BRAUN-BLANQUETIA, vol. 46, 2010 217

LESSONS FOR RESTORATION OF PROTECTED LANDS

Roger DEL MORAL

Department of Biology, Box 351800, University of Washington, Seattle, Washington (USA)E-mail: moral@uw.edu

ABSTRACT

Primary succession requires ame-lioration, dispersal, establishment anddevelopment. Biotic interactions, land-scape effects, changing safe-site quali-ties and chance all affect succession.Studies on many volcanoes (e.g. Etna,Fuji, St. Helens, Ksudach, Taraweraand Kilauea) reveal lessons to improverestoration of damaged habitats withinprotected areas. Successional trajecto-ries develop at different rates due tostress and in alternative directions dueto priority effects and dispersal limita-tions. Natural mosaics are the result.Recovery from major disturbances canbe hastened by alleviating stress throu-ghappropriate fertilization,while takingcare to avoid competitive effects. Evenvery short distances limit dispersal ofmost species, so managers must conti-nually facilitated dispersal. Competi-tion and herbivory can retard and de-flect succession, so active managementmay be needed. Natural processes canproduce several alternative, stable, butstill natural, plant communities, so amosaic of vegetation in protected natu-ral areas should be encouraged. Moni-toring through permanent plots is requi-red to detect invasions by alien speciesand responses to climatic changes. Theinformation from long-term monitoringhas deepened the understanding of suc-cession processes and can lead to moreeffective vegetation management.

KEYWORDS: Mount St. Helens, Plant di-spersal, Primary succession, Volcano-es.

INTRODUCTION

Volcanoes occur throughout theworld, at all latitudes and in most clima-tes. They are concentrated on islandsassociatedwithspreadingseafloors (Ice-land; Rift Valley of Africa), where con-tinental plates meet (the Sunda Arc nearIndonesia; west coast of South Ameri-ca) and over geological hotspots (Gala-pagos Islands, Réunion, YellowstonePark). Volcanoes are very diverse in

their environments. Mt. Etna, formedby the collision of Africa and Europe isin a Mediterranean climate (POLI MAR-CHESE & GRILLO, 2000). Mt. Kilauea, ayoung hot spot volcano is tropical at itsbase, alpine at it summit and supportsrainforests and deserts (K ITAYAMA etal., 1995). Mt. Tarawera occurs in apopulatedwarmtemperatezone(CLARK-SON & CLARKSON, 1995). Japan’s Mt.Fuji reaches to permanent snows from asubtropical habitat (MASUZAWA, 1985)and is climbed by thousands of pilgrimseach year. Mt. Ksudach, at the tip of theKamchatka Peninsula is in a harsh, re-mote continental, sub-boreal zone(GRISHIN et al., 1996) and is rarely visi-ted. Mount St. Helens, in a cool tempe-rate coniferous zone is the most studiedvolcano on the planet (DEL MORAL et al.,2005). The biota of many volcanoes isprotected by regulations, while othersare protected by their isolation. Mosthave no legal or practical protection.

Volcanoes are important to biolo-gical complexity because they can pro-vide refugia for many species, and theirenvironments contrastwith the surroun-dings. They also offer recreation andtourism values. Studies on many volca-noes have led to greater understandingof primary succession (cf. WALKER &DEL MORAL, 2003; DEL MORAL &WALKER, 2007), but more importantly,such studies are leading to greater un-derstanding of restoration (WALKER etal., 2006; DEL MORAL et al., 2006).Without fundamental research into the

mechanisms of succession, restorationefforts will be inefficient and often un-successful.

This paper summarizes lessons le-arned during 26 years of study on MountSt. Helens. I will describe how some ofthese lessons can be applied to retur-ning derelict lands to productive andefficient ecosystems.

SUCCESSIONAL MECHANISMS

The outline of primary successionmechanisms(Fig.1) showscriticalpoin-ts, sometimes called bottlenecks, wherecommunity assembly may be delayedby natural processes or assisted by ma-nagement efforts. Plant establishmentafter a major disturbance normally re-quires physical amelioration before co-lonizing seeds can survive. Nutrientsmay arrive as dust, pollen, non-viableseeds, excreta of passing birds or largeanimals and even in the form of insectsand spiders who soon perish after de-scending into an inhospitable landsca-pe. Erosion may remove unstable sub-strates to reveal suitable terrain, whilecracks may develop to trap seeds. Suchsafe-sites are crucial in the early stagesof community assembly.

