Title: Subsurface architecture of two tropical alpine desert cinder cones that hold water
Abstract: Journal of Geophysical Research: Earth SurfaceVolume 121, Issue 6 p. 1148-1160 Research ArticleFree Access Subsurface architecture of two tropical alpine desert cinder cones that hold water Matthias Leopold, Matthias Leopold School of Earth and Environment, University of Western Australia (M087), Crawley, Western Australia, AustraliaSearch for more papers by this authorAmanda Morelli, Amanda Morelli Institute for Astronomy, University of Hawaii, Honolulu, Hawaii, USA Universidade Federal de São Paulo, São Paulo, BrazilSearch for more papers by this authorNorbert Schorghofer, Corresponding Author Norbert Schorghofer Institute for Astronomy, University of Hawaii, Honolulu, Hawaii, USA Correspondence to: N. Schorghofer, [email protected] for more papers by this author Matthias Leopold, Matthias Leopold School of Earth and Environment, University of Western Australia (M087), Crawley, Western Australia, AustraliaSearch for more papers by this authorAmanda Morelli, Amanda Morelli Institute for Astronomy, University of Hawaii, Honolulu, Hawaii, USA Universidade Federal de São Paulo, São Paulo, BrazilSearch for more papers by this authorNorbert Schorghofer, Corresponding Author Norbert Schorghofer Institute for Astronomy, University of Hawaii, Honolulu, Hawaii, USA Correspondence to: N. Schorghofer, [email protected] for more papers by this author First published: 02 June 2016 https://doi.org/10.1002/2016JF003853Citations: 4AboutSectionsPDF ToolsRequest permissionExport citationAdd to favoritesTrack citation ShareShare Give accessShare full text accessShare full-text accessPlease review our Terms and Conditions of Use and check box below to share full-text version of article.I have read and accept the Wiley Online Library Terms and Conditions of UseShareable LinkUse the link below to share a full-text version of this article with your friends and colleagues. Learn more.Copy URL Share a linkShare onFacebookTwitterLinkedInRedditWechat Abstract Basaltic lava is generally porous and cannot hold water to form lakes. Here we investigate two impermeable cinder cones in the alpine desert of Maunakea volcano, Hawaii. We present the results of the first ever geophysical survey of the area around Lake Waiau, the highest lake on the Hawaiian Islands, and establish the existence of a second body of standing water in a nearby cinder cone, Pu‘upōhaku (~4000 m above sea level), which has a sporadic pond of water. Based on unpublished field notes from Alfred Woodcock (*1905–†2005) spanning the years 1966–1977, more recent observations, and our own geophysical survey using electric resistivity tomography, we find that perched groundwater resides in the crater perennially to a depth of 2.5 m below the surface. Hence, Pu‘upōhaku crater hosts a previously unrecognized permanent body of water, the highest on the Hawaiian Islands. Nearby Lake Waiau is also perched within a cinder cone known as Pu‘uwaiau. Among other hypotheses, permafrost or a massive block of lava were discussed as a possible cause for perching the water table. Based on our results, ground temperatures are too high and specific electric resistivity values too low to be consistent with either ice-rich permafrost or massive rock. Fine-grained material such as ash and its clay-rich weathering products are likely the impermeable material that explains the perched water table at both study sites. At Pu‘uwaiau we discovered a layer of high conductivity that may constitute a significant water reservoir outside of the lake and further be responsible for perching the water toward the lake. Key Points Perennial water is found in two desert craters on Maunakea volcano, Hawaii: Pu‘uwaiau and Pu‘upōhaku Pu‘upōhaku hosts a perennial body of water, the highest on the Hawaiian Islands, which suggests that impermeable cinder cones are not an anomaly Fine-grained material, and not permafrost, is likely the impermeable material that perches the water table at both study sites 1 Introduction and Study Site Impermeable valley bottoms and a sufficient water supply are two preconditions to form lakes, and this is also known from many volcanic areas [Larson, 