Invasion and Recovery of Vegetation after a Volcanic Eruption in Hawaii
NPS Scientific Monograph No. 5
NPS Logo

Discussion (continued)

Probable Reasons for Observed Life-Form Establishment Sequences

The recorded establishment sequences of eight life-form groups are summarized by habitat in Table 15. the term "establishment" rather than "arrival" is used here because there were two cases when a herbaceous seed plant species arrived and then disappeared. Such transient colonizers are not included in Table 15 and the following discussion.

TABLE 15. Summary of establishment sequences of plant life forms in the six habitats. Consistent patterns are blocked out.

Life formHabitat
123 456

Algae1st1st1st 1st1st1st
Lichens2nd2nd2nd 2nd2nd2nd
Native woody seed plants3rd3rd 2nd1st2nda1sta
Exotic woody seed plants4th4th 1st1st1st3rd
Grasses or sedges5th4th 1st3rd1st2nd
Forbs4th1st 1st1st3rd

aSeedlings of surviving woody plants.

Algae were always the first to become established on the new volcanic surfaces. Lichens were never the first; they arrived consistently as the second life form. Mosses and ferns became established either first, along with the algae, or they arrived as the second cryptogamic life-form group together with the lichens. On the two habitats with no remains of an earlier vegetation (i.e., on the crater floor, 1 and the cinder cone, 2), native woody seed plants were the third life form to arrive and exotic woody and herbaceous plants were the last. On habitats 3, 4, and 5, which showed remains of the former Metrosideros forests, exotic woody and herbaceous plants were among the first life forms to become established.

The consistent early arrival of algae supports Doty's (1967b) previous observation on the 1955 Puma lava flow. On the Kilauea crater floor, algae colonies were observed to occupy both the crevices and surfaces of the pahoehoe lava sheets. On the latter, there was no other macroscopic plant life form until arrival of the lichens. During the entire observational period, the lava rock surfaces were occupied only by these two life forms and there was no obvious interaction with any of the other plant life forms.

The reason for the early arrival of algae was not investigated. But, as suggested already by other investigators (Treub 1888; Daly 1961, 1967b), they can probably grow on the volcanic surfaces when organic nitrogen levels are still too low for other plant life forms. As indicated on the crater floor, the algae also seem to have a greater tolerance to high temperatures than the other life forms, except the lichens.

The reasons for the consistently later arrival of lichens is worth a thorough investigation. The explosive, nondirectional invasion pattern of the lichen Stereocaulon volcani suggests that this lichen may be formed on the new lava surfaces in situ by the joining of an alga and fungus. Observations on a number of recent lava flows show a generally uniform distribution pattern of Stereocaulon. If this lichen was established from small wind-carried pieces of lichen thalli, one would expect some indication of a directional invasion pattern related to the location of a major disseminule source. Moreover, of the lichen-initials observed under the microscope, all had the same tiny (1-2 mm) globose thallus. One would not expect wind-transported thallus fragments lobe of such uniform structure. Since lichens are known to be extremely sensitive to toxic substances, their delayed arrival may also be linked to the removal by rain water of lichen inhibitors from new, raw volcanic materials. Such materials could, for example, be traces of sulfurous precipitates. In contrast to algae, mosses, and ferns, lichens were never found at fumaroles or in the pathway of vapor-steam.

The generally subsequent arrival of mosses and ferns after the algae may be only an artifact, insofar as both mosses and ferns develop from gametophytes. The protonema of mosses look like algae and thus were not separately recognized. The fern gametophytes are also easily overlooked. Only the sporophyte generations of mosses and ferns were recorded in this study.

It is interesting that native woody seed plants were the third life-form group to arrive on the totally new primary habitats (1 and 2). They are typically sclerophyllous plants (i.e., Metrosideros, Dubautia, etc.) that indicate a xerophytic adaptation. These native woody plants can probably survive under more severe water stress conditions than the exotic woody plants that appeared as the main pioneers (Buddleja and Rubus species) in habitats 3, 4, and 5. This applies also to the herbaceous plants (grasses and forbs) that were among the last major life forms to arrive on these totally new materials. Somewhat more mesic conditions develop slowly on the new lava flows by entrapments of chipped-off rock flakes, dust, and litter in the crevices. This observation was stressed particularly by Eggler (1941, 1971).

