclastic surge and flow; (2) Strombolian scoria fall; (3) mafic lava flows; (4) silicic lava dome emplacement; (5) Plinian pumice fall; and (6) sector collapse. The graphical display shows the temporal and spatial evolution of particle paths for lava and pyroclasts, their deposition, and the construction of a volcanic edifice in two dimensions. In addition, structural modifications to the simulated volcanic stratigraphy can be added by caldera collapse and normal faulting options, as well as dormant periods of erosion and redeposition. The program can be either run in an auto mode for semi-random eruption evolution or an interactive mode for operator controlled specification of eruption evolution. The basic physics involved includes those of simplified Newtonian/Bingham flow of lava with an arbitrary yield strength for silicic lavas, modified ballistic equations including the effects of turbulence and buoyancy for Strombolian, Plinian, and flow/surge eruptions, and an energy line approximation for conservation of kinetic energy, during runout of pyroclastic flows and surges. With its simple user interface, this program is aimed at students of volcanology as a teaching aid. Research applications include reproduction of the structure and stratigraphy of a known volcanic field to understand its physical volcanologic history and demonstration of various eruptive events at known volcanoes to illustrate volcanic hazards for nontechnical audiences.

INTRODUCTION

Personal computers have become an integral part of geological data representation, modeling, and interpretation. Within the realm of volcanology, many research projects have benefited from computer applications; for example, those of petrological data interpretation and plotting (Carr, 1987), tephra stratigraphic correlation and interpretation (Wohletz and Sheridan, 1979), solution to analytical expressions of Plinian eruption column physics (Woods, 1988), topographic interaction of pyroclastic flows and surges (Malin and Sheridan, 1982; Wadge and Isaacs, 1988), and rheological models of lava flows and domes (Iverson, 1989; Ishihara et al., 1989; Baloga, 1987). Such programs are proving very timely for workers in the ever more quantitative aspect of volcanological research.

Within the realm of physical volcanology, computer simulations can be of tremendous help in evaluating potential volcanic hazards as well as deducing those observed through deposit mapping. Graphical simulations also can greatly simplify the illustration of complex eruption phenomena when communication of volcanological concepts to lay people is required. In general there is a need for teaching aids in illustrating and introducing some basic concepts of eruptive mechanisms as well as the evolution of volcanic edifices of several classical types, including shield volcanoes, silicic domes, calderas, composite volcanoes, tuff cones and rings, and scoria cones, to name but a few. In recognition of such a need, Dehn (1987) developed a parametric model for cinder cone growth, which has in turn inspired the development of the initial algorithm for ERUPT.

Wohletz and Valentine (1990) and Dobran et al. (1993) described results of super computer models of explosive eruptions. Although these models give very precise solutions to the governing equations of the eruptions and allow development of detailed video representations, the complexity of the codes and requirements of very sophisticated computer hardware makes their general application by students and researchers difficult. The intent of the computer program ERUPT, described here, is to offer a streamlined and simplified eruption simulation program that allows students and researchers to easily experiment with various eruptive mechanisms and their lava/tephra emplacement mechanisms to reconstruct typical volcanic landforms.

The computational method of ERUPT is briefly described below, followed by some examples that illustrate its capabilities. Finally we discuss some of the code's possible applications as well as its current limitations. With continued use, we plan for many potential modifications and improvements.

PROGRAM DESCRIPTION

ERUPT graphically portrays four types of explosive eruptive activity (pyroclastic surge and flow, ballistic scoria ejection, pumice fall, and sector collapse) and two types of effusive activity (lava flow and lava dome). Each type has a set of associated parameters such as vent location, wind speed and direction, and relative strength of the eruption. The screen displays a dike intrusion and then the program calls subroutines that simulate the selected eruption. Eruptive episodes portray repeated bursts of lava or pyroclasts by tracing the movement of representative parcels on the screen. Simulated volcanic sounds accompany each eruptive type. At the end of each eruptive burst the various products are displayed as a stratigraphic horizon with a distinct color. Because arrays are written to record the thicknesses and type of emplaced products, the screen can be refreshed at any time and the user can save the results for future simulations.

