ISSN 1860-0387
PhD Disse a ion 01 / 2012
The ole o li e his o y ai s o coexis ence and o es
eco e y a e dis u bance – a modelling pe spec i e
Towa ds a be e unde s anding o species- ich o es s
Claudia Dislich
PhD Disse a ion 01 / 2012 I Claudia Dislich I The ole o li e his o y ai s o coexis ence and o es eco e y a e dis u bance ...
Helmhol z Cen e
o En i onmen al Resea ch – UFZ
Pe mose s aße 15 I 04318 Leipzig I Ge many
In e ne : www.u z.de
The ole o li e his o y ai s o
coexis ence and o es eco e y a e
dis u bance – a modelling pe spec i e
Towa ds a be e unde s anding o species- ich o es s
Disse a ion zu E langung des akademischen G ades
Dok o de Na u wissenscha en (D . e . na .)
an de Fakul ä ü Biologie, Chemie und Geowissenscha en
de Uni e si ä Bay eu h
o geleg on
Diplom Ma hema ike in
Claudia Dislich
aus Hilden
Bay eu h, Mä z 2011
The ole o li e his o y ai s o coexis ence and o es eco e y a e dis u -
bance – a modelling pe spec i e
Towa ds a be e unde s anding o species- ich o es s
PhD Disse a ion, Uni e si y o Bay eu h, 2011.
Au ho ’s add ess: Claudia Dislich, Helmhol z Cen e o En i onmen al Resea ch –
UFZ, Depa men o Ecological Modelling, Leipzig, Ge many.
Homepage: www.u z.de/index.php?de=19206
Supe ised by:
P o . D . And eas Hu h (UFZ Leipzig/Uni e si y o Osnab ück)
P o . D . Bjö n Reineking (Uni e si y o Bay eu h)
De ense commi ee (26.09.2011 in Bay eu h):
P o . D . Be nd Huwe (Uni e si y o Bay eu h)
P o . D . Bjö n Reineking (Uni e si y o Bay eu h, 1s e iewe )
P o . D . And eas Hu h (UFZ Leipzig/Uni e si y o Osnab ück, 2nd e iewe )
P o . D . Michael Hauhs (Uni e si y o Bay eu h)
P o . D . John Tenhunen (Uni e si y o Bay eu h)
Abs ac
T opical o es s a e well known o hei excep ional species ichness – high di e si y o
plan species cons i u e he basis o an equi alen ly ich auna. An as onishing a ie y
o plan li e s a egies has e ol ed, mani es ing i sel also in di e en composi ions o li e
his o y ai s in ees. This hesis in es iga es he ole o ee li e his o y ai s (g ow h,
mo ali y and ec ui men ) on di e en p ocesses s uc u ing species- ich o es s. Ou
s udy sys em is a mon ane ain o es loca ed in he T opical Andes ho spo o biodi-
e si y in sou he n Ecuado . He e, we ind a mosaic o s eep idges and deeply incised
alleys, co e ed wi h p edominan ly b oadlea o es . Fo es s uc u e and species com-
posi ion di e conside ably depending on al i ude and opog aphic posi ion. The o es
co e is equen ly in e up ed by sca s o landslides, which cons i u e an impo an ype
o na u al dis u bance in his ecosys em.
We u ilize ecological models as ools o gain deepe insigh s in o key p ocesses d i ing
he main enance o ee species ichness and a ec ing o es eco e y a e landslides.
The i s pa o his hesis conce ns he ques ion o species coexis ence. We de elop
a heo e ical model o analyze how di e en ade-o s be ween li e his o y ai s ( ee
g ow h, seed dispe sal, ee mo ali y) a ec ee species coexis ence. We ind ha he
conside ed ade-o s alone a e no su icien o explain long- e m species coexis ence.
Addi ional ’s abilizing’ mechanisms seem o be indispensable o acili a e coexis ence in
species- ich o es s. Such mechanisms could esul om bio ic in e ac ions, ha al e
he ela ion be ween in e - and in a-speci ic compe i ion depending on (local) species
abundances (e.g. densi y-dependen mo ali y). O he possible coexis ence mechanisms
likely o be ele an o ou pa icula s udy sys em a e d i en by ex e nal, abio ic ac o s
like a complex opog aphy esul ing in locally di e ing habi a ypes (each suppo ing a
di e en se o species), o he cha ac e o a p e ailing dis u bance egime (e.g. shallow
landslides).
In he second pa o he hesis, we in es iga e he g ow h dynamics o he idge o -
es in ou s udy sys em. To his end, we u ilize he p ocess-based o es g ow h model
FORMIND. We show ha a e calib a ion, he model success ully ep oduces o es
dynamics on di e en le els o complexi y (e.g. basal a ea and s em size dis ibu ion).
We hen use his o es model o in es iga e he in luence o landslide dis u bances on
o es dynamics bo h on he local scale o a single landslide and on he landscape scale.
On landslide si es, changes in en i onmen al condi ions migh lead o changes in di e en
ee li e his o y ai s. We analyze scena ios wi h changes in di e en ai s ( ee ec ui -
men , ee g ow h, ee mo ali y) and ind ha while ee biomass can eco e wi hin he
i s hund ed yea s a e a landslide, he ime un il o es s uc u e and species compo-
si ion is es o ed is conside ably longe (app oxima ely 200 yea s). Changes in di e en
ai s esul in di e ing spa ial dis ibu ions o ee biomass: educed ee g ow h leads
o a mo e homogeneous dis ibu ion o biomass, whe eas educed ec ui men and in-
c eased mo ali y yield a mo e he e ogeneous biomass dis ibu ion (’pa chy’ ege a ion).
On he landscape le el, o e all o es biomass is subs an ially educed by landslides (8-
I
14%), compa ed o only 2 - 3% o he a ea ma ked by isible aces o landslides. Thus
his pa icula ype o dis u bance conside ably in luences he o al o es ca bon balance.
In a complemen a y in es iga ion we s udy abio ic and bio ic ac o s ha po en ially
igge landslide occu ence in ou s udy sys em. Fo his, we de elop an ex ension o
a s anda d physically-based model o slope s abili y. We ind ha due o he p edom-
inan ly shallow ee oo s, some o he obse ed landslides migh be igge ed by he
ege a ion i sel .
This hesis demons a es ha ecological models a e use ul ools o gain deepe insigh s
in o impo an p ocesses shaping o es communi ies. They can be applied o heo e -
ical ques ions such as he ques ion o species coexis ence, as well as o mo e applied,
managemen ela ed ques ions like p edic ing o es eco e y a e dis u bances.
Zusammen assung
T opische Regenwälde sind bekann ü ih en A en eich um – eine hohe Di e si ä on
P lanzen scha eine g oße Viel al an Lebens äumen ü Tie e. Man inde in den T open
eine e s aunliche Mannig al igkei e schiedene Lebenss a egien o , die sich un e an-
de em in un e schiedlichen Eigenscha en de Lebenszyklen on Bäumen ausd ücken. Die
o liegende A bei un e such , welche Rolle solche ’Lebenszyklus-Cha ak e is ika’ (li e
his o y ai s) ü die S uk u ie ung a en eiche Wälde spielen. Unse Fo schungs-
gebie is ein Be g egenwald in Südequado , de zu den opischen Anden, einem de
wel wei en Ho spo s de A en iel al gehö . Dieses Gebie is gekennzeichne du ch
s eile Hänge und ie eingeschni ene Täle , die on a en eichem Laubwald bedeck
sind. Bauma enzusammense zung und Walds uk u a iie en abhängig on de Höhe
übe dem Mee esspiegel und opog aphische Posi ion. Die Waldbedeckung wi d häu ig
du ch E d u sche un e b ochen, die eine wich ige na ü liche S ö ung in diesem Ökosys-
em da s ellen.
Fü die Un e suchung on Schlüsselp ozessen, die die E hal ung de Bauma en iel al
und die Regene a ion on Wälde n nach E d u schen beein lussen, e wenden wi öko-
logische Simula ionsmodelle. De e s e Teil diese A bei konzen ie sich au die Ko-
exis enz on Bauma en. Es wi d ein heo e isches Modell en wickel , um die Auswi kung
e schiedene Zusammense zungen on A eigenscha en (Baumwachs um, Samenaus-
b ei ung und Mo ali ä ) au Koexis enz zu analysie en. Ein Haup esul a diese S udie
is , dass Un e schiede in den be ach e en A eigenscha en ( ade-o s) alleine nich aus-
eichen ü eine lang is ige Koexis enz de Bauma en. Zusä zliche ’s abilisie ende’
Mechanismen scheinen no wendig ü die Koexis enz on Bauma en au langen Zei -
skalen zu sein. Solche Mechanismen könn en zum Beispiel du ch bio ische In e ak ionen
en s ehen, die das Ve häl nis on inne - und zwischena liche Konku enz e ände n
(zum Beispiel dich eabhängige Mo ali ä ). Wei e e mögliche Koexis enzmechanismen
sind abio ische Na u , wie zum Beispiel eine komplexe Topog aphie, die in eine g oßen
II

Viel al e schiedene Lebens äume ü un e schiedliche A en esul ie , ode spezielle
S ö ungs egime, wie zum Beispiel E d u sche in unse em Fo schungsgebie .
De zwei e Teil de o liegenden A bei behandel die Waldwachs umsdynamik des
G a waldes in unse em Fo schungsgebie . Hie ü benu zen wi das p ozess-basie e
Waldwachs umsmodell FORMIND, das nach Kalib ie ung die Walddynamik au e -
schiedenen Komplexi ä ss u en (beispielsweise S ammg und läche und S ammzahl-Du ch-
messe e eilung) ep oduzie . He nach e wenden wi dieses Waldmodell, um den Ein-
luss on E d u schen au den Wald – sowohl au de lokalen Ebene einzelne Ru sch lächen
als auch au de Landscha sebene – zu un e suchen. Wi be ach en Szena ien mi e -
schiedenen Ände ungen in A eigenscha en on Bäumen (Regene a ion, Wachs um und
Mo ali ä be e end) als Reak ion au e ände e Umwel bedingungen nach E d u sch-
S ö ungen. Wäh end die Gesam -Baumbiomasse inne halb de e s en 100 Jah e nach
einem Ru sche eignis egene ie en kann, is die Zusammense zung e schiedene A -
g uppen deu lich länge e ände (ca. 200 Jah e). Ände ungen in den e schiedenen
A eigenscha en üh en zu eine un e schiedlichen äumlichen Ve eilung de Baumbio-
masse au de Regene a ions läche: e inge es Baumwachs um üh zu eine homo-
genen Ve eilung de Biomasse, ge inge e Regene a ion und e höh e Mo ali ä üh en zu
eine he e ogenen Ve eilung de Biomasse. Au de Landscha sebene wi d die Gesam -
biomasse deu lich s ä ke du ch E d u sche eduzie (um 8 -14%), als au Lu bilde n
e kennba is : do sind nu ca. 2- 3% de Fläche on sich ba en Ru schungsspu en
gekennzeichne . Dahe is die Be ücksich igung diese speziellen A on S ö ung une -
lässlich ü die Un e suchung de Kohlens o bilanz in unse em Fo schungsgebie .
In eine wei e en S udie un e suchen wi po en iell wich ige abio ische und bio ische
Fak o en ü das Auslösen on E d u schen in unse em Fo schungsgebie . Hie ü e wei -
e n wi ein e ablie es physikalisch-basie es Modell ü Hangs abili ä . Ein wich iges
E gebnis diese Un e suchung is , dass au g und seh lache Baumwu zeln die Vege a-
ion zum Auslösen on Ru schen be agen könn e.
Diese A bei zeig au , dass ökologische Modelle nü zliche We kzeuge sind, um wich ige
P ozesse in Wälde n zu un e suchen und ih Zusammenwi ken besse zu e s ehen –
sie können sowohl ü ehe heo e ische F ages ellungen, wie die Koexis enz on A en,
als auch ü meh p axisbezogene F agen, wie die Waldendwicklung nach S ö ungen,
angewende we den.
Resumen
Los bosques opicales son conocidos po su excepcional iqueza de especies – una al a
di e sidad de especies de plan as cons i uye la base pa a una auna con una di e sidad
equi alen e. Una imp esionan e a iedad de es a egias de ida ege al ha e olucionado,
a iedad que se mani ies a ambién en las di e en es combinaciones de a ibu os un-
cionales en los á boles. Es a esis in es iga el papel de a ibu os uncionales en los
III
á boles (c ecimien o, mo alidad y eclu amien o) en di e en es p ocesos que es uc u an
los bosques icos en especies. Nues o sis ema de es udio es un bosque de llu ia de mon-
aña si o en el ho spo de biodi e sidad de los Andes opicales en el su de Ecuado .
Aquí, encon amos un mosaico de colinas empinadas y alles p o undamen e exca a-
dos, cubie os p incipalmen e po bosque de ondosas. La es uc u a del bosque y la
composición de especies a ían conside ablemen e en unción de la al i ud y la posición
opog á ica. La cubie a o es al es in e umpida con ecuencia po su cos y co imien os
de ie a, que cons i uyen un impo an e ipo de al e ación en es e ecosis ema.
En es e abajo u ilizamos modelos ecológicos como he amien as pa a consegui un
conocimien o más p o undo de los p ocesos cla e que son esponsables del man enimien o
de la iqueza de especies y que a ec an a la ecupe ación del bosque después de los co -
imien os de ie a. La p ime a pa e de la esis se ocupa de la coexis encia de especies.
Desa ollamos un modelo eó ico pa a analiza cómo di e en es comp omisos ( ade-o s)
en e de a ibu os uncionales de los á boles (c ecimien o del á bol, dispe sión de semilla,
mo alidad del á bol) a ec an a la coexis encia de especies a bó eas. Como esul ados
encon amos que, solos, los comp omisos conside ados no son su icien e pa a explica la
coexis encia de especies a la go plazo. Mecanismos de es abilización adicionales pa ecen
se indispensables pa a acili a la coexis encia en bosques con al a iqueza de especies.
Es os mecanismos pod ían esul a de las in e acciones bió icas, que al e an la elación
en e la compe encia in e - e in a-especí ica, dependiendo de las abundancias (locales) de
especies (p. ej. mo alidad dependien e de la densidad). O os mecanismos de coexis en-
cia p obablemen e ele an es pa a nues o sis ema de es udio pa icula , son conducidos
po ac o es ex e nos y abió icos, como la opog a ía compleja que de i a en ipos de
hábi a localmen e di e en es (cada cual man eniendo un conjun o di e en e de especies),
o el ca ác e del égimen de una al e ación p eponde an e (p. ej. co imien os de ie a
de poca p o undidad).
En la segunda pa e de la esis, se in es iga la dinámica de c ecimien o del bosque en
las zonas de c es a del á ea de es udio. Pa a ello, empleamos el modelo de c ecimien o
o es al basado en p ocesos FORMIND. Mos amos que, as la calib ación, el modelo
es capaz de ep oduci la dinámica del bosque bajo di e en es ni eles de complejidad (p.
ej. á ea basal y dis ibución de amaño del allo). Pos e io men e u ilizamos el modelo
pa a in es iga la in luencia de las al e aciones po co imien os de ie a en la dinámica
del bosque, an o a la escala local del co imien o de ie as en pa icula , como a escala
de paisaje. En los e enos de los co imien os de ie as, cambios de las condiciones
ambien ales conduci ían a cambios de a ibu os uncionales de los á boles. Analizamos
escena ios con cambios en di e en es ca ac e ís icas ( eclu amien o, c ecimien o y mo -
alidad de á boles) y ob enemos que, mien as la biomasa a bó ea puede ecupe a se
den o de los p ime os cien años después del co imien o de ie a, el iempo has a que
la es uc u a del bosque y la composición de especies son es ablecida es conside able-
men e más la go (ap oximadamen e de 200 años). Cambios en las di e en es de a ibu os
uncionales en los á boles conducen a di e en es dis ibuciones espaciales de la biomasa
de á boles: un c ecimien o a bó eo educido lle a a una dis ibución más homogénea de
IV
la biomasa, mien as que un eclu amien o educido y una mo alidad c ecida conducen
a una dis ibución más he e ogénea de la biomasa ( ege ación "en pa ches"). A ni el de
paisaje, la biomasa o al del bosque es educida sus ancialmen e po los co imien os de
ie a (8-14%), en compa ación con an solo un 2-3% del á ea con azas isibles de co -
imien os de ie a. Así, es e pa icula ipo de al e ación in luencia conside ablemen e
el balance o al de ca bono del bosque.
En una in es igación complemen a ia, es udiamos ac o es abió icos y bió icos que
pueden p o oca co imien os de ie a en nues a á ea de es udio. Pa a ello, desa ol-
lamos una ex ensión de un modelo es ánda basado en p ocesos ísicos de es abilidad de
la pendien e. Encon amos que, debido al p edominio de aíces a bó eas de poca p o un-
didad, algunos de los co imien os de ie a obse ados pod ían se desencadenados po
la p opia ege ación.
Es a esis demues a que los modelos ecológicos son he amien as ú iles pa a adqui i
un conocimien o más p o undo sob e p ocesos impo an es que dan o ma a las comu-
nidades o es ales. Los modelos pueden se aplicados a cues iones eó icas ales como
la coexis encia de especies, así como a cues iones más aplicadas y elacionadas con la
ges ión, como la p edicción de la ecupe ación del bosque después de al e aciones.
V
1 In oduc ion
by a single landslide anges om se e al squa e me e s o many squa e kilome e s (Walke and
del Mo al, 2003). En i onmen al condi ions on landslide su aces a e conside ably al e ed com-
pa ed o condi ions in undis u bed o es : ligh is inc eased (Mys e and Fe nandez, 1995), soils
migh be uns able (Walke and Shiels, 2008) and due o he loss o he o ganic soil laye he
soil nu ien con en is educed e en many yea s a e he slide e en (Za in and Johnson, 1995;
Wilcke e al., 2003). One ocus o landslide esea ch is he local scale o single landslides and he e
mos ly he i s phase o succession whe e ea ly landslide colonize s like mosses, lichens, e ns
and bamboo domina e he ege a ion (e.g. Dalling and Tanne , 1995; Ohl and Bussmann, 2004;
Velazquez and Gomez-Sal, 2008). Bu he changed en i onmen al condi ions on landslide si es
may also a ec li e his o y ai s o ees and he e o e in luence he longe e m cou se o o es
eco e y. Fo example, he es ablishmen o ees migh be hinde ed by he ha sh en i onmen al
condi ions and ee g ow h migh be educed due o nu ien limi a ion. Changes in di e en li e
his o y ai s a ec o es s uc u e and p oduc i i y on landslide si es and will hus also a ec
o e all o es p oduc i i y in o es s whe e landslides occu equen ly (Res epo e al., 2003).
Apa om he local scale o a single landslide su ace, landslides a e also an in e es ing
phenomenon o examine on he landscape scale and o e long ime pe iods – hey c ea e a
pa chy dis ibu ion o di e en aged si es wi h di e ing edaphic condi ions and successional
s ages o ege a ion. The e o e, landslides a e assumed o be a d i e o biodi e si y by inc easing
landscape he e ogenei y (e.g. Yamamo o e al., 1995; Gee sema and Poja , 2007; Elias and Dias,
2009). Howe e , in es iga ing a high numbe o di e en aged landslide si es is di icul due o
inaccessibili y o he e ain, and in es iga ing ime pe iods long enough o obse e he ’shaping’
cha ac e o landslides in a o es ed landscape exceeds a human li e span by a . Fo his pu pose,
ecological models can be used.
1.4 Ecological models
The p ocesses and in e ac ions ha shape na u al communi ies a e o en complex and he e o e
di icul o deciphe ia obse a ion. Some o he obs acles a e he ichness o in e ac ions in
ecosys ems which addi ionally migh occu on di e en scales, o he slow speed o p ocesses o
in e es (e.g. o es g ow h). To gain insigh s in hese p ocesses we need o abs ac om eali y
and educe complexi y. Ecological models should se e his pu pose as ools o hinking and
lea ning abou in e ela ionships in ecosys ems and analyzing sys em beha iou unde de ined
condi ions. Ecological models can be use ul o
•ex ending ou unde s anding o (gene al) unc ional ela ionships in ecosys ems
• es ing hypo heses abou he unc ioning o (speci ic) ecosys ems
•p edic ing he sys em beha iou unde di e en (e.g. clima ic o managemen ) condi ions
•in es iga ing scena ios which a e impossible o es in he eal ecosys em
•inspi ing ou hinking abou he unc ioning o ecosys ems (de eloping new hypo hesis)
•ins iga ing ield expe imen s
The complexi y o a model should always be d i en by he ecological ques ions and hypo heses
we ha e in mind (Wissel, 1989). Depending on he se o ques ions ha should be ackled by an
ecological model, he model will look di e en ly. One possible ca ego iza ion o ecological models
is o di e en ia e be ween mo e heo e ical models, ha seek gene al insigh in o ecological
p ocesses and ela ionships and applied models, which aim o desc ibe dynamics o a speci ic
s udy sys em and unde s and in e ela ionships, o en also wi h he pu pose o managing he
5

1 In oduc ion
sys em (Bolke , 2008). Ye , hose a e only he endpoin s o a con inuous scale wi h heo e ical
models on he one end and applied models on he o he (Yodzis, 1989). A heo e ical model will
usually be ela i ely simple and a he concep ual wi h a ela i ely small numbe o pa ame e s.
The coexis ence model which was de eloped o in es iga e he ole o li e his o y ai s o species
coexis ence (pape 1) alls in o he ca ego y o mo e heo e ical models. Applied models which
should ep oduce dynamics o a pa icula sys em o en equi e a highe le el o de ail and can
hus be mo e complex han heo e ical models. Highe complexi y usually in ol es a highe
numbe o pa ame e s and he pa ame iza ion o such models can become a ime-consuming
ask. The o es model ha was u ilized o simula e o es g ow h dynamics in ou s udy a ea
and o in es iga e o es eco e y a e landslide dis u bances (c . pape 2 and 3) alls in he
ca ego y o mo e applied models.
1.5 S uc u e o he hesis
This hesis comp ises ou chap e s. Following he in oduc ion, he second chap e in oduces
concep s and me hodologies ha a e impo an as a b oade backg ound o he esea ch p esen ed
in his hesis. Chap e h ee con ains ou esea ch a icles ha can be ead independen ly om
each o he . The i s a icle in es iga es he ole o ade-o s be ween di e en li e his o y
ai s (mo ali y, g ow h, seed dispe sal) o species coexis ence. We show ha he in es iga ed
ade-o s alone do no suppo long- e m coexis ence – addi ional p ocesses ha egula e he
local compe i ion (e.g. densi y-dependen mo ali y) a e necessa y o enable coexis ence. In he
second a icle, we s udy he g ow h dynamics o he idge o es wi hin ou s udy egion. Fo
ha , we u ilize a p ocess based o es g ow h model and de elop a pa ame iza ion o his model.
The a icle co e s a compa ison o model ou pu and ield measu emen s as well as a de ailed
desc ip ion o he o es model and he pa ame iza ion ha is also employed in he hi d a icle.
We demons a e ha ou o es model is capable o ep oducing he s uc u e and dynamics o
ma u e idge o es on di e en le els o complexi y. The hi d a icle analyzes he impac o
shallow landslides in opical mon ane o es s. Di e en scena ios o o es eg ow h wi h changed
li e his o y ai s o ee species (mo ali y, g ow h, es ablishmen ) a e compa ed ega ding hei
e ec on o es s uc u e and p oduc i i y. We ind ha on he local scale o he landslide,
spa ial s uc u e and p oduc i i y o he successional o es di e s depending on he changes in
li e his o y ai s. On he landscape scale, landslides educe o es biomass and conside ably
modi y he spa ial dis ibu ion o biomass. The ou h a icle in es iga es a complemen a y
aspec o he in e ac ions be ween o es and landslides – i add esses he ques ion whe he
ege a ion ela ed bio ic ac o s migh be one o he igge ing ac o s o landslides. We p esen
an ex ension o a classical slope s abili y model by in oducing an independen o ganic ( oo )
laye a op he mine al soil, which is only loosely connec ed o he soil. Wi h his modi ied slope
s abili y model, we can explain he obse ed shallow landslides in ou esea ch a ea, ha almos
exclusi ely in ol e o ganic ma e ial. The ou h chap e gi es a syn hesizing discussion o he
esul s o his hesis.
6
Chap e 2
Concep s and Me hodologies
The aim o his chap e is o in oduce he concep s and me hods applied in his hesis and o
embed hem in a b oade scien i ic con ex .
2.1 Species coexis ence: concep s and models
The ques ion o biodi e si y and i s main enance can be deal wi h a di e en spa ial and
empo al scales: On he global scale, biogeog aphy and mac oecology y o explain plan di e -
si y by clima ic ac o s, habi a he e ogenei y and his o ical/e olu iona y p ocesses (K e and
Je z, 2007). On he smalle scale o local communi ies, whe e clima ic condi ions a e mo e o less
cons an in space, species coexis ence is hough o be media ed by (in a- and in e speci ic) com-
pe i ion, p eda ion, diseases, dis u bances, spa ial and empo al he e ogenei y (Rickle s, 1987;
Tilman and Pacala, 1993; Chesson, 2000). The coexis ence s udy p esen ed in pape 1 ocuses
on species coexis ence on he communi y le el.
Compe i ion be ween species has caugh a g ea deal o a en ion in ecology o a long ime (e.g.
Da win and Wallace, 1858). One eason o his is ha in communi ies, whe e compe i ion o
esou ces is p esen , i is likely o ha e a majo e ec on species di e si y (Hus on, 1994) – in his
sense he ques ion o species coexis ence is closely ela ed o he ques ion o compe i ion be ween
species. Compe i ion can be iewed a he indi idual scale, whe e o example one ee shades a
neighbou ing ee and he e o e supp esses i s neighbou ’s g ow h, and a he popula ion scale,
whe e he composi ion o all species ai s (e.g. species speci ic g ow h and mo ali y a es) and
hei in e ac ions a ec popula ion dynamics and de e mine whe he a species will domina e,
coexis wi h o he compe i o s o go ex inc in he long un. Mechanisms o species coexis ence
can be di ided in o s abilizing and equalizing mechanisms: s abilizing mechanisms educe
in e speci ic compe i ion in ela ion o in aspeci ic compe i ion whe eas equalizing mechanisms
educe i ness di e ences be ween species (Chesson, 2000; Adle e al., 2007). Examples o s abi-
lizing mechanisms a e densi y-dependen p ocesses, e.g. species speci ic pa hogens o he bi o es
which ensu e ha locally dominan species expe ience nega i e eedbacks ( he Janzen-Conell
hypo hesis: Janzen (1970); Connell (1971)). An example o an equalizing mechanism is he
ade-o be ween g ow h and su i al – slow g owing plan s ha e o en a la ge longe i y han
as g owing plan s (C awley, 1997b; Knei el and Chase, 2004).
In gene al, communi ies can be in an equilib ium o in a non-equilib ium s a e: popula ion
sizes in a communi y which is in i s equilib ium s a e a e s able (apa om s ochas ic luc ua-
7
2 Concep s and Me hodologies
ions): heo e ically he occu ing species coexis o e e . In communi ies which a e no in an
equilib ium s a e, popula ion sizes change. The e can be di e en easons o ha : maybe ime
was no su icien o each equilib ium o maybe he e is no equilib ium s a e. In he la e
case species can ne e heless coexis (e.g. luc ua ing sys em like pa ch cycles in o es gaps),
o species do no coexis o e long ime pe iods bu species’ loss is compensa ed by specia ion
o immig a ion. Chesson dis inguishes s able and uns able coexis ence: s able coexis ence im-
plies ha popula ions can eco e om low densi ies and popula ion densi ies show no long- e m
end. Uns able coexis ence implies ha popula ions do no end o eco e om low densi ies,
i.e. in he long un, species can go ex inc . While s abilizing mechanisms suppo s able coexis-
ence, equalizing mechanisms gi e ise o uns able coexis ence.