Dispersal mechanisms continual-ly feed the new landscape. At first,seeds are rare, and their main effect is toadd nutrients to the substrate. The spe-cies pool is more limited that is usuallyappreciated (cf. DEL MORAL & ECKERT,

Fig. 1 - Mechanisms of primary succession, determined from studies on Mount St. Helens.

218 BENSETTITI A., BIORET F., BOULLET V., PEDROTTI F., Centenaire de la Phytosociologie

2004). A few tens of meters is sufficientto isolate a site from most invaders anddistance alone leads to substantial hete-rogeneity in early communities becau-se the seed rain is sparse and capricious(DEL MORAL & ELLIS, 2004). Landsca-pes differ in the degree to which theyresist dispersal (called permeability). Iffavorable microsites dot the site, thendispersal will be more efficient (JONES

& DEL MORAL, 2005). Wind direction,animal vectors and the local presence ofspecies adapted to stressful conditionsare among factors that affect dispersalrates and hence colonization.

In some sites, patches of survivorsoccur in refugia. On volcanoes, refugiaoccur on high ground that does notreceive lava flows or on slopes pro-tected from eruptive blasts. Survivorsmay contribute disproportionately toearly colonization (FRANKLIN & MAC-MAHON, 2000), but this effect is re-stricted where surviving species arepoorly adapted to the stressful condi-tions found after a major disturbance.

Once physical amelioration anddispersal have set the stage, seedlingscan begin to establish. Survival may beenhanced by other species (facilitation),but the balance between positive andnegative effects of neighbors is com-plex (WALKER, 1993; CALLAWAY &WALKER, 1997; BELLINGHAM et al.,2001). Nutrient status, drought stress,plant density and substrate age are a fewfactors that determine this balance (DEL

MORAL & ROZZELL, 2005). As the firstcolonists mature, they alter conditionsso that additional colonists can esta-

blish. In addition, competitive inhibi-tion, herbivory and seed predation(BISHOP, 2002) affect survivorship andthe relative composition of the assem-bling community.

The rate of succession is normallydeterminedbyenvironmentalstress (DEL

MORAL, 2006), so even when the vege-tation trajectory tends towards a stableequilibrium, a mosaic develops on thelandscape. The mosaic represents thesame community at different stages ofsuccessional development. On MountSt. Helens, permanent plots have recor-ded species composition changes on aridge near the crater since 1984. Twen-ty plots are arrayed between 1218 and1468 m. While the plots suffered thesame devastation, differences in eleva-tion and residual soils have led to diffe-rent rates of development. DCA wasapplied to the matrix of plots by yearsand I aggregated the plots by their com-position in 2005 into four categoriesthat were strongly correlated with ele-vation. During the 22 years of develop-ment, individual plots developed throu-gh several associations. Early in theprocess, groups were sparse and poor inspecies. Through time, plots underwenttransitions so that by 2005, they were inthree groups, and seven groups were nolonger present. Group H was the leastdeveloped and occurred at the highestelevations. Group I occurred in ten in-termediate sites, while Group J, contai-ning the best developed, most diverseplots wasat the lowest elevations. DCA-1 scores are a good indicator of vegeta-tion cover and composition. The mean

scores declined most rapidly in GroupJ, least in Group H (Fig. 2). A linearregression of plot age with DCA wassignificant in each case. The relation-ship was strongest in the two samples ofGroup I in which change was greatest.Group H changes only moderately,while most of the changes in Group Joccurred by 1999. Other studies onMount St. Helens have demonstratedsimilar differences in development ra-tes based on environmental stress.

The mature vegetation that formsafter major disturbances is often similarto that which once occupied the site, butthis is not a requirement of succession.The degree to which there are survivorsfrom before the disturbance may redu-ce initial variation and increases thesimilarities between pre- and post-di-sturbance vegetation. However, manyfactors can cause vegetation to form astable community distinct from its sur-roundings and from earlier vegetationon the site (MCCUNE & ALLEN, 1985).These factors include different availa-ble species pools (due to changes insurrounding land use, invasions by exo-tic species, etc.), changes in climatecompared to the last establishment win-dow, differences in the intensity of theinitiating disturbance, patches of survi-ving plants, seeds or soil, and variousstochastic factors. Development rates,trajectories and stable species composi-tion after primary succession are lesspredictable than after secondary suc-cession (TURNER & DALE, 1998). Themain cause of this phenomenon is thatsurviving organisms are a signal fromthe previous community that conditionsdevelopment and limits developmentaloptions.