ERUPT is fully interactive allowing menu oriented selection of eruption parameters and their editing at any point in the simulation. In addition, a fully automatic mode selects and modifies eruption parameters to show evolution of hypothetical volcanoes and volcanic fields with time. The automatic mode is semi-random while following predetermined evolutionary trends such as those of a composite cone, silicic volcanic field development, and shield volcanoes.

In interactive mode ERUPT prompts the user to choose one of the five different eruption types and its associated parameters for each episode. The duration of the eruptive episode is controlled by the user who can terminate it at any time. This allows the user the option to introduce a tectonic or erosional event, select a new eruptive type, product, or moving into automatic mode.

The automatic mode makes selections of eruptive types and parameters as well as the duration of the eruptive episode. The selections are semi-random, based upon conditional probability of one type of eruption following another.

ERUPT is a modular code written in VisualBASIC^TM, involving 30 subroutines that handle the setup of the simulation, screen representation, and eruptive modes. A simplified flow chart shows the general design of the code (Fig. 1). Compiled versions of the code can easily be run on personal computers with numeric coprocessors; the simulation speed is automatically set to accommodate the higher clock speeds of newer processors. Scaling of eruption simulations is based upon a screen dimensions of 1280x640 pixels, where each pixel by default represents 20 m (vertical and horizontal scales are equal), such that the screen width is equivalent to 24.6 km. The velocities of eruption products are scaled to the screen, with time stepping iterated at artificially high speeds in order shorten simulation times. Because of the range in spatial dimensions of eruptions, (for example between weak Strombolian bursts and caldera forming ignimbrite eruptions) the graphical scale is arbitrary to some degree. While the DOS-version memory limits the number of eruptive phases (changes) that can be simulated (40); up to 500 separate stratigraphic units and vent locations can be displayed in the Windows version. When memory limits are reached, only the static eruption mode is available, which displays parcel trajectories but does not add them to the stratigraphic array.

The modular form employed in ERUPT has developed with the concept of making code modifications simpler. Such modifications are desirable where ERUPT is taken from the teaching applications to research problems, requiring special circumstances of modeling. The physics used by ERUPT are greatly simplified, but retain enough detail to make graphical illustration of eruptions that look real as well as display observed spatial parameters.

General Formulation of Eruption Physics

Pyroclastic Eruptions. A kinematic algorithm is used to display the trajectories of pyroclasts for explosive eruptions. Graphic parcels, representing groups of pyroclasts, are assigned ballistic trajectories with initial conditions selected as a random function of the relative eruptive strength. The number of representative parcels is user designated with a default set proportionally to the eruptive strength chosen. A modified parametric form of the ballistic equation is then used to map out the temporal evolution of each parcel's flight path:

x(i) = [v_x + v_f(i)]t + v_wt , (1)

z(i) = [v_z + v_f(i)]t + (g/2) t² , (2)

where x(i) and z(i) denote incremental lateral and vertical position, vx and vz are the vertical and horizontal velocity components, g is the gravitational acceleration (which can be set to simulate various planetary conditions), and t is time. The magnitude of the lateral and vertical velocity components is scaled by user-specified strength with trajectory angles randomly computed to be between about 45° and vertical. These simple ballistic trajectories may be altered by wind [velocity denoted by vw in Eq.(1)]. For the case of Plinian eruption columns (pumice falls) and pyroclastic surges/flows (collapsing eruption columns and directed blasts) a turbulent component is simulated by addition of incremental random fluctuating velocities [vf(i)] in Eqs. (1) and (2).

The runout distance of pyroclastic flows and surges follow the energy line formulation of Malin and Sheridan (1982):

a(i) = g(sinb - mcosb) , (3)

where a(i) denotes incremental acceleration determined by gravity (g), the local slope of the substrate (b), and the Heim coefficient (m), which is the tangent of the energy line slope. For these eruptions, a simple complementary function of the Heim constant determines the relative eruptive strength, based upon the observations of Sheridan (1979) who shows that large volume pyroclastic eruptions are generally associated with smaller Heim coefficients.