In contrast, the exotic woody and herbaceous plants were among the first invaders on the habitats that contained remnants of the former Metrosideros forests. As explained before, here the snags provided locally more favorable moisture conditions that made it possible for these aliens to arrive on the otherwise xeric substrates.

The sequence of life-form establishment, as shown in Table 15, does not comply with Clements (1916) records of life-form invasion sequences. His observations have dominated textbooks for many years. Clements gave his own observations and quoted those of several other investigators, stating that lichens were the first life forms on xeric rock substrates. These were to be followed by mosses and these, in turn, by herbaceous plants. Woody plants were not recognized in the early pioneer sequence. Moreover, Clements recognized these life-form sequences as real vegetation types or seral stages. However, he also stated (1916:84) that, on sedimentary rocks in climates with moist growing seasons, the pioneers are mostly mosses and liverworts that are often preceded by algae. Yet, he maintained that plant life begins with crustose lichens on igneous rocks. It must be remembered that all of Clements observations relate to the temperate zone, and differences are to be expected. For example, the presence of a fern stage seems to be uniquely tropical (see also Keay 1959). Ferns and lichens dominate the cryptogamic stage in Hawaii.

Table 15 shows that an algal stage may be recognized, but thereafter, there is no single life-form stage. One may speak merely of a cryptogamic stage. It is also significant that on materials without remnants of a former vegetation, woody plants appear definitely before herbaceous plants. This correlates with the pattern of evolution of these life forms, although there may be no logical relationship between present-day primary invasion and plant life-form evolution in the phylogenetic sense.

On new volcanic surfaces in Hawaii, one may, therefore, recognize four stages:

1. An algal stage, which may form the sole stage for 1 year in the humid climate.

2. A cryptogamic stage, which may remain as such for 2 additional years.

3. A native woody seed plant stage, which includes the former life forms, and which may last 1 more year.

4. A stage in which exotic woody and herbaceous plants become associated.

This applies only to new volcanic surfaces of lava or ash without remnants of a former vegetation and to humid climates. No such stages could be observed on the volcanic materials with former vegetation remnants. Here exotic seed plants were among the first pioneer life forms.

Factors Influencing Plant Survival

Depth and nature of pyroclastic deposit

As expected, the survival of original plants decreased with increasing depth of the pyroclastic deposits. But there were great differences in survival with respect to the kind of pyroclastic deposit. Metrosideros trees survived under a pumice blanket slightly deeper than 2.5 m in habitat 5. In habitat 3, all Metrosideros trees were severely damaged by the glowing-hot spatter that became welded upon deposition. Where the original surface was covered with less than 10 cm of spatter, trees recovered fast, but only few survived of those buried under spatter deeper than 10 cm. In this habitat, surviving trees resprouted from the base only and recovery occurred merely up to a spatter-depth of about 50 cm. As mentioned previously, the basal sprouting was probably possible because of small spaces that developed at the contact zone between the spatter and the damaged basal stem. This permitted water penetration from stem runoff into the tree molds and gaseous exchange. In habitat 3, none of the surviving trees showed aerial roots.

Instead, aerial root development was seen only on survival trees in habitat 5 that were buried under pumice deeper than 20 cm. The significance of this aerial rooting is as yet unknown. If it were really essential for survival, one could expect it to have occurred on all reflushing Metrosideros trees that were buried under ash deeper than 50 cm, particularly on those surviving under the deepest deposits > 2.5 m.