FACTORS THAT AFFECTTRAJECTORIES

On Mount St. Helens, similar siteshave not necessarily developed towar-ds similar composition (DEL MORAL &JONES, 2002). Two physically similarsites may retain dissimilar vegetationdue to priority effects related to stocha-stic dispersal events, small differencesin distance from colonists and subse-quent unusual events that are not uni-formly distributed on the landscape.

PRIORITY EFFECTS

To illustrate some of these proces-ses, I selected data from a grid of conti-guous 100 m2 plots that have been sam-pled annually since 1989 (Fig. 3). The

Fig. 2 - Changes in mean DCA-1 scores in four composite samples of vegetation onStudebaker Ridge, Mount St. Helens. Low elevation plots (J) changes most dramaticallyand stabilized by 2000; high elevation plots (H) changed little, and remain sparse.Intermediate plots (I, two samples) continue to develop; lower samples have progressedmore quickly and may be converging with association J.

BRAUN-BLANQUETIA, vol. 46, 2010 219

first five plots in each of 20 rows wereanalyzed in the years shown. Clearly,succession has occurred, as the strongdecline in DCA-I between 1989 and2003; then, while DCA-1 changed litt-le, DCA-2 increased strongly. Therewas a decline in standard deviationsconsistent with convergence from 1989to 1995, but there was an increase in1999 as mosses began to invade. Theshift in DCA-2 was due to the unex-pected crash of Lupinus lepidus. Thiswas the result of an unusual climaticevent and to the resumption of volcanicactivity in late 2004 that has inhibitedmosses. These unforeseen results alsoled to a reduction in the SD. These dataillustrate that long-term monitoring isrequired and that stochastic geologicaland climatic events can influence thetrajectory of recovering vegetation.

A particular site will develop indifferent directions if the first colonistsdiffer. One indication of such effectscomes from comparisons of lupine co-lonieswith adjacent siteswith only spar-se lupines. The colonies were stratifiedby age, based on work by DEL MORAL &ROZZELL (2005), who compared struc-tural differences in lupine colonies withthe adjacent sites (Tab. 1). The data forone site were frequency determinationsin 500 20 by 20 cm quadrats per site.Dominance (1 - Simpson’s index) anddiversity (Hp) were calculated from the-se data. To assess the effects on otherspecies, calculations were also madeafter excluding lupines.

All sparse sites had significantlylower frequencies when lupines wereincluded; but even excepting lupines,vegetation was denser in lupine colo-nies. Dominance was greater only inmature sites with lupines, while exclu-ding lupines showed only stronger do-minance in young, sparse sites. Diversi-ty was greater in mature sparse siteswhen lupines were included. When lu-pines were excluded, diversity was gre-ater in old and young colonies. Theseresults demonstrate that the presence oflupines facilitates vegetation develop-ment and that this effect increases withage. However, the effect is complexbecause mosses are the major benefi-ciaries of lupine presence, and they tendto restrict the development of other spe-cies.

The effect of lupines clearly in-fluenced vegetation trajectories (Fig.4). Lupine colony composition chan-ged with age, as did the adjacent sparsesites.Coloniesdifferedmoderately fromadjacent sites, which is an effect thatwas due primarily to lupines, but also tochanges in species composition among

the other species. Variation among thecolonies was less than among sparsesites (note the very large SD in each setof sparse sites).

These and several other long-termstudies on Mount St. Helens clearlydemonstrate the importance of the spe-cies that first colonize a site. Pioneersalter the subsequent rules of establish-ment and affect the rate at which bio-mass and nutrients accumulate.

GRADIENT EFFECTS

The distance from a source of ve-getation affects species composition.Based on an earlier study (FULLER & DEL

MORAL, 2003), we showed that the ef-fects of dense surviving vegetation atte-nuate sharply with distance (DEL MORAL

& ECKERT, 2005). Beyond 8 m of survi-ving vegetation, there was little effecton percent cover or richness (Fig. 5) asboth measures reached the backgroundlevel. However, in that study, few sur-viving species were capable of inva-ding the bare pumice, and the limitedeffect was due to pioneer species firstestablishing in the refugia, then disper-sing copious seeds into the surroundin-gs. The floristics changed significantlywith distance from refugia, as demon-strated by DCA (Fig. 6). Sites adjacentto the refugia were diverse, but all occuron the right side of the graph, whilethose more distant occur scattered onthe left side of the graph.

