CHAPTER 3: The Study Area Geology The pit crater Kilauea Iki and the area devastated in 1959 are adjacent to the northeastern rim of the Kilauea volcano caldera in Hawaii Volcanoes National Park on the island of Hawaii (Fig. 1). The whole area is nearly 1200 m in elevation. Kilauea is a shield volcano that rises approximately 6100 m from the ocean floor and 1239 m above sea level. Its summit area consists of a caldera that is 4 km long and 3.2 km wide. This caldera formed by subsidence when magma was withdrawn from beneath the summit (Stone 1926). The main vent of Kilauea is a collapsed crater in the floor of the summit caldera, called Halemaumau. Until the 1959 eruption, the volcano had erupted about once every 4 years. A chronology of the historic eruptions for Kilauea and Mauna Loa has been published by Macdonald and Hubbard (1966). Macdonald (1962) reports that two historic eruptions have occurred in or adjacent to Kilauea Iki prior to the 1959 eruption. In 1868 a fissure opened in the southwest wall and lava poured into the crater. In 1832 there was an eruption on Byron Ledge, a rim that separates Kilauea Iki from the main caldera. At that time, lava poured into both Kilauea caldera and Kilauea Iki. The most common lava of the Hawaiian volcanoes is olivine basalt (Macdonald and Katsura 1961).
The December 1959 Eruption Wentworth (1966) described the devastating character of the eruption. As lava poured into the crater from a rift in the crater wall, it consumed the ohia (Metrosideros) forest on the floor and all vegetation on the crater wall up to 115 m from the rim. Prior to the 1959 eruption, Kilauea Iki crater was approximately 250 m deep. After the eruption, the lava lake in the crater rose to a depth of 140 m. Heat from the rising lava lake soon superdried the remaining vegetation and numerous fires occurred which burned vegetation in strips on the upper sides of the crater wall. The devastated vegetation took on a different character where the pyroclastic material fell. Wentworth observed that during the short violent phases of fountaining, cinders and cobble-sized chunks of lava accumulated at the southwestern side at the rate of several decimeters in depth per hour. These ejecta buried a considerable area of forest and stripped the trees in the leeward and surrounding area of their leaves, bark, and small branches. Wentworth recognized four major belts of destruction as the pumice and spatter blanket developed. In the first belt, southeast of the vent, there were no remains of the former vegetation because it was covered with a deep mantle of pyroclastics. This is the area of the pumice cone and its surroundings. Next to this was an area of tree molds, where many trees exploded from overheating and were then consumed by fire. The tree molds are trunk-sized imprints in the spatter blanket that were formed through sudden chilling of the glowing pyroclastic fallout, where it surrounded tree trunks in this zone. This tree-snag zone was replaced at a further distance by another zone where the trees were severely excoriated from fallen pumice and clots of lava, but showed signs of recovery in a few weeks. Their trunks became heavily covered with a new growth of leaves. Within weeks after the eruption, these surviving Metrosideros trees showed abundant flowering. Although Wentworth reported fume damage to some of the vegetation outside the devastated area, he considered the damage minor as compared to past eruptions. However, the fume was surprisingly acid at least locally. Murata (1966) found the steam condensate from the cinder cone during the eruption to have a 1.9 N HCI concentration. Climate The climate of the study area is tropical montane. This implies that the temperature is moderately cool throughout the year. The mean monthly air temperature values remain below 20°C throughout the year, and may vary little, between about 14°C in the winter to about 17°C in the summer. The mean annual temperature at Hawaii Volcanoes National Park Headquarters (near 1200 m elevation) is 15.9°C (Doty and Mueller-Dombois 1966). The park headquarters station is about 2 km northeast of Kilauea Iki, at approximately the same elevation. Frost has never been recorded in the meteorological shelter, but minimum temperatures as low as 3-4°C are not uncommon in February. However, ground frost occurred in December 1969 in all lower lying areas at the Kilauea summit level. The mean annual rainfall at Kilauea Iki is at least as high as at the headquarter's station, for which long-term records are available. The 24-year mean annual rainfall at the headquarters is 2400 mm. For the years 1967 and 1968 the mean was 2873 mm, when the record for the same period in Kilauea Iki crater showed 3280 mm (Fig. 3). Annual rainfall at the headquarters has varied from as high as 4350 mm in 1918 to as low as 1210 mm in 1908. These are the greatest extremes recorded since 1899. Inspection of the annual records (Doty and Mueller-Dombois 1966) shows that during 67% of the years the values are between 2000 and 2500 mm. Similar annual rainfall variations can be expected at Kilauea Iki. Both areas are surrounded by Metrosideros-Cibotitm rain forests. The mean monthly rainfall at the headquarters station is in excess of 100 each month except June, when it may drop to 90 mm. A gradient of about 1000 mm drop in annual rainfall was found to occur from Kilauea Iki to the south end of the study area in the upper Kau Desert. Here, in the upper Kau Desert, the 2-year mean for 1967-68 was 2203 mm as shown in Fig. 3. In accordance with the 2-year and long-term mean difference at headquarters, one may subtract about 500 mm to obtain a long-term mean rainfall of about 1700 mm in the upper Kau Desert at the end of transect AA' (Fig. 2). The long-term mean at Halemaumau crater was 1300 mm (Doty and Mueller-Dombois 1966). Here a summer drought prevails from June through September, which appears to occur also in the southern study area, although of somewhat lessened intensity. Thus, the study area extends from a montane tropical rain forest climate in the north (Kilauea Iki) to a montane tropical seasonal forest climate in the south (upper Kau Desert). This is indicated also by the original vegetation type boundary as shown in Fig. 2.
