Christine Burrill November 19, 2012 AA Flow Dynamics, Mauna Loa 1984 by Peter W. Lipman & Norman G. Banks

I. Introduction
The paper on ‘a’a flow dynamics written by Lipman and Banks is one of the first comprehensive studies regarding detailed observations of a long and complex ‘a’a lava flow. This paper was also one of the first to establish connections between lava flow morphology and changes in lava properties, specifically gas and crystal content. The flow that was studied was generated by the 1984 eruption along the northeast rift zone of Mauna Loa volcano, Hawaii (fig. 1). The major flow reached 27 km from the source vent after four days from the start of the eruption (Lipman and Banks, 1987). A decrease in the eruption rate, related to an increase in microphenocrysts and decrease in gas content of erupted lava, generated lava boats that blocked the flow. Blockages led to lava overflowing channel levees, redirecting the lava supply and leading to a stagnation of the major flow. The purpose of this paper is to summarize the key points from the study of the 1984 eruption, investigate some of the impacts this study has had on subsequent research and give a critical evaluation of the study.
II. Major Points of Paper
Flow Types
Lipman and Banks (1987) observed distinctive types of ‘a’a that developed at certain distances from the vent, with certain changes in slope of the terrain and the way the flow advanced. The four types of ‘a’a are transitional or slabby, scoriaceous, clinkery and blocky. The different types of ‘a’a are also a result in changes of magma rheology, and flow rates. Significant pahoehoe flows were, at first, restricted in extent to within 5 km of the vent (Lipman and Banks, 1987). Most pahoehoe flows were much less than 5 km and were therefore not the main focus of the paper. In well-established channels, the lava advanced as ‘a’a which later permitted the appearance of pahoehoe flows as overflows when blockages of the lava channels downstream caused the lava level to rise above or break through the levees. However, these pahoehoe overflows quickly transformed to ‘a’a.
Transitional or slabby ‘a’a was recognized when the flow contained solid and semisolid slabs and blocks that disrupted flow streamlines in the channel. Observations of this transition at flow 1 were between 3 and 5 km from the vent. Scoriaceous lava was observed at 5 to 12 km from the vent as lava became more viscous and dense, resulting in a frothy red-brown oxidized scoria with a thin glassy surface (Lipman and Banks, 1987). Farther downstream, 12-15 km from the vent, fluid lava was mostly obscured by clinkery ‘a’a that was carried along on top of the flow. Clinkery ‘a’a was distinctive from scoriaceous ‘a’a by its angular surfaces, darker color, lesser degree of oxidation and vesicularity (Lipman and Banks, 1987). Near the flow front, about 20 km from the vent, the lava moved as a mass of gray rubble called blocky ‘a’a with a highly viscous to partly solidified core. As the eruption progressed and the flow progressively stagnated upstream, these distinctive areas of ‘a’a were blurred and mixed, especially at the flow front where the different types of ‘a’a were carried and deposited.
‘A’A Flowage Zones, Levees and Channel Evolution
The major flow (flow 1) was divided into four zones by the authors based on systematic changes that were observed in the structure of the lava channel: a stable channel, transitional channel, dispersive flow and a flow toe or flow front (fig. 2). These zones became established relatively quickly after the eruption began and later changed in length and character as eruption rate decreased and lava properties changed. The major flow changed from a simple, narrow channel to a complex and more widespread flow.
Characteristics of the flow toe were found to vary primarily with the slope of the terrain and distance from the vent. Close to the vent on steep slopes channels were well established and lava moved quickly to the flow toe. The flow toe advanced rapidly and contained scoriaceous and slabby ‘a’a with a near flat surface. These quickly moving flow fronts were also characterized by relatively low heights of 1-3 m compared to slow moving flow fronts traveling on gentle slopes further from the vent that allowed material to pile up more vertically at the flow front. Denser blockier lava and fine granular material from the grinding of the lava blocks against each other were major components of the slower flows (Lipman and Banks, 1987).
The dispersive flow zone immediately behind the flow toe had a quickly moving central flow of blocky ‘a’a with little to no incandescent fluid lava. Closer to the vent more incandescent fluid lava was concentrated in a central channel. The dispersive zone was very limited in extent until the eruption began to wane and later extended back up to 10 km from the flow toe (Lipman and Banks, 1987).
The transitional zone was distinguished by a distinct channel and clinkery ‘a’a. The channel walls consisted of blocky ‘a’a and clinkery ‘a’a that could deform and move. For the first week of the eruption the transition zone was broad but became shorter in extent during the later half of the eruption as the dispersive zone grew in extent (Lipman and Banks, 1987). Behind the transitional zone was the stable channel that had a narrower active channel than the transitional zone and formed when the lens adjacent to the lava channel cooled and would no longer deform. Over the course of the eruption the upper part of the transitional zone transformed into a stabilized channel but due to the decrease in eruption rate, lava was unable to make it to the lower reaches of the flow any more efficiently (Lipman and Banks, 1987).
Lava boats, created by the caving of unstable walls and the straightening of meanders, played an important role in the changes of lava channel morphology and efficiency. Previously these large, rounded masses had been interpreted to form from accumulating layers of lava as pieces of solidified lava rolled down the flow. Lava boats obstructed channels, cutting off the supply of lava down channel and cause overflows of the lava behind the dam, which redirected the lava along new paths. Lava boats were more common in the later half of the eruption as eruption rates decreased and the lava level in the channels dropped, reducing support for channel walls. Lava boats were used to approximate flow velocities and estimate channel depth. It was uncommon to find lava boats within 5 km of the vent (Lipman and Banks, 1987).
