Debris flow during intense rainfall in Snowdonia,North Wales: A preliminary survey

Intense rainfall after the abnormally dry and warm summer of 1983 triggered debris flows in mountainous terrain in North Wales. This preliminary investigation concentrates on a flow which blocked the A5, requiring £56,000 of remedial work. An estimated 118·4 mm of rain fell over steep, rocky catchments in 5 hrs (peak intensity 39·9 mm hr^?1) and water emerging from a rock chute mobilized colluvium on lower slopes, in which pore water pressure was probably already rising fast and bulk properties and other geotechnical conditions, including low shearing resistance, were favourable. Debris flowed in a narrow concave track 585 m long, x slope 27·8°, descending 282 m. The scoured channel, levées and debris lobes typical of documented flows elsewhere suggest that flow was rapid, turbulent, and pulsing.     

The internal structure of lava flows—insights from AMS measurements II: Hawaiian pahoehoe, toothpaste lava and 'a'

We studied the anisotropy of magnetic susceptibility (AMS) of 22 basaltic flow units, including S-type pahoehoe, P-type pahoehoe, toothpaste lava and 'a' emplaced over different slopes in two Hawaiian islands. Systematic differences occur in several aspects of AMS (mean susceptibility, degree of anisotropy, magnetic fabric and orientation of the principal susceptibilities) among the morphological types that can be related to different modes of lava emplacement. AMS also detects systematic changes in the rate of shear with position in a unit, allowing us to infer local flow direction and some other aspects of the velocity field of each unit. 'A' flows are subject to stronger deformation than pahoehoe, and also their internal parts behave more like a unit. According to AMS, the central part of pahoehoe commonly reveals a different deformation history than the upper and lower extremes, probably resulting from endogenous growth.     

Characteristics of intense rainfall over Gujarat State (India) based on percentile criteria

The study focuses on the spatial and temporal variations of intense/extreme rainfall events over Gujarat State (India) during the period 1970–2014. Average monsoon rainfall for the state shows a significant increasing trend, with an increase of 48 mm/decade. Some of the stations in the Saurashtra region show a statistically significant increasing trend but none of the stations in the state show a decreasing trend. The increasing trend in monsoon rainfall is very significant for the past three decades, with an increase of 167 mm/decade. Instead of fixed absolute threshold values, relative threshold values of rainfall corresponding to the 95th, 98th, 99th and 99.5th percentiles for each station have been proposed to represent heavy, very heavy, intense and extreme rainfall, which varied between 70–120, 105–160, 130–210 and 165–280 mm, respectively. Significant increasing trends are observed for the frequency of heavy and very heavy rainfall events over the state.     

Time series analysis of lava flux

We have applied time series analytical techniques to the flux of lava from an extrusive eruption. Tilt data acting as a proxy for flux are used in a case study of the May–August 1997 period of the eruption at Soufrière Hills Volcano, Montserrat. We justify the use of such a proxy by simple calibratory arguments. Three techniques of time series analysis are employed: spectral, spectrogram and wavelet methods. In addition to the well-known ~9-hour periodicity shown by these data, a previously unknown periodic flux variability is revealed by the wavelet analysis as a 3-day cycle of frequency modulation during June–July 1997, though the physical mechanism responsible is not clear. Such time series analysis has potential for other lava flux proxies at other types of volcanoes.     

Factors controlling the lengths of channel-fed lava flows

Factors which control lava flow length are still not fully understood. The assumption that flow length as mainly influenced by viscosity was contested by Walker (1973) who proposed that the length of a lava flow was dependent on the mean effusion rate, and by Malin (1980) who concluded that flow length was dependent on erupted volume. Our reanalysis of Malin's data shows that, if short duration and tube-fed flows are eliminated, Malin's Hawaiian flow data are consistent with Walker's assertion. However, the length of a flow can vary, for a given effusion rate, by a factor of 7, and by up to 10 for a given volume. Factors other than effusion rate and volume are therefore clearly important in controlling the lengths of lava flows. We establish the relative importance of the other factors by performing a multivariate analysis of data for recent Hawaiian lava flows. In addition to generating empirical equations relating flow length to other variables, we have developed a non-isothermal Bingham flow model. This computes the channel and levee width of a flow and hence permits the advance rates of flows and their maximum cooling-limited lengths for different gradients and effusion rates to be calculated. Changing rheological properties are taken into account using the ratio of yield strength to viscosity; available field measurements show that this varies systematically from the vent to the front of a lava flow. The model gives reasonable agreement with data from the 1983–1986 Pu'u Oo eruptions and the 1984 eruption of Mauna Loa. The method has also been applied to andesitic and rhyolitic lava flows. It predicts that, while the more silicic lava flows advance at generally slower rates than basaltic flows, their maximum flow lengths, for a given effusion rate, will be greater than for basaltic lava flows.     