1989]. For lakes where these preconditions are marginal, the subsurface architecture is important to fully understand their genesis and their potential role, especially in desert environments. Perennial lakes only form where a depression is sufficiently impermeable to prevent complete drainage and the water supply exceeds evaporation [Hutter et al., 2011]. Some high-altitude volcanic areas, such as the islands of Hawaii with their coarse block-rich and thus highly permeable basaltic lava flows together with their specific meteorological conditions with low precipitation, do not favor lake genesis. Indeed, bodies of standing water are a rarity on the high volcanoes of Hawaii, or in fact anywhere in Hawaii [Maciolek, 1982], and where present they are potentially a reliable year-round source of water for the fauna and flora in this extreme environment. Here we study two cinder cones with perennial bodies of water near the summit of Maunakea volcano (4207 m above sea level (asl)), where occasional snowstorms are the main source of precipitation (~21 cm/yr) [Giambelluca et al., 2013]. In the saddle region between Maunakea and Maunaloa, deep drilling has recently revealed perched aquifers [Thomas and Haskins, 2013]. Despite the hundreds of cinder cones [Porter, 1972], which form bowl-shaped geomorphic forms to store a body of water, there is only one known lake, Lake Waiau (3970 m asl) on Maunakea, on the island of Hawaii (Figure 1). Even up to now, the lake serves as a spiritual place for local Hawaiian people and as an (unsanctioned) tourist attraction of the Mauna Kea Ice Age Natural Area Reserve. The cinder cone containing Lake Waiau, known as Pu‘uwaiau, has been studied for hydrological [Gregory and Wentworth, 1937; Woodcock and Groves, 1969; Ehlmann et al., 2005], ecological [Wentworth and Powers, 1943], as well as meteorological and bathymetrical aspects [Woodcock, 1980; Patrick and Kauahikaua, 2015]. The reason for the lakes' existence remains uncertain but fine-grained ash, permafrost, massive lava beds, and clays have been proposed to cause low permeability [Ugolini, 1974; Woodcock, 1980; Wolfe, 1987]. Potential permafrost melt has also been discussed in the lakes' recent decline [Woodcock, 1980; Patrick and Delparte, 2014; Patrick and Kauahikaua, 2015]. However, no direct evidence of permafrost under the lake or in its near vicinity has ever been found, despite isolated permafrost occurrence on the summit [Woodcock et al., 1970; Woodcock, 1974]. Figure 1Open in figure viewerPowerPoint (a) Map of Hawaii Island and shaded topography of (b) the summit region, (c) Pu‘upōhaku with its shallow crater, and (d) Pu‘uwaiau. ERT survey lines are shown as straight thick lines, and their starting points are labeled with Greek letters. Pu‘upōhaku pond surpasses the elevation of Lake Waiau. In Figures 1c and 1d elevation contours are spaced 10 m. Pu‘upōhaku, also known as Douglas Cone, is another cinder cone located on Maunakea volcano, at an elevation of approximately 4000 m and part of the Pleistocene formation Laupahoehoe Volcanics [Sherrod et al., 2007]. Pu‘upōhaku has a roughly 1 m deep shallow crater that can hold water, unusual in this barren alpine cold desert environment. It is known to have a sporadic pond. Figure 2 shows Pu‘upōhaku with and without a pond. The water level is limited by an outlet that leads into a sharply incised gully on the northeast side of the cone. Pu‘upōhaku, which has never been described in detail, may be a useful analogue to nearby Lake Waiau, which also sits on an impermeable cinder cone. Lake Mataulano on the Samoan Islands sits in an impermeable cinder cone as well, but in a far warmer climate [Keating, 1992]. Figure 2Open in figure viewerPowerPoint (a) Pu‘upōhaku cinder cone with a shallow crater and a gully on its northeast side. (b) Pond in Pōhaku Crater on 19 Sep 2014 (ranger photo). (c) Ground water in an ~40 cm deep pit at the center of the basin (28 May 2014). (d) The discolored area in the dry crater with a person standing at the center pit. (e) A small puddle of water on 21 Nov 2014. This paper aims to unfold the subsurface architecture of Pu‘uwaiau and Pu‘upōhaku to further explore the causes of the impermeable lake bottoms and thus, the existence of standing water bodies in this harsh volcanic desert environment. 