DEL MORAL & ELLIS (2004) alsodemonstrated the gradient effects of

distance in a different way. Four tran-sects consisting of 100 m2 plots wereestablished at regular distances fromthe edge of intact vegetation on a largelahar deposited 23 years prior to thestudy. Transects were at increasing ele-vation, and thus each was more stres-sful than the previous. Species richnessand plot cover declined with distanceand with elevation. Heterogeneity (me-asured by percent similarity among 25sub-samples and by standard deviationsof these values) also declined signifi-cantly. Within plot similarity declinedfrom57.4%to30%withelevation,whilesimilarity between plots declined signi-ficantly (from 71.2% to 45.4%) and theSD doubled. There is no question that

Fig. 3 - Changes in mean DCA ordination scores among contiguous plots on a sample grid,Mount St. Helens. Samples changed progressively from 1989 to 2003, then an unexpected,dramatic decline in the dominant species (Lupinus lepidus) resulted in a sharp deviation in thetrajectory. Standard deviations remain large, indicating that convergence has not occurred.

Tab. 1 - Structural features within and outsideof lupine patchesestablishedfor over 20 years(Old), 12 to 15 years (Mature) and 5 to 10years (Young). Data are shown for all speciesandwithout Lupinus lepidus. Frequency is thesum of individual percent frequency in thesample; dominance is the complement ofSimpson’sindexanddiversityis theShannon-Wiener function.

220 BENSETTITI A., BIORET F., BOULLET V., PEDROTTI F., Centenaire de la Phytosociologie

distance alone contributes to landscapevariation and to differences in succes-sional trajectories.

UNUSUAL EVENTS

Superimposed on the effects ofarrival order (priority) and landscapeeffects (gradually increasing isolationwith disturbance size) are unpredicta-ble and unique events. Studies by manyindividuals have demonstrated the widearray of these contingent events. Theyrange from the date of the eruption at atime when much of this landscape wascovered in snow and the unexplained

early arrival of Lupinus lepidus at a fewsevere, isolated sites within one year ofthe event. Lupinus has a large seed andis dispersed locally by explosive dehi-scence and by ants, but the first seeds toestablish on pyroclastic material cros-sed at least 4 km of inhospitable terrain.Many cases of unusual, even novel, andunforeseen joint occurrences of speciesrarely found together have been verycommon. The occurrence of Lupinuslatifolius (a species of forest marginsand lush meadows) with L. lepidus, aspecies of dry, subalpine habitats ismost notable. These species continue tocoexist at lower, drier elevations thaneither normally occupies and in expo-

sed habitats at higher elevations.Most recently, an unusual climatic

event occurred. Severe cold occurredabove 900 m on Mount St. Helens befo-re snow blanketed the terrain in 2004.Throughout its range, Lupinus lepidus(an herbaceous perennial) suffered ne-arly total dieback causing it to be sparseuntil mid-summer in 2005. The conse-quences of this event are still beinginvestigated, but it appears that speciessuppressed by lupines expanded sub-stantially.

LESSONS FROM LONG TERMSTUDIES

THE VEGETATION MOSAIC IS NORMAL

Most mature vegetation exists as avariable mosaic. The mosaic is esta-blished early in primary succession bylandscape factors that provide a hetero-geneous propagule rain. Favorable mi-crosites (safe-sites) are initially rare, sothat colonization is not uniform, nor isit random. Spatial heterogeneity thataffects levels of productivity and stressalso contribute to the mosaic. Speciescomposition varies and the rate of deve-lopment is not synchronized on the ter-rain. A large suite of contingent factorsenhances this mosaic. Subsequently,habitats become further differentiatedby the actions of key species, of intro-ducedherbivoresand thenomadicbeha-vior of herbivores such as elk. Theselarge ungulates affect vegetation atMount St. Helens in many ways, inclu-ding, but not limited to, differentialbrowsing, seed predation, internal andexternal seed predation, trampling, wal-lowing and creating safe-sites in theirhoof prints.

Managers of protected landscapesshould recognize that the mosaic is notonly natural, but also desirable if biodi-versity is to be preserved and augmen-ted. A corollary is that there is rarely asingle desirable stable end-point forvegetation. Multiple stable vegetationstates have now been demonstrated inseveral habitats, and may well be therule, not the exception.