Field in es iga ions (e.g. Condi e al., 1996, e c.) and expe imen s (e.g. Tilman, 1994; Hec o
e al., 1999; Sche e -Lo enzen e al., 2005) ha e been used o u he ou unde s anding o he
key p ocesses main aining biodi e si y a he communi y le el. In pa allel, di e en ypes o
heo e ical models in es iga ing species in e ac ions and hei in luence on species coexis ence
ha e been de eloped. As an ou come, wo con as ing iews on biodi e si y ha e es ablished
(Cla k e al., 2007; K a e al., 2008): he i s one assumes ha di e ences in species ai s a e
impo an o communi y di e si y – i in ol es physiological ade-o s along a small numbe o
axes in a ai space, including esou ce compe i ion, ’compe i ion-coloniza ion’ and li e his o y
ai s ( o an o e iew on impo an ade-o s see e.g. S ea ns (1992); C awley (1997b,a); Tilman
(1997)). The second iew, neu ali y, assumes ha species di e ences a e negligible o commu-
ni y di e si y; ins ead, species a e assumed o be unc ionally equi alen , popula ions ollow a
andom d i and go ex inc in he long un (Hubbell, 2001). Thus, neu al heo y iews na u e
om he non-equilib ium pe spec i e. In he sense o Chesson (Chesson, 2000), he equalizing
mechanism o a neu al model is, ha all species ha e exac ly he same i ness ( unc ional equi -
alence); he model does no bea a s abilizing mechanism, di e si y is main ained by immig a ion
( om a species pool) and specia ion. In con as , heo ies assuming ha di e ences in species
ai s ma e o di e si y o en use he equilib ium pe cep ion and ask how di e en ade-o s
can sus ain di e si y. He e, bo h s abilizing and equalizing mechanisms come in o play and i is
no clea pe se, whe he a speci ic ade-o ac s s abilizing o equalizing. P ominen examples
o ade-o s in plan communi ies a e compe i ion-coloniza ion ade-o , su i al- ep oduc ion,
g ow h- ep oduc ion and seed-size e sus seed-numbe (e.g. C awley, 1997b).
Models o coexis ence
One o he ea lies heo e ical model dealing wi h compe i ion o ( wo) species and he ques ion
unde which condi ions hese species coexis is he Lo ka-Vol e a compe i ion model (see e.g.
Townsend e al., 2003). I consis s o wo simple di e en ial equa ions desc ibing a logis ic- ype
popula ion g ow h o wo species depending on g ow h a es, ca ying capaci ies and compe i-
ion coe icien s o bo h species. In he Lo ka-Vol e a model, all ac o s a ec ing compe i ion
a e subsumed in he compe i ion coe icien . Bu compe i ion can occu on di e en axes, i.e.
o di e en esou ces. The e o e, i we a e in e es ed in unde s anding compe i ion in mo e
de ail, an expansion o he simple compe i ion coe icien in o ecologically meaning ul ac o s is
bene icial. Consequen ly, he simple and elegan ma hema ical desc ip ion o compe i ion in he
Lo ka-Vol e a model has unde gone a ious modi ica ions and ex ensions and inspi es ecological
esea ch o da e (e.g. Le ins and Cul e , 1971; Tilman, 1994; Pacala and Rees, 1998; Mu ell
and Law, 2003). One possible way o ca ego izing compe i ion models is o di e en ia e he way
hey add ess space. While he o iginal Lo ka-Vol e a model does no inco po a e any spa ial
8
2 Concep s and Me hodologies
e ec s, many o he models o compe i ion which a e based on di e en ial equa ions ea space
implici ly (e.g. Le ins, 1969; Sla kin, 1974; Has ings, 1980; Tilman, 1994; Pacala and Rees, 1998).
Spa ially implici models assume well mixed-popula ions, i.e. e e y indi idual expe iences he
in luence o all o he indi iduals in he sys em o he same ex en . One bene i o spa ially
implici models is ha hey o en a e analy ically ac able. Bu in many ecosys ems, space
plays an impo an ole since indi iduals in e ac mos ly wi h hei local en i onmen ; his is
pa icula ly ele an o sessile o ganisms like plan s. Fo example, in a spa ially implici o es
model e e y ee would ecei e he same (medium) amoun o ligh as a esou ce. This assump-
ion con adic s he pe cep ion o o es s as a shi ing mosaic o pa ches o di e en successional
s ages (Whi mo e, 1998). Spa ially explici models o compe i ion, whe e indi iduals ha e an
assigned posi ion in he landscape, can accoun o such local in e ac ions o o ganisms. Space
in hese models can be ea ed con inuously o i can be subdi ided in o disc e e uni s (mos ly
squa e la ices). Spa ially explici models a e mo e lexible – hey allow inco po a ing no only
local in e ac ions bu also ac o s like spa ial he e ogenei y o spa ially co ela ed dis u bances
(e.g. Bani z e al., 2008) – bu usually his lexibili y en ails he loss o analy ical sol abili y
(Klausmeie and Tilman, 2002).
Since he ecosys em ha is o in e es o his hesis is a plan communi y, and local in e ac ions
such as compe i ion o ligh and space a e impo an , he model de eloped o in es iga ing
species coexis ence – p esen ed in pape 1 o he hesis – is spa ially explici . Conce ning ade-
o s, we ocus on he basic demog aphic ai s g ow h, seed dispe sal and mo ali y. Addi ionally
o ade-o s be ween hese ai s, we in es iga e wo p ocesses ha modi y local compe i ion:
densi y-dependen mo ali y and ligh -dependen egene a ion. Many exis ing modelling s udies
ocus on single mechanisms o species coexis ence (e.g. Po ho e al., 2006; P onk e al., 2007;
Cla k e al., 2007; Münkemülle e al., 2009; Cla k e al., 2010); ye , he ques ion how di e en
mechanisms in e ac emains open. In addi ion, ecen modelling s udies ha e shown ha some
ade-o s p omo e coexis ence only in na ow pa ame e anges (Lischke and Lö le , 2006; Bani z
e al., 2008). We he e o e in es iga e how di e en ade-o s be ween li e his o y ai s alone
and in combina ion wi h addi ional mechanisms ha modi y local compe i ion a ec species
coexis ence in plan communi ies.
2.2 Unde s anding o es dynamics
Eigh housand yea s ago, p io o he as expansion o human en e p ise ac oss he globe,
ea h’s o es s co e ed an es ima ed a ea o 62 million squa e kilome es (Size e al., 1997).
Fo es s ha e been exploi ed by humans since ancien imes and al eady Pla o no iced he de o -
es a ion o he o es s o A ica (Thomas, 1956). Today’s o es s co e app oxima ely 40 million
squa e kilome es (∼30% o he wo ld’s land co e , FAO (2009)), i.e. abou 40% o he o iginal
o es co e has disappea ed – s onges losses conce n empe a e and opical egions (Malhi
e al., 1999; Hansen e al., 2010).
Fo es s ul il a ious i al unc ions: hey p oduce enewable esou ces such as imbe o con-
s uc ion, uelwood and non- imbe p oduc s like ui s, medicinal plan s e c.; hey mode a e
local clima e and wa e cycles and p ocu e di e se habi a s o animals. In ecen decades g ow-
ing a en ion is gi en o he ole o o es s o he global ca bon cycle in he con ex o clima e
change (Bee e al., 2010) – es ima ions o ca bon s o ed in he wo ld’s o es ange om 220 o
540 giga ons (G ) (Dixon e al., 1994; Hough on e al., 2009), which is a conside able amoun
compa ed o ci ca 750 G ca bon s o ed in he a mosphe e (G ace, 2004). The apid a es o
ongoing de o es a ion, accompanied by losses o o es unc ions and di e si y (Fea nside, 2005;
an de We e al., 2009), b ough along he need o a be e unde s anding o o es ecosys-
9
2 Concep s and Me hodologies
ems and ools o hei managemen . Fo es models a e ins umen s o mee hese challenges –
acco dingly hey belong o he pionee s o ecological models. Rele an a eas o applica ion o
o es models a e (Hu h, 1999)
•ecological esea ch: unde s anding o es dynamics
•p edic ion o o es dynamics unde changed condi ions (managemen , clima e change)
•planning and guidance o o es managemen and moni o ing
Depending on he pu pose o he model, he s a e a iables o desc ibe he o es , as well as
he app op ia e model design may di e . To da e, nume ous o es models o di e en ypes o
o es (bo eal/ empe a e/ opical o es s, e en-/une en-aged, single/mul iple species) exis . One
possible di e en ia ion o o es models is o dis inguish be ween whole-s and models, s and-
class models and single- ee models (Vanclay, 1995). Whole-s and models use agg ega ed
a iables like s and basal a ea, wood olume o he age o a s and o desc ibe a o es (e.g.
Jel sch and Wissel, 1994). The adi ional ype o whole-s and models a e yield ables which
we e i s applied in mono-species, e en-aged s ands. Yield ables p edic he ha es able wood
olume depending on he plan ed species, he age o he o es s and, si e quali y and applied
o es managemen (e.g. Schwappach, 1890). S and-class models a e mo e de ailed han whole
s and models: hey desc ibe he de elopmen o di e en g oups (e.g. diame e size classes, age
classes, species g oups) o ees. Some s and-class models use di e en ial equa ions o desc ibe
he o es s a e (e.g. Mose , 1974); ano he ype o s and-class models a e so-called Ma ko
models, which desc ibe o es dynamics as s ochas ic p ocesses: ees ha e a ce ain ansi ion
p obabili y o mo e om one diame e size class o he nex (Suzuki, 1971; Logo e and Lesnaya,
2000). Single- ee models desc ibe he s a e and g ow h o e e y indi idual ee (abo e a ce ain
size) o he s and, and he e o e hey belong o he class o indi idual-based models (G imm and
Railsback, 2005). These models inco po a e compe i ion be ween ees o esou ces, o example
compe i ion o space can be accoun ed o by a ac o o c own compe i ion (Ek and Monse ud,
1974; Dudek and Ek, 1980; P e zsch, 2001; Shuga , 2003). Fo es gap models belong o he class
o single- ee models; hey desc ibe he g ow h o all ees on a small pa ch o land, he so-called
gap (Bugmann, 2001; Shuga , 2002). The unde lying idea o gap models is ha o es s consis
o mosaic pa ches o di e en successional s ages and ha each gap passes h ough he ull cycle
o succession (c . Figu e 2.1). A de i a i e o single- ee models a e p ocess-o ien ed models,
which explici ly ake in o accoun physiological p ocesses like pho osyn hesis, espi a ion, wa e
and nu ien -cycling (Bossel and K iege , 1991; Moh en and Bu kha , 1994; G o e and E ha d,
1999). Such models allow calcula ing he ca bon balance o each ee, and ca bon can be alloca ed
o di e en compounds o he ee (s em, oo s, lea es).
The mo e de ailed he o es is desc ibed in he model, he mo e complex he model becomes
(e.g. highe numbe o pa ame e s) and compu a ional demand inc eases. Also he ype o da a
used o model pa ame iza ion changes: whole-s and and s and-class models commonly u ilize
census da a; single- ee models o en inco po a e addi ional knowledge on ee allome y and
p ocess-based models also inco po a e knowledge on physiological p ocesses.
As desc ibed in he in oduc ion, we a e in e es ed in unde s anding he dynamics o a opical
mon ane o es in ela ion o dis u bance (landslides); in pa icula , we wan o in es iga e how
changes in di e en li e his o y ai s o ees a ec he egene a ion p ocess on landslide si es.
This is because a ious en i onmen al a iables change on landslides si es, po en ially a ec ing
o es dynamics in he long e m and on la ge spa ial scales. Fo his pu pose, a p ocess-based
app oach ha add esses physiological p ocesses explici ly is mos sui able. To in es iga e he
10

2 Concep s and Me hodologies
Figu e 2.1: Gap-cycle: o es dynamics on a small pa ch o land, a e Shuga (2003).
Impo an p ocesses a e mo ali y (o canopy ees), ec ui men , sel - hinning, compe i ion.
e ec o landslides on he landscape le el i is also ad an ageous o use a spa ially explici
app oach, whe e one can simul aneously compa e dis u bed wi h undis u bed o es zones. The
o es model u ilized in pape 2 and 3 o his hesis is he FORMIND model, which combines
he gap model app oach wi h he p ocess-o ien ed philosophy.
The FORMIND model
The FORMIND model and i s p edecesso model FORMIX a e spa ially explici , indi idual-
based o es models. They a e designed o analyze he dynamics o une en-aged, species- ich
o es s ands wi h a ocus on he impac o na u al o an h opogenic dis u bances on o es
s uc u e and composi ion (Köhle , 2000). They ha e been success ully applied o a ious o es s
h oughou he opics (Hu h e al., 2005; Rüge e al., 2007b; G oene eld e al., 2009; Gu ie ez
e al., 2009; Köhle and Hu h, 2010). The main p ocesses o he model a e ee g ow h, mo ali y,
es ablishmen o young ees and possibly ex e nal dis u bances (e.g. landslides, wind h ows,
i e). On small pa ches (20 x 20 me e ) all ees compe e wi h each o he o ligh and space.
The ligh clima e o e e y pa ch is calcula ed and ees g ow acco ding o pho osyn hesis and
espi a ion a es. FORMIND simul aneously simula es a ce ain numbe o pa ches; his way,
a eas om one hec a e up o se e al hund ed squa e kilome es can be simula ed. Pa ches ha e
explici posi ions and in e ac ia wo p ocesses: big ees shade ees in neighbou ing pa ches
and alling dead ees can damage ees in neighbou ing pa ches.
A common challenge o he pa ame iza ion o o es models o opical o es s is he high
numbe o ee species in combina ion wi h sca ce da a. FORMIND uses he plan unc ional
ype app oach (Smi h e al., 1997; Köhle e al., 2000; Jel sch e al., 2008): species a e g ouped
acco ding o physiological ai s like ligh s a us o maximal diame e . The pa ame iza ion
11
2 Concep s and Me hodologies
Figu e 2.2: Snapsho s om he 3D- isualisa ion o he FORMIND model. Di e en colou s
o ee c owns ep esen di e en plan unc ional ypes.
o such models in ol es analysis o da a om di e en sou ces: census da a (e.g. diame e and
heigh measu emen s), physiological measu emen s (e.g. pho osyn hesis a e, wood densi y) and
measu emen s o en i onmen al a iables (e.g. adia ion abo e canopy). Whe e ield da a is
missing, expe knowledge and alues om he li e a u e o o he o es s can help o de e mine
easonable anges o pa ame e s which can hen possibly be na owed h ough model calib a ion.
In pape 2 we de elop a pa ame iza ion o FORMIND o he idge o es , one o es ype o
ou esea ch a ea. We hen u ilize he model o in es iga e he in luence o landslides on o es
dynamics on he local, as well as on he landscape scale (pape 3). The aim o his in es iga ion
is o de elop a be e unde s anding o he in luence o his special ype o dis u bance on o es
dynamics, pa icula ly on o es s uc u e, biomass and p oduc i i y.
2.3 Slope s abili y models
A common ool o es ima ing landslide isks in mon ane egions a e physically-based slope
s abili y models. Such models combine opog aphical, soil and hyd ological a ibu es, as well as
ege a ion ela ed ac o s (e.g. oo cohesion) in o de o p edic slope s abili y (e.g. Mon gome y
and Die ich, 1994; Bo ga e al., 1998; Guzze i e al., 1999). In combina ion wi h GIS-based
ools, slope s abili y models can be used o p edic landslide loca ions and p oduce landslide isk
maps (e.g. Xie e al., 2004). He e, landslide in en o ies p o ide es s o model pe o mance by
compa ing obse ed landslide loca ions wi h model p edic ions (Bo ga e al., 1998). A p ominen
class o slope s abili y models employ plana in ini e slope analysis, assuming ha slopes a e
con inuous and long, and ha he hickness o he uns able laye is small compa ed o he
slope leng h (e.g. Sidle, 1992). Based on Mou -Coloumb’s ailu e c i e ion, he ac o o sa e y
12
2 Concep s and Me hodologies
– de ined as he a io o s abilizing and des abilizing o ces – desc ibes he s abili y o a slope.
A ac o o sa e y smalle han 1 indica es ins able slope condi ions. This ype o slope s abili y
models ha e been widely applied o in es iga e he ole o de o es a ion and he ole o oads on
slope s abili y (e.g. Sidle and Wu, 1999; Dhakal and Sidle, 2003; Bo ga e al., 2005; Imaizumi
e al., 2008). In pape 4 we adap a s anda d slope s abili y model o he si ua ion o opical
o es s whe e ees ha e p edominan ly shallow oo s. Fo his, we add an abo eg ound laye ,
ep esen ing he o ganic laye which po en ially migh des abilize slopes, o he s anda d model.
This s udy ocusses on he analysis o ac o s ha igge e y shallow ansla ional landslides,
in ol ing almos no ino ganic ma e ial. The eby, we add ess no only he ques ion how landslides
a ec o es dynamics (c . pape 3) bu also how o es ege a ion po en ially a ec s landslide
occu ence.
13
Ecological Modelling 221 (2010) 2227–2236
Con en s lis s a ailable a ScienceDi ec
Ecological Modelling
jou nal homepage: www.else ie .com/loca e/ecolmodel
Wha enables coexis ence in plan communi ies? Weak e sus s ong species
ai s and he ole o local p ocesses
Claudia Dislich∗, Ka in Johs ∗∗, And eas Hu h∗∗
Helmol z Cen e o En i onmen al Resea ch UFZ Leipzig, Depa men o Ecological Modelling, P.O. Box 500136, 04301 Leipzig, Ge many
a icle in o
A icle his o y:
Recei ed 26 Janua y 2010
Recei ed in e ised o m 8 June 2010
Accep ed 10 June 2010
A ailable online 23 July 2010
Keywo ds:
Coexis ence
Compe i ion
T ade-o
Plan communi y
Fo es dynamics
Simula ion model
abs ac
Explaining he coexis ence o species ha basically depend on he same esou ces has been a b ain ease
o gene a ions o ecologis s. Di e en mechanisms ha e been p oposed o acili a e coexis ence in plan
communi ies, whe e space is an impo an esou ce. Using a s ochas ic cellula au oma on simula ion
model we analyze – sepa a ely and in combina ion – he influence o di e en species ai s and p ocesses
which al e local compe i ion on he coexis ence o plan species o e a fixed ime ho izon. We show ha
di e en species ai s ope a e on di e en ime scales in compe i ion. We he e o e sugges he concep
o weak e sus s ong ai s acco ding o sho - o long- e m exclusion o species di e ing in hese ai s.
As a consequence, highly non-linea ade-o s be ween weak and s ong ai s can esul in communi-
ies. Fu he mo e, we ound ha ade-o s based on physiological species ai s such as plan li e ime,
dispe sal ange and plan g ow h, did no suppo b oad and long- e m coexis ence— u he p ocesses
such as densi y-dependen mo ali y and ligh -dependen coloniza ion we e necessa y. This sugges s
ha coexis ence in plan communi ies equi es (s abilizing) local p ocesses o suppo he (equalizing)
ade-o s in species ai s.
© 2010 Else ie B.V. All igh s ese ed.
1. In oduc ion
The ques ion o species coexis ence has been challenging
ecological esea ch o decades. A bundle o heo ies has been
sugges ed o explain coexis ence be ween species, p ominen
examples a e niche heo y (Hu chinson, 1957), neu al he-
o y (Hubbell, 2001) o he in e media e dis u bance hypo hesis
(Connell, 1978; Roxbu gh e al., 2004). Each o hese heo ies has
also been discussed in ela ion o di e si y in plan communi-
ies. In he amewo k o niche heo y one impo an ac o o
species coexis ence in plan communi ies is he exis ence o in e -
specific ade-o s in physiological ai s (Tilman and Pacala, 1993;
W igh , 2002). Fo example, one species could ha e an ad an-
age conce ning one physiological a ibu e, bu his ad an age
could be balanced ia a disad an age conce ning a second a ibu e.
The e is ich li e a u e on di e en ade-o s, mos amous he
compe i ion–coloniza ion ade-o (Tilman, 1994; Holmes and
Wilson, 1998), whe e species di e in compe i i e s eng h and
colonizing abili y. This ade-o has been shown o os e coex-
is ence o a ce ain deg ee (Klausmeie and Tilman, 2002). Fu he
coexis ence mechanisms a e ela ed o spa ial and/o empo al he -
∗P incipal co esponding au ho . Tel.: +49 341 235 1707.
∗∗ Co esponding au ho s.
E-mail add esses: [email p o ec ed] (C. Dislich), [email p o ec ed]
(K. Johs ), [email p o ec ed] (A. Hu h).
e ogenei y. P ocesses ha ope a e on small local/ empo al scales
can ha e he po en ial o enhance species coexis ence, o ins ance
local densi y-dependen p ocesses a ising om species specific
pes s o p eda ion can p e en abundan species o become all-
dominan (Cha e e al., 2002; Molo sky e al., 1999). Wi hin a
heo e ical amewo k Chesson (2000) di ides all hose di e en
p oposed coexis ence mechanisms in o equalizing mechanisms
and s abilizing mechanisms—while equalizing mechanisms educe
fi ness di e ences be ween species, s abilizing mechanisms educe
in e specific compe i ion in ela ion o in aspecific compe i ion
(see also Adle e al., 2007).
Fo analyzing and es ing coexis ence mechanisms, ecological
models play an impo an ole (Du e and Le in, 1994; Ke e al.,
2002; Johs and Hu h, 2005). Bu many model s udies ocus on only
single mechanisms o species coexis ence (e.g. Po ho e al., 2006;
P onk e al., 2007; Es he e al., 2008; Münkemülle e al., 2009);
he ques ion how di e en mechanisms in e ac emains open. In
addi ion ecen model s udies ha e shown ha ade-o s alone
p omo e coexis ence only in na ow pa ame e anges (Lischke and
Lö fle , 2006; Bani z e al., 2008).
In his s udy, we he e o e in es iga e compe i ion o plan s in
a spa ial con ex . We ask, how ade-o s alone and in combina ion
wi h addi ional p ocesses ha modi y local compe i ion (he e-
a e called local p ocesses) a ec coexis ence. As sessile o ganisms
plan s in e ac mos ly wi h hei local en i onmen ; o eflec hese
local in e ac ions we chose a spa ially explici indi idual-based
app oach (Du e and Le in, 1998). Ou model is a s ochas ic cel-
0304-3800/$ – see on ma e © 2010 Else ie B.V. All igh s ese ed.
doi:10.1016/j.ecolmodel.2010.06.011

2228 C. Dislich e al. / Ecological Modelling 221 (2010) 2227–2236
lula au oma on, inspi ed by he Di Game Simula o (Alonso and
Solé, 2000). The gene ali y o he model allows o applica ions
o di e en communi ies wi h sessile o ganisms such as g ass-
lands, o es s o co al ee s. We ocus he e on o es communi ies,
he e o e ou indi idual en i y is a ee. The species ai s we exam-
ine as ade-o a ibu es a e seed dispe sal ange, ee g ow h
a e, and mo ali y a e. We compa e condi ions o species coex-
is ence in h ee models: a basic ade-o model and wo models
whe e local p ocesses a e added o he basic model. These addi-
ional local p ocesses a e densi y-dependen mo ali y (DDM) and
ligh -dependen coloniza ion (LDC). The ou come o compe i ion
be ween wo species o e a gi en ime ho izon is analyzed in wo
s eps: fi s we in es iga e he e ec o single ai s on species
coexis ence. To do so we look a he compe i ion be ween wo
species which a e iden ical in all bu one ai . Secondly we in o-
duce all h ee possible ade-o s be ween he li e-his o y ai s
(dispe sal-mo ali y, dispe sal-g ow h, and g ow h-mo ali y) and
explo e he anges o coexis ence wi h and wi hou local p ocesses.
Finally we s udy he compe i ion be ween mul iple species in a
ade-o communi y and in communi ies wi h he addi ional local
p ocesses.
2. Model desc ip ion
The s uc u e o he model desc ip ion ollows he ODD p o ocol
(G imm e al., 2006). We seek a model o minimal complexi y ha
cap u es he essen ial p ocesses o o es dynamics.
2.1. S a e a iables and scales
Ou model is a spa ially explici , indi idual-based simula ion
model ha includes compe i ion o ees o ligh and space. Space
is di ided in o pa ches on a g id. The g id has 20 ×40 pa ches
(200 ×400 pa ches in he mul i-species compe i ion); each pa ch
(10 m ×10 m) can hos a ma u e ee, hus he simula ed o es co -
e s an a ea o 8 ha. A e compe i ion o space among seeds and
seedlings, especi i ely, a pa ch accommoda es a mos one ee
a a ime. Each ee is cha ac e ized by i s loca ion, he species i
belongs o and i s heigh .
Species di e in he a ibu es seed dispe sal ange, mo ali y
a e and g ow h a e. Maximal ee heigh is fixed o he same le el
o all species and all young ees ecolonize emp y pa ches wi h a
p edefined minimal heigh . Abo e a ce ain h eshold heigh ees
a e ma u e and sp ead seeds wi hin he whole dispe sal ange.
Fo he compe i ion o wo species we s a e ha a species coex-
is s, i i occupies a leas 10% o all pa ches a he end o he
obse a ion pe iod.
2.2. P ocess o e iew and scheduling
The dynamics o he communi y a e modelled wi h an annual
ime s ep. Wi hin one ime s ep, Nsingle s eps a e pe o med, wi h
Nbeing he numbe o pa ches. A single s ep comp ises he p o-
cesses coloniza ion, mo ali y and g ow h comple ed acco ding o
Fig. 1.
2.2.1. Basic ( ade-o ) model
2.2.1.1. Coloniza ion. Species may ha e di e en dispe sal dis-
ances. Emp y si es can be colonized, i seeds each ha si e and
i he mean heigh o neighbo ing ees (8-pa ch-neighbo hood)
is below a h eshold h . Reoccupa ion o emp y pa ches a e a
dis u bance is implemen ed as a lo e y compe i ion: he num-
be o seed-p oducing ees wi hin he dispe sal dis ance o he
emp y pa ch is coun ed and hen, weigh ed acco ding o species
abundances, he colonizing species is chosen a andom.
Fig. 1. O de o p ocesses wi hin a single s ep. This sequence is epea ed wi hin one
yea acco ding o he numbe o pa ches. While coloniza ion and g ow h a e always
influenced by local en i onmen al condi ions, mo ali y is only influenced by local
condi ions in he p esence o densi y-dependen mo ali y (DDM). Ligh -dependen
coloniza ion (LDC) s esses he influence o local en i onmen al condi ions on col-
oniza ion.
2.2.1.2. Mo ali y. Each species has a basic mo ali y a e m, he
p obabili y o each ee o die in one yea . Dying ees a e emo ed
and lea e emp y pa ches o ecoloniza ion.
2.2.1.3. G ow h. Each species has a g ow h a e g ha ep esen s
maximal annual g ow h in he absence o ligh compe i ion. Com-
pe i ion o ligh is asymme ic—in he same en i onmen a big ee
ecei es mo e ligh han a smalle ee. Thus g ow h is educed i
a ee is shaded by neighbo ing ees. The s eng h o g ow h sup-
p ession depends on he mean heigh o he eigh neighbo ing ees
(hm). I his mean heigh is smalle han he heigh o he ocal ee
i, ligh compe i ion is neglec ed and he ee g ows a i s maximal
a e:
Hi( + ):=Hi( )+g· (1)
I he mean heigh o neighbo s exceeds he heigh o he ocal
ee, g ow h is educed by he ac o (Hi( )/hm)0.5, and he new ee
heigh is calcula ed acco ding o
Hi( + ):=Hi( )+g·Hi( )
hm0.5
· . (2)
Fo he educ ion o g ow h unde ligh compe i ion we es ed di -
e en unc ions. P elimina y simula ions showed ha he chosen
unc ion esul s in a easonable dis ibu ion o ee heigh s whe e
all possible ee heigh s a e ep esen ed.
2.2.2. Local p ocesses
2.2.2.1. In aspecific densi y-dependen mo ali y (DDM). When
ees o he same species a e locally clumped hey migh encoun e
inc eased mo ali y due o p opaga ion o pes s o species specific
he bi o es and pa hogens (Cha e e al., 2002). Densi y-dependen
p ocesses ha e been de ec ed in o es s o di e en biomes, e.g. in
opical o es s (John e al., 2002; Pe e s, 2003), in bo eal o es s
(G ay and He, 2009) and in empe a e o es s (Zhang e al., 2009).