Development of lava tubes in the light of observations at Mauna Ulu,Kilauea Volcano,Hawaii

During the 1969–1974 Mauna Ulu eruption on Kilauea's upper east rift zone, lava tubes were observed to develop by four principal processes: (1) flat, rooted crusts grew across streams within confined channels; (2) overflows and spatter accreted to levees to build arched roofs across streams; (3) plates of solidified crust floating downstream coalesced to form a roof; and (4) pahoehoe lobes progressively extended, fed by networks of distributaries beneath a solidified crust. Still another tube-forming process operated when pahoehoe entered the ocean; large waves would abruptly chill a crust across the entire surface of a molten stream crossing through the surf zone. These littoral lava tubes formed abruptly, in contrast to subaerial tubes, which formed gradually. All tube-forming processes were favored by low to moderate volume-rates of flow for sustained periods of time. Tubes thereby became ubiquitous within the pahoehoe flows and distributed a very large proportionof the lava that was produced during this prolonged eruption. Tubes transport lava efficiently. Once formed, the roofs of tubes insulate the active streams within, allowing the lava to retain its fluidity for a longer time than if exposed directly to ambient air temperature. Thus the flows can travel greater distances and spread over wider areas. Even though supply rates during most of 1970–1974 were moderate, ranging from 1 to 5 m³/s, large tube systems conducted lava as far as the coast, 12–13 km distant, where they fed extensive pahoehoe fields on the coastal flats. Some flows entered the sea to build lava deltas and add new land to the island. The largest and most efficient tubes developed during periods of sustained extrusion, when new lava was being supplied at nearly constant rates. Tubes can play a major role in building volcanic edifices with gentle slopes because they can deliver a substantial fraction of lava erupted at low to moderate rates to sites far down the flank of a volcano. We conclude, therefore, that the tendency of active pahoehoe flows to form lava tubes is a significant factor in producing the common shield morphology of basaltic volcanoes.     

The influence of palaeoenvironment and lava flux on the emplacement of submarine, near-shore Late Permian basalt lavas, Sydney Basin (Australia)

Lava flux and a low palaeoslope were the critical factors in determining the development of different facies in the Late Permian Blow Hole flow, which comprises a series of shoshonitic basalt lavas and associated volcaniclastic detritus in the southern Sydney Basin of eastern Australia. The unit consists of a lower lobe and sheet facies, a middle tube and breccia facies, and an upper columnar-jointed facies. Close similarities in petrography and geochemistry between the basalt lavas from the three facies suggest similar viscosities at similar temperatures. Sedimentological and palaeontological evidence from the sedimentary units immediately below the Blow Hole flow suggests that the lower part of the volcanic unit was emplaced in a cold water, shallow submarine environment, but at least the top of the uppermost lava was subaerial with some palaeosol development. The lower lobe and sheet facies was emplaced on a low slope (<2°) in a lower to middle shoreface environment with water depths of 20–25 m. Lava may have transgressed from subaerial to subaqueous and was emplaced relatively passively with lava flux sufficiently high and uniform to form lobes and sheets rather than pillows. The middle unit probably originated from a subaerial vent and flowed into a shallow (10–15 m) submarine environment, and wave action probably interacted with the advancing lava front to form a lava delta. Lava flux was sufficiently high to produce well-developed, subcircular lava tubes, which lack evidence for thermal erosion. In some areas, lava ‘burrowed’ into the unconsolidated, water-saturated lava delta and sand pile to produce intrusive contacts. The upper columnar-jointed unit represents a ponded facies probably emplaced initially in water depths <5 m but whose top was subaerial.     