DEL MORAL & LACHER (2005) re-ported on the vegetation pattern onMount St. Helens after 23 years. It re-mains variable both between habitatsand within habitats. There was substan-tial variation even within each of sevencommunity types. The vegetation pat-tern was only loosely tied to environ-mental factors and these were all spa-tial. This result supports the hypothesisthat at least during early succession, site

Fig. 4 - Lupinus lepidus affects trajectories both by the length of time it is present and byits density. Colonies and adjacent sparse sites differ strongly. Colonies are relativelysimilar, while adjacent sites differ from one other to a much greater degree. Sparse sites aremuch more variable.

Fig. 5 - The effects of refugia on surrounding vegetation decline strongly with distance. Within8 m of a dense colony of vegetation that survived the eruption, both cover and species richnessreacheda level similar to thatatmuch largerdistances.Foreach line, letters indicatemembershipin groups not distinguishable after analysis of variance, followed by a Bonferroni test.

BRAUN-BLANQUETIA, vol. 46, 2010 221

location is the most important factorthat determines species composition.Though it is too early to draw conclu-sions, there was no evidence that vege-tation in similar habitats would conver-ge to a relatively homogeneous state.

STRESS REDUCTION IS REQUIRED FOR INI-TIAL ESTABLISHMENT

Early in primary succession,nothing can occur until stress and infer-tility are reduced to levels where see-dlings can establish. This process isinitially dominated by physical proces-ses (deposition of nutrients from outsi-de the system, minor erosion) and usual-ly includes a rain of such unlikely colo-nists as spiders. These predatory orga-nisms soon perish, but add valuablecarbon and nitrogen to the system. Re-storation ecologists know that fertilitymust be managed for a successful ou-tcome. However, as demonstrated whe-re Lupinus achieves strong dominance,too much fertility can impede develop-ment and reduce diversity.

The conditions for establishmentchange as vegetation develops and theenvironment becomes less stressful.Where at first only a very few, particu-larly favorable, sites could support asuccessful seedling, a greater portion ofthe substrate becomes open to coloniza-tion as amelioration proceeds. As soonas plants establish and mature, the po-tential for biotic stresses to develop alsoincreases. Not only do competitive inte-ractions (seebelow) become prominent,but also herbivores begin to thrive(BISHOP, 2002), with complex implica-tions for succession (FAGAN & BISHOP,2000; FAGAN et al., 2005). Long-termmanagement of developing vegetationin protected areas must account for thisfluctuating biotic environment.

COMPETITION AND FACILITATION OCCUR

THROUGHOUT SUCCESSION

One standard concept of primarysuccession is that biological facilitationis crucial during the early stages andthat competition is limited or balancedbyother factors (KLANDERUB &TOTLAND,2004). However, there is an increasingliterature that indicates that competi-tion is common in early succession andinstressfulenvironments(FRANKS,2003;TOTLAND et al., 2004). The relative mixof positive and negative biotic forceschanges space and time.

In the case of Lupinus lepidus, theaccumulation of diversity and biomass

of other species may be retarded indense patches, but promoted in adja-cent sparse patches. Superimposed onthis pattern is that of herbivory, whichcan be intense on the dense, vibrantmargins of lupines, but restricted bothin sparse populations and in dense, se-nescent colonies.

Herbivores can arrest successionby restricting facilitation effects. Salixspp. could develop moderately densecolonies on pumice at Mount St. He-lens, but a stem borer kills most shrubsbefore they can produce significant fa-cilitation for less stress-tolerant plantspecies. When variations in competi-tion, facilitation and herbivory occur inboth space and time, a vegetation mo-saic is likely to develop and persist. Amajor challenge for maintaining vege-tation in protected lands is to managethe biotic environment to preserve bio-diversity and the vegetation mosaic.