We in oduce an in aspecific densi y ac o d∈[0,1] ha inc eases
he basic mo ali y a e m, i mo e han hal o he neighbo ing
pa ches a e occupied by conspecifics. The new mo ali y a e md
o he ocal indi idual inc eases p opo ionally wi h he numbe o
conspecific neighbo s. The maximum inc emen o he basic mo -
ali y a e is gi en by he densi y ac o , e.g. a densi y ac o o 1
C. Dislich e al. / Ecological Modelling 221 (2010) 2227–2236 2229
leads o a doubling o he mo ali y a e, when all neighbo s a e
conspecifics. Thus mddepends on he numbe o su ounding con-
specifics nand he densi y ac o d:
md=m o n≤4
m·1+n−4
4·d o n>4(3)
2.2.2.2. Ligh -dependen coloniza ion (LDC). I is assumed ha each
species has a p e e ed ligh clima e a which i is mo e likely o
win he ec ui men compe i ion. An index Li o a ce ain ligh
equi emen ,isassigned oeachspeciesi,indica ing he mos a o -
able ligh condi ion o coloniza ion. Lis a numbe be ween 0 and
1, whe e 0 ep esen s a ligh -demanding and 1 a shade- ole an
species. Fo an emp y pa ch, he ligh clima e index lis calcula ed
ia he mean heigh o he eigh neighbo ing ees (i= 1..8):
l=
(1/8) ·hi
h ∈[0,1].(4)
The species iwi h Liclose o ligh clima e lis mo e likely o colo-
nize. We employ he no mal dis ibu ion wi h mean Liand s anda d
de ia ion (he e 0.05) o calcula e weigh ing ac o s wi o he
coloniza ion o each species i:
wi=1
√2 e−(l−Li)2/22.(5)
Acco ding o he weigh s wicoloniza ion is hen implemen ed
s ochas ically.
2.3. Design concep s
The model is o mula ed as a s ochas ic cellula au oma on
wi h pe iodic bounda ies. Coloniza ion and mo ali y a e s ochas-
ic p ocesses, g ow h is de e minis ic and depends on he heigh
o neighbo ing ees (Eqs. (1) and (2)). We ocus on he obse -
a ion o species abundances o e ime. Abundances o species
and hus species coexis ence and exclusion, espec i ely, eme ge
h ough species ai s, spa ial dis ibu ion o species and ini ial con-
di ions. T ees in e ac by supp essing he g ow h o neighbo ing
ees h ough shading and ia densi y dependence. Coloniza ion o
emp y pa ches depends on species composi ion a ound he pa ch,
on ligh clima e and on he ligh demands o species (Eqs. (4) and
(5)).
Wi hin one ime s ep, he pa ches a e chosen a andom; upda -
ing akes place asynch onously, he g id is upda ed a e each single
s ep. This upda e p ocedu e adds s ochas ici y o he (o he wise
de e minis ic) g ow h p ocess.
2.4. Pa ame e s and ini ializa ion
The model pa ame e s (inspi ed by Alonso and Solé, 2000 and
Shuga , 2003) a e summa ized in Table 1. Each ee has an ini-
ial heigh o 0.1 m and can each a maximum heigh o 40 m. Seed
dispe sal se s in a a ee heigh o 1 m and he h eshold o mean
neighbo heigh o coloniza ion (h ) is se o 12 m. A he beginning
o he simula ion, species a e andomly dis ibu ed a low densi-
ies (80 indi iduals pe species), each indi idual has an ini ial ee
heigh o 0.1 m. This ini ializa ion eflec s one possible coloniza ion
si ua ion on ba e g ound. Expe imen s wi h di e en ini ial condi-
ions conce ning he size o ini ial ees and also he densi y o ees
in he ini ial s a e all e ealed simila model beha io and esul s.
We in es iga e he compe i ion o wo species wi h di e en
ai s o e a fixed ime o 1000 yea s. The dispe sal ange a ies
om 10 m (=leng h o one pa ch, only he 4 nea es neighbo s) o
100 m wi h a s ep wid h o 5 m. Annual g ow h a es a y om
0.05 o 1 m/yea (m/y) (s ep wid h 0.05 m/y) and mo ali y a es
Table 1
Pa ame e s o he model.
Pa ame e s Values/ anges
G id cells 20 ×40 (200 ×400a)
Species numbe 2 (196a)
Simula ion leng h [y] 1000 (15 000a)
Species independen
Minimal ee heigh [m] 0.1
Ma u ing heigh [m] 1
Maximal ee heigh [m] 40
Th eshold heigh o coloniza ion [m] 12
Species specific
Mo ali y a e [1/y] [0.005, 0.1]
Dispe sal ange [m] [10,100]
G ow h a e [m/y] [0.05, 1]
Densi y ac o d[0,1]
Ligh equi emen index L[0,1]
aValues o he mul i-species compe i ion.
om 0.005 o 0.1 pe yea (y−1) (s ep wid h 0.005 y−1). Wi h hese
pa ame e anges we co e a la ge a ie y o ecologically eason-
able species ai s (e.g. Phillips and Gen y, 1994; Whi mo e, 1998;
Kohyama e al., 2003; Mulle -Landau e al., 2008).
To in es iga e he ole o single ai s and ade-o s be ween
wo ai s we a y only wo pa ame e s a a ime and fix all emain-
ing ai s o bo h species o medium alues wi hin he conside ed
anges (60 m dispe sal ange, 0.5 m/y g ow h a e and 0.05 y−1mo -
ali y a e).
3. Resul s
To ge a fi s idea o he dynamics o he compe ing popula-
ions, Fig. 2 shows he changes in abundances o e ime and spa ial
snapsho s o h ee si ua ions. The fi s example shows compe i-
i e exclusion o species 1 a e 800 yea s, igge ed by an ad e se
dispe sal-mo ali y ade-o : he disad an age o a highe mo al-
i y a e o species 1 compa ed o species 2 is no balanced by he
la ge dispe sal ange o species 1. The second example shows neu-
al coexis ence: bo h species ha e comple ely iden ical ai s and
we obse e a conside able amoun o fluc ua ion in he abundances
due o s ochas ici y. The las example shows coexis ence h ough
a combina ion o a ade-o and a local p ocess: a shade- ole an
species 2 wi h less dispe sal abili ies and a highe mo ali y coexis s
wi h a ligh -demanding species 1.
3.1. Explo a ion o he basic ( ade-o ) model
3.1.1. T ai s: compe ing species wi h di e ences in only one
a ibu e
We fi s analyze he e ec o single ai s on coexis ence and
examine wo species ha di e only in one ai (dispe sal ange,
mo ali y a e o g ow h a e, Fig. 3). I species di e in dispe -
sal ange o mo ali y a e (Fig. 3a and b), hen al eady a e 1000
yea s compe i i e exclusion has aken place leading o one dom-
inan species ( he species wi h highe dispe sal ange o lowe
mo ali y a e). We only ge coexis ence in he neu al case wi h
iden ical species. In con as , wo species wi h di e en g ow h
a es (Fig. 3c) may s ill coexis a e 1000 yea s. Thus we hypo h-
esize ha , compa ed o g ow h a e, he a ibu es dispe sal ange
and mo ali y a e a e ‘s ong’ ai s. The ‘weak’ ai g ow h a e
influences coexis ence on a slowe imescale; o longe ime pe i-
ods he coexis ence ange in he diag am o di e en g ow h a es
also ‘sh inks’ down o he diagonal (>6000 yea s, simula ed bu
no shown). (No e ha he ime span un il compe i i e exclusion
akes place also depends on he size o he simula ed a ea—on a
bigge g id he dynamics slow down.) Ou esul s show ha he
2230 C. Dislich e al. / Ecological Modelling 221 (2010) 2227–2236
Fig. 2. Species abundance o e ime and snapsho s o he g id a e 1000 yea s o simula ion o h ee pa ame e combina ions. Species 1 is illus a ed in blue (g ey), species
2 is g een (black). Each ci cle is an indi idual ee, he ci cle size indica es ee size. Pa ame e s: (a) espec i e dispe sal anges o 60 and 50 m, g ow h a es o 0.5 m/y o
bo h species and espec i e mo ali y a es o 0.05 and 0.03 y−1. (b) comple ely iden ical species, dispe sal anges o 60 m, g ow h a es o 0.05 m/y and mo ali y a es o
0.05 y−1. (c) Model wi h ligh -dependen coloniza ion (LDC), species 1 ligh -demanding (L1= 0), species 2 shade- ole an (L2= 1); espec i e dispe sal anges o 80 and 60 m,
g ow h a es o 0.5 m/y o bo h species and espec i e mo ali y a es o 0.05 and 0.08 y−1. (Fo in e p e a ion o he e e ences o colo in his figu e legend, he eade is
e e ed o he web e sion o he a icle.)
ime un il compe i i e exclusion akes place may di e consid-
e ably be ween ai s when a ied in an ecologically easonable
ange; we he e o e in oduce he concep o weak e sus s ong
ai s.
3.1.2. T ade-o s: compe ing species wi h di e ences in wo
a ibu es
We now come o ade-o s (Fig. 4), i.e. we conside compe -
ing species ha di e in wo ai s. In he ollowing, one species
(species 1) is chosen o ha e medium p ope ies and we a e in e -
es ed in he ai s a second species mus show o coexis wi h his
‘medium’ species.
Fo all h ee possible ade-o s, he coexis ence anges a e e y
small o an obse a ion pe iod o 1000 yea s. In he dispe sal-
mo ali y ade-o (Fig. 4a) wo species can only coexis , i he e is a
p ope ade-o be ween he wo a ibu es: i species 2 has a la ge
dispe sal ange (e.g. 70 m) han species 1, i mus also ha e a highe
mo ali y a e (0.06 y−1). A sligh change o only one a ibu e (e.g.
species 2 dispe sal ange 70 m, mo ali y a e 0.065 y−1) esul s in
he ex inc ion o one species.
Fo ade-o s in ol ing g ow h a e (Fig. 4b and c), he pa am-
e e combina ions which allow coexis ence show a non-linea
ela ion wi hin he conside ed pa ame e anges compa ed o he
linea coexis ence cu e in Fig. 4a. This is again an indica o o he
a ying s eng hs o species a ibu es: he weak ai g ow h a e
canno balance big di e ences in s ong ai s (in pa icula small
dispe sal anges o high mo ali y a es o species 2). On he o he
hand, he e a e le els o dispe sal ange ( espec i ely, mo ali y
a e) o species 2, whe e di e en g ow h a es o species 2 esul
in coexis ence ( o example g ow h a es be ween 0.75 and 1 m/y
a a dispe sal ange o 55 m, Fig. 4b).
The gene al esul o his in es iga ion o he basic model is
ha coexis ence is only possible o ce ain fine- uned combina-
ions o species ai s. While e en small changes in s ong ai s
(he e dispe sal ange and mo ali y a e) lead o he ex inc ion
o one species, changes in weak ai s (g ow h a e) can p ese e
coexis ence.
C. Dislich e al. / Ecological Modelling 221 (2010) 2227–2236 2231
Fig. 3. Va ia ion o single a ibu es in a wo-species communi y. Each poin is he ou come o a single simula ion un a e 1000 yea s; a yellow (whi e) poin ep esen s
coexis ence, while blue (g ey) s ands o dominance o species 1 and g een (black) o dominance o species 2. Fi s ow: basic communi y wi hou local p ocess. Second ow:
communi y wi h local densi y-dependen mo ali y (DDM); in e, and g bo h species ha e densi y ac o d= 1. Thi d ow: communi y wi h ligh -dependen coloniza ion
(LDC); in i, j and k species 1 is ligh -demanding (ligh equi emen index L1= 0) and species 2 shade- ole an (L2= 1). Column 0: a ia ion o he local p ocesses (d) densi y
ac o and (h) ligh equi emen index. Column 1: a ia ion o dispe sal anges. Column 2: a ia ion o mo ali y a es. Column 3: a ia ion o g ow h a es. The emaining
pa ame e s a e fixed o: 60 m dispe sal ange, 0.5 m/y g ow h a e and 0.05 y−1mo ali y a e. (Fo in e p e a ion o he e e ences o colo in his figu e legend, he eade
is e e ed o he web e sion o he a icle.)
3.2. Explo a ion o he local p ocesses
3.2.1. T ai s: compe ing species wi h di e ences in only one
a ibu e
We in es iga e wo di e en p ocesses which modi y local
compe i ion: densi y-dependen mo ali y and ligh -dependen
coloniza ion. Two species ha ha e comple ely iden ical ai s and
only di e in he s eng h o densi y-dependen mo ali y nea ly
always coexis (Fig. 3d). Only i one species expe iences s ong
densi y dependence while he o he species does no show densi y
dependence, can he la e be a supe io compe i o . The addi ion
o local densi y-dependen mo ali y o bo h species inc eases he
coexis ence ange (Fig. 3e–g). I modifies he compe i i e s eng h
o species ai s: he s ong ai s (dispe sal ange and mo ali y
a e) become weake and he weak ai (g ow h a e) becomes
almos i ele an o coexis ence.
Wi h ligh -dependen coloniza ion (Fig. 3, ligh -dependen
communi y), wo o he wise iden ical species coexis o a b oad
ange o ai combina ions wi h a ious ligh equi emen s
(Fig. 3h). In his way his addi ional species a ibu e ac s in a di -
e en manne o he p e iously conside ed ai s (Fig. 3a–c).
The compe i ion o a ligh -demanding species 1 and a shade-
ole an species 2 esul s in much la ge coexis ence anges o
all h ee single a ibu es (Fig. 3i–k) han in he basic communi y.
Dispe sal anges (Fig. 3i) become almos i ele an o coexis ence.
I bo h species expe ience high mo ali y (Fig. 3j), he ligh -
demanding species domina es he o es , while he shade- ole an
species domina es i i has a low mo ali y a e. We obse e he
mos no able change in he coexis ence pa e n compa ed o he
basic communi y in ela ion o g ow h a es (Fig. 3c and k). In
he in es iga ed ange he e is no combina ion o g ow h a es
whe e he shade- ole an species 2 domina es he compe i ion, and
o some combina ions, whe e in he basic model species 1 was
excluded, i now wins he compe i ion.
3.2.2. T ade-o s: compe ing species wi h di e ences in wo
a ibu es
Conce ning species ha di e in wo ai s (Fig. 4) we obse e
much la ge coexis ence anges o bo h local p ocesses compa ed
o he basic ade-o communi y. The coexis ence anges show a
clea and e en widening i mo ali y depends on he local densi y
o conspecifics (Fig. 4d– ).
Fo he model wi h ligh -dependen coloniza ion we analyze
wo cases whe e species 1 has a medium ligh equi emen
index (L1= 0.5) and species 2 is shade- ole an (L2=1, Fig. 4,
ligh -dependen communi y I) o ligh -demanding (L2=0, Fig. 3,
ligh -dependen communi y II). The bo de s o dominance e sus
coexis ence a eas a e no as sha p as o he o he models. The coex-
is ence anges mainly expand in sec o s whe e in he basic model
species 2 was dominan . Thus he species wi h medium ai s p o -
i s om he in oduc ion o ligh -dependen coloniza ion, because
i can now coexis , whe e p e iously i was excluded.
A shade- ole an species (he e species 2) will only domina e
he compe i ion i i has a low mo ali y a e (Fig. 4g and i) and
a ligh -demanding species will almos ne e exclude he in e me-
dia e species (he e species 1, Fig. 4j–l).
2232 C. Dislich e al. / Ecological Modelling 221 (2010) 2227–2236
Fig. 4. Va ia ion o wo a ibu es in a wo-species communi y. Species 1 has medium p ope ies: dispe sal ange 60 m, mo ali y a e 0.05 y−1and g ow h a e 0.5m/y.
Fi s ow: basic ade-o communi y wi hou local p ocess. Second ow: communi y wi h densi y-dependen mo ali y (DDM) o bo h species (d= 1). Thi d and ou h ow:
communi y wi h ligh -dependen coloniza ion (LDC): species 1 is medium ligh -demanding (ligh equi emen index L1= 0.5). In ligh -dependen communi y I species 2 is
shade- ole an (L2= 1) and in ligh -dependen communi y II species 2 is ligh -demanding (L2= 0). Column 1: dispe sal-mo ali y ade-o , g ow h a e (species 2) = 0.5m/y.
Column 2: dispe sal-g ow h ade-o , mo ali y a e (species 2) = 0.05 y−1. Column 3: g ow h-mo ali y ade-o , dispe sal ange (species 2) =60 m.
A subs an ial e ec o bo h local p ocesses, DDM and LDC, com-
pa ed o he basic model is ha coexis ence is mos ly s able o e
long pe iods o ime. Wi hou such local p ocesses, he coexis ence
anges a e al eady na ow o 1000 yea s and disappea in long-
e m simula ions (15 000 yea s, see Appendix Fig. 7). No e ha e en
iden ical species do no coexis in he long un due o s ochas ici y
in coloniza ion and mo ali y.
4. Discussion
4.1. T ai s ac di e en ly: weak e sus s ong ai s
Ou esul s show ha physiological ai s o plan s may ‘ac ’
on di e en empo al scales, since he pa e ns o coexis ence o
g ow h a es di e om hose o dispe sal anges and mo ali y
a es (Fig. 3). To unde s and his we need o look a how hese p o-
cesses a ec ep oduc ion. Ha ing a la ge dispe sal ange is a clea
ad an age o one species o e ano he since i di ec ly esul s in a
highe p obabili y o colonizing an emp y pa ch. Ha ing a low mo -
ali y a e is also a s ong ad an age: he numbe o dying ees is
educed and he numbe o seed-p oducing ees is inc eased. On
he o he hand, high mo ali y a es allow mo e space o coloniza-
ion esul ing in a highe u no e in he communi y while a he
same ime changing he high s uc u e o he o es ; he e a e mo e
small ees and he numbe o seed-p oducing ees is educed.
Compa ed o hese ‘s ong’ ai s, g ow h a es ac mo e indi-
ec ly on coloniza ion: o ep oduc ion i is only impo an ha a
ee eaches i s ma u e heigh . Secondly, he ac ual g ow h o a ee
is no di ec ly de e mined by he g ow h a e, since g ow h is also
influenced by he heigh o he su ounding ees (Eq. (2)). A high
g ow h a e leads o an inc eased numbe o la ge ees and hus
also o inc eased supp ession o small ees due o shading. Hence
a high g ow h a e is an ad an age o a species, bu one ha wo ks
indi ec ly and hus mo e slowly han he ad an age o a s ong ai

C. Dislich e al. / Ecological Modelling 221 (2010) 2227–2236 2233
such as a la ge dispe sal ange o a low mo ali y a e. The ai s ‘ac ’
on di e en ime scales; his also leads o non-linea coexis ence
ela ions in he ade-o s in ol ing g ow h a es (Fig. 4b and c).
Thus he e migh be h esholds abo e o below which we canno
expec unc ioning ade-o s be ween s ong and weak ai s.
No e ha he dispe sal-g ow h ade-o is one we would no
an icipa e o o es s, whe e gene ally pionee species ha e la ge
dispe sal anges and a e a he same ime a he as g owing
while la e successional species ha e smalle dispe sal anges and
slowe g ow h. A dispe sal-mo ali y o g ow h-mo ali y ade-o
is mo e likely o ma ch o es ecosys ems.
E en o e ela i ely sho imescales (1000 yea s) coexis ence
is only possible o ce ain combina ions o species ai s; e en
small changes in s ong ai s esul in species ex inc ion. This sug-
ges s ha he basic ade-o model is insu ficien o explain species
coexis ence.
4.2. Func ioning o local p ocesses
We explo ed he unc ioning o wo local p ocesses, ha
could po en ially acili a e coexis ence—densi y-dependen mo -
ali y and ligh -dependen coloniza ion. Bo h mechanisms ha e
a simila s ong e ec o inc easing coexis ence by so ening he
s eng h o single species a ibu es. Densi y-dependen mo al-
i y di ec ly p omo es coexis ence by educing he locally abundan
species. We find a simila e ec i locally a e species expe-
ience a educed mo ali y a e (no shown). Bo h phenomena
ha e been discussed o o es s (Janzen, 1970; Wills e al., 2006).
Densi y-dependen mo ali y inc eases he in aspecific compe i-
ion whene e a species becomes locally abundan — his weakens
he s eng h o species a ibu es (Fig. 3, densi y-dependen com-
muni y). Mos combina ions o g ow h a es lead o coexis ence, i.e.
he (weak) ai g ow h a e becomes almos i ele an o coexis-
ence.
The second local p ocess in es iga ed assumes di e en ligh
equi emen s o coloniza ion; his p ocess is di ec ly ela ed o
he he e ogeneous heigh -s uc u e wi hin a o es , which is highly
dynamic o e ime. Wi hou ligh -dependen coloniza ion (LDC)
he local heigh -s uc u e only influences ee g ow h, bu no he
egene a ion. Wi h LDC, emp y pa ches become mo e amenable o
coloniza ion by one o he o he species and spa io- empo al niches
a e hus c ea ed. The e o e wo species wi h di e en ligh equi e-
men s compe e o a lesse ex en o he same pa ches and his
enhances coexis ence. Seen in e ms o he classical Lo ka–Vol e a
compe i ion model, in e specific compe i ion is educed by in o-
ducing LDC.
The dispe sal ange ai becomes almos i ele an o coex-
is ence (Fig. 3e), since he ec ui men lo e y, which p e iously
depended solely on he numbe o po en ial pa en al ees wi hin
he neighbo hood, is eplaced—now he seedling ha is bes
adap ed o he ligh clima e in he emp y pa ch has he bes ec ui -
men chances.
While densi y-dependen mo ali y expands coexis ence anges
symme ically (Fig. 3e–g), adding di e en ligh equi emen
esul s in asymme ic changes in coexis ence anges o mo ali y
and g ow h a es (Fig. 3j and k).
High mo ali y a es (Fig. 3j) o bo h species esul in many
emp y pa ches and small ees. This esul s in high ligh a ailabil-
i y and he ligh -demanding species 1 domina es he compe i ion.
Bo h species can coexis i he ligh -demanding species 1 expe i-
ences low mo ali y and he shade- ole an species 2 expe iences
medium o low mo ali y, because he shade- ole an species p o -
i s om he high abundance o big ees.
A simila a gumen explains he somewha coun e in ui i e
esul (Fig. 3k) ha he ligh -demanding species 1 domina es com-
pe i ion, i i has low g ow h a es: he shade- ole an species
2 simply does no find enough sui able pa ches o coloniza ion,
because o he low ee heigh s o species 1.
As in he case o he single ai s (Fig. 3), coexis ence anges
become conside ably la ge o all h ee conside ed ade-o s
(Fig. 4), when a local p ocess is added o he basic ade-o model.
While he coexis ence ange expands e enly o densi y-dependen
mo ali y (Fig. 4d– ), he majo i y o new coexis ence space caused
by ligh -dependen coloniza ion (LDC) is in a eas whe e species 2
had p e iously domina ed he compe i ion. This sugges s ha he
medium species 1 gene ally does be e han ligh -demanding o
shade- ole an species, because i is no as highly specialized o
ce ain ligh condi ions. The discon inui y o he coexis ence pa -
e ns (Fig. 4, ligh -dependen communi y I and II) s ems om he
ac ha he equilib ium abundances o species can di e so much
(see Fig. 2c), ha one species cons an ly “sc a ches” on he bo de
o coexis ence. Ra e species also ha e a highe p obabili y o going
ex inc due o s ochas ici y.
The only case whe e one, in e ms o ligh demands, mo e spe-
cialized species domina es he medium species is a shade- ole an
species wi h low mo ali y a es (Fig. 4g and i). Such a species
benefi s wo old om low mo ali y—by a low numbe o dying
indi iduals as well as by c ea ing sui able coloniza ion condi ions
o i sel .
4.3. Equalizing e sus s abilizing mechanisms
In he e minology o Chesson (2000), we ound ha he ade-
o s in es iga ed he e in he basic ade-o model a e equalizing
mechanisms ha balance di e ences be ween species, bu do no
p omo e s able, i.e. long- e m coexis ence. Ou esul s show ha
coexis ence is only possible i he species a ibu es show a fixed
ela ion; any de ia ion om his ela ion leads o ex inc ion.
In con as , he local p ocesses (densi y-dependen mo ali y
and ligh -dependen coloniza ion) p oduce b oade coexis ence
anges (Fig. 3d–k, Fig. 4d–l) ha emain s able o e long ime
pe iods (Fig. 7). The e o e hey can be assigned o s abilizing
mechanisms. Wi h hese local p ocesses, ade-o s need no be
fine- uned; in his sense, he local mechanisms conside ed con-
ibu e o s able coexis ence. On he o he hand, local mechanisms
alone do no gua an ee coexis ence when fi ness di e ences
be ween species a e la ge. Fo example, he addi ion o densi y-
dependen mo ali y canno balance a subs an ial di e ence in
dispe sal anges (Fig. 3e). The adjus men o a second a ibu e,
such as mo ali y a es, is necessa y o ob ain coexis ence (Fig. 4d).
Acco ding o he heo e ical findings o Chesson (2000) we con-
clude ha equalizing ade-o s and s abilizing local mechanisms
a e bo h c ucial key p ocesses o acili a ing species coexis ence in
plan communi ies.
Compa ed o he analy ically ac able, spa ially implici pop-
ula ion models in es iga ed in Klausmeie and Tilman (2002) ou
spa ially explici model esembles he local ounde con ol model
in he essen ial poin ha he coloniza ion only akes place in emp y
pa ches. Simila o ou basic ade-o model he coexis ence anges
in he local ounde con ol model a e ma ginal; only he ( heo e -
ical) case o iden ical coloniza ion–mo ali y a ios o species leads
o coexis ence.
The majo i y o simple models dealing wi h he ole o ade-
o s o coexis ence in es iga e ade-o s, whe e one ai is he
compe i i e abili y o species (e.g. Adle and Mosque a, 2000;
Yu and Wilson, 2001; Kisdi and Ge i z, 2003). In hese models
i is assumed, ha he e is a fixed hie a chical o de o compe-
i ion be ween species—a be e compe i o can always in ade
pa ches ha a e occupied by an in e io compe i o . Coexis ence
is enabled because ‘coloniza ion niches’ a e c ea ed: some pa ches
can only be colonized by one o he compe i o s. In his way
ade-o s wi h a fixed compe i i e hie a chy in ol e a s abiliz-
2234 C. Dislich e al. / Ecological Modelling 221 (2010) 2227–2236
ing ace and acili a e coexis ence o a ce ain deg ee. In con as
o he compe i ion–coloniza ion ade-o (Tilman, 1994; Mulle -
Landau e al., 2008), whe e s able coexis ence o a po en ially
unlimi ed numbe o species can be obse ed, he ade-o s in
physiological ai s conside ed in ou s udy do no bea a s abi-
lizing componen . We do no an icipa e an a p io i hie a chy o
local compe i ion; ins ead fi ness di e ences e ol e di ec ly om
physiological species ai s. Only ew s udies examining ade-o s
men ion he insu ficiency o such ade-o s alone o suppo coex-
is ence (Chesson, 2000; Lischke and Lö fle , 2006; Bani z e al.,
2008).
4.4. Ou look: compe i ion o mul iple species
Up o now, we ha e limi ed ou in es iga ion o he compe-
i ion be ween wo species. In ui i ely one migh ask how ou
esul s ansla e i we expand he model om a wo-species o
a mul i-species compe i ion. Fo a fi s es o he e ec o he
local p ocesses on species di e si y we simula ed a communi y
wi h 196 species, choosing all ai combina ions ha lead o coex-
is ence in a wo-species compe i ion wi h he medium species.
We p ese ed he ini ial condi ions (popula ion size and densi y)
o he wo-species expe imen s by enla ging he simula ed a ea.
While we obse e a apid decline o species numbe o he basic
ade-o communi y (leading o a communi y wi h only one dom-
inan species), bo h local p ocesses, densi y-dependen mo ali y
and ligh -dependen egene a ion, clea ly enhance di e si y in he
communi y (Fig. 5). This p elimina y esul unde pins he ele ance
o p ocesses ha modi y local compe i ion o species di e si y.
4.5. Limi a ions
In he s udy p esen ed he e he dispe sal o seeds is modelled in
a simplified way (simila o Alonso and Solé, 2000): seed dispe sal
dis ances a e en angled wi h seed numbe , since a ee dispe ses
seeds in each pa ch wi hin i s dispe sal ange. In eal o es s one
o en finds ha a -dispe sing species p oduce many seeds and
ha e a highe ecundi y compa ed o species wi h sho dispe sal
dis ances (Wes oby e al., 2002; Mulle -Landau, 2010). To simpli y
he model s uc u e, we connec ed dispe sal dis ance o ecundi y,
which is he eason why dispe sal dis ance plays such an impo an
ole o coexis ence. Gene ally, he pa icula s eng h o species
ai s also depends on he specific implemen a ion o p ocesses.
