Introduction

Kīlauea (United States) underwent several changes in 2007. Years of heightened activity, caused by a near-doubling of the magma supply rate to the volcano (Poland and others, 2012), culminated in a magmatic intrusion into Kīlauea’s East Rift Zone that spanned the period June 17–19, 2007. Poland and others (2008) gave an overview of the mid-June 2007 activity, which caused widespread ground cracking in the upper East Rift Zone starting on June 17 and culminated in a short-lived, volumetrically small eruption on the northeast flank of Kānenuiohamo, about 6 kilometers (km) upslope from Pu‘u‘ō‘ō, on June 19 (fig. 1). This activity is informally known as the Father’s Day eruption, since it started on the Father’s Day holiday (United States), and it is categorized as episode 56 of the Pu‘u‘ō‘ō eruption (Orr and others, 2015). The Pu‘u‘ō‘ō eruption lasted in total from 1983 to 2018 (Neal and others, 2019).

Here we describe the Father’s Day eruption and discuss a few aspects of the activity not covered elsewhere, including the ground cracking associated with the intrusion and the results of electromagnetic measurements that were made across the trace of the dike using the very low frequency (VLF) technique (see, for example, McNeill and Labson, 1991). These measurements help us to better understand how the intruding dike splayed as it approached the surface from depth.

Eruption setting

From 1997 to 2007, magma traveled from Kīlauea’s summit reservoir through the East Rift Zone to Pu‘u‘ō‘ō, where it erupted from vents on the southwest flank of Pu‘u‘ō‘ō (fig. 1; Heliker and Mattox, 2003; Orr and others, 2015). Lava tube systems carried the lava to the island’s southeast coast. Ephemeral outgassing spatter cones and small lava lakes were also commonly present on Pu‘u‘ō‘ō’s crater floor during this period.

Starting in 2003, Global Positioning System (GPS) sensors spanning the volcano’s summit and upper East Rift Zone showed gradual extension, interpreted as pressurization of the summit magma reservoir due to an increase in magma supply (Poland and others, 2012). This manifested as an upsurge in discharge from the vent active at Pu‘u‘ō‘ō at that time. The extent of the pressurization was expressed on May 24, 2007, by two M4+ earthquakes and a brief swarm of smaller events (Wauthier and others, 2013) associated with Kīlauea’s outermost caldera faults. The overall pressurization of the volcano’s magmatic system, however, was unaffected by this seismicity, and extension continued.

The pressurized condition of the volcano changed abruptly at 0216 (all times Hawaii Standard Time [HST]) on June 17, 2007, when magma from Kīlauea’s summit began to intrude the volcano’s upper East Rift Zone (fig.1; Poland and others, 2008). Increased seismicity caused nearly continuous ground shaking felt as far away as Volcano, Hawai‘i (fig. 1), about 8 km north of the East Rift Zone. The seismicity was initially centered between Pauahi Crater and Maunaulu but migrated downrift to Makaopuhi Crater over the next two days (fig. 1; Poland and others, 2008).

Borehole tiltmeters at the summit of the volcano recorded the onset of rapid deflation at about the same time as the onset of the earthquake swarm, and tiltmeters at Pu‘u‘ō‘ō began recording deflation a few minutes later (Poland and others, 2008). This was followed by slow subsidence of the floor of Pu‘u‘ō‘ō starting at about 0930 HST on June 17, and spatter cones on the crater floor and on the west and southwest flanks of Pu‘u‘ō‘ō collapsed as they were engulfed by gradually enlarging pits. The crater floor continued to fall over the following days, parts of it dropping about 80 meters (m) by June 21 (Poland and others, 2008; Orr, 2014). The magma supply to the active tube system was severed as the magma column withdrew, causing the Pu‘u‘ō‘ō eruption to cease.