Effusive Eruptions. Simulation of lava flows and lava dome extrusions are based on Newtonian movement of lava parcels over existing topography. Differences in viscosity are simulated by the graphic parcel thickness and flow velocity. These variables produce characteristic long thin lava flows and short stubby domes in a fashion similar to that described by Ishihara et al. (1990). The direction of flow front movement is controlled by surface slope. The flow surface shape of lava is nearly flat whereas the dome surface is convex upward.

Stratigraphic and Structural Representation

Each parcel tracked in explosive eruptive types is assigned a depositional volume as a function the eruption strength parameters. The location where the parcel strikes the substrate is the location where a given volume of tephra will be added to the stratigraphic array. For effusive eruptions, the lava volume of each eruptive episode is a function of the parcel thickness and length of the flow. Because of angle of repose limitations, tephra deposits are smoothed after deposition: scoria accumulations are avalanched to reach steepness of no more than about 35°. The deposit gradient for pyroclastic flows and surges is a function of eruption strength parameter [Heim coefficient; Eq. (3)]; low values of the coefficient result in very low bedding surface slopes whereas high values can retain steep bedding surfaces, which reflects tephra cohesion.

Normal faulting is a morphological control of volcanic fields and edifices. The faulting is either caused by regional tectonic movement or by local readjustment of a volcanic edifice. An example is slumping of the flank of a shield or composite volcano or development of a rift structure. Normal faulting is achieved by user or automatic specification of the location and magnitude of a fault (the amount of dip is randomly selected). The head wall side is then shifted down by the specified amount of magnitude by subtracting y-location pixels from each stratigraphic unit represented.

Along with deposition and avalanching of tephra, explosive eruptions excavate a crater whose width and depth is again a function of the strength or Heim coefficient (Malin and Sheridan, 1982)). Pyroclastic flow/surge eruptions are modeled to produce wider craters than do scoria eruptions. Caldera collapse is modeled as a pair of normal faults centered around the last active pyroclastic flow/surge or pumice fall vent. This option is available for pyroclastic flow/surge eruptions whose Heim coefficient is <0.40 or for pumice fall eruptions of strength >0.5. These option assignments are base upon the general observation that caldera collapse is generally associated with larger-scale eruptions of relatively greater surge/flow mobility or pumice dispersal. Two types of caldera collapse can be selected: one where the crater excavation is not enhanced and another where a large portion of the volcano is removed (called the Crater Lake type). The amount of collapse is proportional to the number of eruptive episodes (duration) and relative strength.

Erosion Model

Erosion is based upon a numerical form of a diffusion equation (Harbaugh and Bonham-Carter, 1970; Pollack, 1969). The temporal change in elevation by erosion or sedimentation is equal to the spatial derivative of an erodibility constant times the topographic gradient:

dz/dt = (d/dx) (K dz/dx) = K d²z/dx² + (dK/dx)(dz/dx) , (4)

where x and z represent lateral and vertical position, respectively. The erodibility (K) is a product of the stratigraphic erosional resistance coefficient and elevation, both of which in turn are functions of x. A spatial averaging technique is applied to transform the diffusion equation into a numerical form and through iterations across the preexisting topography, higher points are eroded and deposition occurs at lower points with erosional preference applied to topography that is high, convex upwards, and/or has a relatively higher erodibility constant. The erosion continues to process each topographic location until the operator halts the function or it reaches the predetermined number of erosional iterations specified by automatic mode operation.

EXAMPLES

The following examples illustrate the main eruptive types while showing the sequential development of a volcanic field stratigraphy. ERUPT designates pumice eruptions as Plinian, scoria eruptions are synonymous with Strombolian, and pyroclastic flow/surge eruptions are referred to as Ignimbrite/surge (MacDonald, 1972).

Plinian Eruption

Our pumice eruption simulates a Plinian event (Walker, 1981) that sustains a vertical column of pumice up to 15 km above the vent. At some height, depending upon the relative eruption strength, the column reaches neutral buoyancy and its upward migration is halted while lateral dispersion begins forming what is called an umbrella cloud. In Fig. 2, a second pumice eruptive phase from a vent located at km 11 models a Plinian column of strength 0.80 which reaches a height of about 10 km. Note the lateral shearing of the umbrella cloud by the wind of 25 m/s directed toward the right. The fallout deposit of pumice shown by the circles thins exponentially away from the vent.