2 Methods 2.1 The Field Notes of Alfred Woodcock The Maunakea summit area is environmentally and archeologically sensitive, and hence, we have made every effort to use preexisting data for our study. A treasure trove turned out to be the unpublished field record books of Alfred Woodcock, further described below. Alfred H. Woodcock (*1905–†2005) carried out numerous field studies in the summit region of Maunakea, based on over 350 field trips. His published works include the sediment and hydrology of Lake Waiau [Woodcock et al., 1966; Woodcock and Groves, 1969; Woodcock, 1980] and the discovery of permafrost near the summit [Woodcock et al., 1970; Woodcock, 1974]. Some of his research on Pu‘upōhaku was mentioned peripherally [Woodcock and Groves, 1969; Woodcock, 1974], but most was never published. We went through his unpublished Record Books (RBs) Nr. 3 to Nr. 16 (1965–1982) to search for information [Woodcock archives, 2007]. Based on his RBs, Woodcock visited Pu‘upōhaku 56 times over 11 years to collect data and measure temperatures with depth. To determine whether there is permafrost beneath the area, Woodcock drove pipes into the ground and measured temperature as a function of depth. The longest of them was 32 ft (~10 m). The pipes were filled with water for faster thermal equilibration, and this water level was maintained by refilling each time after use (RB12, p16). 2.2 Electrical Resistivity Tomography Survey Electric resistivity tomography (ERT) generates data about the subsurface differences in the electrical resistivity. We used a multielectrode system ‘4pointlight hp’ from Lippmann Geophysikalische Messgeräte (20 electrodes [Lippmann, 2008]) in order to create 2-D images of the apparent resistivity. Contact resistance from the electrode into the ground is often high in alpine areas due to the coarse and dry surface conditions in this environment [e.g., Leopold et al., 2013]. However, despite coarse boulders along parts of the survey line, the contact resistance was at a maximum of 4 kΩ and mostly below 1 kΩ due to the fine and moist material below the boulders. In total we ran three lines at Pu‘upōhaku with line α and line α′ running along the same transect but with varying electrode distances and line β at a 90° angle to line α. At Pu‘uwaiau we surveyed four lines. The specific locations and configurations of the lines are displayed in Figure 1 as well as in Table 1. The survey was carried out on 22–25 June 2015. Table 1. Main Parameters and GPS Locations for the Electric Resistivity Survey Lines at Pu‘upōhaku and Pu‘uwaiaua Start End Electrode Survey Area Line No. Lat N Long W Lat N Long W Distance (m) Array Type Pohaku α 19°49′26.6″ 155°29′28.5″ 19°49′29.7″ 155°29′27.9″ 3.5 W, D Pohaku α′ runs between 23 and 42 m along Pohaku line α 1 W, D Pohaku β 19°49′27.2″ 155°29′27.2″ 19°49′27.9″ 155°29′29.2″ 3.5 W, D Waiau α 19°48′38.9″ 155°28′38.4″ 19°48′41.2″ 155°28′36.6″ 4.5 W, D Waiau β 19°48′39.6″ 155°28′41.0″ 19°48′42.7″ 155°28′40.3″ 5 W, D Waiau γ 19°48′38.2″ 155°28′33.0″ 19°48′40.0″ 155°28′37.4″ 5 W, D Waiau γ′ runs between 0 and 9.5 m along Waiau line γ 0.5 W, D Waiau δ 19°48′36.1″ 155°28′39.6″ 19°48′37.1″ 155°28′36.7″ 5 W, D a Note: W = Wenner array, D = dipole-dipole array. We used both a Wenner and a Dipole-Dipole array for data collection. Whereas the Wenner array allows higher current with resulting higher potentials and, accordingly, a better signal-to-noise ratio, the Dipole-Dipole array gives a better near-surface vertical resolution and a deeper survey depth. A joint data set was used for inversion. A measurement frequency of 5 Hz was used with a current of 5–15 mA. Each point was measured between 6 and 10 times depending on the variability of the results (2% limit). The resistivity calculated from these measurements is denoted as apparent resistivity. Subsequently, we generated a 2-D model of the specific resistivity using RES2DINV 3.55.18 [Loke and Barker, 1995] and a least squares inversion. Along the survey line, we measured topography at least every 5 m using a handheld inclinometer. Models generally reached our desired convergence limit of 2.0% after five