The Original Vegetation The vegetation that was present before the 1959 eruption had never been sampled directly. However, a vegetation map at the scale of 1:12,000 was prepared for the whole park area (Dory and Mueller-Dombois 1966). This map was based on air photos taken in 1954. Three major vegetation types were recognized in the study area before the eruption. They form a sequence from north (Kilauea Iki crater) to south (Kau Desert) along the rainfall gradient previously discussed. 1. Closed Metrosideros-Cibotium forest [symbol cM(C) in Fig. 2]. This is the dominant rain-forest type in the park, which occurred formerly in the northern part of the devastated area. The forest is characterized by an upper, almost pure, layer of Metrosideros polymorpha trees and a shrub layer, dominated by Cibotium glaucum tree ferns. Another tall fern, Sadleria cyatheoides, is commonly associated with Cibotium in this area. Scattered small trees and tall shrubs include Myrsine lessertiana, Ilex anomala, Gouldia terminalis, Vaccinium calycinum, among others. The herbaceous undergrowth is usually thin and includes Briza minor, Isachne distichophylla, Gahnia gahniaeformis, Hedyotis centranthoides, and Cibotium seedlings. A more complete description is given by Mueller-Dombois (1966) and Newell (1968). 2. Closed Metrosideros with native shrubs [symbol cM(ns) in Fig. 2]. Cibotium is absent and the most common shrubs are Styphelia tameiameiae, Vaccinium reticulatum, Dodonaea viscosa, Dubautia ciliolata, Wikstroemia sandwicensis, and Coprosma ernodeoides. The tall fern, Sadleria cyatheoides, is often present also in this type, but the shrub composition is otherwise quite different from the first described rain-forest type. This forest joins the Metrosideros-Cibotium forest where rainfall becomes more seasonal. (The boundary is indicated in Fig. 2.) 3. Open Metrosideros forest with native shrubs and Andropogon grass species [symbol cM(ns-A)]. The two Andropogon species are the introduced A. virginicus and A. glomeratus, which occupy a dominant position in the herb layer. Otherwise, this vegetation type is floristically similar to the above. The Andropogon grass species were a minor part of the vegetation in 1954; however, when field checks of the 1954 aerial photos were made in 1966, these species formed a conspicuous component. The New Volcanic Habitats A survey of the devastated area after the eruption revealed several different habitats. It was reasoned that these could not be lumped together at the start, but would have to be studied individually to evaluate effectively the revegetation processes at Kilauea. There were areas where all the original vegetation was completely buried by new volcanic materials, with no residues of organic matter left at the surface. Other areas showed dead standing and fallen trees. These were in part blanketed with different kinds of pyroclastics (spatter and pumice). Still other areas showed plants that survived the ash fallout. Superimposed on these variations in volcanic destruction was the climatic gradient from humid to summer drought. In the latter climate, vegetative growth had always been sparse and of desert-like character. The area is known as the upper Kau Desert. Yet, the desert character of this area is not attributable to the summer-drought climate, but rather to certain peculiarities of the substrate, which will be discussed later. The habitats were classified according to the presence or absence of remnant vegetation and the associated volcanic substrate. It was assumed that differences in succession could be expected in habitats free of organic residues and in those where dead or living plant materials remained. Following is a description of the six habitats recognized. Their geographic outlines are mapped in Fig. 2. The habitats are sketched in profile on the two diagrams (Figs. 3 and 4).