Structures and facies associated with the flow of subaerial basaltic lava into a deep freshwater lake: The Sulphur Creek lava flow, North Cascades, Washington

The ca. 8800 ¹⁴C yrs BP Sulphur Creek lava flowed eastward 12 km from the Schriebers Meadow cinder cone into the Baker River valley, on the southeast flank of Mount Baker volcano. The compositionally-zoned basaltic to basaltic andesite lava entered, crossed and partially filled the 2-km-wide and > 100-m-deep early Holocene remnant of Glacial Lake Baker. The valley is now submerged beneath a reservoir, but seasonal drawdown permits study of the distal entrant lava. As a lava volume that may have been as much as 180 × 10⁶ m³ entered the lake, the flow invaded the lacustrine sequence and extended to the opposite (east) side of the drowned Baker River valley. The volume and mobility of the lava can be attributed to a high flux rate, a prolonged eruption, or both. Basalt exposed below the former level of the remnant glacial lake is glassy or microcrystalline and sparsely vesicular, with pervasive hackly or blocky fractures. Together with pseudopillow fractures, these features reflect fracturing normal to penetrative thermal fronts and quenching by water. A fine-grained hyaloclastite facies was probably formed during quench fragmentation or isolated magma-water explosions. Although the structures closely resemble those developed in lava-ice contact environments, establishing the depositional environment for lava exhibiting similar intense fracturing should be confirmed by geologic evidence rather than by internal structure alone. The lava also invaded the lacustrine sequence, forming varieties of peperite, including sills that are conformable within the invaded strata and resemble volcaniclastic breccias. The peperite is generally fragmental and clast- or matrix-supported; fine-grained and rounded fluidal margins occur locally. The lava formed a thickened subaqueous plug that, as the lake drained in the mid-Holocene, was exposed to erosion. The Baker River then cut a 52-m-deep gorge through the shattered, highly erodible basalt.     

The anisotropy of magnetic susceptibility of lava flows: an experimental approach

Measurements of the anisotropy of magnetic susceptibility (AMS) of natural lavas have shown that AMS varies with depth within a lava flow. We have investigated the reasons for such variation by studying the effects of temperature and strain rate on the AMS of recent lava in the laboratory. Samples of lava from Kilauea were melted and subjected to a range of strain rate and cooling histories. The results show that the degree of anisotropy is a function of both the thermal and shearing history of a sample. High degrees of anisotropy were found only in samples that were deformed at temperatures close to those encountered during eruption and then rapidly quenched. Lavas subjected to similar shear stresses at high temperatures had low degrees of anisotropy if allowed to cool down slowly without further deformation. Additionally, lava subjected to complex shearing yield a lower degree of anisotropy even when high strain rates were imposed on it. These results lead to the conclusion that only the last phase of deformation is detectable using AMS and that high strain rates will not result in high degrees of anisotropy if either deformation ends while lava is still fluid or if the orientation of the maximum shear stress varies with time. The relation between the orientation of the principal susceptibilities and that of shear is less sensitive to variation on shear with time. Consequently, flow directions can be inferred confidently with this type of measurements.     

Loess erosion change modeling during heavy rainfall

A typical gully sub-basin with a complex geomorphological form is used to do a model test of gravity erosion of loess by considering the sequence of slopes in a prototype gully creating a sequence of underlying surface forms in the upper reaches. The results show that the runoff from heavy rainfall is the main external force for the erosion of loess, and also is an important influencing factor to stimulate and intensify the development of gravity erosion. The soil structure and the height of the...     

Stratification response of soil water content during rainfall events under different rainfall patterns

Many researchers have studied the influence of rainfall patterns on soil water movement processes using rainfall simulation experiments. However, less attention has been paid to the influence under natural condition. In this paper, rainfall, soil water content (SWC), and soil temperature at 10‐, 20‐, 30‐, 40‐, and 50‐cm depths were simultaneously monitored at 1‐min intervals to measure the variation in SWC (SWCv) in response to rainfall under different rainfall patterns. First, we classified rainfall events into four patterns. During the study period, the main pattern was the advanced rainfall pattern (38% of all rainfall events), whereas the delayed, central, and uniform rainfall patterns had similar frequencies of about 20%. During natural rainfall, rainwater rapidly passed through the top soil layers (10–40 cm) and was accumulated in the bottom layer (50 cm). When a high rainfall pulse occurred, the water storage balance was disturbed, resulting in the drainage of initial soil water from the top layers into the deeper layers. Therefore, the critical function of the top layers and the bottom layers was infiltration and storage, respectively. The source of water stored in the bottom layer was not only rainfall but also the initial soil water in the upper soil layers. Changes in soil temperature at each soil depth were comonitored with SWCv to determine the movement characteristics of soil water under different rainfall patterns. Under the delayed rainfall pattern, preferential flows preferred to occur. Under the other rainfall patterns, matrix flow was the main form of soil water movement. Rainfall amount was a better indicator than rainfall intensity for SWCv in the bottom layer under the delayed rainfall pattern. These results provide insights into the responses of SWCv under different rainfall patterns in northern China.     