iterations. Resulting RMS errors were mostly below 6%, which are excellent for a harsh alpine environment [Hauck and Vonder Mühll, 2003]. Only line δ at Lake Waiau generated a higher RMS value of 13%. For each line we calculated a depth of investigation index (DOI) as suggested by Oldenburg and Li [1999] using reference models of 0.1 and 10 times background resistivity and 3 times depth of the estimated maximum depth of investigation. A vertical:horizontal damping factor relationship of 1:1 was applied. After normalizing the data to 1 we used a threshold value of 0.2 to select areas below which the model data are no longer sensitive to the physical properties of the subsurface [Hilbich et al., 2009]. All areas in our models were within the 0.2 DOI value and often below the 0.1 value and thus, are all sensitive to the measured physical properties. 2.3 Other Methods Topography at a resolution of 4.5 m × 4.5 m is available from interferometric synthetic aperture radar data from Intermap Technologies. Drainage area is determined from this topography and a program that follows the steepest direction among the eight nearest neighbors. Annual precipitation is based on station data from 1965 to 1981, http://rainfall.geography.hawaii.edu/ [Giambelluca et al., 2013]. Precipitation combined with drainage area provides an estimate for the volume of available water. Potential evapotranspiration is available from http://evapotranspiration.geography.hawaii.edu [Giambelluca et al., 2014]. An extensive collection of ranger photos was made available by the Office of Mauna Kea Management. Infrared images were acquired with a FLIR E30 infrared camera and processed with FLIR Tools. 3 Results and Interpretation: Pu‘upōhaku 3.1 Woodcock's Unpublished Field Notes The first entry about Pu‘upōhaku (Douglas Cone) was on 19 June 1966 (RB3, p137). At this earliest visit, Woodcock removed stones from the lowest point of the crater and found a saturated ash layer 0.6096 m down. There were numerous large rocks on the crater floor, but these rocks were absent below the surface. Table 2 summarizes Woodcock's records about the presence of water. His notes describe a sporadic pond 6 times and perched groundwater 24 times. Not every entry mentions water; sometimes, only temperature measurements are reported, but he never noted the absence of water. Woodcock's observations over 11 years suggest that Pu‘upōhaku crater hosted a permanent body of water, over the duration of his field visits (1966–1977). Table 2. Presence and Depth of Pond or Perched Groundwater at Pu‘upōhakua Date Pond (Depth in m) Ground water (Depth Below Surface in m) Reference 19 Jun 1966 X WB69, RB 3 19 Oct 1968 X (0.4) WB69, RB 6 9 Nov 1968 X (0.35) RB 6 10 Nov 1968 X RB 6 16 May 1969 X (0.3) RB 6 17 May 1969 X RB 6 26 Jun 1969 X (0.3) RB 6 13 Jun 1969 X (0.3) RB 6 8 Aug 1969 X (0.4) RB 6 14 Sep 1969 X RB 6 28 Mar 1970 X (0.46) RB 7 2 May 1970 X (0.4) RB 7 25 May 1970 X (0.61) RB 7 28 May 1970 X (0.02) RB 7 9 Aug 1970 X (0.65) RB 7 27 Sep 1974 X (0.55) RB 7 19 Mar 1972 X (0.3–0.6) RB 10 27 Apr 1972 X (0.3–0.6) RB 10 26 May 1972 X (0.3) RB 10 27 Jun 1972 X RB 10 23 Sep 1972 X (0.6) RB 11 28 Nov 1972 X RB 11 29 May 1973 X (0.45) RB 11 30 Jun 1973 X (0.5) RB 11 20 Aug 1974 X (0.63) RB 12 17 Oct 1974 X RB 12 2 Dec 1974 X RB 12 6 Jun 1975 X (0.2) RB 12 14 Oct 1976 X (0.76) RB 15 1 Jun 1977 X (0.96) RB 15 a Note: WB69 = Woodcock and Groves [1969], RB = Record Book. Figure 3 shows the measured temperatures in the pipes with depth, based on the numerical entries in his notes. Additional temperature measurements from a shorter 12 ft (~3.7 m) pipe are not included in Figure 3. The temperature below 5 m was always above +3°C, and the average at the deepest point (10 m) was +3.7°C. The temperature below 6 m shows little variability throughout the year and hence represents a long-term average. Woodcock often noted the error of each measurement, which varied between 0.1 and 0.54°C. For comparison, the standard deviation of the bottom temperatures in Figure 3 is 0.11°C. In conclusion, all the temperatures were well above freezing. Figure 3Open in figure viewerPowerPoint Temperature profiles in Pu‘upōhaku pipes according to unpublished measurements by Alfred Woodcock. Colors are chosen according to season: shades of brown for the spring season, shades of blue for summer, shades of green for fall, and shades of red for the winter season. Woodcock estimated the maximum overflow depth of the pond to be 1 m (RB6, p19). He also noticed that the temperature of the surface water is often lower than that of the ground, which he attributed to evaporative cooling. 