A decade of recovery of understory vegetation buried by volcanic tephra from Mount St. Helens.
Ecologists have formally studied vegetation change for a century but have developed few widely accepted generalities about it (e.g., McIntosh 1980, Pickett et al. 1987, Glenn-Lewin and van der Maarel 1992, McCook 1994). This failure to find principles regarding succession may be due to insufficient information, to inconsistencies among subdisciplines in the scale and definitions used, to repeated attempts to generalize from studies of a single type of system, and to lack of agreement about the usefulness of different approaches to the study of vegetation change (McIntosh 1980). It is likely, also, that the complexity of vegetation change can account for much of the inability to generalize (McIntosh 1980, Glenn-Lewin and van der Maarel 1992).
Understanding vegetation change is contingent on understanding the effects of the disturbances to which a site is subjected. Disturbances are frequent in many systems and vary greatly in intensity and spatial scale (White 1979, Sousa 1984, Glenn-Lewin and van der Maarel 1992, Attiwill 1994). Local disturbance and patch dynamics influence stand structure and composition, even in fairly stable forests (Pickett and White 1985). Moreover, different types of disturbance can have different impacts (Mcintyre et al. 1995). Disturbances often interact in complex ways with existing vegetation, and conditions at the time of a disturbance can have long-lasting effects (e.g., Webb 1989). Describing how a specific disturbance interacts with existing vegetation at a small spatial scale can increase understanding of both direct effects of the disturbance and subsequent successional changes.
Most plant communities contain more than one stratum, and many strata include more than one growth form. Strata, growth forms, and their constituent species are likely to differ in their response to disturbance and their behavior during succession (Pickett et al. 1987, Falinski 1988, Halpern 1989, Glenn-Lewin and van der Maarel 1992), which complicates understanding vegetation change. A change in any stratum in complex vegetation constitutes vegetation change (Glenn-Lewin and van der Maarel 1992). Forest strata have been studied as separate ecological entities (e.g., synusia, union) since early in this century (Shimwell 1971: 57-58). In Pacific Northwest forests, a single canopy type occurs with a variety of understory types, by which communities are recognized (Daubenmire and Daubenmire 1968:51-52, Franklin and Dyrness 1973).
Here we describe understory vegetation change during the decade following burial by tephra (aerially transported volcanic ejecta) produced by the 1980 eruption of Mount St. Helens (Washington State, USA), for stands that had a surviving tree canopy but differed in the degree of disturbance and in vegetation composition. We interpret differences in vegetation change both within and among sites in terms of initial disturbance and site conditions, plant growth form, environmental variability within forest stands, and the interrelationships among these factors. In addition, we compare different measures of importance and different ways of quantifying vegetation recovery.
Tephra deposition has seldom been included in discussions of ecosystem disturbance (e.g., White 1979, Hemstrom and Franklin 1982, Sousa 1984), although it differs from other disturbances (Antos and Zobel 1987). Airfall tephra has no predictable season or topographic location for its occurrence, unlike fire, windstorm, chemical deposition (salt or air pollutants), and burial by alluvium, sand, or loess. A tephra deposit may affect very large areas (Shipley and Sarna-Wojcicki 1983) more uniformly than other disturbances, with its depth, texture, and chemistry varying gradually with distance from the volcano. The substrate is sterile and low in nitrogen and phosphorus, unlike alluvium and dust. Prompt erosion can remove its consequences, similar to other types of burial, but different from most disturbances. Timing and intensity of the tephra deposit are not affected by people or by properties of the ecosystem, although degree of plant burial can be modified by the presence of pre-eruption vegetation and litter. Plants differ in their ability and mechanisms to cope with burial (White 1979, Zobel and Antos 1982, 1986, 1987a, 1992, Antos and Zobel 1985a, b, 1987, Gecy and Wilson 1990, Sykes and Wilson 1990, Allison 1995), as they do for other disturbances.
Tephra deposition is widespread and relatively frequent in the Cascade Range forests of the Pacific Northwest, and in many other parts of the world (Sheets and Grayson 1979). In the Cascades, 12-22 volcanic events in the last 13 000 yr produced as much tephra as the 1980 eruption (Shipley and Sarna-Wojcicki 1983, Mullineaux 1986), three of those tephra deposits affecting our study area within the life span of the old-growth trees (Waitt and Dzurisin 1981). Although the 1980 eruption of Mount St. Helens destroyed all forest within an area of 61 000 ha (Means et al. 1982), over a much larger area the forest canopy remained intact but the understory was buried by tephra. In subalpine forests, few plants had started to grow on 18 May, the eruption date, so tephra buried the buds of all species with annual shoots, as well as small evergreen and woody plants.
To examine the effects of tephra burial on the forest understory stratum, we used permanent plots in which we monitored vegetation change and environmental factors for a decade following burial by tephra of different depths, and on sites with different initial species compositions. We removed tephra to allow estimates of pre-eruption understory composition, to mimic natural erosion, and to determine the effects of the delay of erosion on vegetation recovery. And we examined responses to burial of major growth forms and species.
We previously reported vegetation change from 1980-1982 in permanent plots, both on unmodified [TABULAR DATA FOR TABLE 1 OMITTED] tephra and with tephra removed (Antos and Zobel 1982, 1985c, 1986). Our primary conclusions include: (1) Tephra 15 cm deep killed most understory plants, but tephra 4 cm deep hardly affected most herb and shrub species. (2) Many mosses emerged from 2 cm of tephra burial, but not from deeper tephra. (3) The degree of damage to shrubs and small conifers was primarily a function of snowpack during the eruption: where woody plants were flattened beneath a snowpack, even 4 cm of tephra deposited on snow buried and severely damaged a 1.5-m-tall shrub layer. We have also reported the morphological responses of surviving species to burial (Antos and Zobel 1985a, b, Zobel and Antos 1982, 1987a, b); the longevity of buried plants (Zobel and Antos 1986, 1992); the process of conifer seedling development (Antos and Zobel 1986, Zobel and Antos 1991b); and variation in tephra properties (Zobel and Antos 1991a). Here we (1) summarize the decade of understory vegetation recovery after the eruption, (2) report the components and sources of that vegetation change, and (3) evaluate how vegetation change varied with local environment, tephra depth, initial snow cover, erosion, and plant growth form. We integrate our studies of tephra properties, the autecology of the species, and our repeated descriptions of vegetation on permanent plots, in order to determine the mechanisms that might account for the pattern of understory recovery from the 1980 eruption of Mount St. Helens.
We established permanent plots at two pairs of sites (Table 1) in the Cascade Range northeast of Mount St. Helens, Washington, USA (46 [degrees] 12 [minutes] N, 122 [degrees] 12 [minutes] W). The pairs of sites differed in the depth of tephra ([approximately equal to]15 and 4.5 cm) deposited by the 18 May 1980 eruption. One member of each pair had a more diverse herb layer prior to the eruption than did the other member; at both herb-rich sites, snowpack was prevalent on 18 May while it was sparse at the herb-poor sites (Table 1). Site designations (DP, DR, SP, SR) are based on these differences: D = deep and S = shallow tephra, R = herb-rich with snowpack and P = herb poor with little snowpack. (In previous papers [Antos and Zobel 1985c, 1986, Zobel and Antos 1991a], the four sites were designated as DP = E or EPX, DR = F or EPH, SP = B or CLX, and SR = C or CLH.)
Vegetation represents the upper limits of the Abies arnabilis zone of Franklin and Dyrness (1973). The old-growth tree canopy at all sites was dominated by Tsuga heterophylla, T. mertensiana, and Abies amabilis; in addition, Charnaecyparis nootkatensis was important at site SR and present at site DP, and Pseudotsuga menziesii was important at sites DP and SP. Large canopy trees of shade-tolerant species exceeded 500 yr in age (usually [greater than]600 yr) at all sites.
The deep tephra had a fine-textured surface stratum, which developed into a hard crust that impeded infiltration of water, seedling root penetration, and emergence of shoots of surviving plants (Antos and Zobel 1985a, Zobel and Antos 1991a, b). The crust of shallow tephra was thin and weak. Most of the tephra was single-grained coarse pumice, whereas the thin basal layer was fine-textured pulverized rock, produced by the initial blast. The basal layer was [less than]7 mm thick (Zobel and Antos 1991a), thinner than the 15-20 mm associated with delayed mortality of canopy Abies amabilis (Segura et al. 1994). Post-18 May eruptions deposited only scattered coarse pumice on the surface, with no significant addition to tephra depth. Our sites were chosen for their relatively fiat topography, to minimize erosion. Site choice was further constrained to areas outside USDA Forest Service logging plans and by safety regulations in force in 1980. The tephra eroded significantly only in the winter of 1980-1981, producing small areas of tephra removal and deposition at sites DP and DR. Natural tephra plots affected by erosion were sampled, but those data are reported separately. Cleared plots filled by deposition were re-excavated in the summer of 1981 unless burial was more than a few centimeters deep.
Vegetation development on undisturbed tephra was described using 100 permanent 1 x 1 m plots, located at 2-m intervals along several transects on relatively fiat terrain at each site. To mimic the effects of tephra erosion, tephra was removed in the summer of 1980, from 50 additional 1 x 1 m plots, located on transects that alternated with transects containing undisturbed plots (Antos and Zobel 1985c). Tephra was removed by hand, using small excavating tools and a vacuum cleaner. Care was taken to minimize damage to plants. Minor amounts of the thin, sticky basal layer were left where their removal would have caused excessive damage to plants. At site DR an additional 50 plots were cleared at the end of the 1982 growing season, alternating within transects with plots excavated in 1980 (Zobel and Antos 1986), and four 2 x 5 m plots were excavated in 1987 (Zobel and Antos 1992) to study the effects of delayed erosion on plant performance.
Plant cover and density within each plot were recorded by species or, for less common and very small bryophytes, by multi-species groups. For surviving trees, only foliage [less than] 1 m above the tephra surface was sampled. Seedlings were identified to species where possible and plants developed from seedlings were recorded separately from plants of the same species that survived the eruption, for the species for which the distinction was possible. At least one author was involved in all sampling, providing consistency of estimation and identification among years and sites. Nomenclature follows Hitchcock and Cronquist (1973) and Lawton (1971).
Sampling occurred in July-early September, after all plant leaves were fully expanded. In a few cases, cover was estimated from the remnants of plants from which leaves had been grazed. All four sites (plus two others [Antos and Zobel 1985c]) were sampled each year from 1980 to 1983. Sites with deep tephra were also sampled in 1984, 1987, and 1990, and those with shallow tephra in 1989. Values from 1989 and 1990 were used to define the degree of recovery after a decade; there should be little difference between values measured in shallow tephra in 1989 and what they would have been in 1990, as many aspects of vegetation in shallow tephra did not change significantly between 1983 and 1989.
Environmental factors measured at each 1 x 1 m plot included mean tephra depth in 1980 (except at site DP); percentage of the plot covered by snow during tephra deposition (based on the cracking of the tephra crust [Antos and Zobel 1982]); and light intensity based on a 1-d exposure of an ozalid paper sensor at the soil surface in each plot center (Friend 1961) in 1981 (sites DP and DR) and 1982 (all sites), expressed as a percentage of light received in a nearby clearcut. (Ozalid paper has recently been shown to measure maximum radiation intensity rather than accumulated radiation [Bardon et al. 1995].) At each sample period, we estimated percent cover of litter in each plot in two categories; "fine litter" included all individual leaves and needles, while "medium litter" included twigs with attached leaves, fallen lichen, twigs [less than]5 cm diameter, and conifer cones (Zobel and Antos 1991a).
Soil water potential was measured 2 cm below the substrate surface at five pairs of cleared and tephra plots at sites DP and DR throughout the summer of 1982, using soil psychrometers (Wescor, Logan, Utah, USA).