New ground cracks developed in four areas from Pauahi Crater to Kānenuiohamo during the Father’s Day eruption (figs. 1–3), and just after sunrise on June 19, gas emissions were discovered emanating from two spots in the area of cracking that crossed the flank of Kānenuiohamo (fig. 4). Specifically, whitish gas emissions were streaming profusely from fissures that formed in 1969 on the north flank of the cone, and a robust plume of bluish fume was rising from its heavily forested northeast flank. Fissures had opened at this second location overnight, erupting sparse tephra and two small pads of lava (figs. 5, 6). A third area of barely perceptible fuming was also observed that same morning on the lava plain just east of Kānenuiohamo, but it was no longer apparent a few hours later, as the day warmed and humidity dropped. Fuming from the third area reappeared and increased dramatically in strength overnight, forming a small white plume when seen in the morning on June 20, whereas the June 19 fuming area at the 1969 fissures to the northeast was barely evident (fig. 4B). The June 20 fuming cracks emitted choking gases and were found to be too hot to approach closely, indicating that magma was close to the surface.

The Father’s Day eruption was followed by a period of repose at Kīlauea, during which there was no discernable eruptive activity. The Pu‘u‘ō‘ō eruption resumed 12 days later, on July 1, with lava erupting in the bottom of the Pu‘u‘ō‘ō (Orr, 2014). After a period of filling, new fissures opened on the east flank of Pu‘u‘ō‘ō on July 21, marking the start of a new episode of eruptive activity (Orr and others, 2022). The trace of the Father’s Day dike on the forested north flank of Kānenuiohamo became more apparent over the following years (fig. 5), as increased heat flow from the intrusion caused considerable tree die-off while leaving the understory ferns unaffected.

Methods

The field observations used in this report were conducted on the ground and during helicopter overflights. Activity at Pu‘u‘ō‘ō was also observed by a telemetered webcam on the north rim of the Pu‘u‘ō‘ō cone (Hoblitt and others, 2008). Another telemetered webcam was temporarily deployed on June 20 to the west flank of Maunaulu looking toward Kānenuiohamo.

Cracks associated with the Father’s Day eruption were identified on the ground by the exposure of fresh crack surfaces where they cut older pāhoehoe and tephra cover (figs. 3B, C) and sometimes duff where forested. Larger cracks were visible from the air (figs. 3A, D). Crack locations were marked by hand-held GPS devices (horizontal accuracy approximately plus or minus (±) 10 m) during foot traverses across the affected area. The mapped extent of cracking is incomplete, especially where these crack zones border forest, which is heavily vegetated with thick undergrowth in most of this area, and where they abut ʻaʻā flows, which have a rubbly carapace that does not preserve small cracks. Crack width and opening direction were measured for a subset of cracks (fig. 2; appendix 1). Crack lengths were not measured during the eruption because of insufficient time and resources during the ongoing volcanic crisis. At later times, it was no longer possible to locate the smaller cracks because they were either covered by soil or had lost their fresh appearance and could not be distinguished from older cracks. In general, the smaller cracks could be traced for up to a few tens of meters before they ended or intersected natural joints in the pāhoehoe surface. The largest cracks could be traced for a few hundred meters, though individual en echelon segments were generally as long as a few tens of meters.

In late June 2007, we re-measured five crack stations on the floor of Nāpau Crater (appendix 2) that had been installed shortly after an eruption there in January 1997 (Thornber and others, 2003). The stations were located along a major East Rift Zone crack system related to the 1997 eruption. Each crack station is marked by a set of three or four hardened-steel survey nails arranged in a network that spans the monitored crack. Distances between the nails, marked from the dimple at the center of each nail head, were measured to the nearest millimeter with a fiberglass surveying tape, allowing changes in crack width to be determined. Prior to our measurements in 2007, the crack stations had last been measured in 1999. During our 2007 survey, we found the survey nails to be heavily corroded, but the center of each nail head could still be approximated to within ±1 mm. A south-to-north foot traverse across Nāpau Crater to look for new ground cracks was also conducted.

The margins of the lava flows and the extents of the eruptive fissures were mapped on the ground using a hand-held GPS with a horizontal accuracy of ±10 m. Lava and tephra samples were collected by hand, after the eruption ended, from points along the fissures and lava pads (fig. 5). In most locations, tephra (spatter) was collected both from the top surface of the lava flows and from the forest floor beneath them. The density of the basalt erupted (appendix 3) was determined by measuring the weight of spatter and lava clasts in air and water following the approach of Houghton and Wilson (1989). Density was converted to vesicularity using a dense rock equivalent density of 2.75 grams per cubic centimeter (g/cm³) (Lesher and Spera, 2015).