Size of plant

Trees that survived in the spatter area had relatively large basal diameters of 20 cm or more. An important survival factor associated with size was probably an epiphytic moss layer surrounding the stem base. Also, under the pumice only the larger trees survived the deeper deposit as seen in Figs. 12.1-12.4. Smaller diameter Metrosideros trees survived under the shallower pumice blanket. In habitat 6, all individuals were small and all survived. Eggler (1948) found on El Paricutin that pine trees of intermediate size survived best under an ash blanket. This observation does not necessarily contrast with the one here where the largest trees were seen lobe the better survivors. Very large and old Metrosideros trees did not occur in the study area. Such individuals may have a decreased survival capacity. However, no prediction can be made from the present observations.

Vegetative regrowth capacity

The observation that all surviving trees in the area recovered only when their trunks were buried to less than half their height stands in contrast to the survival capacity of the other plant life forms—the shrubs and herbs that survived the burial.

Several of the native shrubs and herbs survived under the pumice layer even where their entire shoot system had been broken off or buried. This was particularly true for the native shrubs Vaccinium reticulatum, Dubautia scabra, Styphelia tameiameiae, and Coprosma ernodeoides. But this included also a few shrub-like individuals of Metrosideros polymorpha in the upper Kau Desert (habitat 6). Vegetative regeneration from a basal remnant or root system is probably a characteristic for most shrubs. But in the Hawaiian forms mentioned, it may also be an evolutionary adaptation to this kind of volcanic damage. Among the herbaceous life forms, most of the survivors were plants with underground storage organs. These plants were defined as geophytes although they do not necessarily reduce their shoot systems in a seasonal rhythm as do the typical temperate- and arid-zone geophytes. This group of damage-adapted geophytes included native species, such as Astelia menziesiana (which grows usually as an epiphyte in older rain forests), Dianella sandwicensis, Polypodium pellucidum, Pteridium decompositum, as well as several exotics, Spathoglottis plicata, Hedychium coronarium, and Tritonia crocosmiflora. Among the herbaceous survivors were also a few chamae-hemicryptophytes, such as the native sedges Machaerina angustifolia and Gahnia gahniaeformis. These plants survived usually by regrowth from remnant shoot systems.

Primary Community Formation

According to Poore (1964), most ecologists agree that to qualify as a community a stand of plants must show some form of integration. The kinds of developing integration observed on the new volcanic habitats may be discussed under two subheadings: plant aggregation, and complementation and competition.

Plant aggregation

Initially, invading plants tended to occur as scattered individuals. However, as soon as these first individuals were established, other plant individuals tended to become established next to them. Therefore, aggregation was noticed as an early pattern of plant invasion. This is quite understandable in view of the presence of favorable microhabitats that occurred in all but perhaps habitat 5. On the crater floor (habitat 1), the joint cracks and crevices were occupied by algae, mosses, and ferns. Soon after one fern individual was noticed, others were seen in the same crevices. Often it was difficult to distinguish fern individuals of Nephrolepis exaltata, because new fronds developed relatively quickly from the extending rhizomes. It seemed peculiar in many instances that one crevice was crowded with algae, mosses, ferns, and native woody seed plants, while another adjacent crevice, which looked identical in all aspects, remained vacant. Such vacant crevices were still found in year 9, while others were crowded with individuals of all the above-mentioned life forms. This phenomenon seems to indicate that a favorable event must have facilitated the establishment of a pioneer individual or group. These, in turn, through their own establishment, must have improved the moisture relations in the crevice so that other plant individuals, particularly seed plants, found an easier entrance to the microhabitat. However, Metrosideros seedlings were observed to grow without other plant life forms in certain crevices where rock-flakes had accumulated. This showed that the presence of cryptogams was not necessary for seed plants to become established. What was necessary was merely a somewhat improved retention of precipitation water in the crevices to satisfy the minimum water requirements for establishment of native seed plants.

Plant aggregations developed in habitat 3 at the tree molds, and larger-sized aggregations developed around the snags and surviving trees. A similar snag-associated aggregation pattern developed in habitat 4. In habitat 6 (Kau Desert) plant aggregation was also quite pronounced around surviving Metrosideros trees and shrubs. The environmental relations in these tree- and snag-associated microhabitats have already been discussed.