After disturbance, vegetation changes as species and individuals are added, either from emergence of survivors or persistence of new individuals; as species or individuals are lost to mortality; and as individual shoots change in size. We measured results of these processes, i.e., the importance of each growth form, in three ways: as species richness (species density = number of species per 1-[m.sup.2] plot); shoot density (for vascular plants); and shoot cover. "Shoot size" was calculated as cover/density. Cover and shoot density were measured separately for each species, and species data were aggregated by plant type. We distinguished six major types of plants, based on their growth form and time of establishment: surviving conifer seedlings with foliage [less than] 1 m above the soil surface, shrubs, herbaceous plants, bryophytes, conifer seedlings established after the eruption, and post-eruption non-conifer seedlings (angiosperm trees, shrubs, and herbs). We further classified herbaceous plants using three different criteria: (1) evergreen vs. annual shoots, (2) degree of usual vegetative spread and response to burial by tephra (Antos and Zobel 1985a), and (3) mode of reproduction and change in importance, using a classification modified from Cattelino et al. (1979). Using criterion 2, we recognized three categories: subshrubs (classifications STOL and SS, Antos and Zobel 1985a), long-rhizomatous species (LR, LRS), and immobile species (DB, CR) without substantial spread by stolons or rhizomes. Using criterion 3, herbs were separated by (a) their means of persistence or entry to the site: V = survived and spread vegetatively if at all, D = entered by widely dispersed seed, B = both V and D, and K = did not survive in tephra plots; and (b) whether a significant increase in importance occurred in intact vegetation or only in areas affected by disturbance: T = increased in intact vegetation (i.e., in cleared plots), I = spread only where vegetation had been disrupted (i.e., in tephra plots), N = did not increase significantly in any case. Species were classified based on cover data (or density, for new seedlings); species present in few plots were not classified. A list of species characteristics and changes in importance is available from the authors. At our sites, no species recolonized from seed banks in the soil or canopy.
Significance of differences between years for attributes of a given species or growth form was determined with Wilcoxon's signed-ranks test, and, of differences in plant importance between cleared and tephra plots in a given year, with the Mann-Whitney U test. Analysis of variance and Scheffe's multiple-range test (P = 0.05), a conservative test (Day and Quinn 1989), were used to determine the significance of differences among sites or treatments for environmental factors, seedling numbers, and seedling survival rates.
Beta diversity of species among plots within sites was defined as Sc/S, where Sc = the total number of species at a site, and S = the mean number of species per plot at that site (Whittaker 1972).
To summarize the differences in trajectories of vegetation change, we used the concepts of "inertia" (or resistance) and "resilience" (Westman 1978, 1986, Attiwill 1994). On Fig. 1, 100(b/a) represents inertia, the degree to which the community, growth form, or species maintained its importance following the disturbance. Mean values from tephra plots in 1981 represented the value b, rather than 1980 values; snow beneath tephra did not all melt until late summer in 1980, which could have prevented emergence of plants undamaged by burial. We chose the value a [ILLUSTRATION FOR FIGURE 1 OMITTED] at each site from among the mean annual values for all sample years after 1981 for plots cleared in 1980; the lowest mean was used as a. We assume that, if importance in cleared plots were reduced by plot clearing, it had recovered after 1981. Importance of growth forms in cleared plots varied from year to year, sometimes erratically, but it generally increased with time, perhaps reflecting the use of resources from the unoccupied tephra and buried soil around the cleared plots. The pattern of change in cleared plots did not coincide for different growth forms and taxa. Thus, we considered the most reliable indicator of the pre-eruption condition at a site to be the minimal value in the cleared plots after 1981. The value a [ILLUSTRATION FOR FIGURE 1 OMITTED] was the highest site mean for tephra plots for any year after 1981.
We calculated four measures of resilience, the degree of recovery of vegetation after the disturbance, using the values a, b, c as defined in Fig. 1: (1) importance at the end of the decade as a percentage of pre-eruption importance (100 c/a), indicating the degree to which the original importance has been restored; (2) decadeend importance as a multiple of post-eruption importance (c/b), indicating the degree of expansion of the surviving plants; (3) change in importance during the decade as a percentage of pre-eruption importance [100(c - b)/a], indicating the recovery during the decade relative to the original importance; and (4) change in importance during the decade as a percentage of the importance lost during the disturbance [100(c - b)/(a - b)], indicating the degree to which importance destroyed by the eruption was recovered. These measures of resilience will be referred to by the ratios used to calculate them, e.g., c/a.
In some cases, values of plant importance after the eruption exceeded our estimates of pre-eruption levels, i.e., c [greater than] a [ILLUSTRATION FOR FIGURE 1 OMITTED]. We consider that full recovery requires only that importance reach the pre-eruption value, so we analyzed all resilience values [greater than]100% as 100%, for those measures that used the value a in the definition.
Inertia and measures of resilience could be calculated only as site means, because a was the mean of cleared plots and b and c were means of tephra plot values at a site. For each component of recovery, 32 values were calculated (4 sites x 3 growth forms x 3 attributes of importance, minus 4 values for bryophyte shoot density, which was not measured).
Values of inertia and resilience were transformed to the arcsine of the square root of the proportion before analysis of variance. Attribute (shoot density, shoot cover, and species density), growth form (shrub, herbaceous, bryophyte), tephra depth, and importance of snowpack were the main effects tested in the overall analysis of variance, along with interactions of disturbance level with growth form and with attribute.
We also identified the situations in which a steady state and recovery were reached. "Steady-state" indicates that values at the last two sample dates did not differ statistically; "recovery" was reached when the value c [ILLUSTRATION FOR FIGURE 1 OMITTED] was [greater than or equal to]90% of the estimated pre-disturbance value (a, [ILLUSTRATION FOR FIGURE 1 OMITTED]). "Elasticity" was measured as the inverse of the number of years elapsed before the attribute recovered 50% of the pre-eruption value (to a/2, [ILLUSTRATION FOR FIGURE 1 OMITTED]); fewer years indicate a higher elasticity. Elasticity can be examined only for those attributes for which inertia is [less than]50% and c/a is [greater than]50%.
To summarize conifer seedling establishment and survival, we calculated the mean number of first year seedlings per plot, based on all sample years, and the periodic survival rates between successive sample years. Survival of 1-yr-old seedlings was assessed from 1980-1981, 1981-1982, 1982-1983, and, at sites DP and DR, 1983-1984, and a mean first-year survival percentage was calculated from the percentage survival in each plot at each site. The overall survival of all cohorts from 1984-1987 and from 1987-1990 at sites DP and DR, and from 1983-1989 at sites SP and SR was estimated by dividing the number of [greater than]3-yr-olds in the latter year by the total seedling number in the former year. A model was developed for each species at each site to predict the size of the surviving cohort at the end of 7 yr; the total number of germinants established per year was multiplied by the first-year survival rate, and then by the rates for survival in subsequent periods.
Cover and density in 1983 and in 1989-1990 for each growth form were related to environment (snow coverage in 1980, light, tephra depth, and litter cover in 1983), using Spearman rank-correlation coefficients, which were also calculated among these environmental variables. Only correlations with P [less than or equal to] 0.01 were considered significant, to compensate for calculating a large number of individual correlations.
In order to examine relationships between the extent of damage from tephra burial and easily observable plant characteristics, we developed regression models to relate inertia to environment, plant height, and rate of vegetative spread. Plant size was related to species' initial response to tephra deposition among 137 vascular species (Antos and Zobel 1985a), and is an effective correlate of many adaptive plant characteristics (Chapin 1993). The height of the shoot and the capacity for vertical growth of a buried rhizome or stem are the measures of size most pertinent to survival of burial (Antos and Zobel 1987). We used the potential heights and vegetative spread of the important species at each site (Antos and Zobel 1984, Antos 1988; J. A. Antos and D. B. Zobel; personal observations), along with tephra depth, percent coverage by snowpack, and percentage of light as independent variables for multiple regression equations with inertia as the dependent variable. We developed two types of models: the first type predicted inertia of cover and of species density using means of plant height and potential for spread for each growth form at each site, with the size of each major species being weighted by its cover in cleared plots in 1990. The second type of model predicted inertia of cover using individual size and inertia values for each major species at each site; in total, 46 species-site combinations were used (12 shrub, 23 herb, 11 bryophyte).
To check the generality of the regression models developed for our four sites, we used the equations to predict inertia by growth form for two other sites in the lower elevation Tsuga heterophylla zone for which we have data (Antos and Zobel 1985c; sites A and D; J. A. Antos and D. B. Zobel, unpublished data).
Vegetation change following burial beneath tephra was complex, with the pattern varying among growth forms and species, the attributes used to measure importance, the components of change measured, the degree of site disturbance, and the pre-eruption vegetation. First, we present the patterns of species richness. Next, we describe the variations in importance by growth form, emphasizing the inertia and resilience of surviving plants. Then we summarize seedling establishment and survival. Next, we describe the relationships between environmental variability within sites and performance of plant growth forms. Finally, we present models of inertia as a function of properties of plants and site environment.
Over all sampling dates in all plots, we found 6 tree species in the understory layer, 2 represented only by individuals germinated after the eruption; 12 shrub species; 55 herbaceous species; and 21 field-identifiable groups of bryophytes (plus at least 3 unknowns). In general, there were more herb species than bryophyte taxa in both tephra and cleared plots at each site, more bryophyte taxa than shrub and tree species (P [less than] 0.0001, analysis of variance), and more species in the snowy, herb-rich sites than herb-poor sites (P = 0.056) (Table 2). There was, however, no significant difference in numbers of species between tephra and cleared plots (P = 0.462).
Some local species extinctions occurred after burial by tephra. Our best estimator of species loss was the number of taxa present in cleared plots that were absent from tephra plots at a given site - Site DP (deep tephra, herb-poor, low snowpack): 5 herbs; Site DR (deep tephra, herb-rich, much snowpack): 1 shrub, 18 herbs, 4 bryophytes; Site SP (shallow tephra, herb-poor, low snowpack): 1 herb, 3 bryophytes; and Site SR (shallow tephra, herb-rich, much snowpack): 2 shrubs, 2 herbs, 1 bryophyte. Taxa restricted to cleared plots were all perennial and apparently lacked the capacity to grow through tephra or establish on its surface; most were rare. Among the common species in cleared plots, only Erythronium montanum and Rubus pedatus at site DP remained absent from tephra plots in 1990, even though R. pedatus had [greater than]60% of the herb cover in cleared plots. Several important species absent from tephra in 1981 reappeared in 1982: Tiarella trifoliata and the mosses Rhytidiopsis robusta, Mnium-like species, and Brachythecium-like species at site DR, and the moss Eurynchium oreganum at site SR. Species restricted to tephra plots were rare, established from seed or spores; few lasted more than a year, the main exceptions being mosses.
Species density (mean number of species per plot) increased with time in tephra plots and, to a lesser extent, in cleared plots [ILLUSTRATION FOR FIGURES 2-4 OMITTED]. Exceptions were [TABULAR DATA FOR TABLE 2 OMITTED] the lack of change, or erratic changes, for herbs in both plot types at site DP, and the increases followed by decreases for bryophytes in cleared plots at sites DR and SP. Most increases in cleared plots occurred early in the decade [ILLUSTRATION FOR FIGURES 2-4 OMITTED]. For tephra plots, a linear increase in species density for a growth form was accompanied in general by an exponential increase in its cover [ILLUSTRATION FOR FIGURES 2-4 OMITTED].
Beta diversity, which here expresses the variation in species composition among plots within a site, was higher in tephra than in cleared plots for bryophytes at all sites, and for vascular species in the deep tephra plots (Table 2). In deep tephra, beta diversity of vascular plants was higher than that of bryophytes, but in shallow tephra this pattern was reversed.
Response by growth form
General patterns. - After the eruption, cover of shrubs at sites with snow at the time of the eruption (SR, DR), herbs in deep tephra (sites DP, DR), and bryophytes at all sites was much lower in tephra than in cleared plots ([ILLUSTRATION FOR FIGURES 2-4 OMITTED], Table 3). During the decade following the eruption, the importance of most growth forms increased at most sites in both cleared and tephra-covered plots, although increases were often much larger in tephra plots. Major exceptions to this general pattern of increases were surviving tree cover and density, bryophyte cover in 1989-1990 in cleared plots, herb cover at site DP, and seedling density at site SP. Within tephra plots in 1989-1990, herb importance remained lower in deep than in shallow tephra, and shrub importance was lower at site DR than at other sites (Table 3). Bryophyte cover remained low in tephra plots at all sites.