This eruptive behavior is simulated by vertically directed ballistic trajectories that have superimposed turbulent fluctuations representing the support of the parcels in a buoyantly rising medium. The maximum simulated heights of parcel trajectories is arbitrarily limited by assuming a neutral buoyancy height at which point the parcel's lateral velocity component is displayed. At times when the theoretical ballistic trajectories of the parcels fall below the neutral buoyancy height, the parcels are shown to fall out of the umbrella region along a ballistic path to the ground where they are deposited.

Strombolian Eruption

This simple form of pyroclastic eruption type is based upon the parametric ballistic equation already described; technical description of the eruption physics can be found in McGetchin et al. (1974). Each particle is tracked through time until it intersects a topographic surface or moves out of the field of view. The parcels are traced by small colored circles on the screen; the color of each parcel changes with time to simulate chilling from red-hot lava to dark colored scoria clasts. The strength of the scoria eruption determines the magnitude of the randomly assigned velocity vectors for each parcel. With solution of the parametric ballistic equation, a variable wind factor and gravitational constant modify the appearance of the erupted fountain. Where parcels y-position intersect the topographic surface, deposition occurs subject to later downslope avalanching.

Eruptive phase 9 (Fig. 3) shows a Strombolian eruption and scoria cone at kilometer 21.0. This cone is superimposed on the flanks of the previous scoria cone centered at kilometer 18.0, which rests upon the previously described Plinian deposits. The strength of the eruption is 1.00, which produces a ballistic fountain about 2.5 km high. The eruption has produced a scoria cone with a base diameter of about 4 km and a height of 1.5 km with a large central crater.

Pyroclastic Flow/Surge

Pyroclastic flows or Ignimbrites(Sparks and Wilson, 1976 ) and pyroclastic surges (Wohletz and Sheridan, 1979) are laterally moving density currents of tephra and gases formed during several different eruption types, most notably including Plinian, Peléean, and Vulcanian (MacDonald, 1972). This type of simulation is accomplished by superposition of turbulence on the parcel's ballistic trajectory and conversion of the parcel's potential energy to laterally directed kinetic energy of translation. Ballistic trajectories for each representative particle are randomly generated. Turbulence is introduced by adding a random fluctuation to the trajectories. The ultimate runout of the surge/flow is governed by the energy line concept described by Eq. (3). The initial potential energy of each particle is a function of the height of its ballistic trajectory. This height is converted into a runout distance as a function of the Heim coefficient, which predicts the energy line, a hypothetical line that connects the highest point of origin of the flow/surge with the most distal point of its runout. For example, pyroclastic flows from composite cones display typically high Heim coefficients (>0.4), but caldera-related pyroclastic flows can be highly mobile with runout distances more than 20 times greater than the vertical distance they traverse (Heim coefficient < 0.05). In solving the lateral movement of representative parcels, the parcel's acceleration or deceleration is iteratively calculated by the vertical distance between the parcel and the topography directly below it. If this distance is less than that of the previous iterative position, then deceleration occurs; if the topography is dipping outward from the vent with a slope greater than that of the energy line, then the parcel will accelerate (see Malin and Sheridan, 1982).

Fig. 4 shows a ignimbrite/surge eruption on the flanks of the previous scoria cone at kilometer 20.0. The Heim coefficient is 0.3 and with no wind the total runout is about 15 km. Note that the deposit, shown in gray, is thicker in topographically lower areas but gradually thins away from the source. The eruption has excavated a bowl shaped crater.