Incipient plant aggregations began to develop in some of the cinder cone crevices, although here and in habitat 5 (pumice-with-survival-trees), aggregation patterns were least clearly defined during the period of observation. Nevertheless, it can be said that plant aggregations are typical for the invasion of plants on new volcanic surfaces. This statement is in accord with the observations of Millener (1953) in New Zealand. In all cases, the cause for this early community formation through aggregation seemed to be associated with locally improved edaphic moisture relations. These became effective either through improved retention of soil water, shading, fine particle accumulation, or plant material itself or through locally higher input of water (interception at snags or accumulation of runoff water) or both.

Complementation and competition

The accommodation of seed plants in crevices densely occupied with cryptogams may be viewed as a form of complementation. This form of complementation probably was brought about by an improved runoff-water retention in the established moss mats and among the fern rhizomes. Another form of complementation was observed under Buddleja trees and other shrubs in habitat 3, where several species of rain forest mosses invaded the shaded areas beneath these bushes. Moss invasion in the shade of trees occurred also in some places in habitat 5, under the surviving Metrosideros trees. A third form of complementation was observed in habitat 1 and 3. Here, a herbaceous seed-plant synusia, consisting of grasses and forbs, was spreading under established scattered woody seed plants. This occurred in habitat 1 only at the first 10-15 m inward on the crater floor of transect a. This was also the only place where seed plants had invaded the flat surface of the massive pahoehoe lava. The area was as yet very small in terms of the total crater floor area. The main reason for this form of invasion was probably not the partial shade provided by the scattered woody plants, but rather the morning shade provided by the crater slope rising steeply behind this location. A more advanced development of synusial layering occurred in the spatter-with-snags habitat (3) where the grasses, such as Paspalum dilatatum and Holcus lanatus, advanced in the partial shade of woody plants.

Here also the first signs of competition were noted among the grasses and forbs. Paspalum dilatatum and Setaria geniculata seemed to decline because of the advancing Holcus lanatus, Cyperus polystachyos, and Pennisetum clandestinum. Eupatorium riparium and Lythrum maritimum seemed to be replaced in part by Commelina diffusa. Similar replacement patterns were noted with basally resprouting Metrosideros and Buddleja asiatica. A decline of Rubus rosaefolius and R. penetrans was observed in habitats 3 and 5, while Vaccinium reticulatum and Dubautia scabra increased in spread. It could not be ascertained in all cases whether this replacement pattern was in fact competition. But in habitat 5, several cases were observed where the native Coprosma ernodeoides clearly had invaded patches occupied by the exotic Rubus penetrans, with the result that the latter declined in vigor and then died.

Relationship Between Native and Exotic Invaders

The study has shown that both native and exotic seed plants and ferns participated in the invasion on new volcanic surfaces in Hawaii. New volcanic surfaces without remains of former vegetation were clearly dominated by native species in the early stages of primary invasion. In contrast, pyroclastic fallout habitats with forest remnants (such as habitats 3 and 5) were dominated initially by exotic invaders. However, even in these habitats, native woody seed plants and ferns were in no way affected by competition from exotics, but there were definite indications that exotic seed plants were replaced by competition from native seed plants of similar life form. This was observed in the case of the exotic small-tree Buddleja asiatica, which seemed to be affected by competition from Metrosideros polymorpha snag-resprouts. It was also observed in the case of Rubus penetrans that was replaced in several locations by the native woody creeper Coprosma ernodeoides, which had a life form similar to R. penetrans on these pumice substrates.

The majority of exotic species were herbaceous seed plants, grasses and forbs. Only very few native species are in this life-form group. Thus, the herbaceous exotics fill a practically vacant niche, It is expected that these herbaceous exotics will become even more abundant in time in the continuing process of invasion. However, they do not appear to interfere with the development of native plants, because both are complementary life forms. An early succession was recorded in this group, some of which seemed to be caused by competitive replacement. This form of interaction is expected to increase as the new surfaces become filled more and more with plant life.

<<< Previous <<< Contents >>> Next >>>

Last Updated: 1-Apr-2005