MOSAICS PERSIST

Once formed, vegetation mosaicstend to persist through time. The effectsof priority can be profound. On lahars atMount St. Helens, mosaics have deve-loped that appear to be stable. An openwoodland consisting of several speciesof conifers has invaded where exposedrocks facilitatedseedling establishment.The diversity of conifers, many of themalready reproductive after less than 20years, exceeds that in adjacent matureforests. Vegetation beneath the cano-pies is sparse and distinct from opensurroundings. Where conifers have notbecome established, a carpet of Raco-

mitrium, punctuated by Lupinus and theprostrate shrub Penstemon dominatelarge areas. In addition, there are pa-tches of Populus with a distinct under-story, thickets of the nitrogen-fixingshrub Alnus viridus and large patches ofArctostaphylos nevadensis (related toA. uva-ursi) that cover many squaremeters. While some of these elementsmay not persist, they will alter the localcourse of succession. While the mosaicmay or may not be permanent, it will bean important feature of the vegetationfor centuries. It is neither reasonablenor desirable to suppress this heteroge-neity in an attempt to hasten the deve-lopment of “climax” vegetation.

Based on an exploration of monta-ne meadows on 400 year-old lahars onMount St. Helens, mosaics may indeedpersist. Consider that conifers may in-vade and establish rapidly close to sour-ces of seeds, but in many places, no seedsource occurs. In these places, openmeadows develop. Meadows can resistthe invasion of conifers on Mount St.Helens because summers are dry andsoils infertile. Conifer seeds arrive insmall numbers and a few germinate.Seedlings have been noted to persist fora few years, but slow growth due tonitrogen deficient (GILL et al., 2006)soilspre-disposeseedlings todeath fromthe inevitable extreme drought. Themeadows also become differentiatedinto patches dominated by grasses andthose dominated by low woody shrubs.

MONITORING IS REQUIRED

A major lesson for proper manage-

Fig. 6 - The composition of plots changed significantly with distance as shown by DCA ofthe mean composition. Sites 64 to 128 m distant share less than 50% of the species with thoseadjacent to the refugia.

222 BENSETTITI A., BIORET F., BOULLET V., PEDROTTI F., Centenaire de la Phytosociologie

ment of protected lands is that condi-tions must be continually monitored.Short-term fluctuations in climate, unu-sual disturbances (e.g., volcanic erup-tions, insect plagues) can adversely af-fect vegetation. In some cases, remedialaction may limit the damage. Monito-ring can also provide crucial knowled-ge for restoration of damaged habitatsand to curtail mechanisms that maygenerate undesirable deflection of suc-cession trajectories.

Long-term monitoring also permi-ts testing hypotheses. Short-term stu-dies conducted under a particular set ofconditions can lead to one conclusion.This conclusion may not be valid overthe range of conditions to be expectedover many years on a site. For example,it appeared that vegetation on a lahar atMount St. Helens had reached a stage ofvery slow development and minimumheterogeneity after 20 years. However,subsequently (years 2002 and 2004 inFig. 7), a major shift in the trajectorydirection and sample heterogeneity oc-curred (cf. 1993 to 1999 with 2002 and2004). In thiscase, anunexpectedexplo-sion in Lupinus lepidus in many, but notall, plots of the sample led to dramaticincreases in biomass in several species.The result was dramatic. Had the studyterminated in 1999, the conclusionwould have been that a stable meadowcommunity had been established andthat further significant change wouldrequire the invasion of conifers.

A more important reason for mo-nitoring is to be alerted to events trigge-red by global warming. Global climate

change associated with anthropogenicfactors may lead to average temperatu-re increases of 2-3 °C by 2100 (WE-STOBY & BURGMAN, 2006). Such drama-tic changes will require many species tomigrate away from the pole or upwards(if possible). It is unlikely that specieslinked through facilitation, pollinationor mycorrhizal association will be ableto migrate at the same rates, so disrup-tions are inevitable. Monitoring maypermit early warnings of impendingdisassociations so that intervention mayoccur. For example, species complexescould be translocated into more favora-ble habitats, after due considerationsfor rainfall patterns as well. The targetspecies would require their associates(e.g. mycorrhizae, soil bacteria, polli-nators) and would require assistanceagainstcompetitors.While this isadaun-ting challenge, it would be impossiblewithout long-termmonitoring.Themosteffective method to preserve species inprotected habitats is, however, not bio-logical. Every social and political effortmust be undertaken to reduce CO2 dra-matically… and soon.

CONCLUSION

The study of recovery from theeffects of Mount St. Helens has manylessons for the restoration of damagedprotected lands. Some lessons are cau-tionary, while others provide lessonsfor the kinds of proactive. Successiontrajectories are unlikely to follow pathssimilar to those that previously occur-

red. Climatic factors are changing, lan-dscape conditions differ, the biota maybe different due to recent introductionsand chance events may create unusualinitial conditions. Thus, managers ofprotected lands should not attempt toforce trajectories into predetermined,arbitrary paths.