Tree seedlings comprised the majority of seedlings, especially in tephra plots (Table 3). Tree seedlings were more abundant in tephra plots than in cleared plots, especially in 1990, whereas differences for other seedlings were inconsistent between plot types. Seedling density varied greatly through time.
In a few tephra plots, natural erosion during the winter of 1980-1981 removed most of the tephra. All three plots that eroded at site DP were on a steep drainage channel bank in the only herb-rich part of the stand. The five plots eroded at site DR were in a moist open location near a drainage channel. At both sites, naturally eroded plots had more herbaceous cover than the average cleared plots (Table 4), probably a reflection [TABULAR DATA FOR TABLE 3 OMITTED] of their locations in the most herb-rich parts of the stands. Bryophyte cover in naturally eroded plots was similar to that in cleared plots at site DP and became similar by 1990 at site DR. Surviving tree cover at site DP was greater than for cleared plots, probably a matter of chance location, while shrub cover was similar. In contrast, surviving trees and shrubs on naturally eroded plots were almost eliminated at site DR (Table 4), where they had been trapped beneath snow during the eruption, and were apparently killed by the several extra months of burial before they were released by snow melt in the summer of 1981.
Recovery of surviving plants. - Trajectories of change in vegetation cover [ILLUSTRATION FOR FIGURES 2-4 OMITTED] varied among growth forms and sites, with site differences relating to pre-eruption vegetation, tephra depth, and presence of snow beneath the tephra. Data from plots cleared in 1980 indicated that prior to the eruption, sites SR and DP had the most shrub cover, and site DP had considerably lower herbaceous and bryophyte cover than the other sites [ILLUSTRATION FOR FIGURES 2-4 OMITTED].
TABLE 4. Cover (%) for mature plants and density (number of plants/[m.sup.2]) for seedlings by growth form in plots naturally eroded during the winter of 1980-1981 at sites DP (n = 3) and DR (n = 5). Site type refers to tephra depth and the richness of the herb layer. Values in parentheses are the percentage deviations from values in plots experimentally cleared in 1980 at the same site. Deep/Poor, DP Deep/Rich, DR Growth form 1981 1990 1981 1990 Surviving trees 13.2 (117) 22.3 (238) 0.2 (-99) 0.3 (-98) Shrubs 32.3 (3) 46.9 (9) 0.8 (-94) 7.2 (-77) Herbs 25.3 (423) 18.2 (239) 27.4 (23) 61.6 (72) Bryophytes 9.5 (14) 7.3 (15) 6.6 (-77) 10.8 (-11) Tree seedlings 2.9 (34) 4.5 (-14) 0.2 (-50) 3.0 (-44) Other seedlings 0.9 (736) 5.8 (299) 0.1 (-13) 59.0 (408)
Cover in cleared plots changed with time. Bryophyte cover significantly increased after 1981 in cleared plots, and then significantly declined before 1990 [ILLUSTRATION FOR FIGURE 4 OMITTED]; evergreen herb cover showed the same pattern at sites DP and SR (data not presented).
Inertia and one measure of resilience (c/a) (see Methods: Data analysis, above, for definitions) of non-tree growth forms varied significantly with attribute, growth form, and type of disturbance (Table 5); significant interactions indicated that growth forms behaved differently at different tephra depths for both inertia and c/a, and with and without snow for inertia. Decade-end importance, as a multiple of post-eruption importance (c/b), varied with tephra depth and extent of snow during the eruption (Table 5). Change in importance during the decade of recovery, expressed as a percentage of pre-eruption importance [(c - b)/a], did not vary significantly among attributes, growth forms, or levels of [TABULAR DATA FOR TABLE 5 OMITTED] disturbance (Table 5). Change expressed as a percentage of the loss of importance during the eruption [(c - b)/(a - b)], however, differed significantly between tephra depths, and growth forms behaved differently with tephra depth (Table 5).
Inertia and c/a were closely related, both overall [ILLUSTRATION FOR FIGURE 5A OMITTED], and for individual attributes (species density [r.sup.2] = 0.876, cover [r.sup.2] = 0.946, density [r.sup.2] = 0.964). Thus, the differences in importance among understory growth forms at the end of the decade were primarily those originally imposed by the disturbance: c/b declined sharply as inertia increased, and this relationship was consistent among growth forms [ILLUSTRATION FOR FIGURE 5B OMITTED]; [(c - b)/(a - b)] showed a similar relationship to inertia as did c/a, with the exception of three values for shrubs (cover and density at site DP and species richness at site sP), which have a low value of [(c - b)/(a - b)] for their inertia of 60-80% [ILLUSTRATION FOR FIGURE 5C OMITTED]. The other measure of resilience, (c - b)/a, showed a complex relationship to inertia [ILLUSTRATION FOR FIGURE 5D OMITTED], as discussed later (see Discussion: The nature of the vegetation change).
For non-tree growth forms, inertia was lower for shoot cover than for species density, higher for large growth forms than for small ones, and highest with shallow tephra that did not fall on snow and lowest where deep tephra fell on snow (Table 6). A similar but less distinct pattern was shown by c/a: it was lower for cover than for species density, lowest for bryophytes, and lowest with deep tephra on snow and highest with shallow tephra without snow (Table 6). The expansion of post-eruption importance (c/b) was higher with deep tephra with snow, i.e., at site DR, than elsewhere (Table 6). Change as a percentage of the loss of importance during the eruption [(c - b)/(a - b)] varied among sites similarly to c/a (Table 6).
For surviving trees, neither shoot density nor species density could increase after the eruption, by definition. Changes in cover of surviving trees were small compared to those of other growth forms (Table 3).
The patterns in which steady state and recovery (return to 90% of estimated pre-eruption importance) developed were complex (Table 7). A steady state was sometimes reached at levels below that of the corresponding estimates for pre-eruption values (e.g., bryophyte cover at site DP [ILLUSTRATION FOR FIGURE 4 OMITTED]). Bryophytes did not recover in any situation, but a steady state was reached for species density on shallow tephra, and for cover on [TABULAR DATA FOR TABLE 6 OMITTED] deep tephra (Table 7). Herbaceous plants reached a steady state or recovery, or both, only at sites with shallow tephra. Shrub cover continued to increase at all sites, whereas shrub species density and shoot density reached a steady state in most cases and fully recovered in half of them (Table 7). Surviving tree seedlings reached a steady state in all situations, usually at values less than recovery.
Shoot size (cover per shoot) varied differently from other attributes. For surviving trees it increased by 2.9-fold to 35-fold after the eruption and remained high, because the small individuals were buried and died. Cover per shrub shoot initially decreased where shallow tephra fell on snow (site SR) and did not recover [ILLUSTRATION FOR FIGURE 6D OMITTED]; at the other 3 sites, shrub sizes in cleared and tephra plots overlapped substantially. Cover per herbaceous shoot was higher in tephra than in cleared plots at site DR, where deep tephra fell on snow, but disturbance did not modify herb shoot size substantially at other sites [ILLUSTRATION FOR FIGURE 6 OMITTED]. Delaying clearing until 1982 at site DR reduced both shrub and herb shoot size substantially, but both returned to the range of cleared plot values by 1990 [ILLUSTRATION FOR FIGURE 6B OMITTED]. Values of inertia calculated for shoot size (Table 8) did not vary significantly among either growth forms or sites in an analysis of variance.
TABLE 7. Steady state and recovery of species density, shoot density, and shoot cover in 1989-1990 by growth form and site. Site type refers to tephra depth and the richness of the herb layer. Growth form Bryo- Attribute Site type Tree(*) Shrub Herb phyte Species density Deep/rich E ... ... ... Deep/poor E ER ... ... Shallow/rich E ER ER E Shallow/poor E E ER E Shoot density Deep/rich E E ... ND Deep/poor E E ... ND Shallow/rich E ER ER ND Shallow/poor E ER R ND Shoot cover Deep/rich E ... ... E Deep/poor E ... ... E Shallow/rich E ... ER ... Shallow/poor ER R R ... Note: E = at a steady state (i.e.,showed no significant increase between the two latest sample dates); R = the attribute has recovered (i.e., reached at least 90% of pre-eruption importance, defined as the lowest level reached after 1981 in plots cleared in 1980); ... = neither R nor E; ND = not measured. * Surviving tree.
Elasticity, defined as the inverse of the number of years after 1980 for tephra plots to reach 50% of the lowest post-1981 cleared-plot value, varied by attribute, site, and growth form (Table 9). Species density recovered in fewer years than cover (i.e., elasticity of species density was greater); elasticity on shallow tephra was usually greater than on deep tephra; and elasticity decreased from shrubs to herbs to bryophytes. Exceptions to the general patterns were that shrubs had low elasticity at the sites where tephra fell on snow (DR, SR), and that elasticity of species density in deep tephra was greater for bryophytes than for herbs. Delayed erosion (1982 cleared plots, Table 9) produced elasticity values substantially greater than those on tephra at site DR, and the highest elasticity for bryophytes.
Forms of herbaceous species. - Although the important species of trees and shrubs shared similar forms, herbaceous species included several substantially different types of plants, which may perform differently after burial. Herbaceous species differed in their phenology, degree of usual vegetative spread, growth response to tephra, and relative importance of vegetative vs. seed reproduction after the eruption. In 1981 herbs with only vegetative spread had 39% of the total herb cover in deep tephra, while those with both seedling establishment and vegetative spread had 61% of total herbaceous cover. By 1990 the situation had reversed; species with little or no seedling establishment had 72% of total herb cover, and those with both methods of spread, 28%.
TABLE 8. Values of inertia for shoot size (shoot cover/shoot density) on tephra plots, by growth form and site. Analysis of variance produced insignificant F ratios for both site (P = 0.384) and growth form (P = 0.095). Site type refers to tephra depth and the richness of the herb layer. Growth form Surviving Site type tree Shrub Herb Deep/rich 3496 156 258 Deep/poor 294 102 94 Shallow/rich 1450 68 128 Shallow/poor 582 98 142 Mean 1456 106 156 Standard error of the mean 723 18 36
Among 23 responses analyzed for important herb species (where each response was for one species at one site), (c - b)/a was significantly greater for evergreen species than for those with annual shoots (Tables 10, 11). Cover of herbs with annual shoots had a net change of zero during the decade (cover peaked and then decreased for several species, with important species losing an average of 14% of their maximal cover before 1989-1990); in contrast, cover of species with evergreen shoots increased. Inertia and c/a did not differ significantly with shoot longevity (Table 10). Neither inertia nor measures of resilience differed significantly among forms when herbs were classified by the mechanism of vegetative spread or by the importance of seedlings (Table 10). Among the 23 cases, inertia and c/a were significantly correlated (r = 0.848, P [less than] 0.0001), as with the overall growth form analysis; [(c - b)/a] was not related to inertia (r = -0.138, P = 0.530) and only weakly correlated with c/a (r = 0.408, P = 0.053). Although some means differed substantially [TABULAR DATA FOR TABLE 9 OMITTED] among herb forms, variability within each group was high (Table 11). Among the 23 cases analyzed, inertia and c/a were both higher (P [less than] 0.0001) in shallow tephra than in deep tephra, but other measures of resilience did not differ significantly between tephra depths (P = 0.133-0.176).
Responses of individual species
Although most important species (see Appendix), by definition, had a response pattern similar to the mean for their growth form, a few species differed in response. Here we describe cases in which species with substantial cover responded contrary to their growth form, as shown in Figs. 2-4 and Fig. 6.
Among shrubs, cover of Menziesia ferruginea decreased in tephra plots between 1983 and 1989 and in cleared plots between 1981 and 1989 at both sites SP and SR, and did not change in either cleared or tephra plots at site DE In contrast, the dominant shrubs, Vaccinium spp., increased substantially in all these situations. Shoot size of M. ferruginea decreased more with time than did the mean for shrubs at these three sites. The two dominant shrubs differed in shoot size response, with V. ovalifolium increasing substantially in shoot size in tephra plots at all sites, whereas V. membranaceum, which produced many small sprouts from existing plants, usually decreased or did not change. Most delayed emergence of isolated shoots was by V. membranaceum.