On the effect of lava viscosity on the magnetic fabric intensity in alkaline volcanic rocks

The degree of the anisotropy of magnetic susceptibility (AMS) of basaltic rocks, as is known from the large AMS database of these rocks, is generally very low, while in more acidic volcanic rocks such as andesites, trachytes and phonolites, which have been investigated much less frequently, it is in general much higher. In the present study, the AMS of various volcanic rocks including trachytic and phonolitic rocks was investigated in the Tertiary volcanic region of the eské stedohoí Mts. Viscosities of the respective lavas were calculated from the chemical composition using the KWARE program. A rough correlation was found between the degree of AMS and lava viscosities, probably resulting from different mechanisms orienting the magnetic minerals. In basaltic lava flows this mechanism is traditionally considered to be of a hydrodynamic nature, in trachytic and phonolitic bodies it can also be represented by quasi-intrusive flows resembling, at least partially, ductile flow deformation. This is in agreement with the AMS data predicted by the viscous (liquid flow) and line/plane (ductile flow) models.     

Ground-based thermal imaging of lava lakes at Erebus volcano,Antarctica

Mount Erebus, a large intraplate stratovolcano dominating Ross Island, Antarctica, hosts the world's only active phonolite lava lakes. The main manifestation of activity at Erebus volcano in December 2004 was as the presence of two convecting lava lakes within an inner crater. The long-lived Ray Lake, ~ 1400 m² in area, was the site of up to 10 small Strombolian eruptions per day. A new but short-lived, ~ 1000–1200 m² lake formed at Werner vent in December 2004 sourced by lava flowing from a crater formed in 1993 by a phreatic eruption. We measured the radiative heat flux from the two lakes in December 2004 using a compact infrared (IR) imaging camera. Daily thermal IR surveys from the Main Crater rim provide images of the lava lake surface temperatures and identify sites of upwelling and downwelling. The radiative heat outputs calculated for the Ray and Werner Lakes are 30–35 MW and 20 MW, respectively. We estimate that the magma flux needed to sustain the combined heat loss is ~ 250–710 kg s^− 1, that the minimum volume of the magma reservoir is 2 km³, and that the radius of the conduit feeding the Ray lake is ~ 2 m.     

Multi-instrument remote and in situ observations of the Erebus Volcano (Antarctica) lava lake in 2005: A comparison with the Pele lava lake on the jovian moon Io

The stable, persistent, active lava lake at Erebus volcano (Ross Island, Antarctica) provides an excellent thermal target for analysis of spacecraft observations, and for testing new technology. In the austral summer of 2005 visible and infrared observations of the Erebus lava lake were obtained with sensors on three space vehicles Terra (ASTER, MODIS), Aqua (MODIS) and EO-1 (Hyperion, ALI). Contemporaneous ground-based observations were obtained with hand-held infrared cameras. This allowed a quantitative comparison of the thermal data obtained from different instruments, and of the analytical techniques used to analyze the data, both with and without the constraints imposed by ground-truth. From the thermal camera data, in December 2005 the main Erebus lava lake (Ray Lake) had an area of ≈ 820 m². Surface colour temperatures ranged from 575 K to 1090 K, with a broad peak in the distribution from 730 K to 850 K. Total heat loss was estimated at 23.5 MW. The flux density was ≈ 29 kW m^− 2. Mass flux was estimated at 64 to 93 kg s^− 1. The best correlation between thermal emission and emitting area was obtained with ASTER, which has the best combination of spatial resolution and wavelength coverage, especially in the thermal infrared. The high surface temperature of the lava lake means that Hyperion data are for the most part saturated. Uncertainties, introduced by the need to remove incident sunlight cause the thermal emission from the Hyperion data to be a factor of about two greater than that measured by hand-held thermal camera. MODIS also over-estimated thermal output from the lava lake by the same factor of two because it was detecting reflected sunlight from the rest of the pixel area. The measurement of the detailed temperature distribution on the surface of an active terrestrial lava lake will allow testing of thermal emission models used to interpret remote-sensing data of volcanism on Io, where no such ground-truth exists. Although the Erebus lava lake is four orders of magnitude smaller than the lava lake at Pele on Io, the shape of the integrated thermal emission spectra are similar. Thermal emission from this style of effusive volcanism appears to be invariant. Excess thermal emission in most Pele spectra (compared to Erebus) at short wavelengths (< 3 μm) is most likely due to disruption of the surface on the lava lake by escaping volatiles.     