3.2 Modern Observations Upon visiting Pu‘upōhaku for our own field studies (2014–2015), we saw that the depression at the deepest point of the pond that Woodcock had dug in 1966 is still present and perched groundwater is seen in it (Figure 2c). At times, we saw green algae in the water (as did Woodcock). When a pond is present, it is visible from afar, and photos are sporadically available that document its presence, while perched groundwater is only documented during field visits. Table 3 summarizes recent observations of a pond or perched groundwater that we were able to compile. A pond is still occasionally present, and groundwater was seen within the ~40 cm deep pit at all visits. Hence, the modern observations are qualitatively consistent with Woodcock's. Table 3. List of Recent Documented Observations of Water at Pu‘upōhaku Date Observation Source 28 Mar 2008 pond ranger photo 2 May 2008 pond ranger photo 18 Apr 2009 pond ranger photo 28 May 2014 perched groundwater (~10 cm beneath surface) field visit 23 Aug 2014 pond ranger photo 19 Sep 2014 perched groundwater (~20 cm beneath surface) field visit 22, 28, and 31 Oct 2014 pond ranger photos 2 and 19 Nov 2014 pond ranger photos 21 Nov 2014 pond (~40 cm deep) field visit 20 Apr 2015 pond ranger photo 30 Apr 2015 pond ranger photo 22 Jun 2015 perched groundwater (~20 cm beneath surface) field visit 27 Jun 2015 perched groundwater field visit 6 and 19 Sep 2015 pond ranger photos The surface area on Pōhaku Crater, where the water accumulates, is darker than the rest of the surface, a discoloration due to standing water (Figure 2d). At its widest, the discolored area is 50 m and it stretches about 40 m in the perpendicular direction. From a perspective photograph with markers separated by 30 m along and perpendicular to the direction the photo was taken, the dark discolored area is estimated as 1360 m2. We estimate the gray portion within the dark discolored area (Figure 2d) to be 230 m2. Given that the puddle is very shallow (<1 m) and the drainage area small, it is remarkable that a pond can last for an extended duration in this dry and sometimes windy environment. A contributing factor is that groundwater extends beyond the visible boundary of the pond, while evaporation is primarily limited to the exposed pond surface. Based on 4.5 m × 4.5 m topography, the drainage area of Pohaku pond is about 13,000 m2. The annual precipitation is 21 cm/yr [Giambelluca et al., 2013], although highly variable. That amounts to a supply of 2700 m3 of water per year. The ratio of the drainage area to pond area is about 10, when the dark discolored area is used as typical pond surface area. At the annual mean precipitation, 2 m/yr can be evaporated from the exposed water surface without depleting the pond. For comparison, the Penman-Monteith potential evapotranspiration (PET) in the summit area is estimated as 2.2 m/yr [Giambelluca et al., 2014]. Hence, the water budget estimate is consistent with the presence of an episodic pond. For comparison, Lake Waiau's drainage area is more than 10 times larger, 135,000 m2 [Ehlmann et al., 2005] and the full Waiau lake has about ~7000 m2. The ratio of the drainage area to pond area is 19 for Lake Waiau and 10 for Pu‘upōhaku. Hence, per area half as much water is available for evaporation at Pohaku than at Waiau. At Waiau the available water clearly exceeds the PET rate even for the full lake. If the drainage area at Pu‘upōhaku crater was significantly larger, then the pond would be expected to be permanent even above the surface. The drainage area of the ephemeral water pond is small, but perched water lasts well beyond the duration of a snow cover, so that the crater floor must be impermeable. Hence, Pu‘uwaiau is not the only impermeable cinder cone in the summit region. Our own measurements include an ~50 cm deep temperature sensor and a data logger (HOBO U23-003) near the west rim of the discolored area. On the day it was emplaced (28 May 2014), the soil was wet at this depth. The average temperature over the first 365 days was +3.1°C, consistent with Woodcock's mean values. 