Fig. 5. Mul i-species compe i ion: all ai combina ions (o dispe sal anges, mo -
ali y and g ow h a es), ha coexis wi h he species wi h medium ai s a e
1000 yea s, a e selec ed. This communi y o 196 species is simula ed o he basic
ade-o model, he model wi h ligh -dependen coloniza ion (LDC) and wi h
densi y-dependen mo ali y (DDM). Fo LDC, a andom ligh equi emen index
be ween 0 and 1 is assigned o each species; o DDM, all species ha e densi y ac o
1. Each g aph shows he a e age species numbe o h ee simula ion uns.
In plan communi ies species o en di e in mo e han wo o
h ee physiological cha ac e is ics and a ade-o be ween mul-
iple a ibu es may gi e ise o a highe coexis ence po en ial.
Howe e , we did no find inc eased coexis ence anges o a h ee-
way ade-o be ween he conside ed ai s (see Appendix Fig. 6).
The dominance o a species is go e ned by he ade-o in he
s ong ai s dispe sal ange and mo ali y a e; inco po a ing an
addi ional di e ence in g ow h a es only shi s he dominance
be ween species bu does no enla ge coexis ence a eas. We would
only expec enhanced coexis ence in a ade-o be ween wo o
mo e weak ai s.
The conside ed ime ame and coexis ence c i e ion a e cho-
sen in a way, which allows obse ing he way and di ec ion, in
which he addi ional mechanisms al e coexis ence anges. Fo he
basic model, i is no c ucial, which limi is chosen o he coex-
is ence c i e ion, since o mos pa ame e combina ions, al eady
a e he conside ed simula ion leng h o 1000 yea s, one o he
wo species is gone ex inc . Thus ou esul s would look e y sim-
ila wi h a chosen coexis ence c i e ion o 1% pa ch occupancy
(ins ead o 10%). When he local p ocesses densi y-dependen mo -
ali y o ligh -dependen egene a ion a e added, he coexis ence
Fig. 6. Va ia ion o wo a ibu es in a wo-species communi y. Each poin is he ou come o a single simula ion un a e 15 000 yea s. Species 1 has medium p ope ies:
dispe sal ange 60 m, mo ali y a e 0.05 y−1, g ow h a e 0.5 m/y. Fi s ow: basic ade-o communi y wi hou local p ocess. Second ow: communi y wi h densi y-dependen
mo ali y (DDM) o bo h species (d= 1). Thi d and ou h ow: communi y wi h ligh -dependen coloniza ion (LDC): species 1 is medium ligh -demanding (ligh equi emen
index L1= 0.5). In ligh -dependen communi y I species 2 is shade- ole an (L2= 1) and in ligh -dependen communi y II species 2 is ligh -demanding (L2= 0). Column 1:
dispe sal-mo ali y ade-o , g ow h a e (species 2) = 0.5 m/y. Column 2: dispe sal-g ow h ade-o , mo ali y a e (species 2) = 0.05 y−1. Column 3: g ow h-mo ali y
ade-o , dispe sal ange (species 2) = 60 m.
C. Dislich e al. / Ecological Modelling 221 (2010) 2227–2236 2235
Fig. 7. Dispe sal-mo ali y ade-o o di e en heigh g ow h a es o species 2 (0.05, 0.5 and 1 m/y). Species 1 has medium p ope ies: dispe sal ange 60 m, mo ali y
a e 0.05 y−1and g ow h a e 0.5 m/y.
anges in gene al become la ge wi h a mo e igo ous coexis ence
c i e ion.
4.6. Syn hesis
In es iga ing he pe o mance o di e en plan species ai s in
ade-o s o acili a e coexis ence and he ole o local p ocesses,
we p o ide insigh s on how species di e si y is main ained in plan
communi ies. Ou h ee majo findings a e:
•The conside ed ade-o s be ween physiological a ibu es alone
canno suppo long- e m coexis ence. The a ibu es ha e no
s abilizing e ec , hus ade-o s ac only equalizing.
•The imescale on which di e en species a ibu es ope a e in
compe i ion can di e ; we he e o e sugges he concep o weak
and s ong ai s. As a consequence, highly non-linea coexis ence
cu es in he ai space esul when a s ong and a weak ai a e
in ol ed.
•The na ow coexis ence anges o ade-o communi ies a e
conside ably b oadened by he inclusion o he local p ocesses
densi y-dependen mo ali y and ligh -dependen coloniza ion.
Thus we ha e shown ha hese local p ocesses can make an
impo an con ibu ion o coexis ence in o es communi ies.
Acknowledgemen s
We would like o hank Tama a Münkemülle o help ul
commen s on d a s o he manusc ip . Valuable commen s o
wo anonymous e iewe s helped o imp o e he manusc ip .
We g a e ully acknowledge he financial suppo o he Ge man
Resea ch Founda ion (DFG, Resea ch Uni 816). This wo k was
kindly suppo ed by he Helmhol z Impulse and Ne wo king Fund
h ough he Helmhol z In e disciplina y G adua e School o En i-
onmen al Resea ch (HIGRADE).
Appendix A.
A.1. Long- e m coexis ence
Obse ing longe imescales (15 000 yea s, Fig. 7) o he wo-
species compe i ion u he emphasizes he s abilizing s eng h o
2236 C. Dislich e al. / Ecological Modelling 221 (2010) 2227–2236
he selec ed local p ocesses: while he e is no coexis ence in he
basic ade-o model, he coexis ence anges wi h addi ional local
mechanisms emain almos unchanged compa ed o he sho e
simula ion ime o 1000 yea s (Fig. 4).
A.2. Th ee-way ade-o
In he main ex we in es iga ed species ha di e only in
one o wo o he h ee ai s conside ed. Does a h ee-way
ade-o p oduce la ge coexis ence anges? S a ing om he
dispe sal-mo ali y ade-o , do we gain la ge coexis ence anges
i addi ionally he g ow h a e o he second species di e s om
ha o he o he species? Fig. 6 shows he ou come o compe i ion
o he dispe sal-mo ali y ade-o wi h di e en g ow h a es o
species 2. Fo high as well as o low g ow h a es o species 2 he
coexis ence anges do no inc ease. As in he wo-way ade-o s
hese poin s o coexis ence a e s ill sensi i e o mino changes in
s ong ai s. Mos no ably, a low g ow h a e o species 2 changes
he posi ion o he coexis ence ange (Fig. 6a). He e, a balancing
ade-o be ween mo ali y and dispe sal is only possible i species
2 has a ela i ely low mo ali y a e (<0.03 y−1).
Re e ences
Adle , F., Mosque a, J., 2000. Is space necessa y? in e e ence compe i ion and limi s
o biodi e si y. Ecology 81, 3226–3232.
Adle , P.B., HilleRisLambe s, J., Le ine, J.M., 2007. A niche o neu ali y. Ecol. Le .
10, 95–104.
Alonso, D., Solé, R.V., 2000. The di game simula o : a s ochas ic cel-
lula au oma a model o ain o es dynamics. Ecol. Model. 133,
131–141, h p://www.sciencedi ec .com/science/a icle/B6VBS-415R3W4-
F/2/ 1a5c8250da5960 1802c33aa46d8e22.
Bani z, T., Hu h, A., G imm, V., Johs , K., 2008. Clumped e sus sca e ed: how does
he spa ial co ela ion o dis u bance e en s a ec biodi e si y? Theo . Ecol. 1,
231–240.
Cha e, J., Mulle -Landau, H.C., Le in, S.A., 2002. Compa ing classical communi y
models: heo e ical consequences o pa e ns o di e si y. Am. Na . 159, 1–23.
Chesson, P., 2000. Mechanisms o main enance o species di e si y. Annu. Re . Ecol.
Sys . 31, 343–366.
Connell, J., 1978. Di e si y in opical ain o es s and co al ee s. Science 199,
1302–1310.
Du e , R., Le in, S., 1994. S ochas ic spa ial models: a use ’s guide o ecological
applica ions. Phil. T ans. R. Soc. Lond. B 343, 329–350.
Du e , R., Le in, S., 1998. Spa ial aspec s o in e specific compe i ion. Theo . Popul.
Biol. 53, 30–43.
Es he , A., G oene eld, J., En igh , N., Mille , B., Lamon , B., Pe y, G., Schu , F., Jel sch,
F., 2008. Assessing he impo ance o seed immig a ion on coexis ence o plan
unc ional ypes in a species- ich ecosys em. Ecol. Model. 213, 402–416.
G ay, L., He, F., 2009. Spa ial poin -pa e n analysis o de ec ing densi y-dependen
compe i ion in a bo eal ch onosequence o albe a. Fo . Ecol. Manage. 259,
98–106.
G imm, V., Be ge , U., Bas iansen, F., Eliassen, S., Gino , V., Giske, J., Goss-
Cus a d, J., G and, T., Heinz, S.K., Huse, G., Hu h, A., Jepsen, J.U., J( gensen,
C., Mooij, W.M., Mülle , B., Pe’e , G., Piou, C., Railsback, S.F., Robbins,
A.M., Robbins, M.M., Rossmani h, E., Rüge , N., S and, E., Souissi, S.,
S illman, R.A., Vab(, R., Visse , U., DeAngelis, D., 2006. A s anda d p o o-
col o desc ibing indi idual-based and agen -based models. Ecol. Model.
198, 115–126, h p://www.sciencedi ec .com/science/a icle/B6VBS-4K606T7-
3/2/1dad6192bec683 32 ce6dee9d665b51.
Holmes, E.E., Wilson, H.B., 1998. Running om ouble: long-dis ance dispe sal and
he compe i i e coexis ence o in e io species. Am. Na . 151, 578–586.
Hubbell, S., 2001. The Unified Neu al Theo y o Biodi e si y and Biogeog aphy.
Monog aphs in Popula ion Biology. P ince on Uni e si y P ess.
Hu chinson, G.E., 1957. Concluding ema ks. Cold Sp ing Ha bo Symp. Quan . Biol.
22, 415–427.
Janzen, D.H., 1970. He bi o es and he numbe o ee species in opical o es s.
Am. Na . 104, 501–528.
John, R., Da a aja, H., Su esh, H., Sukuma , R., 2002. Densi y-dependence in common
ee species in a opical d y o es in mudumalai, sou he n india. J. Veg. Sci. 13,
45–56.
Johs , K., Hu h, A., 2005. Tes ing he in e media e dis u bance hypo hesis: when will
he e be wo peaks o di e si y? Di e s. Dis ib. 11, 11–20.
Ke , B., Riley, M.A., Feldman, M.W., Bohannan, B.J.M., 2002. Local dispe sal p omo es
biodi e si y in a eal-li e game o ock–pape –scisso s. Na u e 418, 171–174.
Kisdi, E., Ge i z, S., 2003. On he coexis ence o pe ennial plan s by he
compe i ion–coloniza ion ade-o . Am. Na . 161, 350–354.
Klausmeie , C.A., Tilman, D., 2002. Spa ial models o compe i ion. In: Somme , U.,
Wo m, B. (Eds.), Compe i ion and Coexis ence. Sp inge , Be lin, pp. 43–78.
Kohyama, T., Suzuki, E., Pa omiha djo, T., Yamada, T., Kubo, T., 2003. T ee species
di e en ia ion in g ow h, ec ui men and allome y in ela ion o maximum
heigh in a bo nean mixed dip e oca p o es . J. Ecol. 91, 797–806.
Lischke, H., Lö fle , T., 2006. In a-specific densi y dependence is equi ed o main-
ain species di e si y in spa io- empo al o es simula ions wi h ep oduc ion.
Ecol. Model. 198, 341–361.
Molo sky, J., Du e , R., Dusho , J., G i ea hand, D., Le in, S., 1999. Local equency
dependence and global coexis ence. Theo . Popul. Biol. 55, 270–282.
Mulle -Landau, H., 2010. The ole ance- ecundi y ade-o and he main enance o
di e si y in seed size. PNAS 107, 4242–4247.
Mulle -Landau, H.C., W igh , S.J., Calde on, O., Condi , R., Hubbell, S.P., 2008. In e -
specific a ia ion in p ima y seed dispe sal in a opical o es . J. Ecol. 96,
653–667.
Münkemülle , T., Bugmann, H., Johs , K., 2009. Hu chinson e isi ed: pa e ns o
densi y egula ion and he coexis ence o s ong compe i o s. J. Theo . Biol. 259,
109–117.
Pe e s, H., 2003. Neighbou - egula ed mo ali y: he influence o posi i e and neg-
a i e densi y dependence on ee popula ions in species- ich opical o es s.
Ecol. Le . 6, 757–765.
Phillips, O.L., Gen y, A.H., 1994. Inc easing u no e h ough ime in opical o es s.
Science 263, 954–958.
Po ho , M., Johs , K., Gu , J., Wissel, C., 2006. Clumped dispe sal and species coex-
is ence. Ecol. Model. 198, 247–254.
P onk, T.E., Du inga, H.J., Schie inga, F., 2007. Coexis ence by empo al pa i ioning
o he a ailable ligh in plan s wi h di e en heigh and lea in es men s. Ecol.
Model. 204, 349–358.
Roxbu gh, S.H., Shea, K., Wilson, J.B., 2004. The in e media e dis u bance hypo hesis:
pa ch dynamics and mechanisms o species coexis ence. Ecology 85, 359–371.
Shuga , H., 2003. A Theo y o Fo es Dynamics—The Ecological Implica ions o Fo es
Succession Models. The Blackbu n P ess.
Tilman, D., 1994. Compe i ion and biodi e si y in spa ially s uc u ed habi a s. Ecol-
ogy 75, 2–16.
Tilman, D., Pacala, S., 1993. The main enance o species ichness in plan com-
muni ies. In: Rickle s, R.E., Schlu e , D. (Eds.), Species Di e si y in Ecological
Communi ies. Uni e si y o Chicago P ess, pp. 13–25 (Ch. 2 The Main enance o
Species Richness in Plan Communi ies).
Wes oby, M., Fals e , D., Moles, A., Vesk, P., W igh , I., 2002. Plan ecological s a e-
gies: some leading dimensions o a ia ion be ween species. Annu. Re . Ecol.
Sys . 33, 125–159.
Whi mo e, T.C., 1998. An In oduc ion o T opical Rain Fo es s. Ox o d Uni e si y
P ess.
Wills, C., Ha ms, K.E., Condi , R., King, D., Thompson, J., He, F.L., Mulle -Landau, H.C.,
Ash on, P., Losos, E., Comi a, L., Hubbell, S., LaF ankie, J., Bunya ejchewin, S.,
Da a aja, H.S., Da ies, S., Esu ali, S., Fos e , R., Guna illeke, N., Guna illeke, S.,
Hall, P., I oh, A., John, R., Ki a ip ayoon, S., de Lao, S.L., Massa, M., Na h, C., Noo ,
M.N.S., Kassim, A.R., Sukuma , R., Su esh, H.S., Sun, I.F., Tan, S., Yamaku a, T., Zim-
me man, E., 2006. Non andom p ocesses main ain di e si y in opical o es s.
Science 311, 527–531.
W igh , S.J., 2002. Plan di e si y in opical o es s: a e iew o mechanisms o
species coexis ence. Oecologia 130, 1–14.
Yu, D., Wilson, H., 2001. The compe i ion–coloniza ion ade-o is dead; long li e
he compe i ion–coloniza ion ade-o . Am. Na . 158, 49–63.
Zhang, J., Hao, Z., Sun, I.-F., Song, B., Ye, J., Li, B.W.X., 2009. Densi y dependence on
ee su i al in an old-g ow h empe a e o es in no heas e n china. Ann. Fo .
Sci. 66, 204–213.
351
C. Dislich e al.: Simula ing o es dynamics o a opical mon ane o es in Sou h Ecuado
2009
wi h dbh abo e 5 cm a e sligh ly unde es ima ed
by he model (Fig. 3c); his is mainly due o he
unde es ima ion in s ems o PFT 5. S em numbe s
o he emaining PFTs a e p edic ed ai ly well.
Also o ees wi h dbh abo e 20 cm we obse e a
good i be ween model and da a (Fig. 3d).
To examine model a ia ion we educe he
simula ed a ea o one hec a e and depic se e al
single model uns (Fig. 4). The e is a no able em-
po al a ia ion o o e all basal a ea wi hin single
model uns (and wi hin sho ime) ha gene ally
i s o he a ia ion we ind be ween plo s. The
mean o e all basal a ea is sligh ly o e es ima ed
by he model compa ed o he basal a ea o he
whole in en o y plo (c . Fig. 3b again). This is
due o a sligh o e es ima ion o he as -g owing
species (Fig. 4b). The o e all empo al a ia ion
is mainly caused by he a ia ion o hese species.
The succession o o e all basal a ea shows a
apid inc ease wi hin he i s 40 yea s and a sligh
o e shoo be o e basal a ea eaches a s able le el,
whe e s ems wi h dbh abo e 5 cm comp ise a ba-
10−20 20−30 30−40 40−50 50−60
DBH (cm)
F equency (1/ha)
1
5
10
50
100
500
1000
a
10−20 30−40
DBH (cm)
F equency (1/ha)
PFT 1
1
5
10
50
100
200
10−20 30−40 50−60
DBH (cm)
F equency (1/ha)
PFT 2
1
5
10
50
100
200
10−20 30−40
DBH (cm)
F equency (1/ha)
PFT 3
1
5
10
50
100
200
10−20 30−40
DBH (cm)
F equency (1/ha)
PFT 4
1
5
10
50
100
200
10−20
DBH (cm)
F equency (1/ha)
PFT 5
1
5
10
50
100
200
10−20 30−40
DBH (cm)
F equency (1/ha)
PFT 6
1
5
10
50
100
200
10−20 30−40
DBH (cm)
F equency (1/ha)
PFT 7
1
5
10
50
100
200
b
Fig. 2: S em size dis ibu ion o he whole s and and o each indi idual PFT. Red do s show obse ed equencies;
ba s show he a e ages o he model aken om di e en poin s in ime o one un, e o ba s show minimum and
maximum equencies occu ing o e ime. No e he loga i hmic scale, which is used o be e display abundances
o la ge ees.

352 Vol. 63 · No. 4
sal a ea a ound 25 m2 pe hec a e (Fig. 5a). On
he le el o single PFTs (Fig. 5b) i akes much
mo e ime o basal a ea o each a s able le el
han o he o e all basal a ea. The o e shoo in
o e all basal a ea in he beginning o simula ion
is caused by he o e shoo in basal a ea o PFT 2.
Only he as -g owing species g oups (PFT 1 and
2) display an excess in basal a ea in he ea ly phase
o succession, he o he g oups app oach hei s a-
ble basal a ea wi h di e en paces. I akes a ound
500 yea s o all species g oups o each a s able
basal a ea.
5 Discussion
In his s udy we applied he o es model
FORMIND o simula e he dynamics o he idge
o es o he Rese a Biológica San F ancisco in
sou he n Ecuado . The s eng h o he indi idual-
based model app oach is ha i allows us o dis-
inguish pa e ns on di e en spa ial and empo al
le els, anging om indi iduals o en i e landscapes.
These di e en pa e ns ha e been in ensi ely in es-
iga ed and compa ed wi h a ailable ield da a om
a ious opical si es (c . e.g. hu h and di ze 2000;
Simula ed basal a ea (m2/ha)
Obse ed basal a ea (m2/ha)
Simula ed basal a ea (m2/ha)
Obse ed basal a ea (m2/ha)
●
●
●
●
●
●
●
0 2.5 5 7.5 10 25
0
2.5
5
7.5
10
25
●
●
●
●
●
●
●
PFT 1
PFT 2
PFT 3
PFT 4
PFT 5
PFT 6
PFT 7
S ems > 5 cm DBH
0 2 4 6 8 10
0
2
4
6
8
10
●
●
●
●
●
●
●
PFT 1
PFT 2
PFT 3
PFT 4
PFT 5
PFT 6
PFT 7
S ems > 20 cm DBH
●
●
●
●
●
●
●
Obse ed equency ((1 ha))
ycneuqe de alumiS ((1 ha))
0 500 1000 2300
0
500
1000
2000
●
●
●
●
●
●
●
PFT 1
PFT 2
PFT 3
PFT 4
PFT 5
PFT 6
PFT 7
S ems > 5 cm DBH
0 50 100 150
0
50
100
150
Obse ed equency ((1 ha))
Simula ed equency ((1 ha))
●
●
●
●
●
●
●
PFT 1
PFT 2
PFT 3
PFT 4
PFT 5
PFT 6
PFT 7
S ems > 20 cm DBH
a b
c d
Fig. 3: Simula ed e sus obse ed basal a ea (a, b) and s em numbe pe hec a e (c, d). T ees abo e 5 cm dbh a e shown wi h
illed symbols, ees abo e 20 cm dbh wi h open symbols.
353
C. Dislich e al.: Simula ing o es dynamics o a opical mon ane o es in Sou h Ecuado
2009
Köhle e al. 2003; üGe e al. 2008). These s ud-
ies ha e shown ha in hose cases FORMIND ac-
cu a ely ep oduces pa e ns on di e en le els o
complexi y.
Ou main esul s indica e ha he model p e-
dic s he main s uc u al pa e ns o he idge o es
obse ed in he ield i.e. he ela i e abundance o
di e en PFTs (Fig. 3), a ia ion o model uns (Fig.
4) and s em size dis ibu ion in he ee communi y
(Fig. 2). In addi ion, we used he model o in es i-
ga e he cou se o succession (Fig. 5); his comp ises
he oppo uni y o u he in es iga ion o he di -
e en cha ac e is ics o succession in ela ion o he
ype o dis u bance. As s a ed in he in oduc ion,
shallow landslides o m one majo sou ce o na u al
dis u bance in ou esea ch a ea ( o Pahl e al., sub-
mi ed). The e sion o FORMIND p esen ed he e
did no explici ly include landslides as a dis u bance,
since he in en o y da a used we e de i ed om plo s
in which he dis u bance o landslides is expec ed o
be o mino impo ance.
Two in e es ing insigh s gained om he model
a e (i) ha we did no obse e a species g oup ha
displayed ypical “pionee ” beha iou and (ii) ha
he e is a high empo al a iabili y in he o e all ba-
sal a ea, which occu s wi hin sho ime anges and
does no subside wi h ime (c . Fig. 4). These wo
pa e ns a e u he discussed in de ail below.
As e iden om igu e 5, we obse e wo main
successional esponses conce ning basal a ea in
ime: (1) species exhibi ing apid g ow h o e shoo
hei s able s a e o basal a ea a he beginning
o succession and educe wi h ime ( as -g owing
PFT 1 and 2), and (2) species eaching hei s able
basal a ea a di e en speeds wi hou o e shoo -
ing (medium and slow-g owing PFT 3–7). The
as -g owing g oups, howe e , do no display he
beha iou o ypical “pionee s”, which show high
abundances in ea ly s ages o succession and a e
la e eplaced by o he “climax” g oups (shuGa
1998). Ins ead, he as es g owing PFT 2, which
domina es he i s phase o succession, e ains he
majo sha e o basal a ea h oughou he succes-
sion. Species o Podoca paceae, (Podoca pus olei olius
is he mos abundan species in PFT 4) a e con-
side ed o o m he climax s age o many na u-
al s ands a ound he Podoca pus Na ional Pa k
(lozano 2002). They can each diame e s o up o
100 cm (Gün e , homeie pe s. obse .; ma ín
elez 1998). Howe e , in ou s udy a ea he maxi-
mum diame e obse ed is only app oxima ely 50
cm. The e a e wo possible explana ions o his:
ei he edaphic condi ions p e en he de elopmen
o Podoca pus olei olius (and hence PFT 4) so ha i
does no become as dominan as in o he p ima y
o es s o he egion, o ha he o es has no ye
eached he climax s age. Howe e , he dominance
o PFT 2 migh also ela e o he he e ogeneous
na u e o he idge o es , whe e ees s ay small in
heigh and dis u bances due o na u al landslides
a e equen . Thus, e en in a ma u e idge o es ,
he e exis abundan loca ions wi h sui able condi-
●
Time (y)
Basal a ea (m2/ha)
Time (y)
Basal a ea (m2/ha)
0 100 200 300 400 500 600 700
0
5
10
15
●●
●●
●
●
o e all species
slow g owing species
0 100 200 300 400 500 600 700
0
5
10
15
●●
●
●
as g owing species
in e media e g owing species
a b
Fig. 4: Succession o basal a ea o ees wi h dbh > 20 cm dbh. Ten model uns o a 1 ha a ea (nine o hem in pale shades)
show a ia ion be ween model uns and luc ua ion wi hin single uns. The illed poin s show da a o he whole in en o y
a ea (4.88 ha), open ci cles show da a om 1 ha subplo s o he in en o y a ea. Fo con enience o illus a ion we agg ega-
ed PFT in o as -g owing species (PFT 1 and 2), species wi h in e media e g ow h a es (PFT 3 and 4) and slow-g owing
species (PFT 5, 6 and 7).
354 Vol. 63 · No. 4
ions o hese a he as -g owing species. Apa
om his, i migh also be he case ha in ou lis
o species (Tab. 2) some species wi h ypical pio-
nee beha iou a e missing, since he da a we u i-
lize was collec ed p ima ily on ees g ea e hen
20 cm in diame e .
As e iden om igu e 4, he model p oduces
a high a iabili y o basal a ea in ime. Figu e 4b
shows ha in ac mos o his a iabili y o igi-
na es om he as -g owing ees, due o hei high
abundance and apid esponse; he slowe g owing
species g oups do no compensa e hese luc ua-
ions. A compa ison o model a ia ion wi h em-
po al a iabili y wi hin si es would equi e long-
e m ield measu emen s ac oss se e al decades.
A p esen such measu emen s a e no a ailable o
ou s udy si e. Bu he ex en o a iabili y be ween
si ed i s o he model a ia ion. This means ha
he a iabili y be ween si es can in ac ep esen
empo al and no only spa ial he e ogenei y.
The s uc u al eali y o he o es model used
co esponds wi h i s ela i ely high numbe o pa-
ame e s. Un o una ely, as is he case o mos ap-
plica ions o a he de ailed models, no all pa am-
e e s can be es ima ed wi h empi ical da a om
he speci ic s udy si e. Pa ame e s o mo ali y,
ec ui men and g ow h a e pa icula ly di icul
o ob ain. To ga he eliable in o ma ion on hese
pa ame e s, one needs o collec da a o e long pe-
iods and o e la ge a eas (whi mo e 1998). We
used a combina ion o da a om he li e a u e,
expe knowledge and calib a ion p ocesses o
de e mine hese pa ame e s (see Tab. 1). The ap-
plied pa ame e alues lie wi hin occu ing anges
o opical o es s (PhilliPs and Gen y 1994).
One would also expec highe mo ali y a es o
as -g owing species compa ed o slow g owing
species since hey exhibi mo e “pionee cha ac-
e is ics” (e.g. lowe wood densi y). Howe e , we
no ed ha he ealized simula ed g ow h is sligh ly
as e han he majo i y o obse ed g ow h. As a
consequence, he calib a ed mo ali y and ec ui -
men a es should be conside ed as p elimina y, in
pa icula o he as -g owing species and species
wi h in e media e g ow h a es (see Tab. 1). In he
ma u e o es , only ew indi iduals come close o
eaching hei es ima ed g ow h po en ial. This
calls o mo e empi ical da a om dis u bed locali-
ies o om expe imen al da a, whe e compe i ion
is emo ed o assess eliable alues o po en ial
diame e g ow h.
The p edic ed speed o accumula ion o he
o e all basal a ea is based on he assump ion ha
he un o es ed si e in he ini ialisa ion is equally
sui able o egene a ion o all PFTs and also e-
c ui men a es a e no limi ed, e.g. due o en i on-
men al cons ain s. As a consequence, he ime un-
il he o es eaches i s ma u e s a e in he model
should be conside ed as a lowe limi . Depending
on di e en ini ial en i onmen al condi ions, one
can expec succession o p oceed di e en ly, mos
p obably mo e slowly; also seed dispe sal limi a-
ion migh in luence ec ui men success on la ge
un o es ed a eas. In he s udy a ea, emp y si es
migh occu as a esul o landslides, i e, log-
ging o pas u e abandonmen ; each o hese e en s
0 50 100 150 200
0
5
10
15
20
25
30
Time (y)
Basal a ea (m2/ha)
Time (y)
Basal a ea (m2/ha)
0 100 200 300 400 500 600
0
5
10
15
PFT 1
PFT 2
PFT 3
PFT 4
PFT 5
PFT 6
PFT 7
a b
Fig. 5: Succession o basal a ea. Lines show he a e age o 10 single model uns wi h simula ed a ea o 4 ha. (a) O e all
basal a ea o all ees abo e 5 cm dbh. (b) Basal a ea abo e 5 cm dbh o each PFT. Please no e he di e en ime scales.