Habitat 1. Massive lava with joint cracks (56 ha) This habitat includes the lava lake and high slump scarps on the Kilauea Iki crater floor (Fig. 3, segment 1). The lava lake consists of massive pahoehoe lava, with many joint cracks and crevices. After the eruption, this area was completely barren of life. Many crevices became fumaroles as rainwater permeated the lava lake's thin crust and came in contact with the hot interior. The wall on the northeast side of the crater rises to 115 m. From this direction the prevailing trade wind carries some dust and organic materials (leaves, seeds, etc.) from the adjacent rain forest onto the crater floor. The major disseminule supply can be considered to come from the surrounding undevastated closed Metrosideros-Cibotium rain forest. Habitat 2. Cinder cone (19 ha) Included here are the summit and sides of the cinder cone Puu Puai which is 46 m high above the crater rim (Figs. 3 and 4, segment 2). The cone consists of pyroclastic materials that vary from light-weight, glassy, pumiceous particles to thick, heavy, welded masses of spatter. The hot interior has steamed continuously for several years, and at the summit area large amounts of secondary minerals have been deposited. The summit has numerous deep fissures, and on its northeast side, large slough-offs have occurred. Only the southwestern part of the summit and slopes exhibit stability. The climate here is similarly humid as in habitat 1. However, the trade winds blow strongly and continuously over the cone. During some periods, winds of 65-75 kph have been measured. Similar to habitat 1, the cone habitat consists entirely of barren volcanic materials. Its major disseminule source can also be considered to come from the northeast from the closed Metrosideros-Cibotium rain forest. Habitat 3. Spatter area with tree snags (6 ha) The surface substrate consists of thick welded spatter (a form of pyroclastic material with a density>1) which piled up to a meter or more around the forest trees, killing most of them (Fig. 4, segment 3). Where the spatter was thin at the margin of this fallout, a few trees survived. They later resprouted from buried trunk bases or root systems. The rainfall on this habitat was about 3000 mm per year for the years 1967-68, similar to habitats 1 and 2 (see Figs. 3 and 4). The position of habitat 3, next to a relatively undisturbed, species-rich closed Metrosideros-Cibolium forest, provided for a more abundant supply of disseminules than those arriving in habitat 2. A board walk for park visitors was constructed through habitat 3. This also may have contributed to an occasional arrival of man-carried disseminules. Many of the dead trees in habitat 3 remained standing. They acted as local interceptors of wind-driven rain. In this manner, they provided favorable microhabitats for pioneer plants and animals. Habitat 4. Pumice area with tree snags (31 ha) The pumice depth varied from approximately 12 m to 30 cm in this habitat. Pumice is defined as pyroclastic material with a density <1. Much of this area is on the leeward side of the cinder cone (Fig. 3, segment 4). This resulted in some protection from the prevailing winds. Rainfall is less over this habitat compared to the other three (see rainfall gradient indicated in Fig. 3). Here, the original forest was deeply buried, so that only dead trees remained with their top parts above the ash blanket. They seldom fell under the wind pressure except at the eastern and western edges of the habitat where the ash blanket was thin. This habitat probably received a less abundant supply of disseminules than the other three habitats because of greater distance from the surrounding undestroyed vegetation. In several places along the deepest part of the fallout (Fig. 3, transect AA') in this habitat, and that of habitat 5, several small crater-like slumps occurred where ash filtered, after the blanket had formed, into earthquake cracks in the old substratum. These areas of subsidence vary from one to several meters in width or diameter and they can be up to 5 m deep. Habitat 5. Pumice area with surviving trees (125 ha) The pumice layer varies from approximately 3 m to 30 cm in depth along transect AA' (Figs. 2 and 3). Here, nearly all Metrosideros trees survived. Because it is lee of the cone and slopes gently in a southwesterly direction, this habitat is somewhat protected from the prevailing wind. However, it receives greater insolation in the lower sectors because of decreased cloud cover. The joint boundary with habitat 4 across transect AA' coincides approximately with the mapped boundary between the original rain and seasonal forest (Fig. 2). Habitat 6. Thin fallout area (263 ha) Here, in the upper Kau Desert, the pumice begins with a depth of 30 cm and decreases at he south end of the fallout area to 2.5 cm or less (Fig. 3). Because of the thin cover of pumice with a density of <1, the pumice particles maybe moved about easily during heavy showers. This movement is further facilitated by a partially cemented ash-crust that formed the former habitat surface. The partial cementation of this former surface was the result of phreatic explosions at Halemaumau (Powers 1948). The substrate is thus very unstable. The pumice is of finder particle size than in other habitats and is made up largely of fine, black, glass-like fragments and some long, thin threads of glass, called Pele's hair. The vegetation is characterized by widely scattered, short Metrosideros trees and native shrubs. The woody plants grow in cracks and crevices and in deposits of the hard-crusted ash. A cloud cover seldom extends to this habitat, while cloud cover is a regular phenomenon over habitats 1 through 5. Therefore, habitat 6 receives a larger amount of sunshine than the others. Many deformed shrubby Metrosideros trees are bent toward the southwest away from the direction of the trade wind that blows from the summit area of Kilauea.
chap3.htm Last Updated: 1-Apr-2005 |