TABLE 10. Analysis of variance results (P values) for inertia and two components of resilience of shoot cover [c/a and (c - b)/a] for contrasting forms of herbaceous plants, classified by their shoot longevity, mechanism of vegetative spread, and importance of seedlings in post-eruption spread. The analysis of variance was run including tephra depth (shallow, deep) as a second source of variation; tephra depth was significant for all criteria for inertia (P = 0.0001-0.0012) and for c/a (P [less than or equal to] 0.0001), but not for (c b)/a (P = 0.165-0.625). P Criterion for c - b classification df Inertia c/a a Shoot longevity 1, 20 0.0755 0.7383 0.0272 Vegetative spread 2, 19 0.3556 0.5084 0.6356 Importance of seedlings 1, 20 0.9468 0.6816 0.6624
Perennial herbaceous species varied substantially in their properties, with several differing significantly from the mean herb response to burial by tephra. In tephra plots at site SR, the robust herbs Valeriana sitchensis and Trautvetteria caroliniensis decreased significantly in cover, while the herb layer mean did not change. At site DR, cover of neither V. sitchensis nor Erythronium montanum changed in tephra or cleared plots, while other major species increased substantially. Shoot size of V. sitchensis increased in both tephra and cleared plots at site DR, in contrast to the mean and to the dominant, E. montanum, which decreased in size. A prostrate subshrub, Gaultheria humifusa, showed no increase in cover over time in either treatment at site SP, while the herb layer in general, including three other low-stature subshrubs, increased. Pyrola secunda increased in both cleared and tephra plots between 1983 and 1989 at site SR after overall herb layer cover had reached a steady-state.
TABLE 11. Mean values of inertia and two measures of resilience [c/a and (c - b)/a] (in %, with 1 SE in parentheses) for different forms of herbaceous plants, classified by shoot longevity, mechanism of vegetative spread, and importance of seedling vs. vegetative spread. Significance of differences is given in Table 10. Inertia 100x 100x Criterion/Form n(*) (%) (c/a) (c - b)/a Shoot longevity Annual 5 55.8 55.8 0.0 (23.0) (23.0) (0.0) Evergreen 18 38.3 64.8 26.5 (8.9) (10.1) (5.5) Method of vegetative spread Subshrubs, stolons 9 38.8 58.1 19.3 (14.5) (15.3) (8.1) Long rhizomes 6 66.5 84.2 17.7 (16.4) (12.0) (10.4) Immobile 8 27.5 52.1 24.6 (11.2) (17.4) (8.4) Importance of seedlings Seedlings common 9 28.9 49.3 20.4 (12.9) (15.6) (7.6) Vegetative spread only 14 50.6 71.5 21.0 (10.8) (11.0) (6.6) * n = number of situations represented.
[TABULAR DATA FOR TABLE 12 OMITTED]
The most important bryophyte taxa (see Appendix) on the forest floor were large species; leafy liverworts and small mosses were important only on large woody debris or the bark of tree bases. At the end of the decade, cover of all major forest floor mosses had declined significantly in cleared plots, but not on tephra [ILLUSTRATION FOR FIGURE 4 OMITTED]. The large decrease in moss cover in cleared plots at site SR was almost entirely due to the dominant, Rhytidiopsis robusta; Dicranum and Eurynchium spp. declined only slightly. On the surface of tephra, bryophytes that differed from those in cleared plots made up a substantial proportion of cover. Ceratodon purpureus and Pohlia annotina were identified from the dominant bryophytes on the tephra surface at site DR; the other sites supported similar-looking mosses.
General patterns. - Seedlings of herbs, shrubs, and trees established on deep tephra. In 1990, establishment was generally greater on undisturbed tephra than on cleared plots in terms of density and, to a lesser extent, species richness (Table 12). Seedling cover remained low in 1990, [less than]1.1% for tree seedlings in all situations except on undisturbed tephra at site DR. Densities of seedlings did not increase with time except for trees. Most first-year seedlings did not survive, especially herb species. For several dominant herbaceous species we saw no seedlings.
Repeated sampling indicated that seedlings of a few herbaceous species did contribute substantially to their importance, but those seedlings do not show up in Table 12 because seedlings soon became indistinguishable from surviving plants. In 1990, individuals of several species were identified as being older than 1 yr and having established from seed on the deep tephra, a distinction not made in earlier sampling; the majority were Vaccinium spp., and others included Achlys triphylla, Anaphalis margaritacea, Epilobium sp., Erythronium montanum, Hieraceum albiflorum, Mitella sp., Rubus sp., Salix sp., Sorbus sitchensis, Tierella trifoliata, Valeriana sitchensis, Veratrum viride, and Viola spp. The total density of these seed-origin plants was 1.27 individuals/[m.sup.2] at site DP and 2.24 individuals/[m.sup.2] at site DR; their cover was very small. For Vaccinium spp., young plants of seed origin remain identifiable for a long time, as described by Alaback and Tappeiner (1991:536). At our sites the largest Vaccinium from seed were producing their first mature-type leaves at the end of the decade and no individual occupied an area greater than 100 [cm.sup.2]. At the end of the decade, Tiarella trifoliata, Mitella spp., and Viola spp. populations included a substantial proportion of seedorigin plants. The origin of the seed in all these cases was probably from surviving plants, rather than dispersal from outside the stand.
Conifer seedling establishment and survival. - Tree seedlings survived in large numbers and have considerable potential to change the nature of the understory. Conifer seedling establishment after the eruption and subsequent survival (Table 13) varied with species, tephra depth, and understory type within a tephra depth. Seedlings of Abies amabilis established less frequently than those of Tsuga spp. except at site DR, but Abies survived better at all sites. Establishment of Abies was greater at sites with herb-rich than with herb-poor understories, but the reverse was true for Tsuga (Table 13). Survival of tree seedlings beyond their first year was greater with deep than with shallow tephra. Survival of Abies was very high, especially in deep tephra (Table 13), a cumulative 22-53% after 7 yr for the large 1983 cohort. Chamaecyparis nootkatensis established well at site SR but survival past year 1 was nil; at site DP, in contrast, the few established seedlings survived well (Table 13).
Conifer establishment in cleared plots was significantly below that on natural tephra for all three conifer species in deep tephra and for Tsuga in shallow tephra (Table 13). In deep tephra, survival was consistently [TABULAR DATA FOR TABLE 13 OMITTED] greater on natural tephra only at site DR, the herb-rich site. Survival was not consistently different between cleared and natural plots in shallow tephra. Plots cleared in 1982 showed similar patterns to plots cleared in 1980 at site DR, with somewhat higher survival (Table 13).
Numbers of [greater than]3-yr-old conifer seedlings were weakly related to numbers of 1st-yr seedlings (r = 0.50). Predictions of the total number of seedlings after 7 yr, from a model using number of germinants and periodic survival rates for each site, were related only marginally better to number of [greater than]3-yr-old conifer seedlings at the end of the decade (r = 0.51).
Soil water potential in 1982 reached minima of -0.20 [+ or -] 0.03 MPa (mean [+ or -] 1 SD) in tephra plots and -0.35 [+ or -] 0.19 MPa in cleared plots at site DP, and -0.20 [+ or -] 0.07 MPa in tephra plots and -0.18 [+ or -] 0.18 MPa in cleared plots at site DR. There was no significant difference between means for paired cleared and tephra plots at either site. Tephra became much drier in the late 1980s, but water potentials were not measured then.
Environmental conditions differed substantially among sites. Measurements in individual plots confirm that there were major differences among sites in tephra depth and snow cover at the time of the eruption (Table 14). Light intensity was lower at site DP than at other sites, in both tephra and cleared plots (Table 14). Litter cover was greater on cleared than on tephra plots. Most litter cover was fine (i.e., composed of individual conifer leaves). Differences among sites in medium litter cover were small. For most environmental factors on most sites, there was substantial variability among individual plots (Table 14), enough potentially to produce differences among plots in vegetation recovery.
Among environmental factors, medium litter was significantly correlated with fine litter only in the tephra plots at sites DR and SP (Table 15). At sites with the least snow cover (DP, SP), initially snowy plots later accumulated less litter (Table 15). Light and tephra depth were correlated with litter cover only sporadically.
[TABULAR DATA FOR TABLE 14 OMITTED]
Correlation of plant importance with environment within sites
Density and cover of a growth form within a site were significantly (P [less than] 0.01) related to environmental factors: (1) more often for undisturbed tephra than for cleared plots; (2) more often for cover than for shoot density; (3) more often in herb-rich than in herb-poor sites; and (4) more often for litter values than for other factors (Tables 16 and 17). Consistent negative correlations of cover with medium litter in cleared plots at site DR was an exception to the first generalization.
Both categories of litter were related to vegetation in tephra plots (Table 16). Litter cover was negatively related to cover of bryophytes at all sites and to herb cover and density at herb-rich sites (DR, SR). Higher litter cover was usually associated with higher seedling density in 1983 but with lower seedling density in 1989-1990. Shrub importance showed few, and inconsistent, associations with litter.
Most relationships of herbs, bryophytes, and seedlings to litter in cleared plots were similar to those in tephra plots. Surviving-tree cover and density in cleared plots decreased as medium litter increased, as did shrub cover at site DR.
In tephra plots, growth-form importance was correlated with the presence of snow repeatedly only at site DP, and with light intensity and tephra depth only sporadically (Table 17). In cleared plots, light was related to vegetation only at site DR, where herb and shrub importance were higher in darker plots (Spearman rank correlation = -0.42, -0.39, -0.37, and -0.36 for herb density in 1983, herb cover in 1983 and 1990, and shrub cover in 1983, respectively).
TABLE 15. Significant (P [less than] 0.01) Spearman rank-correlation coefficients (r) among environmental variables (light, snow cover in 1980, tephra depth in 1980, and litter cover in 1983) measured at individual plots. For tephra plots, n = 88-102; for cleared plots, n = 42-50. Site type refers to tephra depth and the richness of the herb layer. Site type Treatment Variables r Deep/poor, DP Tephra Fine litter x Snow -0.34 Fine litter x Light -0.29 Deep/rich, DR Tephra Fine litter x Light -0.29 Medium litter x Fine litter 0.52 Shallow/poor, SP Tephra Fine litter x Snow -0.64 Fine litter x Tephra depth -0.27 Medium litter x Snow -0.45 Medium litter x Fine litter 0.44 Shallow/rich, SR Cleared Medium litter x Light 0.45
In some cases, the sign of the correlation of an environmental factor with growth-form importance differed consistently between years and plot types. In 1983, seedling density, where significantly correlated, was higher with more litter; in 1990, it was lower (Table 16). All seven significant correlations with tephra depth and snow cover were negative in 1983, but all four were positive in 1989-1990 (Table 17).
Models relating environment and growth form to plant survival
A regression model to relate inertia for species density to tephra depth and plant height, using mean values for growth forms, was significant:
Inertia for species density (%)
= 70.076 - 0.486(Tephra depth) (mm) (1)
+ 0.247(Potential plant height) (cm);
[R.sup.2] = 0.633, df = 9, P = 0.011.
No model predicting inertia for cover was significant [TABULAR DATA FOR TABLE 16 OMITTED] [TABULAR DATA FOR TABLE 17 OMITTED] with growth-form mean values, even though inertia for cover at the growth-form level was closely related to inertia for species density (r = 0.918, P [less than] 0.0001).
For the 46 important species-site combinations representing all three growth forms, the most effective equation predicting inertia for cover used tephra depth and the indicator variable that separated growth forms (1 = shrub, 2 = herb, 3 = bryophyte):
Inertia for cover (%)
= 96.653 - 0.402(Tephra depth) (mm) (2)
- 16.576(Growth form indicator value);
[R.sup.2] = 0.377, df = 43, P [less than] 0.0001.
When the growth form indicator variable was removed, the best model was:
Inertia for cover (%)
= 66.180 - 0.431(Tephra depth) (mm); (3)
[r.sup.2] = 0.270, df = 44, P = 0.0002.
Species height and potential for vegetative spread did not contribute significantly to equations that predict inertia for cover.