Lava Flows and Domes

Effusive activity is simulated by Newtonian fluids that move down slope under gravity (Hulme, 1974; Fink, 1983). The volume of each eruptive episode is randomly determined, with large volume flows having a greater likelihood of producing a longer flow. Lateral movement of the flow is dictated by the difference in height of the flow top and adjacent topographic elements. A Bingham-like yield stress (Hulme, 1974) is simulated for viscous dome lavas by an increased thickness of flow parcels and higher viscosity is represented by a slower translation rate of those parcels. Downslope movement will proceed as long as lava parcels continue to be extruded. Where a flow meets a topographic obstacle, the ability of the flow to overtop that obstacle is determined by the relative height of the obstacle to the lava's upstream height. Because lavas become more viscous with cooling, the effective hydraulic head of the lava is not necessarily the difference in height between its vent and flow front. For long flows the effective head is determined by the difference in height of the flow front and some effective source location upstream. With these methods of rheological simulation, lava flows tend to pond in topographically low areas while dome lavas will pile up into mounds.

A lava vent shown in Figure 5 at kilometer 17.0 has produced a shield volcano on the flanks of a previous lava flow that ponded in the crater of the pyroclastic surge/flow eruption discussed earlier (Fig. 4). The shield has a collapse crater. In this program collapse craters form when large volumes of lava have been erupted. In contrast, Fig. 6 shows the extrusion of a viscous lava dome at kilometer 11.0 on the flanks of a previously erupted dome at kilometer 10.0.

Erosion, Faulting, and Caldera Collapse

Figure 7 (phase 18) shows the dormant eruptive phase 10 where their has been a normal faulting event followed by a period of erosion upon the earlier developed volcanic stratigraphy of Fig. 6. The fault is located near kilometer 15.0 and is of moderate magnitude with the right-hand side down-dropped several hundred meters. As described earlier, the erosion model takes into account differential erosion based upon elevation and erosional resistance. Figure 7 illustrates the effects of erosion by the truncation of the lava shield; its top is smoothed preferentially with respect to the younger lava domes. Horizontal stripes portray the resulting sedimentary deposits that cover the flanks and bury the fault scarp.

In order to illustrate caldera collapse (Williams, 1941) a large composite volcano was generated on top of the previously eroded volcanic stratigraphy. The composite cone eruptive phases (Fig. 8) comprise a repeated sequence of lava flows and scoria eruption with dome extrusion in the crater formed by the last scoria eruption, and finally a erosional phase. Caldera collapse is simulated by a large summit pyroclastic flow/surge eruption (Heim coefficient = 0.2), and the resulting truncated cone with a summit caldera depression is shown in Fig. 9. A blanket of tephra has been deposited over all the previous stratigraphic units and the caldera (Crater Lake type) is some 4 km in diameter and about 2 km deep. Down-drop of the central portion of the volcano is displayed between a set of inwardly dipping faults about 4 km apart. Finally, a small scoria cone has been erupted in the caldera (Fig. 9).

Volcanic Sector Collapse

Since the 1980 eruptions of Mount St. Helens, sector collapse has been increasingly recognized as an important type of eruptive activity especially displayed during the evolution of composite cones and domes (Siebert, 1984). Sector collapse results from gravitational instability of rocks comprising the slopes of the volcanic edifice as a result of a variety of geological processes including erosive oversteepening of the slopes, progressive fumarolic alteration and weakening of the slopes, and intrusive displacement of the edifice. These processes and others cause a portion (sector) of the volcano to break loose and avalanche downslope producing debris flows that extent several or more km away from the base of the volcano. ERUPT simulates this activity by a combination of low-angle faulting and erosion that produce both a dust cloud and debris avalanche deposit as a portion of the volcanic edifice is removed (Fig. 10). The user specifies the position of the collapse head scarp thus determining its magnitude. We note that some unusual results can occur by poor choice of the location of a sector collapse and advise caution in its application.

DISCUSSION

ERUPT is a versatile program for the personal computer that allows the visualization of several common volcanic eruptive processes. The program provides a level of detail sufficient for researchers in volcanology but simple enough for casual users to appreciate. To be able to realistically portray volcanic processes on the personal computer screen, several limitations or simplifications were utilized in the algorithms. For example, the spatial and temporal scales are approximate for the sake of ease in visualization. Also, the numerical representation of volcano physics has been greatly simplified such that simulations may be considered only semiquantitative.

One advantage of the program is the ease by which a user can interact to modify or test various sequences of events. This process can lead to a better understanding the physical processes involved and can help to make accurate reconstructions of complex volcanic structures.