Interpretation of porosity in dacite lava domes as ductile–brittle failure textures

Pores in dacite lavas of three volcanoes in Japan (Unzen, Yakedake and Daisen) have characteristics unlike vesicles. Most of the pores have ragged walls marked by the protrusion of groundmass crystals and phenocrysts. The shapes of these pores range from irregular to planar. Many pores occur in ‘pressure shadows’ adjacent to phenocrysts. Some pores are crack-like and pass around phenocrysts and have blunt terminations. Most of the pores are interpreted as the result of failure of the magma by flow during ductile–brittle transition. This phenomenon, known as cavitation, is well established in hydraulics and the material sciences. Cavitation of the lavas in this study was promoted by steep slope of emplacement, subtle mechanical contrasts between bands, and strain intensification around rigid crystals and lithic fragments. Cavitation in lava can extract volatiles from magma causing premature embrittlement and weakening. Cavitation is inferred to be critical to the initiation of lava-front pyroclastic flows.     

Assessing changes in rainfall erosivity in Sicily during the twentieth century

Changes in rainfall erosivity are an expected consequence of climate change. Long‐term series of the single storm erosion index, EI, may be analysed to detect trends in rainfall erosivity. An indirect approach has to be applied for estimating EI, given that long series of rainfall intensities are seldom available. In this paper, a method for estimating EI from the corresponding rainfall amount, h_e, was developed for Sicily. This method was then applied at 17 Sicilian locations, representative of different climatic zones of the region, to generate a long series (i.e. from 1916 to 1999 in most cases) of EI values. Linear and step (step located at 1970) trends in annual and seasonal erosivity were detected by both classical approaches (Mann–Kendall test, Wilcoxon‐Mann‐Whitney rank‐sum test) and a new empirical approach (quantile approach, QA), based on the determination of the erosivity values corresponding to selected probability levels. A power relationship between EI and h_e with a space‐ and time‐variable scale factor and a time‐variable process parameter yielded the most accurate predictions of EI. However, a simpler model, using a time‐variable scale factor and a constant process parameter, yielded reasonably accurate EI estimates. Annual erosivity did not increase in Sicily during the twentieth century. At the most, it decreased at a few locations (three of the 17 considered locations). Significant trends were observed more frequently for winter erosivity (six locations) than for summer erosivity (two locations), suggesting that the erosive storms of winter determined the occasional occurrence of a negative trend in annual erosivity. In general, the QA compared reasonably well with more classical approaches. The QA appears promising since step trends for different return periods may be detected but efforts are needed to statistically formalize the proposed approach. Copyright © 2007 John Wiley & Sons, Ltd.     

Earthen barriers to control lava flows in the 2001 eruption of Mt. Etna

Preceded by four days of intense seismicity and marked ground deformation, a new eruption of Mt. Etna started on 17 July and lasted until 9 August 2001. It produced lava emission and strombolian and phreatomagmatic activity from four different main vents located on a complex fracture system extending from the southeast summit cone for about 4.5 km southwards, from 3000 to 2100 m elevation (a.s.l.). The lava emitted from the lowest vent cut up an important road on the volcano and destroyed other rural roads and a few isolated country houses. Its front descended southwards to about 4 km distance from the villages of Nicolosi and Belpasso. A plan of intervention, including diversion and retaining barriers and possibly lava flow interruption, was prepared but not activated because the flow front stopped as a consequence of a decrease in the effusion rate. Extensive interventions were carried out in order to protect some important tourist facilities of the Sapienza and Mts. Silvestri zones (1900 m elevation) from being destroyed by the lava emitted from vents located at 2700 m and 2550 m elevation. Thirteen earthen barriers (with a maximum length of 370 m, height of 10–12 m, base width of 15 m and volume of 25 000 m³) were built to divert the lava flow away from the facilities towards a path implying considerably less damage. Most of the barriers were oriented diagonally (110–135°) to the direction of the flow. They were made of loose material excavated nearby and worked very nicely, resisting the thrust of the lava without any difficulty. After the interventions carried out on Mt. Etna in 1983 and in 1991–1992, those of 2001 confirm that earthen barriers can be very effective in controlling lava flows.     