3.3 ERT Survey Results of our ERT survey at Pu‘upōhaku are given in Figure 4. The lines at Pu‘upōhaku are characterized by a sharp gradient of specific electric resistivity values with depth. Due to the high moisture content in the near surface (partly standing water) we expected low surface values, however, as documented in Figures 4b and 4d the low resistivity zone of values between 50 and 200 Ωm reaches to a depth of approximately 2.5 m. The low values were also documented outside of areas with visible standing water; thus, we conclude that they most likely represent a moist, most likely completely water saturated zone. Below 2.5 m depth the specific resistivity values rise over a short distance of less than 1 m to over 3 kΩm reaching a maximum of 5 kΩm at roughly 6–7 m depth. Resistivity images indicate another change toward lower values of around 1.2 kΩm at 10 m depth; however, the transition is less pronounced. Figure 4Open in figure viewerPowerPoint Electric resistivity survey at Pu‘upōhaku. (a) View from the survey area toward Maunakea summit across the survey line. (b) ERT inversion results of line α and (c) depth of investigation (DOI) calculation of line α. (d) ERT inversion of line α between 23 and 42 m with 1 m electrode distance for higher horizontal and vertical resolution. (e) ERT inversion results of line β with adopted topography. (f) Depth of investigation (DOI) calculation of line β. Based on the notes from Woodcock we know that soft ash-rich sediments were drilled down to a total depth of 10 m and temperature measurements were always above freezing. Thus, we conclude that the increase of resistivity between 2.5 and 10 m depth is most likely caused by a decrease in water content. However, above 2.5 m depth the extremely low values can only be explained by a completely saturated aquifer at this depth. It is unknown if the lower boundary of the perched water body varies in depth over the seasons. 4 Results and Interpretation: Pu‘uwaiau 4.1 ERT Survey Interestingly, the subsurface in the vicinity of Lake Waiau shows a slightly different electric resistivity compared to Pu‘upōhaku. Lines α and ß run on the southeast, and the west sides of the lake with line α being only 1 m beside the lake's shoreline (compare Figure 1). The ERT results are given in Figure 5. Line α starts with a layer that is characterized by electric resistivity values between 150 to 250 Ωm. At roughly 6–8 m depth the inversion model shows a sharp gradient toward lower values of ~20 Ωm. A similar distribution can be observed in line ß, which runs in a 30° angle toward the lake's shore line and crosses the overflow channel at 73 m along the line. The transition from higher to the very low resistivity values gets closer to the surface with less distance to the lake's shoreline, where it can be found as shallow as 3–4 m deep. In accordance, the high resistivity layer on top thins out toward the overflow channel, which was dry during our survey. However, line ß documents a third layer between 60 and 75 m along the line and at a depth between 15 and 20 m. Here the resistivity values rise again to over 1 kΩm. Figure 5Open in figure viewerPowerPoint Electric resistivity survey at Pu‘uwaiau. (a) ERT inversion results of line α. (b) ERT inversion results of line β. (c) ERT inversion of line γ and adopted topography. (d) Depth of investigation (DOI) calculation below. (e) ERT inversion results of line δ. Line γ stretches more or less from the rim of the cinder cone toward the lake from southeast to northwest, close to the trail. Here the three divisions in the resistivity image that were already indicated in line ß are clearly visible. Values start between 300 and 400 kΩm in layer one (starting at 40 m along the line), drop along a sharp gradient at 6–8 m depth to resistivity values as low as 10–40 Ωm in layer 2 and rise again in layer 3 to 200–400 Ωm. Layer 3 starts on the surface at the beginning of the survey line and