Intrinsic habitat heterogeneityshouldbewelcomedandpreserved.Thisallows for variations in the vegetationthat support biodiversity and offer abuffer to climate variation and change.Competition favors established plants,so the importance of arrival order can-not be overemphasized. The biotic va-riation imposed by strong competitorspersists until life history events or di-sturbance permits transitions to otherspecies (CONNELL & SLATYER, 1977).Initially sites are quite variable, withpatterns dictated in large part by stocha-stic events (DEL MORAL, 1999). Even-tually, deterministic processes lead tostronger relationships among speciesand between species and their environ-ment. For example, in the early stages,many species can establish on a barrensite, but after many years, only thosethat are more efficient under the deve-loping regime can survive competitionand herbivory. These processes resultin associations that are more predicta-ble, yetheterogeneity, onceestablished,appears to persist.

Rehabilitation, restorationandpro-tection of vegetation all require subtleand profound understanding of howcommunities assemble, where the bott-lenecks occur and how trajectories canbe modified to meet specific goals. Mo-nitoring natural recovery from devasta-ting disturbances has provided some ofthis understanding.

ACKNOWLEDGEMENTS

I thank Prof.ssa Emilia Poli Mar-chese for the kind invitation to partici-pate in the Symposium on Ecology,Conservation and Management of Eco-systems in Protected Areas at the XVIIth

International Congress, Vienna. I alsothank the U.S. National Science Foun-dation for funding since 1981 and theUniversity of Washington GraduateSchool for travel funds.

BIBLIOGRAPHY

BELLINGHAM P.J., WALKER L.R, WARDLE

D.A., 2001 - Differential facilita-tion by a nitrogen-fixing shrubduringprimarysuccession influen-

Fig. 7 - Long term monitoring can provide surprises. Repeat sampling of 64 plots on a lahardemonstrated moderate, consistent change from 1987 to 1999. Variation (standard error)decreased. However, the trajectory abruptly shifted direction and rate in 2002, and by 2004,variation was large. The years 2001 to 2004 saw more change than in all the previous changes.

BRAUN-BLANQUETIA, vol. 46, 2010 223

ces relative performance of ca-nopy tree species. J. Ecol., 89: 861-875.

BISHOP J.G., 2002 - Early primary suc-cession on Mount St. Helens: theimpact of insect herbivores on co-lonizing lupines. Ecology, 83:191-202.

CALLAWAY R.M., WALKER L.R., 1997 -Competitionand facilitation:asyn-thetic approach to interactions inplant communities . Ecology, 78:1958-1965.

CLARKSON B.R., CLARKSON B.D., 1995 -Recent vegetation changes onMount Tarawera, Rotorua, NewZealand. New Zeal. J. Bot., 33:339-354.

CONNELL J.H., SLATYER R.O., 1977 -Mechanisms of succession in natu-ral communities and their roles incommunity stability and organiza-tion. Amer. Natur., 111: 1119-1144.

DEL MORAL R., 1999 - Plant successionon pumice at Mount St. Helens,Washington. Amer. Midl. Natur.,141: 101-114.

DEL MORAL R., 2006 - Vegetation dyna-mics in space and time: an exam-ple from Mount St. Helens. J. Veg.Sci., 17: in press.

DEL MORAL R., ECKERT A.J., 2005 -Colonization of volcanic desertsfrom productive patches. Amer. J.Bot., 92: 27-36.

DEL MORAL R., ELLIS, E.E., 2004 - Gra-dients in heterogeneity and struc-ture on lahars, Mount St. Helens,Washington,USA. PlantEcol.,175:273-286.

DEL MORAL R., JONES C.C., 2002 - Earlyspatial development of vegetationon pumice at Mount St. Helens.Plant Ecol., 162: 9-22.

DEL MORAL R., LACHER I.L., 2005 - Ve-getation patterns 25 years afterthe eruption of Mount St. Helens,Washington. Amer. J. Bot., 92:1948-1956.

DEL MORAL R., ROZZELL L.L., 2005 -Effects of lupines on communitystructure and species association.Plant Ecol., 181: 203-215.

DEL MORAL R., WALKER L.R., BAKKER

J.P., 2006 - Insights gained fromsuccession for the restoration oflandscape structure and function.In press, in: Walker L.R., WalkerJ., Hobbs R.J., Linking Restora-tion and Succession in Theory andin Practice. Springer Environmen-tal, New York.