On March 29, the first significant blockage of the channel occurred, diverting the lava from flow 1 to form flow 1A . On April 5, a second large blockage and overflow resulted in the creation of flow 1B (fig. 1). After March 29 blockages became larger and smaller blockages were more frequent. They also began to form higher up towards the vent turning the simple, stable channel zone into a more complex feature (Lipman and Banks, 1987). The authors observed a dynamic and simple feedback relationship between the lava boats and blockages; lava boats created upstream would cause a blockage that reduced lava level farther downstream creating more lava boats downstream.
The lava behind blockages sometimes overflowed the channel banks as the level of the lava rose or sometimes broke through the blockage and created a surge of lava down the pre-existing channel. Small obstructions in the channel would begin to deform and move downchannel and the lava would eventually break through a low or weak section. Sometimes this would create overflows because the preexisting channel would not be able to contain all the lava surging out from behind the blockage. Much of the volume of the eruption was lost to overflows, leading to the stagnation of the major flow.
Levees, the sidewalls of a lava channel, were found to be closely tied to channel evolution and yield strength of the lava as previously concluded by Hulme (1974). Four major types of levees were observed during the 1984 flow: early levees including lateral and nested lateral levees, overflow levees, deformation levees and accreted levees. Accretionary, overflow, rubble and initial levees were types recognized during a study of lava flows on Mount Etna, Sicily in 1975 [Sparks et al., 1976] and minor differences were discussed.
Levees formed early and were complex because they changed as conditions of the channel changed. As the eruption rate decreased and flow in the channel became narrower, a second set of levees would be created within the first levees. Overflows, surges and gradual deformation would also modify the levees but not until later in the eruption. These features, according to the authors, may be useful for determining whether or not prehistoric lava flows were long-lived. Levees tended to be broader than the channels and had high points several meters from the active channel and could indicate the size and movement history of prehistoric flows.
Relationship Between Eruption Rate, Lava Properties and Lava Channel Morphology
Walker (1973) proposed that eruption rate was the most important factor determining the length of a lava flow. This idea countered the long held belief that lava flow length was governed by viscosity. Lava temperature, gas and crystal content, and density were key in Lipman and Banks (1987) interpretation of the changes observed in lava channels.
Temperatures were measured with a thermocouple inserted into fluid lava and remotely with HOTSHOT 2-color infrared pyrometer aimed at active fountains or any fluid lava within a channel (Lipman and Banks, 1987). Near the vents at 2,850 m the lava was 1,140 3 C throughout the eruption and temperatures, as far as 10 km away from the vent, measured 1,135 5 C, showing little change in temperature over distance (Lipman and Banks, 1987). Densities were measure by taking samples of lava and quenching them. Densities of lava in the channels increased down channel and spatter material erupted at the vent decreased in density gradually over the course of the eruption from 1.0-1.2 g/cm3 to 0.5 g/cm3 (Lipman and Banks, 1987). The change in density down channel was attributed to degassing which was evident by high SO2 levels that made field observations impossible downwind of the channel. As spatter density decreased, Lipman and Banks noticed the height of lava fountains increased and channel level dropped. They hypothesized that the drop in channel level was caused by a decrease in bulk eruption rate or that more gas may have been lost during high fountaining and resulted in denser lava in the channels.
Observations and measurements made on the 1984 Mauna Loa flow had important implications for a well established relationship between the pahoehoe-’a’a transition and cooling of lava in open channels. Temperatures changed little over 10 km and thought to have been partly sustained by heat of crystallization and frictional deformation. The remaining fluid lava in the lower parts of the channel eventually transitioned to ‘a’a but the slope and flow velocities were low. The remaining explanation was the observed increase in microlite content and solid debris that led to an increase in viscosity. However, the authors knew a decrease in vesicularity had been shown experimentally to increase viscosity and could be the case with the 1984 Mauna Loa flows. A decrease in gas and vesicles would cause an increase in density and increased thickness of vesicle walls. Thicker vesicle walls would be more resistant to any deformation whereas thin vesicle walls may be responsible for the high fluidity of pahoehoe lava (Lipman and Banks, 1987). Lava tubes were present during the 1984 eruption and thought by Lipman and Banks to greatly delay the transition of pahoehoe to ‘a’a. Lava tubes could only be created when lava flow was low and steady or decreasing and could prevent the loss of gases from the lava. The slower moving lava would be more laminar, keeping vesicles from breaking and would maintain a lower bulk viscosity (Lipman and Banks, 1987).
Lava Crystallinity
Three types of crystal textures were present in the 1984 lava: olivine phenocrysts, dusty microlites and microphenocrysts of augite, plagioclase and olivine. Microlites were absent in vent spatter but became increasingly more abundant downchannel. The olivine phenocrysts did not vary significantly during the eruption and were interpreted by Lipman and Banks (1987) to be disequilibrium crystals present in the magma reservoir before eruption. The microphenocrysts on the other hand increased significantly as the eruption progressed. Microphenocrysts increased from 0.5% to 30% and simultaneously increased in size (Lipman and Banks, 1987). Most microphenocrysts had textures and shapes typical of rapid crystallization from undercooled magma. No changes were detected with the bulk major-oxide composition therefore Lipman and Banks (1987) believed the increase in microphenocrysts to be a result of a separation of a gas phase during the eruption instead of a decrease in magma temperatures.