Equations for single growth forms incorporated a single, different environmental factor in each, again with no plant characteristics being significant:
For shrubs: Inertia for cover (%)
= 74.079 - 0.599(Snowpack cover) (%); (4)
[r.sup.2] = 0.542, df = 10, P = 0.006;
For herbs: Inertia for cover (%)
= 94.459 - 0.713(Tephra depth) (mm); (5)
[r.sup.2] = 0.590, df = 21, P [less than] 0.0001;
For bryophytes: Inertia for cover (%)
= 6.730 - 0.326(Percentage of light); (6)
[r.sup.2] = 0.542, df = 9, P = 0.010.
Eqs. 1-6 were used to predict the inertia for growth forms at two lower elevation sites with 2.3 and 7.5 cm tephra and no snowpack. Success varied. Relationships between predicted inertia and measured inertia values for growth forms at the low elevation sites (D. B. Zobel and J. A. Antos, unpublished data) were:
Using Eq. 1: for species density (n = 6): [r.sup.2] = 0.845, P = 0.005. Eq. 1, despite the high [r.sup.2], underestimated inertia in the low-elevation plots, the difference between predicted and measured inertia being -21.8 [+ or -] 12.1% (mean [+ or -] 1 SD).
Using Eq. 2: for cover (n = 6): [r.sup.2] = 0.632, p = 0.030;
Using Eq. 3: for cover (n = 6): [r.sup.2] = 0.205, not significant;
Using Eqs. 5 and 6: for cover (n = 4): [r.sup.2] = 0.127, not significant. Eq. 4 was not used, as lower elevation sites had no snow cover.
Tephra deposition and its interactions with plant phenology and snowpack produced substantial, long-lasting changes in vegetation composition. Such changes could appear inexplicable without a detailed knowledge of the disturbance and how it affects plants.
Nature of the disturbance
Volcanic disturbance regimes in the Cascade Range. - Plant burial by tephra in 1980 was not an isolated event, but part of a disturbance regime that has affected most of the Cascade Range, and adjacent areas, repeatedly during the development of these largely volcanic mountains. This regime included 12-22 major eruptions in the last 13 000 yr, as well as myriad smaller eruptions (Shipley and Sarna-Wojcicki 1983, Mullineaux 1986). During the late Pleistocene and Holocene, the most consistent tephra producer has been Mount St. Helens, which began erupting about 40000 yr ago; significant tephra has also come from Mount Rainier, Mount Baker, Glacier Peak, and the climactic eruptions of Mount Mazama. The area east and northeast of Mount St. Helens has received tephra deposits more often than other areas of the Cascades, because of prevailing wind directions and the frequent eruptions of Mount St. Helens (Shipley and Sarna-Wojcicki 1983). Following a very deep deposit in 1480, the area received 10 shallow tephra strata between 1482 and 1800 (Mullineaux 1986). In late 1799 or early 1800 (Yamaguchi 1983, Mullineaux 1986), a tephra deposit with texture, depth, and chemistry similar to the 1980 deposit (Waitt and Dzurisin 1981) impacted our study sites. This recent high frequency of tephra deposition has been typical of 3000 of the last 4000 yr (Mullineaux 1986). This frequency is greater than that of wildfire in Abies amabilis-dominated forest types at Mount Rainier, 85 km north of Mount St. Helens, where the fire return interval was 295-616 yr (Hemstrom 1982).
The current soil-vegetation mosaic of the Pacific Northwest reflects the influence of repeated tephra deposits (Franklin and Dyrness 1973). Burial by tephra has probably eliminated those species least adapted to it from the areas most frequently affected. For example, the robust herb Xerophyllum tenax occurs in our plots 44 km from the volcano, but not at or near our deep-tephra plots. It has little capacity to escape tephra once the vegetative shoot is fully buried (Antos and Zobel 1982).
The 1980 eruption. - The 1980 tephra deposition from Mount St. Helens was a large, well-documented disturbance (Waitt and Dzurisin 1981, Zobel and Antos 1991a). Over 1000 [km.sup.2] of forest east of the devastated area received tephra [greater than]3 cm deep (Waitt and Dzurisin 1981: Fig. 355). Disturbance magnitude can be described in terms of its (1) intensity, defined by the characteristics of the external disturbing force (here, tephra depth, degree of crust formation, and presence of snow), and (2) severity, defined by the degree of damage to the biota (here, as inertia), all of which varied among and often within our sites.
Effects of the 1980 tephra deposit emphasize the critical importance of conditions at the time of disturbance. Timing of the 1980 eruption, before plants had expanded their 1980 tissues, probably minimized damage to herbs and, except where tephra fell on snowpack, to shrubs. The co-occurrence of snowpack with tephra fall was the only circumstance at our sites in which tephra killed woody plants taller than the tephra was deep.
A mosaic of tephra types occurred in 1980, even within a forest stand (Antos and Zobel 1985c, 1986); that mosaic, in modified form, persisted for the decade (Zobel and Antos 1991 a). At tree bases, single-grained, coarse tephra accumulated but the fine-textured crust was thinner than average (Zobel and Antos 1991a). Microsites with little tephra remained beneath large woody debris and where tephra slumped off large wood and steep microtopography. Woody litter broke the integrity of the tephra (especially the crust), providing high-probability paths for growth of shoots through the tephra. In deep tephra these refugia increased manyfold the inertia of herbs, bryophytes, and prostrate buried shrubs. Areas without large woody debris would have had lower inertia than our sites; this represents an important role of large woody debris not previously recognized (Harmon et al. 1986). Thus, even for this truly exogenous disturbance, the final phase, the settling of the tephra onto the soil and plant surfaces, was modified by living and dead plant material to a degree that substantially increased survival of understory plants.
The tephra changed after it settled; pH, Ca, K, Mg, and S declined rapidly after deposition. Changes in tephra chemistry at a site from 1980 to 1987 were greater than variation among sites at a given date (Zobel and Antos 1991a). Nitrogen and phosphorus accumulated slowly in the tephra, and chemical differences associated with canopy cover, the original richness of the understory, and microtopography developed. Although our data do not allow us to relate differences in plant performance to the changing tephra chemistry, there was substantial potential for such an influence, particularly affecting seedling establishment. For example, in 1980 at site DR, seedling roots would have encountered tephra crust with 4.0 times the sulfur, 2.7 times the calcium, 1.6 times the magnesium and potassium, but only 75% of the phosphorus and 17% of the nitrogen as in 1982 (Zobel and Antos 1991a).
Contrary to the assertion that "The ecological effects of this  ash fall proved ephemeral . . ." (del Moral and Bliss 1993:5), disturbance of understory by deep tephra was severe and recovery has been slow. After a decade, cover of bryophytes in our plots remained [less than]20% of pre-eruption values at all sites and did not increase significantly between 1987 and 1990 on deep tephra; cover of herbs remained [less than] 13% of the pre-eruption level in deep tephra; and cover of shrubs was 10% of its pre-eruption value where tephra fell on snow. In addition, significant, delayed mortality of one canopy tree species was associated with depth of the initial component of the tephra deposit (Segura et al. 1994).
The nature of the vegetation change
Several systems of classification describe vegetation change. The classic distinction has been between primary succession, on substrates not previously occupied by plants, and secondary succession, on substrates on which previous vegetation has been destroyed (McIntosh 1980). The 1980 tephra produced characteristics of both primary and secondary succession: a new substrate available for rooting, but much organic legacy (Franklin et al. 1988) from the previous stand, i.e., a surviving tree canopy producing shade, litter, and seeds and, in the buried soil, live roots and perennating organs, microorganisms, organic matter, and nutrients (Zobel and Antos 1991a). Both the tephra and buried soil contributed to the success of conifer seedlings. Tephra supported greater conifer seedling establishment and survival, and a higher proportion of Tsuga, than on cleared plots (Tables 12 and 13; Zobel and Antos 1991b). The success of Tsuga on tephra and the increase in its relative importance also occurred where a forest canopy was absent (del Moral et al. 1995). Even though seedling roots had not penetrated below the tephra after 6 yr (Zobel and Antos 1991b), the buried soil contributed to conifer seedling success; the mycorrhizae on their roots probably allow rapid nutrient uptake from below the tephra.
Another way to classify vegetation change is by the sources of the plants that make up the new vegetation: survivors of the disturbance, in situ propagules of the seed/spore bank, and propagules dispersed from outside the disturbed area (Grubb and Hopkins 1986). Survivors were the predominant source of post-eruption vegetation in our area. The seed/spore bank was buried deeply enough to exclude emergence of germinants except where tephra was removed (Zobel and Antos 1986, 1992). Post-eruption seedlings were predominantly of species already on site, and were important for some herbs and shrubs, but especially for conifers. Seedling success of Tierella trifoliata was contrary to its behavior in clearcuts in Alaska (Tappeiner and Alaback 1989). All vascular species important in our plots appear well adapted to old-growth forests with Tsuga heterophylla, a conclusion drawn for four of them in southeastern Alaska (Tappeiner and Alaback 1989). A few vascular species initially absent from the plots invaded, but most died within a year or two, and plants that survived to reproductive maturity contributed little to cover. This system, with a major influence of surviving, established plants, appears to change composition slowly because invaders are limited to species that can establish in shade and on low-nitrogen substrates. The major exception to survivors as sources for post-disturbance plants was colonization of the tephra by weedy widespread mosses. Recovery from burial in a salt marsh also involved primarily vegetative regrowth of a subset of the predisturbance flora (Allison 1995).
Vegetation change can also be described by the inertia-resilience dichotomy (Westman 1986, Attiwill 1994), although few values of inertia and resilience of terrestrial plants appear in the literature (Attiwill 1994). In our situation, inertia (b/a, [ILLUSTRATION FOR FIGURE 1 OMITTED]) varied with disturbance, growth form, and attribute of importance measured, and growth forms were damaged to differing degrees (i.e., their inertia differed) with different amounts of disturbance (Tables 5 and 6). The four measures of resilience (the degree of recovery after the disturbance) provided different information and had different patterns of significance (Tables 5 and 6): measures that used the decade-end importance (c/a, c/b) both responded significantly to environment, and c/a varied with growth form and attribute, as did inertia; in contrast, resilience measures that used the change in importance after the eruption [(c - b)/a, (c - b)/(a - b)] varied little, except for a relationship between the latter and tephra depth. The relationships of measures of resilience to inertia also varied: (a) for c/a [ILLUSTRATION FOR FIGURE 5A OMITTED] recovery was often complete above 40% inertia, but it was very limited below 5% inertia; (b) c/b was usually high at inertias below 5%, while above 5% c/b declined slowly as inertia increased [ILLUSTRATION FOR FIGURE 5B OMITTED]; (c) at high inertia, (c - b)/a is limited by the small initial reduction in importance (c - b) [ILLUSTRATION FOR FIGURE 5D OMITTED], from 5% to 60% inertia there was little relationship of (c - b)/a to inertia, and, most importantly, recovery from below 5% inertia appears to be limited by the severe damage.
Further understanding of vegetation change comes from examining other components of resilience (Westman 1978, 1986). For those growth form x site combinations that reached a steady state, we can determine elasticity (Table 9), amplitude (the threshold level beyond which the property will not return to its initial level), and malleability (the degree to which the eventual steady state differs from the predisturbance level). Amplitude for shrub density lies between the level of disturbance at deep and that at shallow tephra, as shrub density reached a steady state at all sites, but without recovery in deep tephra (Tables 3 and 7). The malleability for shrub density at deep tephra sites was 25% and 83% for sites DP and DR, respectively. Such estimates may be inaccurate when a slow increase does not show significance and is thus perceived as a steady state. Shrub density at DP was stable after 1983, whereas at site DR it increased 22% from 1987 to 1990, although the increase was not statistically significant. Also, the eventual development of a dense conifer layer in the understory may modify the other layers, making any long-term estimates based on the growth form response to burial alone inaccurate.
Understory response to tephra and other disturbances. - The only other study of response of forest understory plants to airfall tephra beneath an intact canopy was on Kodiak Island, Alaska, after the Katmai eruption in 1912 (Griggs 1918, 1922). Even though tephra was twice as deep at Kodiak as at our deep-tephra sites, most vegetation recovery was from surviving plants, many of which produced a new set of roots in the tephra, as did ours. Cracking of tephra increased survival on Kodiak Island. In our situation, cracking of tephra crust occurred only in tephra deposited on snow, and may have increased seedling numbers and bryophyte cover, given the correlations of plant importance with snow cover (Table 17). Even after 3 yr burial on Kodiak Island, plants responded rapidly to erosion, as in our sites. Seedlings in the forest on Kodiak Island were rare until 3 yr after the eruption, and they had produced little cover after 5 yr.