An important use of the program is as a teaching aid. Because of its simplicity it can be operated at nearly all grade levels from elementary school through graduate school. At the most sophisticated levels the program can help researchers to reconstruct the most complex volcanic structures and to understand the evolution of a single volcano or a volcanic field.

ERUPT can also be used in simulation of volcanic hazards. By accurately reconstructing the form and stratigraphy of a potentially dangerous volcano, the user can animate future eruptive activity, which is useful for illustrating hazards to non-technical audiences. This feature is very important in areas where the menace of volcanism is poorly known.

ACKNOWLEDGMENTS

This work done under the auspices of the U. S. Department of Energy with support by Los Alamos National Laboratory's Laboratory Directed Research and Development funds.

Software distribution of ERUPT is planned to be handled by RockWare Scientific Software, who now distribute a shareware version. Please send requests to:

RockWare, Inc.

2221 East St.

Golden, CO 80401 USA

(303) 278-3534

or use your web browser to download ERUPT at www.rockware.com

REFERENCES

Carr, M., 1987. IGPET2. Copyright, M. Carr, 25 Moran Ave., Princeton, New Jersey 08542.

Dehn, J., 1987. Model of cinder cone formation. M.S. thesis, Arizona State University, Tempe, Arizona, 85 pp.

Dobran, F., Neri, A., and Macedonio, G., 1993. Numerical simulation of collapsing volcanic columns, J. Geophys. Res., 98: 4231-4259.

Fink, J. H., 1980. Structure and emplacement of a rhyolitic obsidian flow: Little Glass Mountain rhyolitic obsidian flows, northern California. Tectonophys., 66: 147-166.

Harbaugh, J. W. and Bonham-Carter, G., 1970. Computer Simulation in Geology. Wiley-Interscience, New York, 575 pp.

Hulme, G., 1974. The interpretation of lava flow morphology. Jour. Roy. astron. Soc., 39: 361-383.

Ishihara, K., Iguchi, M., and Kamo, K., 1989. Numerical simulation of lava flows on some volcanoes in Japan. In: J. Fink (Editor), Lava Flows and Domes. IAVCEI Proceedings in Volcanology, Vol. 2, Springer-Verlag, Berlin, 174-207.

MacDonald, G., 1972. Volcanoes. Prentice-Hall, Inc., Englewood Cliffs, New Jersey, 510 pp.

Malin, M. C. and Sheridan, M. F., 1982. Computer-assisted mapping of pyroclastic surges. Science, 217: 637-639.

McGetchin, T. R., Settle, M., and Chouet, B., 1974. Cinder cone growth modeled after Northeast Crater, Mount Etna, Sicily. Jour. Geohpys. Res., 79: 3257-3272.

Pollack, H. N., 1969. A numerical model of the Grand Canyon. In: D. C. Baars (Editor), Geology and Natural History of the Grand Canyon Region, Four Corners Geol. Soc. Guidebook to Fifth Field Conf., 61-62.

Sheridan, M. F., 1979. Emplacement of pyroclastic flows: A review. In: C. E. Chapin and W. E. Elston (Editors), Ash Flow Tuffs. Geol. Soc. Amer. Spec. Pap., 180: 125-136.

Siebert, L., 1984. Large volcanic debris avalanches: characteristics of source areas, deposits, and associated eruptions. J. Volcanol. Geotherm. Res., 22: 163-197.

Sparks, R. S. J. and Wilson, L., 1976. A model for the formation of ignimbrite by gravitational column collapse. Jour. Geol. Soc. London, 132: 441-451.

Wadge, G., 1988. The potential of GIS modeling of gravity flows and slope instabilities. Internat. Jour. Geograph. Inform. Sys., 2: 143-152.

Wadge, G. and Isaacs, M. C., 1988. Mapping the volcanic hazards from Soufriere Hills Volcano, Montserrat, West Indies using an image processor. Jour. Geol. Soc. London, 145: 541-551.

Walker, G. P. L., 1981. Plinian eruptions and their products. Bull. Volcanol., 44: 223-240.

Williams, H., 1941. Calderas and their origin. California Univ. Pubs. Geol. Sci., 25(6): 239-346.