The calculated liquidus temperature for most crystal phases at one atmosphere is at least 20-30C higher than the measured temperature of lava during the 1984 eruption. However, the measured temperatures do correlate with the liquidus that contains about 1% volatiles (Lipman and Banks, 1987). The pressurization induced by a higher volatile content would prevent the growth of crystals. After the 1984 eruption began, the loss of volatiles likely undercooled the magma and initiated the crystallization of microphenocrysts. The authors considered a zoned magma chamber scenario but due to the compositional uniformity of the lava and its textures, this option did not seem to be the best fit. The exact amount of gas loss necessary for to initiate crystallization was not known and could not be determined at the time. The increase in crystal content of the erupted lava was thought to be the cause of the observed decrease in eruption rate in addition to factors such as the release of elastic-strain energy from magma stored in the reservoir, the availability of magma from depth and conduit evolution from dike to plug geometry (Wadge, 1981). The increase in crystal content led to an increase in viscosity and was likely the cause of the end of the eruption and the upward migration of channel stagnation (Lipman and Banks, 1987).
III. Impact on Research Since Publication
The paper by Lipman and Banks (1987) has been cited over 100 times by papers that cover a variety of topics such as lava flow dynamics on other volcanoes and other planets, hazard assessment, monitoring and remote sensing of volcanoes and studies of lava properties and the transition from pahoehoe to ‘a’a. The paper has been cited by many studies specifically on Mount Etna, Sicily due to similar types of lava flows and the large population at risk from these flows (Wright, 2001).
A study by Cashman et al. (1999) used the qualitative and semi-quantitative observations from Lipman and Banks (1987) to determine rates of cooling, crystallization, and changes in surface morphology of a lava channel produced from Kılauea Volcano in May 1997 but could also be used at other volcanoes. Flow velocity profiles and depth measurements were necessary to approximate rates and could only be found at the time from two sources including the 1984 lava flow on Mauna Loa.
Temperature measurements of the 1984 lava flow have been used in studying lava flows remotely on Earth and other planets. A study by Wright et al. aimed to establish how the surface temperature distributions of terrestrial lavas vary as a function of eruption style and lava composition. On the Jovian moon Io, ‘a’a lava surface temperatures of between 1400 and 1700 °C were recorded during the NASA’s Galileo mission (Wright et al., 2011). During the 1984 eruption on Mauna Loa, no temperatures in excess of 1150°C were determined by any method (Lipman and Banks, 1987). Any individual temperature retrieved from any individual pixel on remote images was found unreliable as a proxy for lava eruption temperature and composition because of the nature of ‘a’a flows. For channel-fed ‘a‘a flows, motion of the fluid core causes cracks in the crust (Crisp and Baloga, 1990), and hot material is continually exposed to a spacecraft at some point along the flow. Now this can be accounted for and better temperatures estimates can be made when remotely observing lava flows.
IV. Critical Evaluation
I thought that Lipman and Banks’ explanation of the different zones of the lava channel and how they evolve over time was not well laid out. I think major points were weakened by the lack of a more logical structure. I had a hard time distinguishing what observations were important to the key ideas. I think the length of the paper could have been reduced by keeping the observations and interpretations separate and would have cut out some of the redundancy. I also feel that the paper would be more complete if it followed through with how all the different types of ‘a’a were distributed at the end of the eruption. Also lava tubes and pahoehoe, though present during this eruption, were not talked about as important parts of the lava flow but were mentioned in the interpretations. Though the relationship they were investigating between flow, degassing and viscosity does relate to lava tubes and pahoehoe, I thought the authors should have just focused on the ‘a’a.
I would have liked an identifiable background section and explanation of how the research was solving a particular problem .The paper seemed to stand alone without being in the context of other research. Some comparison is drawn between levees from previous research and the significance of the new information from this study was mentioned but only briefly. Other related research mentioned but the paper gave me the impression that much of the categorizing, diagramming and relationships they were observing was being done for the first time. Without more background, I think the impact of the paper is slightly lost to me. When I researched papers citing Lipman and Banks 1987 paper, the citations do not appear as often or as significantly as I expected within some papers.
V. Conclusions
The 1984 eruption on Mauna Loa produced one of the longest ‘a’a flows to be observed in real time. The key idea gained from this study is that the evolution of the lava flow is closely tied to changes in lava density, gas content, eruption rate and crystal content as well as topography and lava temperature and not simply just the viscosity or eruption rate as was previously thought. This paper highlights that lava flows are dynamic and complicated, but the length and other features of the lava flow can help provide some insight into certain aspects such as the eruption rate and lava properties. Based on observations of lava temperature, density, gas emissions, and crystal content and the relationship of these properties to lava flow morphology, it was thought monitoring these parameters would be important for hazard assessment and define conditions that produce to lava flow fields of predominantly pahoehoe or ‘a‘a.
The paper has helped build a foundation of qualitative and quantitative observations on relationships between lava properties and channel morphology that has been useful in subsequent research. However, the impact may have been different if there had been a more logical structure to the presentation of the study and its conclusions.
VI. Figures
Mauna Loa