was identified at a depth of >20 m toward the lake. Line δ runs from the top of the elevated mound like structure south west of the lake toward line γ. It follows a geomorphic form that looks similar to initial stages of rock glaciers as found in many alpine areas of the world apart from a missing steep rock face to provide source material for talus development [Barsch, 1996]. The line starts with a moderate high resistivity layer of 0.5 to 1 kΩm followed by a sharp gradient toward much higher values in roughly 2 m depth (Figure 5e). This second layer has specific resistivity values between 4 and 7 kΩm. A third layer can be identified starting at 38 m along the line in a depth of approximately 15 m and showing much lower resistivity values of 0.5 to 1 kΩm. It is important to note that again layers 2 and 3 are separated by a sharp gradient. The schematic in Figure 6 is a conceptual model of the stratigraphy of Pu‘uwaiau. The three-layer subsurface architecture as identified in our ERT survey in lines α, β, and γ is a key in discussing the lake's existence. Layer 1 can be interpreted as typical relict periglacial slope deposits of former solifluction lobes as described by Woodcock mixed with abundant fine material of either ash and/or aeolian origin. These lobes originate mostly from the south wall of the cinder cone from a geomorphic form that most likely represents a relict rock glacier (Figure 7a). Layer 2 is a very fine deeply weathered ash layer, which consists of a mainly silty matrix where it is exposed on the surface. Noteworthy, percentages of clay were not identified during several finger tests at various sites in the field but this observation is restricted to the surface. No direct information concerning texture is available for deeper parts. We assume that layer 2 acts as a structure that helps retain and perch water in the vicinity of the lake. Due to its volcanic or aeolian origin a rather large pore volume but small pore size can be assumed, whether thin internal clay layers within layer 2 or a compacted layer 3 help perch the water is unclear. Since the lake sediments were dated to a minimum age of 15 ka [Wolfe et al., 1997; Pigati et al., 2008] layer 3, which is stratigraphically deeper, may represent older compacted periglacial material or ice compacted material from one of the older glacial drifts described by Porter [1986, 2005]. Alternatively, layer 3 could already represent the in situ volcanic cinder cone material. Figure 6Open in figure viewerPowerPoint Sketch model of the subsurface architecture of Pu‘uwaiau. Layer structure is based on ERT survey and geomorphic field observation. No signs of permafrost were found during older field observations and our ERT surveys. The fine-grained and clay-rich zone functions as an aquifer and guides the interflow for water from snowmelt and occasional rain fall, which is protected from evaporation by the coarser relict periglacial solifluction lobes. Impermeability of the lower boundary is most likely generated by clays and potentially compacted older periglacial material. The ash layer might function as an artesian aquifer causing seeps of water along coarser ash layers at the deeper parts of the cinder cone. (Note that this sketch is not a true transect. It summarizes known subsurface information from various studies for a better geomorphic and hydraulic understanding of Lake Waiau.) Figure 7Open in figure viewerPowerPoint (a) Lake Waiau basin and (b) infrared-derived surface temperatures at noon. The lake level was near a historic low when these images were taken. Line δ also shows a three-layer model (Figure 5e) but different from the other three lines. It starts with a layer that included lots of fine materials within polygonal structures. Their rims consist of cracks partly filled with larger boulders on the surface. This layer represents the geomorphic structures of a typical, however, now inactive, periglacial active layer as found on many other rock glaciers [e.g., Leopold et al., 2011]. The depth of this layer is roughly 2 m, but the model resolution based on electrode distances of 5 m is coarse. Interestingly, an approximately 15 m thick second layer shows much higher resistivity values. The sharp grad