355
C. Dislich e al.: Simula ing o es dynamics o a opical mon ane o es in Sou h Ecuado
2009
p obably esul in e y di e en condi ions o e-
gene a ion conce ning e.g. he size o he dis u bed
a ea o nu ien and myco hiza limi a ions o he
g ound. Following a landslide o example, i akes
se e al yea s be o e ees s a ecolonising a si e
(Bussmann e al. 2008).
Ano he ca ea o igina es om he ac ha
all empi ical da a ha was used o his s udy
was aken om “ma u e” o es si es. Resul s o a
s udy on a 38-yea -old seconda y o es (Gün e
e al. 2007) and obse a ions on a 15-yea -old si e
(homeie , pe s. comm.) con i m ha mos species
in ea ly successional s ages belong o he as -g ow-
ing g oup PFT 2 (pa icula ly Alcho nea g andi lo a,
Alza ea e icilla a, Hie onyma mo i ziana). Howe e ,
on he 38-yea -old si e, also slow-g owing species
as G a en ieda ema gina a (PFT 6) and Pu diaea nu ans
(PFT 7) whe e ound. Thus slow- and as -g owing
species can simul aneously be membe s o he same
successional s age. To in es iga e ansien dynam-
ics o hese o es s, addi ional da a om dis u bed
si es would be use ul in o de o analyse ea ly-s age
succession in mo e de ail.
Ou s udy demons a es ha FORMIND is a
p omising ool o he simula ion o opical mon-
ane o es s dynamics. This model will help o u -
he ou unde s anding o ce ain aspec s o he
complex dynamics o hese highly di e se and ul-
ne able ecosys ems.
6 Pe spec i es
Cu en ly, e o s a e being made o link he
model o landslide dis u bances, which a e one
main cause o na u al dis u bance in ou esea ch
a ea (Bussmann e al. 2008). In he u u e, we in-
end o de elop pa ame e isa ions o he emain-
ing o es ypes: he a ine o es , which di e s
subs an ially om he idge o es in e ms o
s uc u e, g ow h dynamics and species ichness,
and he o es ypes a highe ele a ions. Co e ing
all main o es ypes o he Rese a Biológica will
allow de eloping a model ha simula es o es dy-
namics on a egional scale. Such an in eg a ion o
models o di e en o es ypes o e la ge al i u-
dinal g adien s is a unique exe cise which has no
been a emp ed be o e. The model has a numbe
o po en ial applica ions anging om in es iga -
ing he impac o di e en na u al dis u bances on
o es s uc u e and ee species di e si y, o ana-
lysing di e en po en ial managemen s a egies.
The la e aspec is o g ea impo ance due o he
high p essu e on Andean mon ane o es s and he
need o de elop ecologically sus ainable, economi-
cally a ac i e s a egies as an al e na i e o li e-
s ock a ming.
Acknowledgmen s
We g a e ully acknowledge he inancial sup-
po o he Ge man Resea ch Founda ion (DFG,
Resea ch Uni 816). This wo k was kindly sup-
po ed by he Helmhol z Impulse and Ne wo king
Fund h ough he Helmhol z In e disciplina y
G adua e School o En i onmen al Resea ch
(HIGRADE). The i s au ho would like o hank
Guy Pe’e and Flo ian Ha ig o ui ul discus-
sions on he manusc ip .
356 Vol. 63 · No. 4
Pa ame e Desc ip ion Uni PFT 1
En i onmen al Pa ame e s
kLigh ex inc ion coe icien m2g ound m-2lea
I0A e age i adiance abo e canopy μmol(pho ons) m-2s-1
sdLeng h o daily pho osyn he ic ac i e pe iod h
Rec ui men Pa ame e s
dsDiame e o ing owing ees m
ImMinimum ligh in ensi y o es ablishmen % o I010
NMaximum ec ui men a es o small ees ha-1y-1 160
Mo ali y Pa ame e s
mbBasic mo ali y y-1 0.05
muMaximum mo ali y o small ees y-1
dsDiame e up o which mo ali y is inc eased m
d Minimum diame e o alling ees m
p Falling p obabili y o dying ees %
T ee Geome y Pa ame e s
H1Diame e -heigh ela ionship
H20.56
F1Fo m ac o
F2
C0
C own diame e as unc ion o diame e -cu esC1
C2
gmMaximum diame e g ow h mm y-1 10
dmMaximum diame e cm 40
hmMaximum heigh m 20
c1C own leng h ac o as unc ion o heigh -cu es
L1Lea a ea index pe ee
L2
σF ac ion o s em wood biomass o o al biomass
Biomass P oduc ion Pa ame e s
pmMaximum pho op oduci i y o di e en lg p μmol(CO2)m-2s-1 7
αSlope o ligh esponse cu e o di e en lg p μmol(CO2)
μmol(pho ons)-1 0.2
ρWood densi y o di e en lg p m-3 0.40
gPa ame e o g ow h espi a ion
G3
Pa ame e o maximum diame e g ow h cu e
-0.02344
G2-0.12500
G10.03458
G00.00767
mT ansmission coe icien o lea es
φPa ame e o con e sion in o ganic d y ma e μmol(CO2)-1
Technical Pa ame e s
aPa ch size m²
ΔhS ep wid h o e ical disc e iza ion m
Table 1: Pa ame e s o FORMIND o he idge o es o he opical mon ane ain o es o he Rese a Biológica San
F ancisco, Sou h Ecuado .

357
C. Dislich e al.: Simula ing o es dynamics o a opical mon ane o es in Sou h Ecuado
2009
PFT 2 PFT 3 PFT 4 PFT 5 PFT 6 PFT 7 Re e ence
0.6 es ima ed
700 Bendix e al. 2008
12 üGe 2007
0.01 echnical pa ame e
10 5 5 1 1 1 üGe 2006
300 150 50 200 280 50 calib a ed
0.09 0.05 0.05 0.006 0.016 0.008 calib a ed
0.1 üGe 2007
0.1 üGe 2007
0.1 es ima ed
20 es ima ed
2.5 de i ed om in en o y da a
0.54 0.56 0.59 0.55 0.56 0.53
0.77 calib a ed
-0.18
0.2 es ima ed
0.1
0.5
20 6 6 2 2 2 S. Gün e and J. homeie pe s.
Comm.
70 40 50 25 40 50 de i ed om in en o y da a
25 20 25 15 20 20 de i ed om in en o y da a
0.25 üGe 2007
2.2 es ima ed
0.1
0.6 es ima ed (nenninGe 2006)
7 5 5 3 3 3 es ima ed
0.2 0.25 0.25 0.4 0.4 0.4 üGe 2006
0.40 0.57 0.57 0.60 0.60 0.60 es ima ed (nenninGe 2006)
0.2 yan 1991
0.36581 0.27328 -0.03537 0.42989 -0.07212 -0.03673
calib a ed
-0.48239 -0.21630 -0.02452 -0.18816 0.01995 0.01261
0.14906 0.03076 0.01112 0.01260 -0.00147 -0.00114
0.00656 0.00481 0.00499 0.00189 0.00201 0.00201
0.1 la Che 2001
0.63 *44e(-12) la Che 2001
400 echnical pa ame e
0.5 echnical pa ame e
358 Vol. 63 · No. 4
G oup Maximum
diame e a b eas
heigh (cm)
Maximum annual
diame e g ow h
(mm/yea )
Species
PFT 1 40 10 Hie onyma aspe i olia Pax & K. Ho m.
Hie onyma duquei Cua ec.
My cia sp. no .
Oco ea aciphylla (Nees) Mez
Vismia c . omen osa Ruiz & Pa .
PFT 2 70 20 Alcho nea g andi lo a Müll. A g.
Alza ea e icilla a Ruiz & Pa .
Cle h a e olu a (Ruiz & Pa .) Sp eng.
Clusia c . ducuoides Engl.
Hie onyma mo i ziana (Müll. A g.) Pax & K. Ho m.
Nec and a linea i olia (Ruiz & Pa .) Mez
Pe sea e uginea Kun h
Pe sea sp.4
Pe sea sp.5
Tapi i a guianensis Aubl.
PFT 3 40 6 Aba ema killipii (B i on & Rose ex B i on & Killip)
Ba neby & J.W. G imes
Aniba muca (Ruiz & Pa .) Mez
Calyp an hes c . pulchella DC.
Elaeagia pas oense L.E. Mo a
Elaeagia u ilis (Goudo ) Wedd.
Endliche ia g iseo-se icea Chande bali
Eugenia sp.
Hedyosmum goudo ianum Solms
Ilex hippoc a eoides Kun h
Inga s ia a Ben h.
Ladenbe gia s enoca pa (Lamb.) Klo zsch
Lau aceae sp.
Ma ayba inelegans Sp uce ex Radlk.
Nec and a memb anaceae (Sw.) G iseb.
Ruagea glab a T iana & Planch
PFT 4 50 6 Ficus cua ecasana Dugand
Meliosma sp.
Mic opholis guyanensis (A. DC.) Pie e
My sine co iaceae (Sw.) R. B . ex Roem. & Schul .
Naucleopsis ancisci Be g & Homeie (ined.)
Podoca pus olei olius D. Don ex Lamb.
PFT 5 25 2 Alcho nea ipline ia (Sp eng.) Müll. A g.
Aniba sp.
Clusia sp. 1
Eschweile a sessilis A.C. Sm.
Fa amea coe ulescens K. Schum. & K. K ause
Gua e ia sp. 1
Hedyosmum anisodo um Todzia
Miconia c . calophylla T iana
Table 2: G ouping o ee species in o plan unc ional ypes. Common species a e p in ed in bold ype.
359
C. Dislich e al.: Simula ing o es dynamics o a opical mon ane o es in Sou h Ecuado
2009
Miconia ini olia Naudin
Miconia heaezans (Bonpl.) Cogn.
Oco ea sp.1
Pe sea a eola ocos ae (Allen) an de We
Pe sea subco da a (Ruiz & Pa .) Nees
Pe sea webe baue i Mez
Sche le a sp.
Siphoneugena sp. 1
Sloanea sp. 1
Weinmannia c . sp.1
Weinmannia ellip ica Kun h
Weinmannia haenkeana Engl.
Weinmannia so bi olia Kun h
PFT 6 40 2 Ch ysophyllum lana um T.D. Penn.
G a en ieda ema gina a (Ruiz & Pa .) T iana
Ilex c . ambo oica Loes.
Lica ia subsessilis an de We
Me iania anciscana Ulloa & Homeie
Miconia punc a a (Des .) D. Don ex DC.
Oco ea ben hamiana Mez
Oco ea sp.2
Oco ea sp.4
Roupala mon ana Aubl.
PFT 7 50 2 Endliche ia o eocola Chande bali
Nec and a subbulla a Rohwe
Pu diaea nu ans Planch.
S ilpnophyllum oellgaa dii L. Ande sson
360 Vol. 63 · No. 4
Appendix
We e o mula ed pa s o he model s uc u e
compa ed o p e ious applica ions o FORMIND
(e.g. Appendix in G imm e al. 2006; Köhle 2000).
We use powe laws o he diame e -heigh ela ions
as well as o he o m ac o ; as a esul o hese
modi ica ions, he biomass is also a powe law o
he diame e . Hence we can di ec ly ecalcula e he
main enance espi a ion om he maximal g ow h
cu e and he biomass (see below). P e iously, hese
calcula ions whe e pe o med using look-up a-
bles, which is mo e ime-consuming han a di ec
calcula ion.
T ee geome y
T ee heigh h is calcula ed as a powe law om he
diame e a b eas heigh d
h(d) = H1 · d H2.
C own leng h c1 is a cons an ac ion o ee heigh
c1(h) = C· h.
C own diame e cd is calcula ed as
cd(d) = C0 · d + C1 · exp(-C2 · d).
C own a ea ca is calcula ed as
The o m ac o is he co ec ion ac o o he de-
ia ion o s em o m om a cylind ical shape. I is
calcula ed as a powe law
(d) = F1· d F2.
(Fo his s udy, we calib a ed F1 and F2, such ha
he esul ing ee biomass sui s obse ed diame e -
biomass ela ionships (nenninGe 2006; Cha e e
al. 2005))
Abo eg ound ee biomass b is a cen al a iable o
he model; i is connec ed o d, he diame e a b eas
heigh , ia he equa ion
whe e is he o m ac o , ρ is he wood densi y and
σ he ac ion o s em wood biomass o o al abo e-
g ound ee biomass. Using he abo e s a ed ela-
ions we a i e a
(1)
Submodels o FORMIND
Wi hin one yea he ou submodels – es ablishmen ,
mo ali y, ecalcula ion o ligh clima e and ee
g ow h – a e applied in he ollowing o de .
Es ablishmen
I he i adiance on he o es loo in a pa ch ex-
ceeds he minimum ligh Im o es ablishmen o a
PFT, a new coho o small ees wi h dbh=1 cm
es ablishes. The numbe o ec ui s pe hec a e is
calcula ed as he maximum numbe o ec ui s pe
hec a e Nm di ided by he numbe o pa ches pe
hec a e (25). Addi ionally, i is checked ha he laye
o seedling c owns is no comple ely c owded p io
o es ablishmen .
Mo ali y
The e a e di e en sou ces o mo ali y:
1. No mal mo ali y: each species g oup (PFT) has
a speci ic basic mo ali y a e mb
2. Mo ali y o small ees: ees wi h diame e d<ds
a e a ec ed by an addi ional size dependen mo -
ali y ms
whe e mu is he maximum size- dependen mo -
ali y o small ees. (Fo coho s wi h less han
100 indi iduals o diame e d ≥ 10 cm, mo ali y
is s ochas ically de e mined o each ee o he
coho . O he wise, he numbe o dying ees is
calcula ed de e minis ically.)
3. Sel - hinning: i heigh laye s in a pa ch a e o e -
c owded wi h ee c owns, i.e. c own a ea ex-
ceeds pa ch a ea, mo ali y o ees wi h c owns
in hese laye s is inc eased due o compe i ion
o space. T ees a e andomly emo ed un il ee
c owns i in o he pa ch (c own a ea ≤ pa ch
a ea).
4. Gap building: La ge alling ees kill a p opo -
ion o he ees in he pa ch whe e hei c own
hi s he g ound. When a ee wi h diame e d > d
dies, i alls wi h p obabili y p . The alling di ec-
ion is de e mined andomly and he p obabili y
ha a ee in he a ge pa ch is killed is p opo -
ional o he a io be ween he c own p ojec ion
Impac o shallow landslides on o es s uc u e in opical mon ane o es s
Claudia Dislicha,b,∗
, And eas Hu ha,c,∗∗
aHelmol z Cen e o En i onmen al Resea ch UFZ Leipzig
Depa men o Ecological Modelling
P. O. Box 500136, 04301 Leipzig, Ge many
bBiogeog aphical Modelling, Uni e si y o Bay eu h
Uni e si ä ss aße 30, 95440 Bay eu h, Ge many
cFacul y o Ma hema ics and In o ma ics, Uni e si y o Osnab ück
Ba ba as aße 12, 49076 Osnab ück, Ge many
Abs ac
Shallow landslides a e a main cause o na u al ecosys em dis u bance in opical mon ane o es s. Due o landslides,
ege a ion and o en also he uppe soil laye a e emo ed, lea ing space o a p ima y succession unde al e ed en i-
onmen al condi ions. We u ilize a p ocess-based o es simula ion model and de elop possible scena ios o in es iga e,
how changes in di e en li e his o y ai s o ees in luence o es eco e y on landslide si es. We hen apply he model
o an e e g een opical mon ane o es in Sou he n Ecuado . Fo all eg ow h scena ios, i akes a leas 200 yea s
un il he pos -landslide o es eaches i s ma u e s uc u e. On he local scale o landslides o es p oduc i i y is educed
conside ably o mos eg ow h scena ios. Landslides p oduce dis inc spa io- empo al a ia ion in ee biomass wi hin
he i s decades o eco e y, which could possibly be compa ed o emo ely sensed da a. On he landscape le el o e all
ee biomass is educed by 13% due o landslide dis u bances, o es p oduc i i y is only sligh ly educed (∼5%). The
e ec o landslides on landscape he e ogenei y is p onounced: hey c ea e a mosaic o o es pa ches o di e en ages,
simila o he well-s udied gap-building p ocess bu on a la ge spa ial scale. Landslides p oduce ho spo s o biomass
loss and po en ially also o o es p oduc i i y.
Keywo ds: o es model, landslide, o es dynamics, opical mon ane o es , FORMIND
1. In oduc ion
Landslides a e a majo cause o na u al ecosys em dis-
u bance in opical mon ane o es s (Res epo e al.,
2009). Vege a ion and o en he uppe soil laye s a e e-
mo ed om he slide su ace, lea ing space o a p ima y
succession (c . Figu e 1). Landslides al e en i onmen al
condi ions on he slide su ace conside ably: pho oac i e
adia ion is inc eased (Mys e and Fe nandez, 1995), soils
migh be uns able (Walke and Shiels, 2008) and due o
he loss o he o ganic soil laye he soil nu ien con en is
educed e en many yea s a e he slide e en (e.g. Za in
and Johnson, 1995; Wilcke e al., 2003). These changed
condi ions can a ec di e en li e his o y ai s o ees
and he e o e in luence o es eco e y on landslide si es
(c . Figu e 2).
Se e al obse a ional s udies o ege a ion eco e y on
landslides si es ha e ocused on he i s yea s o succession
(e.g. Ohl and Bussmann, 2004; Velazquez and Gomez-Sal,
2008; Shiels e al., 2008). These s udies ound a high
∗P incipal Co esponding Au ho
∗∗Co esponding Au ho
Email add esses: clau[email p o ec ed], Phone:+49
(0)341-235-1707 (Claudia Dislich), [email p o ec ed]
(And eas Hu h)
a iabili y in species assemblage and spa io- empo al
pa hways o egene a ion, ha a e likely in luenced by
small scale e osion and sca e ed subs a es and a e hus
ha d o p edic . S udies in es iga ing long e m e ec s
o landslides o en u ilize ch onosequences o landslides,
assuming space- o - ime subs i u ion (Picke , 1989). In-
es iga ed ecosys em a ibu es a e o example empo al
changes in biomass (Reddy and Singh, 1993; Res epo
e al., 2003), species and s uc u al di e si y (Dalling,
1994; Elias and Dias, 2009) and soil nu ien s (Za in
and Johnson, 1995; F izano e al., 2002; Wilcke e al.,
2003). Es ima ed eco e y ime o di e en a ibu es
a ies conside ably - while soil nu ien s a e assumed o
eco e wi hin se e al decades (Res epo e al., 2009),
es o a ion o biomass can ake mo e han 100 yea s
(Dalling, 1994; Res epo e al., 2003). Di e en gene al
mechanisms a e sugges ed o in luence he pa hways o
egene a ion, including acili a ion and inhibi ion (S e n,
1995; Kessle , 1999; Walke and del Mo al, 2003; Walke
e al., 2009), as well as eedbacks o ege a ion bo h
on on abio ic ac o s and on ege a ion de elopmen
i sel . Nu i ion expe imen s suppo he hypo hesis,
ha nu ien limi a ion (mos ly N limi a ion) is a key
limi ing ac o o plan g ow h on landslide si es, bu hey
mos ly ocus on ew s udy species and sho e m e ec s
P ep in submi ed o Else ie Ma ch 22, 2011

Figu e 1: Moun ain idge wi h se e al aces o shallow landslides (le ) and ace o a ecen landslide (middle). Bo h pic u es a e aken
in ou s udy si e, he Rese a Biológica San F ancisco in Sou he n Ecuado . Righ : Visualiza ion o he FORMIND model wi h a ecen
landslide; di e en colo s ep esen di e en species g oups.
(e.g. Dalling and Tanne , 1995; Fe che e al., 1996).
Mos empi ical s udies ocus on a local scale, i.e. he
scale o single landslides, and in es iga e di e en zones
wi hin landslides (e.g. Wilcke e al., 2003; Velazquez and
Gomez-Sal, 2008). Bu landslides a e also an impo an
phenomenon o examine on he landscape scale whe e
hey pose a eoccu ing dis u bance ha in luences
o es dynamics and p oduces a pa chy dis ibu ion
o di e en aged si es. Howe e , in es iga ing a high
numbe o di e en aged landslide si es is di icul since
old landslides a e unde ec able on ae ial pho og aphs
and ha d o ind and access wi hin closed o es s in
complex e ains. Remo e sensing echniques o e new
possibili ies o in es iga e high numbe s o landslides and
p o ide ools o add ess ques ions abou landslide e ec s
on he landscape le el (Da is e al., 2004; Lin e al., 2004).
Howe e , ou knowledge abou eco e y p ocesses
o o es on olde landslide si es s ill emains limi ed
and in ol es unce ain, possibly in e ac ing, pa ame-
e s. In his s udy, we he e o e p opose a modelling
amewo k o in es iga e dynamics o landslide-a ec ed
o es s. P ocess-based o es models a e sui able ools o
in es iga e changes o o es dynamics a e na u al o
an h opogenic dis u bances (Shuga , 1998). Pa icula ly
indi idual based spa ially explici models ha e he ad-
an age, ha one can obse e dynamics on small spa ial
scales, such as he su ace o a single landslide, as well
as on he la ge spa ial scales o he landscape le el. We
u ilize he indi idual based, spa ially explici simula ion
model FORMIND o in es iga e he in luence o landslides
on o es s uc u e, succession and abo eg ound ca bon
cycle. The model simula es o es g ow h and has been
used o analyze di e en ypes o dis u bances in a ious
opical o es s si es (e.g. Köhle e al., 2003; Rüge e al.,
2008; G oene eld e al., 2009). The p ocess-based design
o FORMIND allows o change single li e his o y ai s
and analyze hei impac on pos -landslide succession.
In his s udy we in es iga e he e ec s o landslides on
o es dynamics in an e e g een mon ane o es in sou he n
Bio ic ac os
loss o seedbank
seed dispe sal
seed p eda ion/pa hogens
absence o myco hiza ungi
es ablishmen o dense
ege a ion (sh ubs, e ns)
Fo es egene a ion
ec ui men o ees
ee g ow h
ee mo ali y
Abio ic ac o s
high sola adia ion
low nu ien a ailabili y
soil ins abili y
soil mois u e
empe a u e
exposi ion o wind
Landslide
Figu e 2: O e iew o abio ic and bio ic ac o s ha a e in luenced
by landslides and po en ially a ec o es egene a ion.
Ecuado . Since empi ical knowledge on o es egene a ion
in ou s udy egion is sca ce, we de elop di e en scena ios
o o es eg ow h wi h changed li e his o y ai s o ees.
This s udy is di ided in o wo pa s: i s we concen a e
on he local scale o he landslide su ace and p ocesses
ha in luence o es eco e y and o es s uc u e du ing
succession: we analyze he spa io- empo al o es eco e y
p ocess a he landslide si e and compa e he ca bon bal-
ances o di e en scena ios. In he second pa we upscale
o he landscape le el - he e we in es iga e he impac o
landslides on spa ial he e ogenei y and ca bon budge o
he o es and compa e he dis u bance egime o land-
slides wi h he mo e equen bu less se e e dis u bance
egime o gap-building (due o alling ees).
2
2. Me hods
2.1. S udy si e
Ou s udy a ea is he Rese a Biológica San F ancisco
(RBSF), pa o he biosphe e ese e Podoca pus - El
Condo , loca ed on he eas e n slopes o he Andes in
sou he n Ecuado wi hin one o he wo ldwide ho spo s
o biodi e si y (B ummi and Lughadha, 2003). The o -
es ese e s e ches om 1800 up o 3200 me e abo e
sea le el (asl) and is cha ac e ized by s eep slopes (on a -
e age 40◦) and deeply incised alleys. The RBSF is e y
ich in ee species, wi h mo e han 280 species iden i ied
so a in he 1000 ha a ea (Homeie and We ne , 2007).
The o es can be classi ied as e e g een mon ane o es
and di ided in o ou o es ypes wi h dis inc s uc u e
and species composi ion (Homeie e al., 2008). Be ween
1900 and 2100 m asl we dis inguish idge and a ine o -
es . The idge o es has lowe basal a ea and also a lowe
canopy heigh (15 −20 m) bu highe ee densi y com-
pa ed o he a ine o es , whe e he canopy eaches up o
35 me e . Wi h app oxima ely 70 ee species, he idge
o es is no as species- ich as he a ine o es . The wo
o he o es ypes a e loca ed a highe ele a ions; in hese
si es canopy heigh and species ichness dec ease (Homeie
e al., 2008; Mose e al., 2008). In his s udy we ocus on
he idge o es (1900-2100 m asl).
Wi hin he RBSF, shallow landslides a e a main sou ce
o na u al dis u bance (Bussmann e al., 2008); app ox-
ima ely 2.6% o he a ea is co e ed wi h isible aces
o landslides. Mos slides a e shallow ansla ional slides
(S oyan, 2000); in some slide e en s only he abo eg ound
ege a ion is emo ed - hese slides can be e med ege-
a ion slides (Vo pahl e al., submi ed). Landslide e en s
al e he o es s uc u e in a d ama ic way - usually all
ege a ion on op o he landslide su ace is emo ed (see
Figu e 1). Na ow bands o ege a ion slip downwa ds
and lea e ba e a eas o app oxima ely 10-30 m wid h and
20-100 m leng h.
2.2. The FORMIND model
To s udy he in luence o landslides on o es dynam-
ics we u ilize he indi idual based, spa ially explici o es
g ow h model FORMIND (c . Figu e 1 igh ). In o de o
handle he high di e si y o ee species in opical o es s
(many o hem ex emely a e), ee species wi hin he
model a e g ouped in o plan unc ional ypes (PFT) ac-
co ding o physiological a ibu es like maximum a ain-
able diame e . All ees wi hin small pa ches, he ea e
called plo s (we use a plo size o 20 x 20 me e ), compe e
o ligh and space; ee g ow h is ealized on he basis o
ca bon balance acco ding o pho osyn hesis and espi a-
ion a es. The main p ocesses included in he model a e
he es ablishmen o young ees, ee mo ali y and ee
g ow h. Dying ees can all o e and damage o he ees;
we e e o his p ocess as gap-building. A de ailed de-
sc ip ion o he cu en model e sion used o his s udy
can be ound in Dislich e al. (2009).
As a i s s ep owa ds his s udy, FORMIND has been
pa ame e ized o he idge o es o he RBSF. In his
de ailed pa ame iza ion, 70 ee species o he idge o -
es we e g ouped in o se en plan unc ional ypes ( o
de ails see Dislich e al., 2009). Fo his s udy, we sub-
sumed hese se en g oups in o h ee: pionee ( as g ow-
ing species: 10-20 mm maximal annual diame e g ow h),
mid-successional (6 mm maximal g ow h) and climax (slow
g owing: 2 mm maximal g ow h) species. We sligh ly
adap ed he p e ious e sion o he model and included
landslides as a special ype o dis u bance in o he model
(see sec ion 2.5 and Appendix A).
2.3. Di e en pa hways o egene a ion a e landslides
Accoun ing o sca ci y o da a on p ocesses and pa-
ame e s a ec ing o es dynamics ollowing landslides in
ou s udy a ea, we compiled esul s om o he s udies o
de elop di e en possible scena ios o o es eg ow h ac-
coun ing o changed en i onmen al condi ions which a e
desc ibed in mo e de ail below.
The i s scena io se es as a e e ence scena io, whe e
he only e ec o landslides is he emo al o ees wi h
no addi ional e ec on li e his o y ai s. In he emaining
scena ios he new es ablishing ees expe ience changes in
li e his o y ai s due o landslides ( educed g ow h, e-
duced ec ui men , inc eased mo ali y). We implemen ed
changes in li e his o y ai s o ees wi h a eedback mech-
anism, which con ols he magni ude o changes in li e his-
o y ai s depending on he amoun o al eady es ablished
ege a ion. This eedback mechanism is inspi ed by he
gene al idea o indi ec acili a ion on he communi y le el
(Walke and del Mo al, 2003): he mo e he al eady es-
ablished communi y esembles he p e-dis u bance com-
muni y, he mo e adequa e a e he condi ions o o es
eco e y, i.e. ees expe ience less disad an ages. Em-
pi ical con i ma ion o such eedback mechanisms exis s
o nu ien accumula ion: Reddy and Singh (1993) ound
ha he accumula ion o soil nu ien s on landslides si es
in he Himalaya p oceeds in a non-linea way and is in e -
dependen wi h he eco e y p ocess o ege a ion. The
spa ial le el a which we conside changes in li e his o y
ai s a e plo s o 20 x 20 m size. We aim a a simple and
compa able o mula ion o he scena ios.