Vegetation recovery from burial by tephra outside forest has also been dominated by survivors, except where deposits are very deep (del Moral 1983, Antos and Zobel 1987, Tsuyuzaki 1987, Pfitsch and Bliss 1988, Halpern et al. 1990). Survivors also produced most cover on debris flows in the Oregon Cascades (Gecy and Wilson 1990), and following fire in a Tsuga heterophylla forest (Stickney 1986). Even some mosses survive tephra deposition. In Iceland, tephra layers up to 2 cm deep were immediately incorporated into the 5-10 cm deep moss carpet, and layers 2-5 cm were rapidly overgrown by mosses (Bjarnason 1991); with 2 cm tephra from Mount St. Helens, only the smallest forest bryophytes had not recovered their importance after 2 yr (Antos and Zobel 1985c). In contrast, 2 cm of silt-sized tephra from Mount St. Helens killed steppe mosses within 4 mo in 1980 (Harris et al. 1987).
Vegetation change at our sites shared some characteristics with post-fire succession, the most extensive form of pre-logging vegetation change in the Washington Cascades (Hemstrom and Franklin 1982). The first decade of vegetation change after fire in a moist Tsuga heterophylla-dominated forest habitat in Idaho depended primarily on what survived or initially established, with the successional pattern reflecting differential development among initial species rather than species replacement (Stickney 1986). In contrast, species turnover was more important during early succession on clearcuts in the Cascade Range, although survivors were also important (Halpern 1988). Rapid turnover is especially important during succession in some systems, for example, in fallow fields (e.g., Bornkamm 1988, Myster 1993). In contrast to most secondary successions, including those after fire (Stickney 1986) and logging (Halpern 1988) in the Pacific Northwest, herb cover at our deep tephra sites was still very low relative to pre-disturbance values after 10 yr. The substrate surface did support ruderal moss cover and a heavy colonization by conifer seedlings, as is common on other tephras (Griggs 1922, Bjarnason 1991) and after fire in Tsuga heterophylla forest (Stickney 1986).
The residual tree canopy and the new, primary substrate on the surface distinguish vegetation change after tephra fall in the forest from most secondary successions. The virtual absence of early seral vascular species at our sites probably results from tree canopy effects; numerous local species did establish widely on tephra in the open (del Moral et al. 1995). Low nutrient levels in tephra could prevent establishment of seral species typical of relatively nutrient-rich, early secondary successions, but we recognized no symptoms of nutrient deficiency in tree seedlings in tephra, which is consistent with observations in subalpine areas at Mount St. Helens (del Moral 1983, Pfitsch and Bliss 1988).
Pattern and explanation of vegetation change
Ecologists have long sought repeatable patterns and widely applicable explanations for the recovery of vegetation from disturbance (e.g., Mcintosh 1980, Finegan 1984, Pickett et al. 1987, Glenn-Lewin and van der Maarel 1992, Attiwill 1994). Some are sure that ". . . it is impossible to create general theories on vegetation dynamics . . ." (van der Maarel 1988:16, McCook 1994); others state such theories (Pickett and McDonnell 1989). Whatever the potential for generalization, attempts to generalize should be based on detailed, long-term case studies that represent a range of types of disturbance and ecosystems.
Vegetation recovery following burial of forest understory plants by tephra included several characteristics that must be accommodated by any general description of patterns and mechanisms of vegetation change. For example:
1) The pattern of vegetation changes that were identified varied depending on whether species diversity, plant density, or cover was measured (Tables 3, 5, and 6).
2) Growth forms behaved differently both at a given site and at different sites (Tables 3, 5, and 6). Effects of burial by the same tephra depth would also change with time of year, as plant phenology and snow cover vary.
3) Inertia for species density increased with plant size; an equation relating inertia and size was useful ([r.sup.2] = .85) for predicting inertia for species density of growth forms in stands in a lower elevation vegetation zone, which were not used to develop the equation. Plant size was not significant, however, in equations predicting inertia of cover for either growth forms or important species.
4) Components of vegetation change can differ. Although inertia and one measure of resilience (c/a) showed similar significance patterns, were sensitive to all sources of variation, had the same significant interaction (Table 5), and showed the same patterns among attributes, growth forms, and sites (Table 6), c/b, (c - b)/a, and (c - b)/(a - b) did not vary with attribute or growth form.
5) A decade after burial by tephra, vegetation recovery (described by c/a) was closely related to conditions just after the eruption (described by inertia) [ILLUSTRATION FOR FIGURE 5A OMITTED]. Therefore, to understand better the recovery in our system, one should emphasize those factors that modify the initial damage caused by tephra deposition, rather than the differential expansion after the disturbance. It is not clear how general this conclusion may be.
Our conclusion about the overriding influence of the damage caused by the disturbance may become less tenable as slow changes accumulate. Total recovery, if it occurs, obviously will take much more than a decade for some growth forms and species, given the major differences still present between tephra and cleared plots. Some early events, in particular conifer seedling establishment on deep tephra (Table 13), may deflect the system to a trajectory different from that present prior to the disturbance, producing a dense conifer sapling layer, similar to that in some stands closer to Mount St. Helens prior to the eruption (J. Antos, personal observation, 1979).
6) Where inertia controls the pattern of vegetation change, an early, detailed description of the disturbance is critical. Immediate, detailed attention to properties of the disturbance and the number and condition of survivors is required to maximize understanding the response of the vegetation, as was noted also by del Moral and Bliss (1993) and del Moral et al. (1995).
Vegetation change results from interactions among ecosystem components (Pickett et al. 1987), including the tree canopy, substrate, and understory (Table 18). Litter, snow, tephra deposition patterns, and light intensity are all some complex function of the location, species, and size of overstory tree canopies, constituting what Glenn-Lewin and van der Maarel (1992) call a "third party" effect. This may explain the consistency of the sign of the correlations of plant importance with the various environmental factors at one time (Table 17). Among years, though, correlations often varied in sign.
Usefulness of our data for understanding vegetation change
Advantages and limitations of this study. - Many features of our system facilitated understanding effects of disturbance on vegetation. The disturbance was completely natural, with its origin outside the ecosystems affected, essentially instantaneous, well documented, widespread, and probably the most homogeneous of disturbances over large areas. Measurements were on permanent plots, with their associated advantages (Glenn-Lewin and van der Maarel 1992:35-36). Analysis by growth forms provided a more general result than a species-level analysis, yet retained the details that determine physiognomy.
Other factors constrained both data collection and interpretation. Plot choice was limited by post-eruption safety rules and timber-harvest plans, and to topography where erosion would be minimal. We sampled the shallow tephra sites only once after 1983, although this should cause little difficulty: most vegetation characteristics in shallow tephra did not increase significantly between 1983 and 1989 (Table 7), so that 1989 values should be good estimates of what we would have measured in 1990. The co-occurrence of extensive snowpack during the eruption with herb-rich understories, and of limited snowpack with herb-poor understories, resulted in the confounding of the effects of snowpack with those that control herb richness. It seems reasonable, however, to conclude that differences in inertia between herb-rich and herb-poor sites are a function of snowpack at the time of disturbance, whereas differences in resilience reflect other more permanent differences in site characteristics.
Our sites were chosen to clarify processes associated with recovery from burial by tephra in a single habitat. They were not necessarily typical of surrounding areas receiving the same amounts of tephra. Much of the region is in the lower elevation Tsuga heterophylla vegetation zone, and much had been clear-cut or was in younger stands, with different species compositions, tree canopies, and presumably amounts of damage and rates of recovery than our old-growth stands. Most of the region has moderate to steep slopes, where erosion was more common, and thus recovery probably more rapid than at our sites. In contrast, concave topography accumulated deep water-eroded tephra, which retarded recovery.
Estimation of pre-eruption vegetation. - Pre-disturbance vegetation data are seldom available for unpredictable natural disturbances, precluding calculation of inertia and components of resilience (Attiwill 1994). The nature of the tephra allowed an effective post-eruption estimate of pre-disturbance vegetation. Removal of tephra, mostly large, loose, single grains, was possible with minimal damage to surviving plants (Antos and Zobel 1985c). We observed no major differences in processes of vegetation change between naturally and experimentally eroded plots, once tephra was removed. Our use of the lowest post-1981 value in cleared plots for the value a [ILLUSTRATION FOR FIGURE 1 OMITTED] in our calculations maximized inertia, measures of resilience, and the chance of reaching recovery, and minimized values for the number of years to 50% recovery, compared to alternative choices. Thus, our estimates of the severity of the disturbance are probably conservative. This is particularly true for bryophytes for which cleared plot cover declined substantially before 1989-1990.
Extrapolating study of vegetation change. - The outcome of this tephra deposit cannot be used directly to predict what will happen after future similar deposits. What may seem to be minor differences in conditions at the time of a disturbance may produce major discrepancies in vegetation lasting for decades, perhaps for the life of the stand (McCune and Allen 1985, Glenn-Lewin and van der Maarel 1992, del Moral and Bliss 1993). In our case the timing of burial relative to plant phenology modified its effects. For example, expanded shoots of Veratrum viride, a tall herb with annual shoots, collapsed under a thin tephra deposit (Mack 1981). In contrast, tephra deposited where V. viride was dormant had little effect; plants penetrated 15 cm of natural tephra the following spring, although few shoots escaped a 40-cm-deep experimental addition of tephra (Zobel and Antos 1987a). Veratrum was most resistant to burial for 1-2 wk in spring, as the compact, folded shoots were emerging; most escaped an additional 40 cm of tephra deposited at this time.
The timing of the 1980 eruption was optimal for herb and shrub recovery where snow had melted but deciduous herb shoots and shrub leaves had not yet emerged. In contrast, where tephra fell on snowpack, which persists for 6-9 mo each year, conditions for recovery of the woody plants buried beneath the snow were highly unfavorable (Antos and Zobel 1982). With patchy snowpack, a condition that lasts several weeks each spring, the extremes of favorability for recovery coexisted just centimeters apart, and the consequences of snow remained obvious a decade later [ILLUSTRATION FOR FIGURES 2B AND D OMITTED]. In contrast, snowpack provided the most favorable conditions in other circumstances. Near the edge of the blast zone, where all exposed shoots were killed by heat from the initial eruption, the only surviving woody plants were those buried beneath snow (Means et al. 1982, Halpern et al. 1990), despite their burial by deeper tephra than that received at any of our sites.
Autecological responses of plants to burial by tephra
Most species in our plots survived shallow burial, and some individuals of many species survived deep burial. We sought to explain differences in survival and expansion based on plant structure and physiology. Ours and other data from a decade of study at Mount St. Helens allow us to update and add to our generalizations from the limited pre-1980 literature about plant burial by tephra (Antos and Zobel 1987:251).
1) After a decade, recovery of previously established plants remains more important than seedling establishment in many habitats (Halpern et al. 1990). The vegetative responses allowing survival of burial may have developed as a response to the frequent tephra deposition in the Cascades, but more likely as a response to the common small incidents that bury forest plants beneath leaf litter, bark, logs, and colluvium (Antos and Zobel 1985a).
2) Erosion facilitates plant recovery, especially in deep tephra, where differences in importance were often 10-fold or more between tephra and cleared plots (Tables 3 and 4, [ILLUSTRATION FOR FIGURES 2-4 OMITTED]). Where erosion is common, the pattern of erosion appears to determine what survives more than do plant characteristics (Griggs 1922, del Moral 1983, Franklin et al. 1988). In shallow tephra, the major importance of erosion was to increase survival of bryophytes and, where there was snow, shrub recovery (Tables 3 and 4). Even erosion that was delayed by 2 yr increased recovery. The substantial longevity of plants while buried by tephra (Griggs 1922, Zobel and Antos 1986, 1992), and the capacity of the long-term survivors to expand after release ([ILLUSTRATION FOR FIGURES 2-4, 6 OMITTED]: plots cleared at site DR in 1982), enhance the effect of erosion.