Mauna Loa

Figure 1. Location of Hawaiian Islands and the 1984 lava flows along the Northeast Rift Zone (modified from Lipman and Banks, 1987; and USGS Hawaii Volcano Observatory website: http://hvo.wr.usgs.gov/maunaloa/hazards/vents.html)

Figure 2. Flowage zones from vent to toe comparing the different structural features of an ‘a’a channel (Lipman and Banks, 1987)

VII. References Cited
Cashman, K. V., C. Thornber, and J. P. Kauahikaua (1999), Cooling and crystallization of lava in open channels, and the transition of Pahoehoe Lava to 'A'(a)over-bar, Bulletin of Volcanology, 61(5), 306-323.

Crisp, J.A., and Baloga, S.M., 1990, A model for lava flows with two thermal components, Journal of Geophysical Research, 95, 1255–1270.

Hulme, G. (1974), The interpretation of lava flow morphology, Geophysical Journal of the Royal Astronomical Society, 39, 361 -383.

Lipman, P.W., and Banks, N.G. (1987), ‘A’a flow dynamics, Mauna Loa 1984, in Decker, R., et al., eds., Volcanism in Hawaii: U.S. Geological Survey Professional Paper 1350, 1527–1567.

Macdonald, G.A. (1953), Pahoehoe, ‘a’a, and block lava, American Journal of Science, 251, 169-191.

Sparks, R. S. J., H. Pinkerton, and G. Hulme (1976), Classification and Formation of Lava Levees on Mount Etna, Sicily, Geology, 4(5), 269-271.

Wadge, G. (1981), The variations of magma discharge during basaltic eruptions, Journal of Volcanology and Geothermal Research, 11, 139-168.

Walker, G.P.L. (1973), Length of lava flows, Philosophical Transactions of the Royal Society of London, A274, 646-656.
Wright, R., L. P. Flynn, and A. J. L. Harris (2001), Evolution of lava flow-fields at Mount Etna, 27-28 October 1999, observed by Landsat 7 ETM+, Bulletin of Volcanology, 63(1), 1-7.
Wright, R., L. Glaze, and S. M. Baloga (2011), Constraints on determining the eruption style and composition of terrestrial lavas from space, Geology, 39(12), 1127-1130.