Scena io 1: undis u bed eg ow h
In his i s scena io we assume ha a e a landslide all
li e his o y ai s o ees a e as in he undis u bed o es .
Consequen ly, he only en i onmen al change due o land-
slides ha is sensed by he ees is he inc ease in ligh
le els.
Scena io 2: educed g ow h (due o nu ien lim-
i a ion)
I is a well es ablished ac in landslide esea ch ha nu-
ien limi a ion causes slow o es eg ow h on landslides
si es (e.g. Dalling and Tanne , 1995; Za in and Johnson,
1995; Singh e al., 2001; Shiels e al., 2008). Expe imen s
in a Pue o Rican o es ound an inc ease in biomass o
3
ee seedlings a e nu ien addi ion in he o de o magni-
udes (Fe che e al., 1996), and Chaudh y e al. (1996) e-
po up o 90% g ow h educ ion o plan ed ee seedlings
on a young landslide compa ed o g ow h in undis u bed
himalayan o es . On he o he hand, ege a ion is likely
o eed-back on he cou se o nu ien accumula ion on
slide su aces ia decomposi ion o li e and dead wood,
and impeding u he leaching o soils by educing e osion.
We he e o e de elop a scena io o educed ee g ow h
whe e he amoun o g ow h educ ion depends on he so
a accumula ed dead biomass on he slide su ace.
T ee g ow h is exp essed in biomass inc emen pe yea .
We assume a 90 % educ ion o g ow h in he beginning o
succession and educed g ow h un il he accumula ed dead
biomass (bdead) on he plo equals he minimum s anding
biomass in a ma u e plo (bma ). The e o e, he educed
biomass inc emen (binc ed) is calcula ed om he biomass
inc emen unde undis u bed g ow h (binc) as
binc ed =0.9·bdead
bma
+ 0.1
| {z }
educ ion ac o
·binc.
Scena io 3: educed ec ui men (due o hicke
o ming ege a ion)
The i s ege a ion ha es ablishes on landslide su aces
a e mosses, lichens and g asses. I has been obse ed ha
subsequen ly a co e o a hicke -like ege a ion o e ns
can es ablish on landslides (Gua igua a, 1990; Walke ,
1994; Russell e al., 1998). Such dense ege a ion migh
inhibi he es ablishmen o ees. The i s ees es ablish-
ing wi hin his hicke p oduce shade, ha slowly esul s
in a dieback o e ns and li le by li le gene a es mo e
sui able condi ions o ee ec ui men .
We he e o e de elop a scena io wi h educed ec ui -
men o ees whe e he amoun o educ ion depends on
he biomass o al eady es ablished ees on he slide.
The model uses ixed ec ui men a es o ee seedlings
(ing ow h) pe hec a e and yea and plan unc ional ype.
Fo consis ency be ween scena ios, we assume a 90% e-
duc ion o ec ui men a es in he beginning o succession
and educed ec ui men un il he s anding biomass bon
he plo equals he minimum s anding biomass in a ma u e
plo . The e o e he educed ec ui men a e ing ow h ed
is calcula ed as
ing ow h ed =0.9·b
bma
+ 0.1
| {z }
educ ion ac o
·ing ow h.
Scena io 4: inc eased mo ali y (due o ins able
physical condi ions)
Di e en ac o s migh accoun o an inc eased mo ali y
on young landslides: ins abili y o soil, ampli ied changes
in soil mois u e and empe a u e, exposi ion o wind, p e-
da ion and pa hogenes. We assume ha hese h ea s di-
minish wi h eg owing ee biomass.
We he e o e de elop a scena io wi h inc eased mo ali y
a es o ees whe e he amoun o mo ali y inc emen
depends on he biomass al eady es ablished on he slide.
In he o es model he e a e di e en sou ces o mo al-
i y (size-, densi y- and PFT-speci ic), which sum up o he
o e all mo ali y a e mo . We assume a 90% inc emen
o his o e all mo ali y in he beginning o succession and
inc eased mo ali y un il he s anding biomass bon he
plo equals he minimum s anding biomass in a ma u e
plo .
The e o e he inc eased mo ali y a e mo inc is calcu-
la ed as
mo inc =




1 + 1−0.9·b
bma
+ 0.1
| {z }
inc emen ac o





·mo
=1.9−0.9·b
bma ·mo .
Combina ion o Scena io 2 and Scena io 3
In addi ion o he abo e desc ibed scena ios, whe e only
one li e his o y ai is changed a a ime, we es a com-
bina ion o educed g ow h and educed ec ui men o
ees; bo h ai s a e educed by 90 % in he beginning o
succession.
2.4. Time lag o egene a ion a e landslides
The o es model conside s ees wi h s em diame e >
1 cm a b eas heigh . Since a e a landslide ees a e
emo ed om he su ace, he e is a ime lag be ween slide
occu ence and he ime un il he i s ees each he size
o 1 cm diame e a b eas heigh . Based on he po en ial
g ow h o ees in he undis u bed o es (S. Gün e , pe s.
comm.), we es ima e hese ime lags as 3 yea s o he as
g owing species, 5 yea s o in e media e g owing species
and 12 yea s o he slow g owing species. This is only
a ough es ima ion, since a ia ion in g ow h dynamics o
seedlings is high, bu mode a e changes in hese ime lags
do no ha e a s ong e ec on ou esul s. The same ime
lags a e applied o all scena ios.
2.5. Implemen a ion o landslides in o FORMIND
Fo he i s pa o his s udy, we assume ha he whole
simula ion a ea (1 hec a e) is a ec ed by a landslide. In
he second pa we in es iga e dynamics on he landscape
scale, whe e landslides a e a eoccu ing dis u bance ha
a ec s only small pa s o he simula ed a ea (c . Figu e 1
igh ). To es ima e landslide equency and sizes, we u i-
lize ae ial pho og aphs o ou s udy a ea, which ha e been
p ocessed and e alua ed by S oyan (2000). In he pe iod
be ween 1989 and 1998, 183 landslides ha e been obse ed;
his esul s in a dis u bance a e o app oxima ely 0.02
slides pe hec a e and yea . The size dis ibu ion o land-
slides was de i ed om he 1998 ae ial pho og aph using
4
A cGIS; slide sizes ange om 80 o 8317 m2, wi h an
a e age slide size o 1120 m2.
We implemen landslides on he landscape le el in o
FORMIND in he ollowing way: e e y yea a andomly
d awn numbe de e mines, i a landslide occu s. The land-
slide size is d awn om he size dis ibu ion o landslides,
ounded o he model plo size o 20 x 20 me e . The
di ec ion o landslides is always he same, he s a ing lo-
ca ion o he landslide is a bi a y. Neighbo ing plo s a e
a ec ed un il he slide eaches he p ede e mined size. To
a oid edge e ec s he landscape is modelled as a o us.
Fo es eco e y hen p oceeds acco ding o one o he sce-
na ios desc ibed abo e.
3. Resul s
3.1. Landslide le el
A i s we ocus on he succession o ees on he slide
su ace and compa e di e en scena ios o o es eg ow h.
Figu e 3 (le column) depic s he buildup o biomass
di ided in o pionee , mid-successional and climax ee
species. The landslide emo es he comple e ee biomass
om he slide su ace and o all scena ios he subsequen
accumula ion o new ee biomass is a i s domina ed by
he as g owing pionee species be o e mid-successional
species and e en ually also climax species inc ease in bio-
mass. The cou se o biomass eco e y di e s be ween he
scena ios; in he i s scena io, whe e ees do no su e
any d awback due o changed en i onmen al condi ions,
we obse e a apid eg ow h o ee biomass wi hin he
i s 30 yea s o succession. In all emaining scena ios,
succession is slowed down. I ee g ow h is educed (sce-
na io 2), ee biomass is e y low wi hin he i s 40 yea s
a e he landslide, ollowed by a simila s eep inc ease in
biomass as in he undis u bed eg ow h scena io. Fo sce-
na io 3 and 4 ( educed es ablishmen and inc eased mo -
ali y) he inc emen in he buildup phase is less s eep.
In all scena ios, a ee biomass o app oxima ely 100 ons
pe hec a e has es ablished a e 100 yea s. Fo he com-
bina ion o scena io 2 and 3 he eco e y o ee biomass
is s ongly delayed and se s in only a ound 180 yea s a e
he landslide e en .
All scena ios p oduce a dis inc pa e n o spa ial a ia-
ion in biomass wi hin he i s decades o succession (Fig-
u e 3, second column). While in he i s scena io spa ial
a ia ion is almos he same as in ma u e o es , he spa-
ial a ia ion is lowe han in he ma u e o es , i g ow h
a es o ees a e educed (scena io 2). Spa ial a ia ion
in biomass inc eases s ongly i ec ui men on he slide
su ace is educed (scena io 3), i.e. ege a ion is mo e
pa chy; also o inc eased mo ali y (scena io 4), spa ial
a ia ion in biomass inc eases, bu o a lesse ex en and
in a humped-shaped o m. A much s onge inc ease in
spa ial he e ogenei y o biomass is p oduced in he com-
bined scena io o educed g ow h and educed ec ui men ;
he e i akes mo e han 300 yea s un il a ia ion e u ns o
Time [y]
]%[ a o eoC
0 20 40 60 80 100
0
0.1
0.2
0.3
0.4
0.5
0.6
Time [y]
]ah/ [ ssamoiB
0 20 40 60 80 100
0
20
40
60
80
100
120
140
Scena io 3:
educed ec ui men
Time [y]
]ah/ [ ssamoiB
0 20 40 60 80 100
0
20
40
60
80
100
120
140
Scena io 4:
inc eased mo ali y
Time [y]
]ah/ [ ssamoiB
0 60 120 180 240 300
0
20
40
60
80
100
120
140
Combined e ec s:
educed g ow h and ec ui men
Time [y]
]ah/ [ ssamoiB
0 20 40 60 80 100
0
20
40
60
80
100
120
140
Scena io 2:
educed g ow h
Time [y]
]ah/ [ ssamoiB
0 20 40 60 80 100
0
20
40
60
80
100
120
140 Pionee
In e media e
Climax
Scena io 1:
undis u bed eg ow h
Time [y]
]%[ a o eoC
0 20 40 60 80 100
0
0.1
0.2
0.3
0.4
0.5
0.6
Time [y]
]%[ a o eoC
0 20 40 60 80 100
0
0.25
0.5
0.75
1
1.25
1.5
1.75
2
Time [y]
]%[ a o eoC
0 20 40 60 80 100
0
0.1
0.2
0.3
0.4
0.5
0.6
Time [y]
]%[ a o eoC
0 60 120 180 240 300
0
0.5
1
1.5
2
2.5
3
3.5
4
4.5
5
Figu e 3: Le column: succession o ee biomass a e landslide o
di e en scena ios; mean o 50 simula ion uns o one hec a e. Righ
column: spa ial a ia ion o biomass on one hec a e (be ween plo s o
20 x 20 me e ) o he di e en scena ios: Coe icien o a ia ion o
plo biomass o e ime. Black lines depic he mean o 50 simula ions
(each one hec a e); g ey a eas ma k ±2- imes s anda d de ia ion
om mean. No e di e en scales o scena io 3 and he combined
e ec s scena io.
5
he le el o spa ial a ia ion in ma u e o es . To summa-
ize, while he di e ences in biomass succession o he ou
scena ios a e ela i ely small (especially o scena io 3 and
4), we ind majo di e ences in he spa ial a ia ion o bio-
mass wi hin he i s decades o succession. A e 100 yea s
o succession, o all ou scena ios, o e all ee biomass is
only sligh ly lowe han in he ma u e o es be o e he
slide; howe e , species composi ion is s ill di e en om
ma u e o es , wi h a highe p opo ion o pionee and a
lowe p opo ion o climax species. Spa ial a ia ion in
biomass is simila o he a ia ion in ma u e o es and
di e ences be ween he ou scena ios ha e le elled ou .
In he ollowing, we will in es iga e in mo e de ail how
o es p oduc i i y is changed a e landslides; o his pu -
pose we analyze he annual p oduc ion o successional and
ma u e o es . We calcula e biomass gains (due o g ow h
and ec ui men ) and losses (due o mo ali y) o e ime
o o es wi hou landslides and o he di e en scena -
ios a e landslide dis u bance (Figu e 4, small panels). In
all scena ios, biomass losses display a highe luc ua ion
han biomass gains, since losses a e de e mined by mo -
ali y o (some imes la ge) ees, while biomass gains a e
d i en by smalle en i ies, namely g ow h o single ees
and biomass o newly ec ui ed ees. Wi hou landslides,
annual biomass gains as well as biomass losses luc ua e
a ound 7 ons (o ganic d y ma e ) pe hec a e. A e a
landslide e en , he e a e no ees on he slide su ace,
hence biomass p oduc ion and losses a e se o ze o. In
he cou se o o es succession, biomass p oduc ion needs
o exceed biomass losses so ha biomass accumula es. We
obse e his in all ou scena ios; only o he combina-
ion o educed g ow h and ec ui men bo h p oduc ion
and losses emain e y low (<0.2 ons pe hec a e) wi hin
he i s 100 yea s a e landslide. Du ing o es eco e y,
biomass p oduc ion empo a ily exceeds he p oduc ion in
ma u e o es , mos p onounced in scena io 1 and 2, whe e
he p oduc ion cu e eaches a peak o mo e han 10 ons
pe hec a e. The accumula ed di e ence be ween he wo
cu es (biomass gains and losses) desc ibes he biomass
accumula ion (c . Figu e 3, le column); since biomass
losses a e on a e age p opo ional o s anding biomass,
he iming o biomass accumula ions in Figu e 3 coincide
wi h he p oduc ion cu es in Figu e 4.
In eg a ed o e he i s 100 yea s o succession a e
landslide, annual biomass gains and losses a e, compa ed
o ma u e o es wi hou landslides, educed o all scena -
ios, excep o he undis u bed eg ow h scena io, whe e
biomass gain is sligh ly highe (Figu e 4, ba plo ). The
ma u e o es is in an equilib ium s a e, whe e biomass
gains equal biomass losses and hus ne p oduc ion is ze o.
A e a landslide he a e age annual ne p oduc ion is pos-
i i e and almos equal o he ou scena ios (∼1 on pe
hec a e), whe eas he combined scena io yields almos no
biomass. G oss biomass p oduc ion is educed by 37% o
he educed g ow h scena io, by 24% pe cen o educed
ec ui men , by 9% o inc eased mo ali y and by 99% o
he combina ion o educed g ow h and ec ui men .
Scena io Pionee Mid-successional Climax
species species species
undis u bed eg ow h 28 (0) 50 (2) 202 (11)
educed g ow h 65 (1) 85 (2) 218 (11)
educed ec ui men 37 (2) 53 (3) 209 (9)
inc eased mo ali y 28 (1) 47 (2) 199 (11)
educed g ow h & 227 (9) 247 (10) 401 (14)
educed ec ui men
Table 1: Ma u a ion ime (yea s) o successional o es a e land-
slide o di e en species g oups (columns) and di e en eg ow h
scena ios ( ows). Calcula ion is based on s em size dis ibu ions o
ma u e and successional o es . Resul s gi en a e mean (and s an-
da d de ia ion) o 50 simula ion uns on 1 ha.
In all scena ios, landslides ha e an e ec on o es p o-
duc i i y - wi hin he i s decades a e he landslide he
o es becomes a empo a y ca bon sink; he magni ude
o changes in biomass gains and biomass losses, howe e ,
depends on he eg ow h scena io.
So a , we ha e ocused a he i s 100 yea s o o es
succession a e landslides. Howe e , species composi ion
is s ill di e en om ma u e o es a e his pe iod (c .
Figu e 3 le ). To analyze s uc u al di e ences be ween
ma u e and ansien o es in mo e de ail we u ilize
s em size dis ibu ions o ma u e and successional o es
( o de ails, see Appendix B). Table 4.4 p o ides ime
spans needed o he di e en species g oups o each
a ma u e s uc u e, i.e. s em size dis ibu ion. Fo all
scena ios, he o de in which species g oups each he
ma u e s a e is he same: pionee species a e he i s
a e only ew decades o eco e y (28 - 65 yea s), ollowed
by he mid-successional species (47-85 yea s) and, a e
a longe pe iod, he climax species (199-218 yea s).
Di e ences be ween he ou scena ios a e ela i ely small,
he educed g ow h scena io p oduces sligh ly slowe
ma u a ion imes. As al eady obse ed (c . Figu e 3
and 4), he combina ion o educed g ow h and educed
ec ui men esul s in a s ongly delayed eco e y o
o es which is also exp essed in la e ma u a ion o
s em size dis ibu ions (227-401 yea s). Fo all scena ios,
he ime-lag be ween ma u a ion ime o he di e en
species g oups is almos he same: app oxima ely 20 yea s
be ween pionee and mid-successional species and 150
yea s be ween mid-successional and climax species. This
means ha once succession se s in and he pionee species
eco e , he model p edic s simila pa hways o succession.
In all cases i akes a leas 200 yea s un il all species
g oups ha e eached a ma u e s em size dis ibu ion, i.e.
a o es s uc u e and species composi ion simila o he
p e-landslide si ua ion is es o ed.
3.2. Landscape le el
Un il now we ha e es ic ed ou iew o he su ace
o he landslide. In he ollowing we will b oaden ou
6

0
5
10
15
20
Time [y]
a
h
/
0 20 40 60 80
Scena io 1
undis u bed eg ow h
0
5
10
15
20
Time [y]
a
h
/
0 20 40 60 80
Scena io 2
educed g ow h
0
5
10
15
20
Time [y]
a
h
/
0 20 40 60 80
Biomass gain
Biomass loss
Wi hou landslide
0
5
10
15
20
Time [y]
a
h
/
0 20 40 60 80
Scena io 3
educed ec ui men
0
5
10
15
20
Time [y]
a
h
/
0 20 40 60 80
Scena io 4
inc eased mo ali y
0
5
10
15
20
Time [y]
a
h
/
0 20 40 60 80
Combinded e ec s
educed g ow h and ec ui men
0
1
2
3
4
5
6
7
e
u
a
M
s
e
o
:
1
c
S
d
e
b
u
s
i
d
n
u
h
w
o
g
e
:
2
c
S
d
e
c
u
d
e
h
w
o
g
:
3
c
S
d
e
c
u
d
e
n
e
m
i
u
c
e
:
4
c
S
d
e
s
a
e
c
n
i
y
i
l
a
o
m
n
o
i
a
n
i
b
m
o
C
)
3
c
S
+
2
c
S
(
)
y
a
h
(
/
Biomass gain
Biomass loss
Figu e 4: Small panels: Biomass gains (due o ee g ow h and ec ui men ) and biomass losses (due o mo ali y) a e landslide o he
di e en scena ios in one exempla y simula ion un. Righ : Mean annual build up o biomass and loss o biomass. The i s ba depic s
dynamics o ma u e o es whe e biomass gain equals biomass losses. Remaining ba s show dynamics a e aged o e he i s 100 yea s o
o es succession a e landslide o he di e en scena ios ( esul s a e based on 50 simula ions o one hec a e each).
pe spec i e o he landscape le el whe e landslides a e a
eoccu ing dis u bance ha a ec s only small pa s o
he o es . Fo he es o his s udy we selec he scena io
wi h educed ec ui men o ees (scena io 3); esul s
o he emaining scena ios a e p o ided in Appendix C.
Tempo al dynamics o an exempla y simula ion un a e
illus a ed in Figu e 5. Repea ed landslides o di e en
sizes lead o an ab up educ ion in biomass (le panel);
wi h each landslide e en , spa ial a ia ion in ee biomass
inc eases s eeply and e u ns slowly owa ds he le el o
undis u bed o es when no landslide occu s o a longe
pe iod (middle panel). Each landslide is ma ked as a
s ong peak in biomass losses and biomass p oduc ion
dec eases sligh ly in he yea s a e a landslide ( igh
panel).
Fo he analysis o landslides on he landscape le el, we
chose a ealis ic slide equency o 0.02 slides pe hec a e
and yea . We hen compa e landslides wi h he dis u -
bance o gap-building, by swi ching on and o he gap-
building p ocess in FORMIND. While he abo eg ound
biomass dec eases wi h inc easing dis u bance egime -
om 140 ons pe hec a e wi hou gap-building and land-
slides o 117 ons wi h gap-building o 100 ons wi h gap-
building and landslides - o es p oduc i i y displays a di -
e en dynamic (Figu e 6, le and middle). We ind he
lowes p oduc i i y o o es wi hou gap-building (5.6
ons biomass gain pe hec a e and yea ), a conside ably
highe p oduc i i y o o es wi h gap-building (7.3 ons)
and a sligh educ ion o p oduc i i y ( o 6.9 ons) when
in addi ion o o es gaps also landslides occu . In all
cases, he o es is in a quasi equilib ium s a e whe e bio-
mass gains equal biomass losses. The equency dis ibu-
ion o biomass (on 20 x 20 me e plo s) is changed con-
side ably by he di e en dis u bance egimes (Figu e 6,
igh ). Wi hou landslide dis u bance, he equency dis-
ibu ion o biomass is unimodal, wi h plo biomass ang-
ing be ween 50 and 225 ons pe hec a e o o es wi hou
gap-building, and be ween 25 and 200 ons pe hec a e o
o es wi h gap-building. The in oduc ion o landslides
esul s in an inc ease o plo s wi h low biomass: 13% o
plo s main ain less han 50 ons ee biomass compa ed
o only 0.1% o plo s when only gap-building dis u bances
a e p esen . In exchange, he ac ion o plo s wi h high
biomass dec eases when landslides a e p esen .
On he landscape scale, landslides educe abo eg ound
s anding biomass and conside ably change he spa ial he -
e ogenei y o o es biomass.
4. Discussion
4.1. Plausibili y o scena ios
We u ilized he o es g ow h model FORMIND and an-
alyzed di e en hypo he ical scena ios o changes in li e
his o y ai s o ees o p o ide new insigh s on he ole
o landslides o o es s egene a ion and o es s uc u e.
Due o sca ci y o da a om ou s udy sys em, he de el-
oped scena ios a e based on indings om s udies in o he
o es s and heo e ical conside a ions (see Me hods). Ve y
likely, ees on landslide si es in ou s udy a ea will e-
spond o he sho age o nu ien s: i is well known ha
7
Time
]ah/ [ ssamoiB
0 20 40 60 80 100
0
20
40
60
80
100
120
140 Pionee
In e media e
Climax
Landslide
0
10
20
30
40
Time [y]
ah/
0 20 40 60 80 100
Biomass gain
Biomass loss
Time [y]
]%[ a o eoC
0 20 40 60 80 100
0
0.2
0.4
0.6
0.8
1
1.2
1.4
Figu e 5: Exempla y simula ion o o es dynamics on one hec a e wi h landslide equency 0.02 pe hec a e and yea unde eg ow h scena io
3 ( educed ec ui men ). Le : Biomass dynamics o di e en species g oups. Middle: spa ial a ia ion o biomass (be ween plo s o 20 x 20
me e ): Coe icien o a ia ion o plo biomass o e ime. Righ : Biomass gains and losses.
nu ien accumula ion o p e-landslide le els akes decades
(Wilcke e al., 2003) and esul s o a nu i ion expe imen
(NUMEX) poin ou ha ee g ow h is nu ien limi ed
in ou esea ch a ea (Wullae e al., 2010). Howe e , he
amoun o which ee g ow h will be educed in ou s udy
sys em emains specula i e. Reduced ec ui men can pos-
sibly occu due o a dense ege a ion laye o ea ly land-
slide colonis s (e.g. e ns). The ole o hese ea ly colo-
nize s is ambiguous, hey migh p omo e he eco e y o
soil in e ms o s abiliza ion and nu ien accumula ion,
and in his sense ac acili a i e - on he o he hand hey
migh inhibi es ablishmen o ees. Ohl and Bussmann
(2004) sugges a combina ion o ole ance (sugges ing all
species being equally capable o es ablish bu ha ing di -
e en success in popula ion g ow h) and acili a ion o
sou he n Ecuado (Connell and Sla ye , 1977). Bu e en i
hese ea ly colonize s do no hinde ee es ablishmen , e-
duced ec ui men can also esul om con inued soil e o-
sion and he e ogeneous soil condi ions (Walke and Shiels,
2008). E osion, wind, pho o inhibi ion o inc eased le els
o pa hogenes and he bi o y could accoun o inc eased
mo ali y a es on landslide si es. Empi ical e idence o
his plausible e ec is s ill missing, in ac Mys e (2002)
did ind high le els o he bi o y and pa hogenes on land-
slides bu no e ec on ee mo ali y and Fe che e al.
(1996) epo mode a e le els o pho oinhibi ion o only
one o ou s udy species in Pue o Rico. A plan ing ex-
pe imen in a Jamaican o es (Dalling and Tanne , 1995)
ound e en highe mo ali y a es in he unde s o ey com-
pa ed o mo ali y on landslide si es, bu his esul migh
be in luenced by he selec ion o gap-demanding species.
P esumably, di e en li e his o y ai s o ees will be
a ec ed by landslides a he same ime. We es ed a com-
bina ion o educed g ow h and ec ui men and ound a
s ongly delayed eco e y p ocess o ege a ion. The com-
bina ion o he wo educ ion mechanisms ampli ies hei
impac (explana ion ollows below). Such slow o es e-
co e y has been obse ed in some o es sys ems: Res epo
e al. (2003) ound only a ound 25 % o he ma u e biomass
on a 124 yea old landslide in Hawai’i and Dalling (1994)
hypo hesized ha i may ake a ound 500 yea s un il bio-
mass eaches he p e-landslide le el o a Jamaican o es .
The ou scena ios p edic a eco e y ime o o es bio-
mass close o he p e-landslide le el be ween 30 and 80
yea s (c . Figu e 3, le column). In he undis u bed e-
g ow h scena io he o es eco e s apidly. Biomass es-
ima ions om a single landslide in he s udy a ea (C.
Dislich, unpublished da a) indica e ha he scena io wi h
undis u bed eg ow h (scena io 1) is oo as : on a 39-
46 yea old landslide we ound app oxima ely 49 ons o
biomass pe hec a e, which is oughly 40% o he ma u e
biomass. This biomass es ima ion sugges s ha he e-
duced g ow h scena io unde es ima es biomass eco e y
while scena io 3 and 4 ( educed ec ui men and inc eased
mo ali y) seem o mo e o less p edic an adequa e speed
o biomass eco e y. The combined scena io o educed
g ow h and ec ui men p edic s an un ealis ically slow e-
co e y o biomass.
Bu o none o he scena ios, he s uc u e o he mod-
elled successional o es ully co esponds o he o es
s uc u e o his ield obse a ion since he model unde -
ep esen s he equency o small ees and o e es ima es
he equency o la ge ees: in he ield, almos all indi-
iduals had a diame e < 10 cm, and only ew indi iduals
we e la ge han ha . This sugges s ha he ma u a ion
imes gi en in Table 4.4 should be conside ed only as lowe
bounds. The unde - ep esen a ion o small ees, which
does no occu in simula ions o he ma u e o es (see
Dislich e al., 2009), may s em om spa ial cons ain s in
c own sizes o small ees (allome ic ela ionships).
In summa y, a educed ec ui men scena io, a educed
g ow h scena io wi h a mo e mode a e educ ion in he be-
ginning o succession, o a combina ion o educed ec ui -
men and educed g ow h wi h less educ ion, all seem o
be plausible scena ios. The applied changes in li e his o y
ai s a e qui e s ong and may be modi ied, i obse a ions
indica e mo e mode a e changes. Cu en ly he quan i i-
ca ion o he educ ion is di icul due o sca ci y o da a,
howe e , ou app oach allows he gene a ion o a ange o
plausible egene a ion imes.