Wohletz, K. H. and Sheridan, M. F., 1979. A model of pyroclastic surge. In: C. E. Chapin and W.E. Elston (Editors), Ash Flow Tuffs. Geol. Soc. Amer. Spec. Pap., 180: 177-194.

Wohletz, K. H. and Valentine, G. A. 1990. Computer simulations of explosive volcanic eruptions. In: M. P. Ryan (Editor), Magma Transport and Storage. John Wiley & Sons, New York, 113-136.

Woods, A. W., 1988. The fluid dynamics and thermodynamics of eruption columns. Bull. Volcanol., 50: 169-193.

FIGURES

(note: Images from DOS version do not show full resolution)

Fig. 1. Simplified flow chart of ERUPT.BAS (version 1.0) with only the primary subroutines shown. Subroutine ERUPT loops recursively until up to 40 eruptive stratigraphic units have been recorded. After dike intrusion, one of four eruptive-type subroutines is called; pyroclastic types require trajectory velocity calculation, deposit emplacement, cratering, and if chosen, caldera collapse; lava types call a flow adjust subroutine instead of deposit, crater, and collapse ones. The eruptive type subroutine loops through as many bursts as desired or predetermined by auto-mode selection. After completion of eruptive-type bursts, stratigraphic thicknesses are determined and then added to the graphic screen before dike cooling occurs. The ERUPT subroutine is completed by a set of CHANGE subroutines that give the user options for faulting, eroding, screen refreshing, choice of a new eruptive type and it vent position. With establishment of a new eruptive center and screen labels, CHANGE loops back to the beginning of ERUPT.

Fig. 2. DOS screen representation of a Plinian pumice eruption at kilometer 11.0. Note the column rising about 8 km before spreading laterally with a simulated wind of 25 km/hr blowing to the right. Fallout of pumice occurs on previously deposited layers represented by blue colors.

Fig. 3. A Strombolian eruption at kilometer 21.0 produces a ballistic fountain in phase-4 eruptive bursts. The scoria layers (green) show phase 9 deposits onlapping the older phase 8 scoria cone.

Fig. 4. Pyroclastic flow/surge eruption (Ignimbrite/surge) of phase 12 occurs in an evolving crater at kilometer 20.0. The eruptive column rises about 10 km above the vent and collapses to cause runout of density currents over 10 km outward from the vent. The deposits of this eruptive phase are shown in gray. Note the crater excavation in the phase 9 scoria cone.

Fig. 5. Phase 13 through 15 eruptions are basalt-like lava flows (red and purple) that have built a lava shield centered near kilometer 17.0. Because these lava eruptions continued sufficiently long, crater collapse is shown above the vent areas.

Fig. 6. Lava domes at kilometers 10.0 and 11.0 have resulted from phase 16 and 17 eruptions. The dome lavas are shown as brown colors. Note the relatively steep-sloped flanks of these domes compared to those of the lava flows of previous phases.

Fig. 7. Phase 18 is a dormant period marked by formation of a normal fault near kilometer 15.0 (note downdrop of pyroclastic layers to the right of the fault). In addition, this dormant phase shows an erosional break marked by relatively greater truncation of the previous lava shield that is more erodible dome lavas. The sedimentary deposit resulting from the erosion break is shown as horizontal lines.

Fig. 8 Eruptive phases 19 through 33 have produced a large composite cone consisting of alternating layers of lava and scoria with a small summit dome.

Fig. 9. Caldera collapse (phase 37) has truncated the previously formed composite cone, resulting in a Crater-Lake-type caldera of about 4 km diameter and 2 km deep. The blanketing pyroclastic deposit from the caldera eruption (gray) is not shown within the caldera; however, if the Crater-Lake-type option were not chosen, the pyroclastic deposit would be thickest inside the caldera, but the topographic expression of the caldera would not be so pronounced. Note the inwardly dipping caldera faults along the caldera walls that cause downward displacement of precaldera units under the caldera.

Fig. 10. Volcanic sector collapse of the previously formed composite cone truncates the left-hand side of the cone producing a debris avalanche (white) that covers downslope topograph.