Very low frequency electromagnetic energy in the 15–28 kilohertz (kHz) range is transmitted nearly continuously by a global network of high-powered military transmitters, producing a vertical electric field and tangential magnetic field. These components of the VLF wave are affected by local conductivity heterogeneities, such as those caused by shallow intrusive bodies, which can generate an electromagnetic contrast that is orders of magnitude higher in conductivity than the surroundings (Zablocki, 1978; Kauahikaua and others, 1998) that can last for decades (Davis, 2015). For a shallow, narrow conductor, such as a dike, the VLF wave generates an induced magnetic field detectable with a portable VLF receiver, which measures the elliptically polarized sum of the transmitted and induced magnetic fields. This technique has long been used in volcanic settings (for example, Zablocki, 1978; Jackson and others, 1987; Kauahikaua and others, 1996; Davis, 2015), for instance to estimate the location of new dikes (Zablocki, 1978). We used the technique here to locate the trace of the Father’s Day dike.

Results

Discussion

The location and orientation of ground cracks between Pauahi Crater and Pu‘uhuluhulu and near Kānenuiohamo define zones oriented approximately northeast–southwest. The northeast end of the crack zone between Pauahi Crater and Pu‘uhuluhulu is offset about 1 km north of the southwest end of the crack zone near Kānenuiohamo. This location suggests that the Father’s Day dike splayed as it approached the surface, perhaps along structural weaknesses, forming separate dike segments, as supported by geodetic modeling (Montgomery-Brown and others, 2010). Although the dike segment between Pauahi Crater and Pu‘uhuluhulu reached a level shallow enough to crack the overlying ground surface, it was still too deep to be detected using the VLF technique. The opening of new cracks in the eastern extension of the Koa‘e fault system could indicate that some magma was diverted there from the East Rift Zone, but it might also be the result of the seaward slip of Kīlauea’s south flank associated with the Father’s Day eruption (Montgomery-Brown and others, 2010).

The dike segment near Kānenuiohamo, however, did reach a level shallow enough to be detected using the VLF technique. Although it is possible that the Father’s Day dike near Kānenuiohamo approached the surface as a single curvilinear dike, we believe it more likely that the dike in this area splayed further and approached the surface along at least three en echelon segments as suggested by the VLF receiver results in concert with the locations of the fuming cracks and the eruption site (fig. 2).

The seismicity related to the Father’s Day eruption migrated down the East Rift Zone in four pulses (Poland and others, 2008) that may correspond to the four dike segments we identify. If so, the migrating seismicity suggests the dike segments formed sequentially from west to east and culminated with the dike reaching the surface at the point of lowest elevation near the dike’s eastern tip, where lava erupted. Magma rose close enough to the ground surface at the June 20 fuming area on the lava plain west of Kānenuiohamo to expose the top of the magma column to atmospheric pressure, releasing heat and volcanic gases but failing to erupt.

Conclusions

The June 17–19, 2007, Father’s Day eruption was the result of the forceful intrusion of magma from Kīlauea’s summit magma reservoir into the volcano’s upper East Rift Zone. The intruding dike splayed as it migrated downrift and ascended, forming at least four separate dike segments that caused cracking of the overlying ground surface. Measurement of very low frequency magnetic fields using a handheld very low frequency receiver, along with the location of outgassing cracks and eruptive fissures, help to define the traces of the eastern three dike segments near Kānenuiohamo. The Father’s Day eruption culminated in a brief (approximately 30 minutes long) eruption early in the morning on June 19, discharging two small pads of lava totaling about 1,525 cubic meters of lava near the eastern end of the dike. Adjusting volume based on the average vesicularity of eruptive products, which is 52 percent, equates to a very low dense-rock equivalent eruption rate of about 0.4 cubic meters per second.