8
No
gap−building Wi h
gap−building Gap−building
& Landslides
0
20
40
60
80
100
120
140
0
1
2
3
4
5
6
7
8
9
No
gap−building
Wi h
gap−building
Gap−building
& Landslides
Biomass gain
Biomass loss
Biomass [ /ha]
/ha
%
0.001
0.1
1
10
100
0 25 50 75 100 125 150 175 200 225
No gap−building
Gap−building
Gap−building & landslides
(ha y)
Figu e 6: Abo eg ound biomass pe hec a e (le ) and biomass gains and losses (middle) o o es wi hou gap-building dis u bance, o
o es wi h gap-building ( alling p obabili y 0.2 o dead ees abo e 10 cm DBH) and o o es wi h gap-building and landslide dis u bance
( equency 0.02 pe hec a e and yea ). Ba s show he mean o 50 simula ion uns on one hec a e o a pe iod o 1000 yea s; e o ba s display
he s anda d de ia ion. Righ : F equency dis ibu ion o biomass on plo s (20 x 20 m) wi hou gap-building (black) wi h gap-building (da k
g ey) and wi h gap-building and landslides (ligh g ey) espec i ely. Resul s wi h landslides a e based on scena io 3 ( educed ec ui men ).
4.2. Spa io- empo al pa e ns as an oppo uni y o iden-
i ying mechanisms
The analysis o spa ial a ia ion in biomass a he
scale o 20 x 20 me e (c . Figu e 3 igh column) e eals
dis inc pa e ns o he di e en scena ios. Su p isingly,
he spa ial biomass a ia ion in he undis u bed eg ow h
scena io is almos s able h oughou succession - one
migh ha e assumed ha in he ea ly phase o succession
all plo s ha e a e y simila and low amoun o biomass,
which would cause a educ ion in spa ial a ia ion o
biomass dis ibu ion. Howe e , his is no he case since
he as g owing ee species quickly es ablish a di e se
heigh s uc u e ha di e s be ween plo s and leads
o a ia ions in biomass. I g ow h a es a e educed
a e a landslide, we ind a low a ia ion o plo biomass
wi hin he i s decades o succession. In his case, a
as di e si ica ion o he canopy s uc u e is supp essed
by g ow h educ ion, esul ing in a mo e homogeneous
dis ibu ion o biomass. The inc ease in spa ial a ia ion
o ee biomass o educed ec ui men and inc eased
mo ali y is no unexpec ed: wi h educed ec ui men
he ew es ablished ees a e sca e ed une enly by chance
a he pos -landslide su ace, ha e less compe i ion wi h
o he ec ui s and consequen ly can g ow as . Due o
he eedback o al eady es ablished ege a ion, di e ences
be ween plo s wi h ew ec ui s and many ec ui s a e
u he enhanced. Inc eased mo ali y ac s in a simila
way: in he beginning o succession, di e ences be ween
plo biomasses a e a he small, he he e ogeniza ion o
biomass dis ibu ion only se s in when some plo s build
up enough biomass o b ing he mo ali y a e almos
back o he no mal le el. Plo s wi h less es ablished
ees loose mo e biomass (due o accele a ed mo ali y
a es) compa ed o plo s wi h mo e es ablished ees;
his eedback enhances di e ences in plo biomass. Apa
om delayed biomass eco e y, he combina ion o e-
duced g ow h and ec ui men a es leads o a s ong
inc ease in spa ial he e ogenei y o biomass dis ibu ion:
he combina ion o he wo eedback e ec s ampli ies
hei impac . Due o ec ui men educ ion only ew
plo s ecei e su icien ec ui s which hen addi ionally
expe ience a s ong g ow h educ ion. Once hese plo s
ca y enough biomass so ha g ow h educ ion educes,
hey build up a ma u e s uc u e. Fo a long ime he
majo i y o plo s emain in he phase whe e ee biomass
is no high enough o o e come g ow h educ ion.
Na u ally, one migh expec ha di e en mechanisms
in luence o es eco e y on landslide si es a he same ime
and hus in e ac wi h each o he like in he abo e consid-
e ed combined scena io. Ne e heless, he iden i ied (di -
e en ) pa e s o spa ial a ia ion in ee biomass aise
hope ha one could possibly iden i y he mechanisms ha
a ec o es eg ow h wi h he aid o emo e sensing ech-
niques. One possibili y would be o i he pa ame e s o
he landslide module wi h emo ely sensed da a. In com-
9
bina ion wi h olde ae ial pho og aphs ha aid in de e -
mining he ages o landslides, in o ma ion on he empo al
dynamics o egene a ion can be de i ed. High esolu ion
echniques like Lida image y could be used o a ain small
scale in o ma ion on he spa ial dis ibu ion o biomass
on single slide su aces and could hus be compa ed o
model ou pu s om he di e en scena ios and combina-
ions. T us ing he s uc u al ealism o ou model, his
compa ison could indica e which li e his o y ai s a e a -
ec ed by landslides and ad ise u u e ield expe imen s o
es hese hypo heses.
4.3. E ec o landslides on landscape he e ogenei y
The dis u bance egime o landslides a ec s la ge
pa ches o o es han o es gaps ha a e p oduced by
alling ees; while ypical landslide sizes in ou sys em
ange be ween 200 and 1500 m2(mean ∼1100 m2), sizes
o o es gaps ange be ween 20 and 700 m2, bu a e mos
equen ly smalle han 200 m2(B okaw, 1985; Yamamo o,
1992). Landslides a e also a mo e se e e dis u bance han
o es gaps since hey a ec he o es ecosys em abo e
and below g ound; hey p oduce a cha ac e is ic inge -
like signa u e on he landscape (c Figu e 1). The mosaic
landscape s uc u e wi h o es pa ches o di e en succes-
sional s ages ha is c ea ed due o landslide dis u bances
(c . Figu e 6, igh ), in pa icula he mo e open a eas,
will likely ha e a posi i e impac on he di e si y o lo a
and auna (landslide specialis species), as was sugges ed
in se e al p e ious s udies (e.g. Yamamo o e al., 1995;
Gee sema and Poja , 2007; Elias and Dias, 2009). Addi-
ionally, due o landslides, ho spo s o pa icula ly low and
high p oduc i i y eme ge, which do no occu o he wise -
young landslides a e cha ac e ized by low p oduc i i y, bu
du ing he eco e y p ocess, p oduc i i y can occasionally
exceed he p oduc i i y o ma u e o es (c . Figu e 4).
4.4. Fo es p oduc i i y and dis u bances a he landscape
scale
The ole o dis u bances o he ca bon balance o o es s
is no in ui i ely clea - due o a dis u bance, ees die
and ca bon s o ed in ege a ion will pa ly be eleased
and pa ly be s o ed in he soil du ing decomposi ion. In
he case o landslides, ege a ion and soil a e emo ed and
migh come o es a he lowe end o he landslide o be
anspo ed ou o he sys em ia s eams. Soil e osion
migh con inue a e landslides (Walke and del Mo al,
2003; Walke and Shiels, 2008). The dis u bed a ea o -
e s space whe e new biomass can accumula e, he e o e
one can hypo hesize ha dis u bances inc ease o es p o-
duc i i y. This hypo hesis is con i med in he case o he
gap-building dis u bance. Wi hou gap-building, s anding
biomass is high bu p oduc i i y is low (c . Figu e 6, le
and middle panel) since he canopy is pe manen ly closed
and ligh le els in lowe s a a a e low. When dying ees
all o e and c ea e gaps, usually no all ees in he gap
a e damaged, o es soil and seed bank emain in ac and
he e o e emaining ees can u ilize newly a ailable space
and apidly ill in he gap. Consequen ly, o es p oduc-
i i y inc eases in he p esence o o es gaps.
Ou simula ion expe imen s show ha he addi ional
dis u bance o landslides do no u he inc ease o es p o-
duc i i y. Depending on he eg ow h scena io, he e is
only a sho pe iod du ing succession whe e g oss p oduc-
i i y exceeds p oduc i i y o ma u e o es (c . small pan-
els in Figu e 4); du ing his pe iod he es ablished young
ees can p o i om inc eased a ailable ligh and space.
Bu in eg a ed o e he i s cen u y o succession a e
landslide, abo eg ound o es p oduc i i y is educed o
all scena ios (excep he undis u bed eg ow h scena io)
compa ed o p oduc i i y in ma u e o es (c . Figu e 4,
igh ). Consequen ly, o es p oduc i i y on he landscape
scale is also sligh ly educed by landslides (c . Figu e 6,
middle panel and Appendix C).
The o e all s anding biomass is educed by 13 % due o
landslide dis u bances (c . Figu e 6 le ). This ela i ely
high educ ion migh pa ly be caused by ou simpli ying
assump ion o andom landslide loca ions. In he eal land-
scape, he loca ion o landslides depends on many ac o s
like s eepness o he e ain, geomo phological and hyd o-
logical ac o s - he e o e ce ain pa s o he o es will be
a ec ed s onge by landslides han o he s.
Measu ing o es ne p ima y p oduc ion (NPP) is a
labo ious ask - i in ol es es ima ing di e en abo e
and below g ound componen s like abo eg ound biomass
inc emen , li e all and below g ound p oduc ion (Cla k
e al., 2001a). A syn hesis o 39 opical o es si es shows
a ela i ely wide a ia ion o NPP e en a e aken in o
accoun di e en empe a u es and p ecipi a ion egimes
(Cla k e al., 2001b). The p edic ed biomass gains (c .
Figu e 4 and 6) ep esen abo e-g ound coa se wood
p oduc i i y, which is an impo an componen o NPP,
since i domina es abo e-g ound ca bon s o age dynamics
(Chambe s e al., 2001). Unde he assump ion ha d y
biomass is 50 % ca bon, he biomass gain o 5.6-7.3 ons
pe hec a e and yea p edic ed by ou model (c . Figu e
6, middle panel) co esponds o 2.8-3.7 ons ca bon;
hese es ima ed amoun s lie well wi hin he ange (1.5-
5.5 Mg C ha−1a−1) o es ima ed p oduc i i ies o 104
neo opical o es si es (Malhi e al., 2004).
4.5. Limi a ions
P esumably, di e en ee species will show di e en e-
ac ions o he dis u bance egime o landslides, o exam-
ple, Gün e e al. (2007) showed ha only a limi ed num-
be o o es species could egene a e on a 38-yea old aban-
doned pas u e in ou s udy a ea. In his s udy, we applied
he same changes o li e his o y ai s o all species g oups.
Po en ially one could o cou se apply di e en changes o
di e en species g oups, bu his would equi e de ailed
knowledge on how di e en species eac o di e en en-
i onmen al condi ions, especially nu ien limi a ions. In
10
Shallow ansla ional landslides in opical mon ane
o es s - a hin owa ds bio ic con ol?
Pe e Vo pahl∗Claudia Dislich†Helmu Elsenbee ∗Michael M¨a ke ‡
Bo is Sch ¨ode ∗§
Ma ch 17, 2011
Abs ac
We in es iga ed landslides in a Sou h Equado ean mon ane o es o gain insigh s in possible bio ic and abio ic
ac o s ha igge hese slides and ound e idence o he occu ence o e y shallow ansla ional landslides
ha do no in ol e appa en quan i ies o mine al soil. This suppo s p esump ions o a s ong coupling and
in e ac ion o bio ic and abio ic p ocesses in opical mon ane en i onmen s and implies he necessi y o ega d
ege a ion dynamics in shallow slope s abili y models o hese a eas o a s onge ex en .
In opical mon ane en i onmen s, oo s o plan s end o g ow in a massi e o ganic laye a op he mine al
soil a he han pene a ing i . Thus hei con ibu ion o slope s abili y di e s om o he egions. Conside ing
hese di e ences, we in oduced an independen o ganic laye a op he mine al soil in o a s anda d model o
shallow slope s abili y.
We applied he model o ou own measu emen s on and close o ele en landslides in he Andes o Sou he n
Ecuado , among which h ee we e e y shallow. Being able o ep oduce ou indings, he model implies ha in
case o e y shallow landslides, he apid mass mo emen e en is likely o ha e been caused by he ege a ion
i sel .
In oduc ion
Apa om hei haza dousness o human li e and in as uc u e, landslides can p o ide a bene icial ecological
e ec : In opical mon ane ain o es s, landslides ep esen one o he mos impo an ecosys em dis u bances
(Lozano e al., 2005; K¨ohle and Hu h, 2007; Bussmann e al., 2008). Thei size and equency con ibu e o
he high le els o ascula plan di e si y in hese a eas, since landslide sca s p o ide habi a s o pionee species
(Connell, 1978; Sheil and Bu slem, 2003). F om an ecological poin o iew, knowledge on he d i ing mechanisms
o landslides is a p e equisi e o unde s anding and p edic ing po en ial u u e changes in he gene al condi ions
o biodi e si y in opical mon ane ecosys ems, which can aid in he planning o conse a ion measu es.
To coun e ac he h ea landslides ep esen o human ac i i y and o p o ide ools o planning o sus ainable
in as uc u e measu es, he p edic ion o he spa ial occu ence p obabili y o landslides has u ned in o a majo
esea ch e o (e.g Mon gome y and Die ich, 1994; Wu and Sidle, 1995; Guzze i e al., 1999; Guzze i, 2004;
Gui ie es e al., 2010). The concep ual abs ac ion unde lying mos p ocess-based models o shallow ansla ional
landslides is ha o an in ini e slope segmen co e ed wi h ege a ion (Wu e al., 1979; Buchanan and Sa igny,
1990; Sidle and Wu, 1999; Xie e al., 2004). Moh -Coulomb’s ailu e c i e ion p esc ibes ha he ac o o sa e y
(F oS) o a slope segmen is gi en by he a io o s abilizing and des abilizing o ces, whe e F oS < 1 indica es
ins abili y condi ions.
F oS =S abilizing o ces
Des abilizing o ces (1)
∗Uni e si y o Po sdam, Ins i u e o Ea h- and En i onmen al Sciences, Ka l-Liebknech -S . 24/25, 14476 Po sdam, Ge many
†Helmhol z Cen e o En i onmen al Resea ch - UFZ, Depa men o Ecological Modelling, Pe mose s . 15, 04318 Leipzig, Ge many
‡Heidelbe ge Akademie de Wissenscha en, c/o Geog aphisches Ins i u de Ebe ha d Ka ls Uni e si ¨a T¨ubingen, R¨umelins 19-21,
72070 T¨ubingen, Ge many
§Leibniz-Cen e o Ag icul u al Landscape Resea ch (ZALF) e.V., Ebe swalde S . 84, 15374 M¨unchebe g, Ge many
1

Mos implemen a ions o his concep include ege a ion by an inc ease o soil cohesion due o oo ne wo ks
ha g ow pe pendicula h ough a po en ial sliding plane, while he des abilizing e ec s o abo eg ound biomass
o he e ec o wind o ces as ans e ed by ees in o he g ound ia a u ning momen mos ly a e assumed o
be negligible. As a consequence, he s abilizing e ec o oo s o di e en plan species has been subjec o ield
expe imen s (Wu e al., 1988a; Abe ne hy and Ru he u d, 2001; an Beek e al., 2005) and labo a o y es s (Wu
e al., 1988b; De Bae s e al., 2008) and lead o slope s abili y models ha , o example, allow o in es iga e he
in luence o di e en o ms o land use on landslide dis ibu ion and equency (Schmid e al., 2001; Siddle and
Dahkal, 2003). O he in es iga ions addi ionally ega ded he in luence o ege a ion on slope hyd ology (Casadei
e al., 2003; Keim and Skaugse , 2003).
Howe e , all modeling app oaches implici ly assume ha oo s g ow in o he mine al soil and hence enhance i s
mechanical p ope ies. P andini e al. (1977) poin ed ou ha his implici assump ion does no apply uni e sally,
and i ce ainly does no in hose opical mon ane en i onmen s whe e ee oo s g ow p e e en ially wi hin a
massi e o ganic laye abo e he mine al soil. Thus, in hese egions, oo s do no necessa ily ac as slope s abilize s
by inc easing he soils shea esis ance o cohesion. Ins ead, hey may be conside ed pa o a sepa a e laye wi h
i s own mechanical p ope ies.
Endea o ing o con ibu e o he unde s anding o his di e en si ua ion in opical mon ane en i onmen s,
ou s udy is loca ed in a opical mon ane ain o es in he Andes o Sou he n Ecuado (Fig. 1), whe e we ound
e y shallow ansla ional landslides ha did no in ol e appa en quan i ies o mine al soil. Shallow ansla ional
slides, mainly o exclusi ely consis ing o o ganic ma e ial imply a s ong bio ic con ol as has been p oposed by
Rich e e al. (2009) and hin owa ds an addi ional sel o ganiza ion mechanism in opical mon ane ecosys ems.
Thus, based on a sligh ly modi ied s anda d model o shallow slope s abili y, ou wo k ocuses on he explana ion
o he e y shallow ansla ional landslides we obse ed.
S anda d models o shallow slope s abili y do no allow o landslides ha do no in ol e mine al soil, no
do hey accommoda e a hick o ganic laye . Thus, as a i s s ep, we ex end a s anda d model o shallow slope
s abili y by he addi ion o an o ganic laye a op he mine al soil and in oduce he mass o his laye as a
des abilizing componen . Using he model o ep oduce he c i ical minimum soil dep h, as necessa y o ailu e,
we explo e he condi ions ha lead o single slope ailu es in he s udy a ea.
Hypo hesizing ha a) he si ua ion close o a landslide e lec s he si ua ion on he landslide be o e he e en and
b) ou model will ep oduce he obse ed dep h o ailu e i pa ame ized wi h measu emen s om he landslide,
we apply he model o ou measu emen s, conduc ed on and close o ele en landslides in he s udy a ea (Fig. 1).
Based upon ou measu emen s, we u he apply a sensi i i y analysis o he model o explain he na u e o he
e y shallow ansla ional slides we obse ed.
S udy a ea
The da a we use he e we e collec ed in he Rese a Bi´os e a de San F ancisco (RBSF), pa o he biosphe e
ese e Podoca pus - El Condo , in he Andes o Sou he n Ecuado (3o58’S, 79o04’W, Fig. 1). The s udy a ea
consis s o se e al low-o de ca chmen s sou h o Rio San F ancisco and comp ises 8.4 km2, anging in al i ude
om 1,870 o 3,165 m a.s.l. S eep slopes (up o 70o) a e co e ed by an e e g een lowe (<2,150 m a.s.l.) and
uppe b oad-lea ed mon ane ain o es up o he ee line be ween 2,700 m and 3,000 m a.s.l. A highe ele a ions,
a sub p´a amo sh ubland eme ges (Beck e al., 2008).
In his moun ainous ecosys em, shallow ansla ional landslides a e a equen , na u al phenomenon (Lozano
e al., 2005; Bussmann e al., 2008; Dislich e al., 2009; Res epo e al., 2009; Rich e e al., 2009) and isible
landslide sca s pe manen ly co e app oxima ely one o h ee pe cen o he s udy a ea, as deduced om ae ial
pho og aphs ( aken in 1962, 1969, 1976, 1989 and 1998, espec i ely). Mos landslides wi hin he s udy a ea
can be classi ied as ockslides, ea h lows and shallow ansla ional landslides. Rockslides and ea h lows occu
close o an h opogenic in e ence (such as oads), while wi hin he undis u bed pa s o he a ea all apid mass
mo emen s ha e been classi ied as shallow ansla ional landslides (Bussmann e al., 2008).
In addi ion o his, massi e o ganic laye s abo e he mine al soil exis , ha mainly consis o dead o ganic
ma e wo en wi h plan oo s, while con aining only e y small amoun s o mine al soil. We ound hese laye s,
whose mass may each o up o 700 ha−1(Wilcke e al., 2002), p e e en ially on s eep slopes in he in e media e
al i udinal anges wi hin he s udy a ea (2,100 o 2,700 m a.s.l.).
2
Figu e 1: Loca ion o he s udy a ea in Sou he n Ecuado eas o Loja. Ele en in es iga ed landslides a e ma ked
by yellow ec angles, su eys o ege a ion ela ed pa ame e s by ligh blue ci cles and loca ions o
o ganic laye ensile esis ance measu emen s by g een hexagons. Anno a ed UTM-WGS84 coo dina es.
Field in es iga ions
Du ing h ee ield campaigns om Sep embe 2008 o No embe 2010, we in es iga ed a o al o ele en landslides
in he s udy a ea (numbe ed yellow ec angles in Fig. 1) a h ee dis inc al i udinal anges (below 2,150 m a.s.l,
2,150 o 2,350 m a.s.l. and abo e 2,350 m a.s.l., espec i ely). Soil physical pa ame e s, such as in si u and
d y densi y (ρs), wa e con en (Θ), soil cohesion (Cs) and in e nal angle o ic ion (ϕ) we e assessed in soil
p o iles, ha we e c ea ed a he op edge o en landslides. We assessed ege a ion ela ed pa ame e s, such as
abo eg ound biomass (MB), o ganic laye hickness (ho, ligh g een ci cles in Fig. 1) and measu ed he o ganic
laye ’s esis ance (Co, g een hexagons in Fig. 1). Geog aphic coo dina es as well as al i ude abo e sea le el we e
eco ded in UTM WGS84 o ma by he aid o a hand-held GPS de ice o ype Ga min e ex is a HCx wi h an
in eg a ed ba ome ic al ime e .
Landslide su eys
We based a ough es ima e o landslide age on he successional s a e o he ege a ion on he landslide sca
and used he e ms ’ esh’ o landslides wi hou ege a ion and ’old’ o o he s since an exac de e mina ion o
landslide age was no possible.
Wi h he aid o a o al s a ion, as well as by manual measu emen s wi h compass, clinome e and ape measu e,
3
ela i e 3D-coo dina es o app oxima ely hund ed poin s on each landslide we e eco ded. These poin s se ed
as da a sou ce o a h ee-dimensional econs uc ion o he landslides’ ailu e planes, om which opog aphic
in o ma ion, such as landslide leng hs (L) and median wid hs (W), we e de i ed by p ojec ion on o a ho izon al
plane. A e age slope angels (α) we e calcula ed by a linea eg ession h ough a p ojec ion o all poin s on o a
e ical plane.
D y and in si u soil bulk densi ies (ρs) and soil wa e con en (Θ) we e calcula ed om undis u bed soil samples
om di e en dep hs o soil p o iles, which we c ea ed a he op edge o en o he ele en landslides. Whe e
possible, we ook h ee samples pe soil laye , which we e weigh ed, d ied (24h a 80oC), and weigh ed again.
Landslide dep hs (hs) we e es ima ed by isually ixing he in e sec ion o he landslide su ace wi h he e ical
soil p o ile.
Soil cohesion was measu ed in si u by a Geono H-60 hand-held ane es e h ee imes pe iden i iable soil laye
wi hin each p o ile. D y soil bulk densi ies (ρs), wa e con en (Θ) and cohesion measu emen s (Cs) we e used o
c ea e dep h p o iles o hese pa ame e s o each landslide. The me hod o measu ing soil shea esis ance wi h a
small o sion p obe (like he Geono H-60) ends o o e es ima e soil cohesion i he soil con ains skele on, which
was he case in ou soil p o iles, whe e soil ex u es anged om clays o e sil y sands o ine sand, while g ain
size o he soil skele on anged om millime e s o decime e s in diame e .
To compensa e he e ec o ela i ely small s ones in he soil, we applied a me hod o simul aneously measu e
soil cohesion and he in e nal angle o ic ion. This app oach ollows a common me hod o in-si u measu emen
o soil shea esis ance as applied by se e al o he esea ches (c . Wu e al., 1988a; Comino and Du e a, 2009),
bu uses a much smalle de ice.
A s eel cylinde (d= 72 mm) was used o ho izon ally shea ou soil o a ep oducible c oss sec ion, while
measu ing he maximum o ce applied by a sp ing balance wi h a d ag indica o (Fig. 2). Di e en e ical loads
we e supe imposed o he soil in he cylinde by a o ce , whose diame e was sligh ly smalle han he inne
diame e o he cylinde . The o ce was a ached o he bo om o a bucke , allowing o a y he supe imposed
load. This p ocedu e was epea ed a leas wice pe supe imposed load and wi h a leas wo di e en loads.
Figu e 2: De ice o measu ing soil shea esis ance unde di e en loads.
A e Moh -Coulomb, he soil’s shea esis ance esul s om a no mal load applied o he soil and i s cohesion:
τ=σ an ϕ+Cs(2)
Whe e τis he shea esis ance [N m−2]; σ he no mal load [N m−2]; ϕ he in e nal angle o ic ion and Cs he soil cohesion
[N m−2].
Thus, a linea eg ession h ough all measu emen s is used o de e mine he soil’s in e nal angle o ic ion (ϕ)
and i s cohesion (Cs):
Fshea
2π=Fload
2π an ϕ+Cs(3)
Whe e Fshea is he applied maximum o ce o shea ou he p obing cylinde [N]; he adius o he p obing cylinde [m]
and Fload he addi ional weigh o ce applied o he soil in he p obing cylinde [N].
Ou me hod s ill is sensi i e o coa se s ones in he sample. Thus we we e able o use i on se en o he ele en
landslides (i.e. #1, #2, #5, #6, #7, #9 and #11 in Fig. 1). On ou landslides (i.e. #2, #5, #6, #7 in Fig. 1),
we applied he me hod in he soil p o ile close o he es ima ed dep h o ailu e. On wo landslides (#1 and #9
4
in Fig. 1), we we e able o apply he me hod in se e al dep hs o he soil p o ile and on he oldes landslide (i.e.
#11 in Fig. 1), we conduc ed measu emen s on he su ace i sel .
As has been men ioned by Ande son and Howes (1985), ield measu emen s o soil cohesion and in e nal ic ion
angle do pe haps no accu a ely e lec he s eng h o he esidual soils examined since hey depend on ma ix
suc ion. Thus on landslide #11, we addi ionally assessed changes o soil cohesion and he in e nal ic ion angle
as esul om sa u a ion wi h wa e .
Vege a ion ela ed pa ame e s
In he absence o knowledge abou he eal si ua ion p eceding a sliding e en , we in es iga ed ege a ion- ela ed
pa ame e s close o landslides, assuming ha he si ua ion he e is qui e simila o ha on he slide be o e he
e en . S anda d app oaches o assess abo eg ound biomass, equi e su eys in homogeneous o es s ands o
a leas 1 ha (100 m x 100 m) in ex en (B own e al., 1989; B own, 1997; Leuschne e al., 2007). Shallow
ansla ional landslides, howe e a e ini ia ed a a smalle scale (10 o 30 m). Thus we we e in e es ed in possible
local biomass agg ega ions a his scale and assessed ege a ion biomass and e ical hickness o he o ganic laye
in ci cula a eas, wi h a diame e o 10 m a ound se e al single poin s close o landslides (poin in es iga ions,
ligh g een ci cles in Fig. 1). Since ou measu emen s we e conduc ed on s eep slopes, we addi ionally measu ed
he upslope and downslope slope angle (α) by a clinome e .
A special sampling design was no applied and as a consequence o he ela i ely small su ey a ea (78.5 m2
pe si e), we expec ed ou esul s o highly a y be ween he single poin in es iga ions. Thus, we agg ega ed
all measu emen s om poin in es iga ions in one al i udinal le el and used he esul ing mean alues in u he
calcula ions.
To calcula e abo eg ound ee biomass, we es ima ed he a e age ee heigh om measu emen s wi h ape
measu e and clinome e and calcula ed s em diame e s a b eas heigh (1.3 m, DBH) o all ees wi hin he a ea
( adius = 5 m) om s em pe ime e s, measu ed by a ape measu e.
Following Leuschne e al. (2007), we applied wo di e en allome ic equa ions o he es ima ion o a e age
ee biomass, which, a e B own and I e son (1992) and B own (1997), a e sui able o ame ican opical we
o es s:
Ma
=e−3.375+0.948 ln(DBH2h)(4)
Mb
= 21.297 −6.953DBH + 0.740DBH2(5)
Whe e M is he biomass pe ee [kg], DBH is he a e age s em diame e [cm] a b eas heigh (130 cm) and his he
a e age ee heigh [m] in he ci cle ( adius = 5 m).
We addi ionally calcula ed a basic es ima e by assuming all ees o be cylinde s o mean DBH and mean
heigh wi h a cons an wood densi y o ρwood = 600 kg m−3, which, a e B own (1997), is an app op ia e a e age
alue o opical mon ane o es s. In his manne we a i ed a an es ima e on a e age ee mass:
Mc
=ρwood hDBH
200 2
π(6)
Whe e ρwood is he bulk densi y o wood [kg m−3].
The abo eg ound biomass densi y was hen calcula ed by mul iplica ion o he a e age ee mass by he numbe
o ees in he ci cle, di ided by he in es iga ed a ea, p ojec ed on o a ho izon al plane.
MB=M n
π 2cos α(7)
Whe e MBis he biomass densi y [kg m−2], n he numbe o ees in he ci cle [1], αis he slope angle [o] and he adius
o he in es iga ed a ea [m].