Formation of clastogenic lava flows during fissure eruption and scoria cone collapse: the 1986 eruption of Izu-Oshima Volcano, eastern Japan

The 1986 eruption of B fissure at Izu-Oshima Volcano, Japan, produced, among other products, one andesite and two basaltic andesite lava flows. Locally the three flows resemble vent-effused holocrystalline blocky or aa lava; however, remnant clast outlines can be identified at most localities, indicating that the flows were spatter fed or clastogenic. The basaltic andesite flows are interpreted to have formed by two main processes: (a) reconstitution of fountain-generated spatter around vent areas by syn-depositional agglutination and coalescence, followed by extensional non-particulate flow, and (b) syn-eruptive collapse of a rapidly built spatter and scoria cone by rotational slip and extensional sliding. These processes produced two morphologically distinct lobes in both flows by: (a) earlier non-particulate flow of agglutinate and coalesced spatter, which formed a thin lobe of rubbly aa lava (ca. 5 m thick) with characteristic open extension cracks revealing a homogeneous, holocrystalline interior, and (b) later scoria-cone collapse, which created a larger lobe of irregular thickness (<20 m) made of large detached blocks of scoria cone interpreted to have been rafted along on a flow of coalesced spatter. The source regions of these lava flows are characterized by horseshoe-shaped scarps (<30 m high), with meso-blocks (ca. 30 m in diameter) of bedded scoria at the base. One lava flow has a secondary lateral collapse zone with lower (ca. 7 m) scarps. Backward-tilted meso-blocks are interpreted to be the product of rotational slip, and forward-tilted blocks the result of simple toppling. Squeeze-ups of coalesced spatter along the leading edge of the meso-blocks indicate that coalescence occurred in the basal part of the scoria cone. This low-viscosity, coalesced spatter acted as a lubricating layer along which basal failure of the scoria cone occurred. Rotational sliding gave way to extensional translational sliding as the slide mass spread out onto the present caldera floor. Squeeze-ups concentrated at the distal margin indicate that the extensional regime changed to one of compression, probably as a result of cooling of the flow front. Sliding material piled up behind the slowing flow front, and coalesced spatter was squeezed up from the interior of the flow through fractures and between rafted blocks. The andesite flow, although morphologically similar to the other two flows, has a slightly different chemical composition which corresponds to the earliest stage of the eruption. It is a much smaller lava flow emitted from the base of the scoria cone 2 days after the eruption had ceased. This lava is interpreted to have been formed by post-depositional coalescence of spatter under the influence of the in-situ cooling rate and load pressure of the deposit. Extrusion occurred through the lower part of the scoria cone, and subsequent non-particulate flow of coalesced material produced a blocky and aa lava flow. The mechanisms of formation of the lava flows described may be more common during explosive eruptions of mafic magma than previously envisaged. Received: 30 May 1997 / Accepted: 19 May 1998     

Discovery of a recurrent lava lake on Saunders Island (South Sandwich Islands) using AVHRR imagery

Monitoring of the remote South Sandwich Islands volcanic arc, using advanced very high resolution radiometer (AVHRR) data, has identified a radiant pixel on channels 3 (3.55–3.93 μm) and (rarely) 4 (10.3–11.3 μm). The position of the pixel is coincident with Mount Michael, Saunders Island, the active summit crater of a snow-covered basaltic stratovolcano with a persistent steam plume. The radiant pixel was continuously present in successive AVHRR images acquired during intervals of several months in the period examined (March 1995–February 1998), although apparently disappearing for similar time intervals. More than 5000 images were examined during this study. The radiant pixel is interpreted to indicate that the crater has contained a lava lake, the first to be recorded in the South Sandwich Islands. The lake appears to persist in the crater for several months at a time, at least. It may have drained completely at times and was probably absent when the crater was viewed briefly during an overflight in January 1997. Persistent or recurring lava lakes are very uncommon world-wide and that at Mount Michael is one of only two recorded in the Antarctic region.