3) Some exhumed plants recover after [greater than or equal to]3 yr of burial. Although buried Pinaceae and aerial shoots of shrubs survived less than a year, most understory species retained survivors after three seasons of burial (Zobel and Antos 1986), and at least two shrub, one herb, and three moss species grew when uncovered at the end of their eighth season of burial (Zobel and Antos 1992). Individuals of all species exhumed in 1987 survived 9 yr after they were released (D. Zobel, personal observation, 1996). Longevity during burial is not universal for plants that dominate sites where burial is common; sand dune plants die in darkness in [less than]5 mo (Sykes and Wilson 1990).
4) Plants may emerge from tephra as long as 10 yr after burial. Many individuals of Vaccinium membranaceum, a dominant shrub, escaped the tephra for the first time in the 6th through 10th growing seasons after burial, both where tephra had been thinned by erosion and in spots with no sign of erosion. Excavation confirmed that many delayed emergent shoots came from buried belowground tissue that was not connected to any previously emerged shoot (Zobel and Antos 1992). Although we recorded no certain instances in which herbs first penetrated the tephra after the second year of burial, Rubus pedatus in Alaska emerged up to 4 yr after burial (Griggs 1922).
5) Plant height and rate of vegetative spread were less strongly related to success after the eruption than we had initially estimated (Antos and Zobel 1985a). Using plant size did not improve our ability to predict inertia for cover in our forested permanent plots. This may be due to the plasticity of plant form (Antos and Zobel 1985a, b): after burial, long-rhizomatous herbs slightly altered the orientation of rhizome growth, whereas short-rhizomatous dicots produced nearly vertical rhizomes with long internodes, which grew to near the surface in one season (Antos and Zobel 1985b). Erythronium montanum, which survived well but did not move perennating organs upward, did not increase its importance in tephra during the decade of recovery, while associated herbs did. A few species with little capacity to expand were severely injured by burial, such as Gaultheria humifusa and, in deeper tephra, Xerophyllum tenax.
Important species of trees and shrubs varied less than herbs.
6) Woody plants survive deeper deposits than herbaceous plants, except (1) when stems are trapped beneath the tephra where it fell on snowpack, and (2) for some low-statured subshrubs with prostrate, slender woody stems (treated here in the herb layer).
7) Growth of woody plants usually declines but sometimes increases after tephra deposition (Hinckley et al. 1984, Zobel and Antos 1985, Segura et al. 1994). Growth increases of saplings in clearcuts (Zobel and Antos 1985) reversed with time (J. Antos and D. Zobel, personal observations, 1987).
8) Most, but not all, surviving plants produce adventitious roots in the tephra. Many small conifers rooted adventitiously in 15 cm of tephra (Zobel and Antos 1982), as had larger trees in earlier eruptions (Lawrence 1954, J. Antos and D. Zobel, personal observations, 1982). Although many shrubs produced abundant adventitious roots within 2 yr after the eruption, Ericaceae dominant in our plots produced few (Antos and Zobel 1985a). Most herbaceous survivors produced roots in tephra before 1982 (Antos and Zobel 1985a).
9) Species that normally colonize nutrient-poor substrates should have an advantage on tephra, where the concentration of available nutrients is often low. In our area, probably only N and P would be low (del Moral and Clampitt 1985, Zobel and Antos 1991a). This generalization does not apply well to our sites, because most species establishing on the tephra surface were already present, except for a few widespread mosses. This generalization seems more appropriate in the absence of a residual canopy (Wood and del Moral 1987, del Moral and Bliss 1993:38, del Moral and Wood 1993), as the canopy inhibits establishment of ruderal vascular species and supplies nutrients via litter. Tolerance of herbs to low nutrient levels is associated with high inertia after frost, fire, and drought, but with low resilience (MacGillivray et al. 1995); in our case, there was no inverse relationship of inertia with measures of resilience [ILLUSTRATION FOR FIGURE 5 OMITTED].
10) The spatial pattern of survivors and initial colonizers is reinforced by plant growth and reproductive properties. Several successful species in our plots, including Rubus lasiococcus, R. pedatus, Mitella spp., Vaccinium spp., and some bryophytes, produced clumped distributions on tephra by vegetative spread. Seedling establishment around survivors also produced patchy distribution patterns, such as for Erythronium montanum. In non-forest habitat, patterns of plant distribution also reflect both vegetative and reproductive spread by the initial inhabitants, both survivors and colonizers (Andersen and MacMahon 1985, Tsuyuzaki 1987, del Moral and Bliss 1993). This pattern of plant spread reinforced the already patchy pattern of surviving plants resulting from the influence on tephra distribution of the tree canopy, woody litter, and surviving understory plants.
11) Stress tolerance is critical to survival in tephra; reproductive mode is less critical. All successful post-eruption vascular species were perennials already present in the forest understory. These species spread little outside forest even where they survived (del Moral et al. 1995). Within this group, however, reproductive and growth patterns varied (Antos and Zobel 1984, 1985a), with success both by relatively short-lived, short-rhizomatous species reproducing by seed, and by longer-lived species that spread vegetatively but produced few or no seedlings. The vegetatively spreading form was even more successful at lower elevations in 1980-1983 (J. A. Antos and D. B. Zobel 1985c, unpublished data). In contrast, among mosses the vagile colonizing species, not obvious in pre-eruption forests (although their spores apparently were present in the buried forest floor [Zobel and Antos 1992]), formed most cover on the tephra surface. Two major factors appear to differentiate the behavior of vascular plants and bryophytes that spread on tephra: (a) the tephra constituted a primary successional, low-nutrient habitat for the unrooted bryophytes, but not for rooted, mycorrhizal species, and (b) unlike pioneer vascular plants, the widespread weedy bryophytes tolerated shade.
Numerous authors emphasize the importance of life-history traits for understanding and predicting vegetation change. In subalpine habitats at Mount St. Helens, colonizers require stress tolerance of seedlings and, for isolated sites, long-distance dispersal of seeds (del Moral and Bliss 1993). It remains to be demonstrated how widely general correlations among life-history traits can allow prediction of vegetation change, as outlined, e.g., by Cattelino et al. (1979), Grime (1979), and McCook (1994). Among our species, inertia for species density was related to plant dimensions, but inertia for cover was not. The life-history requirements for success after disturbance seem likely to be specific to the situation. In our study, properties limiting damage by burial were more critical than those governing post-disturbance spread.
The variation patterns of forest understory vegetation a decade after the deposit of 4-15 cm of airfall tephra from Mount St. Helens were a complex function of the growth form, tephra depth, presence of snowpack under the tephra, and pre-eruption flora. The pattern also varied depending on the measure of importance used. If this instance is representative of vegetation responses to disturbance, generalizations about vegetation change will remain difficult to identify.
One conclusion from our results, if widespread, may simplify study of early succession where survivors are important: the major differences in importance of nontree growth forms at the end of the decade were established at the time of disturbance, i.e., the proportion of pre-eruption importance of a growth form that was present at the end of the decade was highly correlated with inertia for that growth form. The amount of change in importance during the decade of recovery did not differ among growth forms or measures of importance. To clarify plant responses to tephra and possibly other disturbances, attention should be focussed on the factors that control the initial damage caused to the vegetation by the disturbance, rather than on differences that control how plants expand their importance following the disturbance.
Our results emphasize the importance of a detailed description of the disturbance, the importance of survivors among vascular plants, the slow recovery of herbs and bryophytes from deep tephra, the continuing impact of initial conditions, and the effects of the substantial plasticity and longevity of the subalpine forest understory species following burial.
TABLE 18. Interactions between the overstory canopy, substrate, and understory plants, which constitute mechanisms of vegetation change (Pickett et al. 1987) for our situation.
A. Canopy effects on substrate
Produces variable tephra texture and thickness Produces litter Supports tree root entry into new tephra from below Modifies precipitation
B. Canopy effects on understory
Produces variable light intensity Produces litter Modifies precipitation
C. Substrate effects on understory
Damages plants Regulates survival, initially and through time Regulates seedling establishment Provides rooting medium Supplies nutrients and water
D. Understory effects on substrate
Produces roots and rhizomes in tephra Produces litter
Sampling from 1980 through 1984 was supported by the National Science Foundation and the U.S. Department of Agriculture's Science and Education Administration. The Natural Sciences and Engineering Research Council of Canada and Oregon State University provided partial support for sampling in 1987-1990. Cooperation of the U.S. Forest Service was critical to establishing and protecting our plots. Help in the field from Charles Halpern, Brad Smith, Susan Seyer, Michael Ryan, Joanna Smith, and especially Tom Hill made this study feasible. Brenda Castanzo, Ornella Cirella, and Gerry Allen helped produce the figures. We thank Bruce McCune, Mark V. Wilson, and the anonymous referees for detailed reviews of the manuscript.
Alaback, P. B., and J. C. Tappeiner II. 1991. Response of western hemlock (Tsuga heterophylla) and early huckleberry (Vaccinium ovalifolium) seedlings to forest wind-throw. Canadian Journal of Forest Research 21:534-539.
Allison, S. K. 1995. Recovery from small-scale anthropogenic disturbances by northern California salt marsh plant assemblages. Ecological Applications 5:693-702.
Andersen, D. C., and J. A. MacMahon. 1985. Plant succession following the Mount St. Helens volcanic eruption: facilitation by a burrowing rodent, Thomomys talpoides. American Midland Naturalist 114:62-69.
Antos, J. A. 1988. Underground morphology and habitat relationships of three pairs of forest herbs. American Journal of Botany 75:106-113.
Antos, J. A., and D. B. Zobel. 1982. Snowpack modification of volcanic tephra effects on forest understory plants near Mount St. Helens. Ecology 63:1969-1972.
Antos, J. A., and D. B. Zobel. 1984. Ecological implications of belowground morphology of nine coniferous forest herbs. Botanical Gazette 145:508-517.
Antos, J. A., and D. B. Zobel. 1985a. Plant form, developmental plasticity, and survival following burial by volcanic tephra. Canadian Journal of Botany 63:2083-2090.
Antos, J. A., and D. B. Zobel. 1985b. Upward movement of underground plant parts into deposits of tephra from Mount St. Helens. Canadian Journal of Botany 63:2091-2096.
Antos, J. A., and D. B. Zobel. 1985c. Recovery of forest understories buried by tephra from Mount St. Helens. Vegetatio 64:103-111.
Antos, J. A., and D. B. Zobel. 1986. Seedling establishment in forests affected by tephra from Mount St. Helens. American Journal of Botany 73:495-499.
Antos, J. A., and D. B. Zobel. 1987. How plants survive burial: a review and initial responses to tephra from Mount St. Helens. Pages 246-261 in D. E. Bilderback, editor. Mount St. Helens 1980: botanical consequences of an explosive eruption. University of California Press, Berkeley, California, USA.
Attiwill, P. M. 1994. The disturbance of forest ecosystems: the ecological basis for conservative management. Forest Ecology and Management 63:247-300.
Bardon, R. E., D. W. Countryman, and R. B. Hall. 1995. A reassessment of using light-sensitive diazo paper for measuring integrated light in the field. Ecology 76:1013-1016.
Bjarnason, A. H. 1991. Vegetation on lava fields in the Hekla area, Iceland. Acta Phytogeographica Suecica 77.
Bornkamm, R. 1988. Mechanisms of succession on fallow lands. Vegetatio 77:95-101.
Cattelino, P. J., I. R. Noble, R. O. Slatyer, and S. R. Kessell. 1979. Predicting the multiple pathways of plant succession. Environmental Management 3:41-50.
Chapin, F. S., III. 1993. Functional role of growth forms in ecosystem and global processes. Pages 287-312 in J. R. Ehleringer and C. B. Field, editors. Scaling physiological processes, leaf to globe. Academic Press, San Diego, California, USA.
Daubenmire, R., and J. B. Daubenmire. 1968. Forest vegetation of Eastern Washington and Northern Idaho. Washington Agricultural Experiment Station Technical Bulletin 60.