DEL MORAL R., WOOD D.M., TITUS J.H.,2005 - Proximity, microsites, andbiotic interactions during early

succession. Pp 93-110, in DaleV.H.,SwansonF.J, CrisafulliC.M.(eds.), Ecological responses to the1980 eruption of Mount St. He-lens. Springer Science, New York.

DEL MORAL R., WALKER L.R., 2007 -Environmental disasters, naturalrecovery and human response.CambridgeUniversityPress,Cam-bridge UK.

FAGANW.F.,BISHOP J.G.,2000 -Trophicinteractions during primary suc-cession: herbivores slow a plantreinvasion at Mount St. Helens.Amer. Natur., 155: 238-251.

FAGAN W.F., LEWIS M., NEUBERT M.G.,AUMANNC.,APPLEJ.L.,BISHOPJ.G.,2005 - When can herbivores slowor reverse the spread of an inva-ding plant? A test case from MountSt. Helens. Amer.Natur., 166:669-685.

FRANKLIN J.F., MACMAHON J.A., 2000 -Messages from a mountain. Scien-ce, 288: 183-1185.

FRANKS S.J., 2004 - Competitive andfacilitative interactions within andbetween two species of coastaldune perennials. Can. J. Bot., 81:330-337.

FULLER R.F., DEL MORAL R., 2003 - Therole of refugia and dispersal inprimary succession on Mount St.Helens, Washington J. Veg. Sci.14: 637-644.

GILL R.A., BOIE J.A., BISHOP J.G., LAR-SEN L., APPLE J.L., EVANS , R.D.,2006 - Linking community andecosystem development on MountSt. Helens. Oecologia, 148: 312-324.

GRISHIN S.Y., DEL MORAL R., KRESTOV

P., VERKHOLAT V.P., 1996 - Suc-cession following the catastrophiceruption of Ksudach volcano(Kamchatka, 1907). Vegetatio,127: 129-153.

JONES C.C., DEL MORAL R., 2005 - Ef-fects of microsite conditions onseedling establishment on the fo-reland of Coleman Glacier,Washington. J. Veg. Sci., 16: 293-300.

KITAYAMA K., MUELLER-DOMBOISD.,VI-TOUSEK P.M., 1995 - Primary suc-cession of Hawaiian montane rainforest on a chronosequence of ei-ght lava flows. J. Veg. Sci., 6: 211-222.

KLANDERUB K., TOTLAND O., 2004 - Ha-bitat dependent nurse plant effectsof the dwarf-shrub Dryas octope-talaonalpineand arcticplantcom-munity structure. Ecoscience, 11:410-420.

MASUZAWA T., 1985 - Ecological stu-

dies on the timberline of MountFuji. I. Structure of plant commu-nity and soil development on thetimberline. Bot. Mag. Tokyo, 98:15-28.

MCCUNE B., ALLEN T.F.H., 1985 - Willsimilar forests develop on similarsites? Canad. J. Bot., 63: 367-376.

POLI MARCHESE E., GRILLO M., 2000 -Primary succession on lava flowson Mt. Etna. Acta Phytogeograp.Suecica, 85: 61-70.

TOTLAND O., GRYTNES J.A., HEEGAARD

E., 2004 - Willow canopies andplant community structure alongan alpine environmental gradient.Arctic Antarctic Alpine Res., 36:428-435.

TURNER M.G., DALE V.H., 1998 - Com-paring large infrequent disturban-ces: what have we learned. Ecosy-stems, 1: 496-496.

WALKER L.R., 1993 - Nitrogen fixersand species replacement in pri-mary succession. Pp. 249-272, inMiles J., Walton D.W.H. (eds.)Primary succession on land,Blackwell, Oxford.

WALKER L.R, DEL MORAL R., 2003 -Primarysuccession and ecosystemrestoration. Cambridge Universi-ty Press, Cambridge.

WALKER L.R., WALKER J., DEL MORAL

R., 2006 - Forging a new alliancebetween succession and restora-tion. In press, in Walker L.R.,Walker J., Hobbs R.J., Linking re-storation and succession in theoryand in practice. Springer Environ-mental, New York.

WESTOBY M., BURGMAN M., 2006 - Cli-mate change as a threatening pro-cess. Aust. Ecol., 31: 549-550.