Day, R. W., and G. P. Quinn. 1989. Comparisons of treatments after an analysis of variance in ecology. Ecological Monographs 59:433-463.
del Moral, R. 1983. Initial recovery of subalpine vegetation on Mount St. Helens, Washington. American Midland Naturalist 109:72-80.
del Moral, R., and L. C. Bliss. 1993. Mechanisms of primary succession: insights resulting from the eruption of Mount St. Helens. Advances in Ecological Research 24:1-66.
del Moral, R., and C. A. Clampitt. 1985. Growth of native plant species on recent volcanic substrates from Mount St. Helens. American Midland Naturalist 114:374-383.
del Moral, R., J. H. Titus, and A. M. Cook. 1995. Early primary succession on Mount St. Helens, Washington, USA. Journal of Vegetation Science 6:107-120.
del Moral, R., and D. M. Wood. 1993. Early primary succession on the volcano Mount St. Helens. Journal of Vegetation Science 4:223-234.
Falinski, J. B. 1988. Succession, regeneration and fluctuation in the Bialowieza Forest (NE Poland). Vegetatio 77:115-128.
Finegan, B. 1984. Forest succession. Nature 312:109-114.
Franklin, J. F., and C. T. Dyrness. 1973. Natural vegetation of Oregon and Washington. USDA Forest Service General Technical Report PNW-8.
Franklin, J. F., P.M. Frenzen, and F. J. Swanson. 1988. Recreation of ecosystems at Mount St. Helens: contrasts in artificial and natural approaches. Pages 1-37 in J. Cairns, editor. Rehabilitating damaged ecosystems. Volume II. CRC Press, Boca Raton, Florida, USA.
Friend, D. T. C. 1961. A simple instrument for measuring integrated light values in the field. Ecology 42:577-580.
Gecy, J. L., and M. V. Wilson. 1990. Initial establishment of riparian vegetation after disturbance by debris flows in Oregon. American Midland Naturalist 123:282-291.
Glenn-Lewin, D. C., and E. van der Maarel. 1992. Patterns and processes of vegetation dynamics. Pages 11-59 in D. C. Glenn-Lewin, R. K. Peet, and T. T. Veblen, editors. Plant succession. Theory and prediction. Chapman & Hall, London, England.
Griggs, R. F. 1918. The recovery of vegetation at Kodiak. Ohio Journal of Science 19:1-57.
-----. 1922. The Valley often Thousand Smokes. National Geographic Society, Washington, D.C., USA.
Grime, J. P. 1979. Plant strategies and vegetation processes. John Wiley & Sons, Chichester, England.
Grubb, P. J., and A. J. M. Hopkins. 1986. Resilience at the level of the plant community. Pages 21-38 in B. Dell, A. J. M. Hopkins, and B. B. Lamont, editors. Resilience in Mediterranean-type ecosystems. Junk, Dordrecht, The Netherlands.
Halpern, C. B. 1988. Early successional pathways and the resistance and resilience of forest communities. Ecology 69:1703-1715.
-----. 1989. Early successional patterns of forest species: interactions of life history traits and disturbance. Ecology 70:704-720.
Halpern, C. B., P. M. Frenzen, J. E. Means, and J. F. Franklin. 1990. Plant succession in areas of scorched and blown-down forest after the 1980 eruption of Mount St. Helens, Washington. Journal of Vegetation Science 1:181-194.
Harmon, M. E., J. F. Franklin, F. J. Swanson, P. Sollins, S. V. Gregory, J. D. Lattin, N. H. Anderson, S. P. Cline, N. G. Aumen, J. R. Sedell, G. W. Lienkaemper, K. Cromack, Jr., and K. W. Cummins. 1986. Ecology of coarse woody debris in temperate ecosystems. Advances in Ecological Research 15:133-302.
Harris, E., R. N. Mack, and M. S. B. Ku. 1987. Death of steppe cryptogams under the ash from Mount St. Helens. American Journal of Botany 74:1249-1253.
Hemstrom, M. A. 1982. Fire in the forests of Mount Rainier National Park. Pages 121-126 in E. E. Starkey, J. F. Franklin, and J. W. Matthews, technical coordinators. Ecological research in National Parks of the Pacific Northwest. Proceedings of the second conference on scientific research in the National Parks, 1979. Oregon State University Forest Research Laboratory, Corvallis, Oregon, USA.
Hemstrom, M. A., and J. F. Franklin. 1982. Fire and other disturbances of the forests in Mount Rainier National Park. Quaternary Research 18:32-51.
Hinckley, T. M., H. Imoto, K. Lee, S. Lacker, Y. Morikawa, K. A. Vogt, C. C. Grief, M. R. Keyes, R. O. Teskey, and V. Seymour. 1984. Impact of tephra deposition on growth in conifers: the year of the eruption. Canadian Journal of Forest Research 14:731-739.
Hitchcock, C. L., and A. Cronquist. 1973. Flora of the Pacific Northwest. An illustrated manual. University of Washington Press, Seattle, Washington, USA.
Lawrence, D. B. 1954. Diagrammatic history of the northeast slope of Mt. St. Helens, Washington. Mazama 36:41-44.
Lawton, E. 1971. Moss flora of the Pacific Northwest. The Hattori Botanical Laboratory, Nichinan, Miyazaki, Japan.
MacGillivray, C. W., J. P. Grime, and the Integrated Screening Programme (ISP) Team. 1995. Testing predictions of the resistance and resilience of vegetation subjected to extreme events. Functional Ecology 9:640-649.
Mack, R. N. 1981. Initial effects of ashfall from Mount St. Helens on vegetation in eastern Washington and adjacent Idaho. Science 213:537-539.
McCook, L. J. 1994. Understanding ecological community succession: causal models and theories, a review. Vegetatio 110:115-147.
McCune, B., and T. F. H. Allen. 1985. Will similar forests develop on similar sites? Canadian Journal of Botany 63: 367-376.
McIntosh, R. P. 1980. The relationship between succession and the recovery process in ecosystems. Pages 11-62 in J. Cairns, Jr., editor. The recovery process in damaged ecosystems. Ann Arbor Science Publishers, Ann Arbor, Michigan, USA.
McIntyre, S., S. Lavorel, and R. M. Tremont. 1995. Plant life-history attributes: their relationship to disturbance response in herbaceous vegetation. Journal of Ecology 83: 31-44.
Means, J. E., W. A. McKee, W. H. Moir, and J. F. Franklin. 1982. Natural revegetation of the northeastern portion of the devastated area. Pages 93-103 in S. A. C. Keller, editor. Mount St. Helens: one year later. Eastern Washington University Press, Cheney, Washington, USA.
Mullineaux, D. R. 1986. Summary of pre-1980 tephra-fall deposits erupted from Mount St. Helens, Washington State, USA. Bulletin of Volcanology 48:17-26.
Myster, R. W. 1993. Tree invasion and establishment in old fields at Hutcheson Memorial Forest. Botanical Review 59: 251-272.
Pfitsch, W. A., and L. C. Bliss. 1988. Recovery of net primary production in subalpine meadows of Mount St. Helens following the 1980 eruption. Canadian Journal of Botany 66:989-997.
Pickett, S. T. A., S. L. Collins, and J. J. Armesto. 1987. Models, mechanisms and pathways of succession. Botanical Review 53:335-371.
Pickett, S. T. A., and M. J. McDonnell. 1989. Changing perspectives in community dynamics: a theory of successional forces. Trends in Ecology and Evolution 4:241-245.
Pickett, S. T. A., and P.S. White, editors. 1985. The ecology of natural disturbance and patch dynamics. Academic Press, New York, New York, USA.
Segura, G., L. B. Brubaker, J. F. Franklin, T. M. Hinckley, D. A. Maguire, and G. Wright. 1994. Recent mortality and decline in mature Abies amabilis: the interaction between site factors and tephra deposition from Mount St. Helens. Canadian Journal of Forest Research 24:1112-1122.
Sheets, P. D., and D. K. Grayson, editors. 1979. Volcanic activity and human ecology. Academic Press, New York, New York, USA.
Shimwell, D. W. 1971. The description and classification of vegetation. University of Washington Press, Seattle, Washington, USA.
Shipley, S., and A. M. Sarna-Wojcicki. 1983. Distribution, thickness, and mass of late Pleistocene and Holocene tephra from major volcanoes in the northwestern United States: a preliminary assessment of hazards from volcanic ejecta to nuclear reactors in the Pacific Northwest. United States Geological Survey Miscellaneous Field Studies Map MF-1435.
Sousa, W. P. 1984. The role of disturbance in natural communities. Annual Review of Ecology and Systematics 15: 353-391.
Stickney, P. F. 1986. First decade plant succession following the Sundance forest fire, northern Idaho. USDA Forest Service General Technical Report INT-197.
Sykes, M. T., and J. B. Wilson. 1990. Dark tolerance in plants of dunes. Functional Ecology 4:799-805.
Tappeiner, J. C., II, and P. B. Alaback. 1989. Early establishment and vegetative growth of understory species in the western hemlock-Sitka spruce forests of southeast Alaska. Canadian Journal of Botany 67:318-326.
Tsuyuzaki, S. 1987. Origin of plants recovering on the volcano Usu, northern Japan, since the eruptions of 1977 and 1978. Vegetatio 73:53-58.
van der Maarel, E. 1988. Vegetation dynamics: patterns in time and space. Vegetatio 77:7-19.
Waitt, R. B., Jr., and D. Dzurisin. 1981. Proximal air-fall deposits from the May 18 eruption - stratigraphy and field sedimentology. Pages 601-616 in P. W. Lipman and D. R. Mullineaux, editors. The 1980 eruptions of Mount St. Helens, Washington. United States Geological Survey Professional Paper 1250.
Webb, S. L. 1989. Contrasting windstorm consequences in two forests, Itasca State Park, Minnesota. Ecology 70: 1167-1180.
Westman, W. E. 1978. Measuring the inertia and resilience of ecosystems. BioScience 28:705-710.
----- 1986. Resilience: concepts and measures. Pages 5-19 in B. Dell, A. J. M. Hopkins, and B. B. Lamont, editors. Resilience in Mediterranean-type ecosystems. Junk, Dordrecht, The Netherlands.
White, P. S. 1979. Pattern, process, and natural disturbance in vegetation. Botanical Review 45:229-299.
Whittaker, R. H. 1972. Evolution and measurement of species diversity. Taxon 21:213-251.
Wood, D. M., and R. del Moral. 1987. Mechanisms of early primary succession in subalpine habitats on Mount St. Helens. Ecology 68:780-790.
Yamaguchi, D. K. 1983. New tree-ring dates for recent eruptions of Mount St. Helens. Quaternary Research 220:246-250.
Zobel, D. B., and J. A. Antos. 1982. Adventitious rooting of eight conifers into a volcanic tephra deposit. Canadian Journal of Forest Research 12:717-719.
Zobel, D. B., and J. A. Antos. 1985. Response of conifer shoot elongation to tephra from Mount St. Helens. Forest Ecology and Management 12:83-91.
Zobel, D. B., and J. A. Antos. 1986. Survival of prolonged burial by subalpine forest understory plants. American Midland Naturalist 115:282-287.
Zobel, D. B., and J. A. Antos. 1987a. Survival of Veratrum viride, a robust herbaceous plant, when buried by volcanic tephra. Northwest Science 61:20-22.
Zobel, D. B., and J. A. Antos. 1987b. Composition of rhizomes of forest herbaceous plants in relation to morphology, ecology, and burial by tephra. Botanical Gazette 148: 490-500.
Zobel, D. B., and J. A. Antos. 1991a. 1980 tephra from Mount St. Helens: spatial and temporal variation beneath forest canopies. Biology and Fertility of Soils 12:60-66.
Zobel, D. B., and J. A. Antos. 1991b. Growth and development of natural seedlings of Abies and Tsuga in old-growth forest. Journal of Ecology 79:985-998.
Zobel, D. B., and J. A. Antos. 1992. Survival of plants buried for eight growing seasons by volcanic tephra. Ecology 73: 698-701.
|Printer friendly Cite/link Email Feedback|
|Author:||Zobel, Donald B.; Antos, Joseph A.|
|Date:||Aug 1, 1997|
|Previous Article:||Effects of drought and prolonged winter on Townsend's ground squirrel demography in shrubsteppe habitats.|
|Next Article:||Species assemblages and indicator species: the need for a flexible asymmetrical approach.|