O ganic laye p ope ies
Wi hin each poin in es iga ion, we conduc ed up o 20 measu emen s o e ical o ganic laye hickness by
e ically p obing he laye wi h a pole a di e en , andom loca ions wi hin he ci cle ( adius = 5 m). By using
5
ou slope angle measu emen s, we hen calcula ed o ganic laye hickness pe pendicula o he slope (ho).
In o de o es ima e o ganic laye bulk densi ies, 64 samples we e aken a 5 di e en loca ions in he s udy
a ea (close o landslides #4 and #5 in Fig. 1 a an al i ude o ≈2,300 m a.s.l.) by exca a ing ec angula blocks
(30 cm x 30 cm) h ough he comple e o ganic laye down o he mine al soil and measu ing hei olume. The
exca a ed o ganic laye ma e ial was weigh ed in si u. Fi een samples ( h ee pe loca ion) we e sa u a ed wi h
wa e and weigh ed again in o de o es ima e he maximum o ganic laye bulk densi y unde ain all condi ions.
We assessed o ganic laye ensile esis ance in si u close o landslides (g een hexagons in Fig. 1) by epea ed
applica ion (n= 903) o up u e es s. Following De Bae s e al. (2008), a la ge numbe o ine oo s con ibu e
mo e o he o al ensile esis ance han a small numbe o big oo s. Thus we measu ed he ensile esis ance o
he ine oo ma ix in he o ganic laye , no conside ing plan oo s wi h diame e s signi ican ly g ea e han 5
mm.
O ganic laye up u e es s we e conduc ed in e ical p o iles h ough he o ganic laye , s a ing a he op o
he p o ile and we e subsequen ly execu ed down o he mine al soil (see. Fig. 3).
Figu e 3: O ganic laye ensile es .
Each up u e es s a ed wi h he p epa a ion o a block o o ganic laye ma e ial o ep oducible size (10 cm
X 10 cm) by sawing wo e ical, pa allel cu s. A small ake se ed as g ippe and was hooked in o he ma e ial
om abo e in a way ha i co e ed a dep h o 10 cm. While ho izon ally pulling ou he specimen, we measu ed
necessa y o ces wi h a sp ing balance equipped wi h a d ag indica o . We assumed he up u e plane o oughly
ollow a ci cula pa h wi h a adius o 10 cm (Fig. 3) and calcula ed i s a ea o 157.1 cm2.
While conduc ing es s a di e en dep hs wi hin he o ganic laye , we dis inguished be ween measu emen s in
he laye and a he bounda y o he mine al soil. Expec ing high a ia ions in ensile esis ance, we conduc ed
up o 100 epe i ions pe p o ile.
Model
Following Gabe and Dunne (2002) and Casadei e al. (2003), we idealize a slope ailu e by a shallow, ec angula
block on an inclined plane (Fig. 4). The des abilizing o ce esul s o he downhill componen o he block’s weigh
o ce, while he block is s abilized by shea esis ance a i ’s basal plane and a he sides and by ensile esis ance
a he uppe pe ime e . A shallow slope s abili y model ha includes la e al o ces, implici ly iola es he in ini e
slope assump ion since i is sensi i e o he ailu e dimensions. Thus we espec hese dimensions in ou model,
namely by he leng h (L) and wid h (W) o he ailu e.
Du ing ou ield su eys, we ound massi e o ganic laye s no con aining ob ious quan i ies o mine al soil.
Usually he ansi ion o he unde lying mine al soil was e y sha p and he numbe o oo s, we ound in he
uppe mos laye o he mine al soil was negligible. Thus, ou mechanical se up consis s o wo, non-o e lapping
laye s (Fig. 4), which a e assumed o be homogeneous. An o e lapping o hese laye s, indica ing a oo ing dep h,
howe e , would no change he ma hema ical desc ip ion. The di e ence o o he app oaches is ha we explici ly
allow he hickness o he o ganic laye o exceed ha o he mine al soil. By mechanically decoupling he wo
laye s we a e able o desc ibe e en e y shallow ansla ional slides ha in ol e no mine al soil.
6

Figu e 4: Mechanical se up. Slope-pa allel leng h (L) and wid h (W) o he sliding block; hs: Heigh o mine al
soil; ho: Thickness o o ganic laye .
The ac o o sa e y o ou se up, again, is he a io o s abilizing and des abilizing o ces.
F oS =Fb+Fl
FGsin α(8)
Whe e F oS is he ac o o sa e y [1]; Fb he basal esis ance o ce [N]; Fl he la e al esis ance o ce [N]; FG he block’s
o al weigh o ce [N] and α he slope angle [o].
We in oduce he weigh o he o ganic laye a op he mine al soil as addi ional, des abilizing componen . Thus,
he sys em’s o al weigh o ce comp ises abo eg ound biomass, he mass o he o ganic laye and he mass o soil
down o he dep h o ailu e:
FG= (MB+hoρo+hsρs)g L W (9)
whe e gis he ea h’s accele a ion [m s−2]; MB he abo eg ound biomass densi y [kg m−2]; hs he hickness o he mine al
soil [m]; ho he hickness o he o ganic laye [m]; ρo he bulk densi y o he o ganic laye [kg m−3] and ρs he bulk densi y
o soil [kg m−3].
We subs i u e he weigh o ce pe a ea ((MB+hoρo+hsρs)g) by a p essu e load (G).
FG=G L W (10)
As p esc ibed by Moh -Coulomb’s ailu e c i e ion, he esis ance o ce in he basal a ea o he block esul s om
in e nal ic ion, soil cohesion and he cohesion p o ided by oo s, pene a ing he basal plane. We did no ind
many o n oo s on he slip su ace o he landslides, we su eyed. Thus we assume he addi ional oo cohesion
a he slip su ace o be a ac ion o he maximum cohesion, as p esen in he o ganic laye . In mos cases, his
ac ion can be assumed o be ze o.
Fb=FN an ϕ+L W (Cs+xCo) (11)
whe e FNis he e ec i e no mal componen o he block’s weigh o ce [N]; ϕis he in e nal ic ion angle o he soil [o]; Cs
soil cohesion [N m−2]; Cois he maximum oo cohesion in he o ganic laye [N m−2]; xis he ac ion o oo s con ibu ing
o basal cohesion [1] and αis he slope angle [o].
The sa u a ed ac ion o he e ical soil column causes a po e wa e p essu e (U) a he basal plane, which
educes he e ec i e no mal o ce.
U=ρw
m hs
cos αg(12)
whe e Uis he po e wa e p essu e a he ailu e plane [N m−2]; mis he sa u a ed ac ion o he soil column [1] and ρw
is he bulk densi y o wa e [kg m−3].
7
Thus he e ec i e no mal o ce esul s o:
FN=MB+ρoho+ρs−m ρw
cos αhsg L W cos α(13)
FN= (G−U)LW cos α(14)
The la e ally s abilizing o ce esul s om soil cohesion and om he o ganic laye ’s cohesion. I is e ec i e a
bo h sides and a he uppe pe ime e o he block.
Fl= (2L+W)(hsCs+hoCo) (15)
whe e Cois he o ganic laye ’s cohesion [N m−2].
Subs i u ing in o equa ion 8 leads o he inal ac o o sa e y:
F oS =[MB+ρoho+ (ρs−mρw
cos α)hs]gcos α an ϕ+Cs+x Co+2L+W
LW (hsCs+hoCo)
(MB+hoρo+hsρs)gsin α(16)
Fo ou analysis, we a e especially in e es ed in he c i ical minimum soil dep h, as is necessa y o ailu e. Thus
we se he ac o o sa e y (Eq. 16) o one and sol e he equa ion o hs.
hc i .
s=(MB+hoρo)g(cos α an ϕ−sin α) + Cs+x Co+hoCo2L+W
LW
[mρw an ϕ+ρs(sin α−cos α an ϕ)] g−Cs2L+W
LW
(17)
Model pa ame iza ion and sensi i i y analysis
As a i s s ep, we assume a mos ins able si ua ion as gi en by comple e sa u a ion o he soil column (m= 1)
and calcula e minimum ac o s o sa e y (F oS, Eq. 16) and he minimum c i ical soil hickness (hc i
s, Eq. 17)
o all landslides, whe e we ha e a comple e se o pa ame e s and compa e he esul s o he obse ed landslide
dep hs.
Then, ega ding he spa ial a iabili y a he scale o me e s in pa ame e s ela ed o soil s eng h, we es ima e
alue anges o soil cohesion (Cs), soil densi y (ρs) and in e nal ic ion angle (ϕ) based upon ou measu emen s
and use hese alue anges o calcula e anges o F oS and hc i .
s o all landslides we su eyed. Again, we compa e
he esul s o ou obse a ions.
Finally, using plausible mean alues o model pa ame e s, we apply a sensi i i y analysis o he model and
explo e he equi emen s o e y shallow ansla ional landslides by a ying hs, while all o he pa ame e s a e
held cons an .
Resul s and discussion
We e alua ed ela i e 3D coo dina es, measu ed on landslide su aces and de i ed landslide dimensions and a e age
slope angles. Plana landslide leng hs (Lp) a ied be ween 7 and 62 m and median wid hs (W) anged om 8 o
23 m, while we ound a e age slope angles (α) om 31o o 55o(Tab. 2). We isually es ima ed he dep h o each
o he ele en landslides and ound hs o ange om below 10 cm up o 1.3 m (Tab. 2); wo o he landslides (#4
and #5) we e ex emely shallow and one landslide (#8) exposed a dep h o ailu e o abou 20 cm.
We assessed d y soil densi y and in si u soil wa e con en in undis u bed samples, aken om di e en ho izons
o en soil p o iles, which we c ea ed a op o landslides #1 o #10 and measu ed in si u soil shea esis ance wi h
a o sion p obe (Tabs. 3 and 4). Soil cohesion (Cs), as ob ained om hese measu emen s anged om 8.1 o 70
kPa (30.9±16.1 kPa, mean ±s anda d e o o mean, n=119). We calcula ed d y soil densi ies (ρs) o 1,342±220
kg m−3(mean ±s anda d e o o mean, n=33). Acco ding o Jahn e al. (2006), we conduc ed ield assessmen s
o soil ex u e and soil skele on in all soil ho izons (Tabs. 3 and 4).
As we suspec ed, measu emen s wi h a small o sion p obe o e es ima ed he soil cohesion. We calcula ed
much smalle alues by ou me hod o simul aneously de e mine soil cohesion and in e nal ic ion angle, which
we applied on se en landslides (Tab 5). He e we ound soil cohesion (Cs) o ange om 4.5 kPa up o 20.8 kPa
8
(10.7±5.0 kPa, mean ±s anda d e o o mean, n=17) and in e nal ic ion angles (ϕ) anging om 25oup o
51o(37o±8o, mean ±s anda d e o o mean, n=17).
Bo h me hods a e sensi i e o small scale soil inhomogenei ies, while his e ec is smalle in ou me hod due
o he compa ably bigge shea plane. We hink ha ou me hod s ill o e es ima es soil cohesion i coa se s one
agmen s a e p esen in he shea plane. Thus we ecommend i o soils wi hou o a leas wi h only small s ones
(d < 5 mm) in he soil skele on. The esul s o ϕ, howe e should no be a ec ed by his sys ema ic o e es ima ion
o shea esis ance. Due o he low weigh and small size o he de ice, he me hod p o ed especially use ul o
he applica ion in ough, moun ainous e ain.
On landslide #11, we assessed changes in soil shea esis ance and in e nal ic ion angle on he ailu e su ace
due o sa u a ion by conduc ing wo se ies o measu emen s. We ound no ob ious change o Cs, bu a educ ion
o ϕby 39% ( om 40.2 o o 24.7 o). E en hough, he wo measu emen s ha e been conduc ed in he same soil
laye close o each o he , i is possible ha small scale soil inhomogenei ies ha e caused his dec ease o ϕ. Thus
u he epe i ions o he expe imen a e equi ed o quan i y he e ec .
We conduc ed 39 poin in es iga ions o ege a ion ela ed pa ame e s (Tab. 7) and calcula ed abo eg ound
ee biomass es ima es o 49 o 92 ha−1in he lowe al i udinal ange (<2,050 m a.s.l.), o 35 o 66 ha−1in
he in e media e al i udinal ange (2,050 o 2,350 m a.s.l.) and o 13 o 24 ha−1in he uppe al i udinal ange
(>2,350 m a.s.l.).
As we expec ed, we ound s ong a ia ions in ee densi ies and ege a ion composi ion e en by poin in es i-
ga ions conduc ed in one and he same al i udinal ange. None heless ha e we been able o ep oduce he dec ease
o abo eg ound biomass wi h al i ude a.s.l., as has been epo ed by o he s udies. Leuschne e al. (2007), o
example, es ima ed abo eg ound ee biomass o 132 up o 199 ha−1a an al i ude o 1890 m a.s.l. and o 74
up o 127 ha−1a 2,380 m a.s.l. by using he same allome ic equa ions. Mose e al. (2008) used a di e en
allome ic equa ion o he es ima ion o an abo e g ound ee biomass o 173 ha−1a 1,890 m a.s.l. and o
100 ha−1a 2,380 m a.s.l. Ou es ima es o abo eg ound ee biomass, lie below alues o bo h s udies, which
is no su p ising, since we conduc ed ou measu emen s close o landslides, mainly on s eep open slopes and no
in small alleys o go ges, whe e due o sedimen and nu ien accumula ion a highe abo eg ound biomass can
de elop (Oeske e al., 2008).
O ganic laye hickness a ied s ongly be ween he single si es and amoun ed o 35±21 cm (mean ±s anda d
e o o mean; n= 306). In si u o ganic laye densi y, as calcula ed om 64 samples and a o al sample olume
o 2.81 m3 esul ed o 208.1±89.0 kg m−3(mean±s anda d e o o mean). Densi y o sa u a ed o ganic laye ,
as de i ed om 15 samples and a o al sample olume o 0.59 m3 esul ed o 278±86.7 kg m−3(mean±s anda d
e o o mean). Using hese alues, we calcula e he a e age mass o he o ganic laye o 728±187 ha−1(up o
973±182 ha−1unde we condi ions), which compa es p e y well o Wilcke e al. (2002), who es ima ed he
mass o he o ganic laye in he s udy a ea o up o 713 ha−1.
We conduc ed a o al o 903 in si u o ganic laye up u e es s in nine o ganic laye p o iles a h ee al i udinal
le els o he s udy a ea (Tab. 6) and ound o ganic laye ensile esis ance o a y om 1.84 kPa o 2.69 kPa
(2.17 ±1.16 kPa, mean ±s anda d e o o mean). Tes s conduc ed in he laye yielded 2.26 ±1.2 kPa and es s
conduc ed a he bounda y o he mine al soil yielded 2.07 ±1.09 kPa, which is a signi ican educ ion by 8.5%
(Welch wo sample - es , p- alue=0.012, signi icance le el: 95%). A compa ison o ensile es s conduc ed in he
dis inc al i udinal anges, howe e , did no e eal a signi ican di e ence.
Model applica ion
The numbe o o n oo s on he ailu e su aces o he in es iga ed landslides in ou s udy a ea was negligible.
Thus, we assume ha he e is no oo con ibu ion o basal esis ance (xCo= 0). A i s calcula ion o he ac o
o sa e y (Eq. 16) and o he c i ical soil dep h (hc i .
s, Eq. 17) om he comple e se o pa ame e s, we ha e o i e
o he ele en landslides (i.e. #1, #2, #6, #7 and #9, Tab. 1), e ealed a po en ial slope ins abili y on landslides
#1 and #9. He e, he calcula ed minimum soil dep hs lie wi hin cen ime e s in he ange o ou obse a ions. In
he o he h ee cases, he ac o s o sa e y we e g ea e han 2 and he minimum c i ical soil dep hs, as necessa y
o ins abili y exceeded he obse ed landslide dep hs by up o 2.8 m.
A i s sigh , wo ou o i e is a bad esul . Bu i ou measu emen esul s we e pu ely andom, he almos
pe ec p edic ion o ailu e dep hs o landslides #1 and #9 canno be explained. Thus, we can assume ha ou
me hod was app op ia e bu applied a he w ong spo o a he w ong ime in h ee ou o i e cases and we owe
9
Table 1: Model applica ion o i e landslides, model pa ame e s, c i ical minimum soil dep hs (hc i .
s) and ac o s
o sa e y (F oS).
# L W MBα ϕ ρsCshobs.
shc i
sF oS
[m] [m] [kg m−2] [o] [o] [kg m−3] [kPa] [m] [m] [1]
1 30.8 7.8 1.9 53.1 25.8 1,265 4.7 0.42 0.45 1.06
2 35.4 7.0 1.9 49.1 50.9 1,504 10.2 0.39 1.29 2.43
6 31.3 10.6 5.2 37.1 41.4 1,200 14.3 0.60 3.37 3.45
7 43.2 10.1 5.2 37.0 46.4 1,120 12.9 0.40 2.74 4.46
9 34.9 22.6 7.0 33.0 27.0 1,326 7.7 1.33 1.36 1.02
Co= 2.17 kPa; h0= 0.35 m
he compa ably high soil cohesions and in e nal ic ion angles, we measu ed on landslides #2, #6 and #7 ei he
o he high spa ial he e ogenei y o hese pa ame e s o o hei dependency on soil wa e con en (Ande son and
Howes, 1985).
Sidle and Swans on (1981) s a ed ha conse a i e alues o appa en soil cohesion should be used o ac o
o sa e y calcula ions, i he measu ed accu acy is ques ionable. Facing a simila si ua ion, namely by a high
a iabili y o ou pa ame e s ela ed o soil s eng h, we es ima ed app op ia e alue anges o calcula e F oS and
hc i .
s o all ele en landslides. Ou measu emen s o Cs a ied om 4 up o 20 kPa, while we suspec a sligh
o e es ima ion o Csby he me hod we applied. Values o ϕ anged om 25oup o 51oand alues o ρs anged
om 800 up o 1800 kg m−3. Thus, we es ima ed a alue ange o soil cohesion o Cs= 5 ±3 kPa and o he
in e nal ic ion angle o ϕ= 30 ±4owhich lie wi hin ou measu emen anges and comply wi h alues used by
Collison and Ande son (1996). Simila , we assumed a alue ange o soil densi y o ρs= 1,400 ±100 kg, as
de i ed om mean alues and s anda d e o s o mean o ou measu emen s.
As a esul , calcula ed alue anges o he c i ical soil hickness co e he obse ed dep h o ailu e in all bu
h ee cases (#4, #5 and #8 in Fig. 5, igh ), which we e he shallowes landslides in ou s udy. He e ac o s o
sa e y also did no each a c i ical alue below one.
Figu e 5: Fac o s o sa e y (F oS, le ) and c i ical minimum soil hickness (hc i .
s, igh ). The ho izon al g ey line
ma ks he c i ical F oS o one and he black do s ep esen obse ed landslide dep hs (hs).
10
Table 3: Soil p o iles (pa 1). Numbe o associa ed landslide, e ical soil dep h (z), Ho izon, Tex u e and
Consis ency sho cu s a e Jahn e al. (2006), size (F: 2..6 mm; M: 6..20 mm; C: 20..60 mm; S: 60..200
mm) and abundance (N-none; V: 0..2%; F:2..5%; C:5..15%; M:15..40%; A:40..80%; D:<80%; S - S one
line) o soil skele on, Soil cohesion as measu ed by o sion p obe (Cs; Geono H-60; mean ±s anda d
e o o mean), d y soil densi y (ρs; mean ±s anda d e o o mean; n= 3) and g a ime ic soil wa e
con en (Θ; mean ±s anda d e o o mean; n= 3).
#zSymbol Tex u e Consis . Skele on CsρsΘ
[cm] Size/Abundance [kPa] [ m−3] [%]
1 25 O - - -
1 20 Oe - - -
1 -20 Ah Sil ko4 FM/F 25.3±3.1; n=3 1.04±0.15 34.4±4.0
1 -50 A CL ko4 FM/F 37.7±7.5; n=3 1.18±0.02 26.8±3.1
1 -63 Bw Sil ko4 FM/F 23.0±3.6; n=3 1.39±0.04 19.0±0.8
1 -80 B LS so FM/V 23.7±10.8; n=3 1.45±0.10 18.6±2.5
1 -115 Cb LS so FM/S - 1.50±0.09 17.8±0.9
1 -135 Cw LS lo FM/M 23.3±8.1; n=3 - -
1 -200 C ko2 S 47.7±7.1; n=3 - -
2 15 O - - -
2 8 Oe - - -
2 -20 Ah SL ko4 N 18.7±4.0; n=3 1.41±0.03 24.3±0.5
2 -80 Bw Sil ko2 M/V 33.3±11.9; n=3 1.50±0.09 19.2±1.4
2 -90 B SL ko1 N 29.7±5.1; n=3 1.60±0.03 17.8±1.0
2 -100 Cb S 43.3±6.1; n=3 - -
2 -145 Bwb LS ko1 FM/C 45.7±9.5; n=3 1.51±0.08 16.2±1.9
2 -160 Bw LS ko1 M/V 41.7±3.2; n=3 1.50±0.06 17.6±0.2
2 -225 Cw S 40.7±22.1; n=3 - -
3 15 O - - -
3 10 Oe - - -
3 -5 Ah 16.0±2.8; n=2 - -
3 -30 Ah FS MC/D 41.3±5.0; n=3 - -
3 -57 Bw FS ko1 FM/F 46.0±1.7; n=3 1.68±0.04 16.3±0.3
3 -102 Cw VFS ko1 CS/A 63.3±6.1; n=3 1.56±0.07 17.9±1.3
3 -130 C FS ko1 N 46.0±2.0; n=3 1.72±0.04 16.8±0.1
4 -10 Ah L ko4 N 44.3±13.1; n=3 1.44±0.15 20.6±5.1
4 -140 Cw FS ko3 N 63.7±22.2; n=3 1.49±0.02 20.0±0.3
5 10 O - - -
5 -20 Ah SiL ko4 N 14.5±2.8; n=3 - -
5 -100 Cw - S - - -
17

Table 4: Soil p o iles (pa 2). Numbe o associa ed landslide, e ical soil dep h (z), Ho izon, Tex u e and
Consis ency sho cu s a e Jahn e al. (2006), size (F: 2..6 mm; M: 6..20 mm; C: 20..60 mm; S: 60..200
mm) and abundance (N-none; V: 0..2%; F:2..5%; C:5..15%; M:15..40%; A:40..80%; D:<80%; S - S one
line) o soil skele on, Soil cohesion as measu ed by o sion p obe (Cs; Geono H-60; mean ±s anda d
e o o mean), d y soil densi y (ρs; mean ±s anda d e o o mean; n= 3) and g a ime ic soil wa e
con en (Θ; mean ±s anda d e o o mean; n= 3).
#zSymbol Tex u e Consis . Skele on CsρsΘ
[cm] Size/Abundance [kPa] [ m−3] [%]
6 15 O - - -
6 -20 Ah L ko4 FM/C 22.7±1.6; n=3 0.80±0.17 45.3±7.9
6 -75 Cw SCL ko5 MC/A 41.3±25.3; n=3 1.30±0.00 24.1±0.1
6 -110 C - S 70.0±17.8; n=3 1.50 16.8
7 30 O - - -
7 5 Oe - - -
7 -9 Ah L ko4 N 13.3±1.5; n=3 1.04±0.17 36.0±4.6
7 -40 Bw SiL ko4 M/F 10.8±3.8; n=3 1.00±0.34 37.7±12.7
7 -56 Cw L ko3 CS/A 27.8±2.0; n=3 1.31±0.31 27.7±10.1
7 -95 C S 48.7±17.8; n=3 - -
8 150 O - - -
8 -22 Ah L ko4 FM/F 32.7±3.2; n=3 1.20±0.13 26.5±4.7
8 -75 Cw S - - -
9 30 O - - -
9 -20 Ah SC ko4 C/A 15.8±3.4; n=3 1.04±0.18 29.9±8.9
9 -45 Bw LS ko1 MC/C 14.5±8.2; n=3 1.47±0.08 13.1±2.3
9 -55 Bs LS ko1 MC/C 9.7±4.7; n=3 1.35±0.12 14.4±1.6
9 -150 Cb LS ko1 MC/C 20.0±7.8; n=3 1.43±0.08 10.8±1.3
9 -158 Ab LS ko1 F/M 43.3±13.3; n=3 1.35±0.06 12.1±1.0
9 -190 Bwb Si ko1 S 19.0±3.0; n=3 1.38±0.02 38.1±1.6
9 -230 Bsh Si ko1 F/F 15.3±2.8; n=3 1.15±0.25 43.4±6.0
9 -235 Cw Si ko1 F/F 11.8±1.3; n=3 1.31 39.0
9 -290 C Si ko1 /M 17.3±2.6; n=3 1.44±0.03 35.5±5.7
10 5 O - - -
10 -72 Ah L ko3 C/M 8.1±2.1; n=3 0.90±0.13 36.5±1.7
10 -105 Cw SCL ko4 C/M 23.7±1.6; n=3 1.32±0.06 19.1±2.1
18
Table 5: In si u soil shea es s. Associa ed landslide numbe , e ical dep h wi hin he soil p o ile (z), o al
numbe o shea es s (n), numbe o di e en loads (nLoads), soil cohesion (Cs; mean ±s anda d e o
o mean) and in e nal ic ion angle (ϕ; mean ±s anda d e o o mean) calcula ed by linea eg ession
(Eq. 3) and mean soil cohesion, measu ed wi h o sion p obe Geono H-60 (C ane).
# Landslide z n nLoads Csϕ C ane
[cm] [kPa] [o] [kPa]
1 1 40 8 4 11.16±2.70 37.5±6.9 37.67
2 1 70 8 4 7.08±0.93 46.3±2.1 23.67
3 1 90 8 4 5.22±2.11 43.0±4.7 16.00
4 1 95 7 2 12.44±1.39 40.0±1.6 28.29
5 1∗125 14 2 4.65±0.97 25.8±1.9 11.64
6 2∗30 6 3 10.21±2.19 50.9±4.3 33.33
7 5∗20 30 5 4.96±6.32 46.9±7.7 14.50
8 6∗30 6 3 14.31±1.13 41.4±5.1 31.67
9 7 23 6 3 15.06±1.97 32.8±8.6 14.00
10 7∗40 6 3 12.86±1.50 46.4±4.8 20.00
11 9 6 6 2 20.79±4.99 27.7±9.2 15.92
12 9 18 6 2 14.61±0.86 30.1±1.6 16.08
13 9 130 8 3 4.53±2.10 32.6±3.5 20.00
14 9 160 9 3 5.31±1.88 37.2±2.7 31.17
15 9∗185 10 4 7.69±2.43 27.0±4.4 42.00
16 11∗120 15 3 16.17±1.91 40.2±3.1 55.00
17 11∗,a 120 12 3 15.23±3.26 24.7±7.1 -
∗Measu emen in es ima ed ailu e plane; aSa u a ed condi ions
19
Table 6: In si u o ganic laye up u e es s. Loca ion id, minimum (τmin) and maximum (τmax) ensile esis ance,
mean alue (τ, )
Loca ion τmin τmax τlog(τ)n
[kPa] [kPa] [kPa] log([kPa])
a 0.89 3.25 2.41±0.88 0.79±0.49 11
Ø 0.89 3.25 2.41±0.88 0.79±0.49 11
b 0.32 5.47 1.89±1.01 0.50±0.53 205
b 0.32 6.24 2.06±1.11 0.58±0.54 101
b 0.89 7.45 2.69±1.04 0.92±0.39 150
b 0.76 11.84 2.29±1.28 0.72±0.44 150
Ø 0.32 11.84 2.22±1.15 0.67±0.50 606
c 0.76 5.60 2.63±1.15 0.87±0.46 39
c 0.48 7.48 2.09±1.36 0.55±0.64 97
c 0.64 3.82 2.04±0.71 0.65±0.37 51
c 0.32 7.51 1.84±1.17 0.45±0.57 99
Ø 0.32 7.51 2.07±1.19 0.58±0.57 286
20
Table 7: Poin in es iga ions, o de ed by ele a ion. Geog aphic coo dina es (UTM WGS84), ele a ion a.s.l. (El .),
slope angle (α), hickness o o ganic laye (ho; mean ±s anda d e o o mean; numbe o samples),
es ima ed mean ee heigh (h), ee diame e in b eas heigh (130 cm; DBH; mean ±s anda d e o o
mean; numbe o samples) and abo eg ound ee biomass densi y es ima es a e equa ion 4 (Ma
B, B own
and I e son, 1992), equa ion 5 (Mb
B, B own, 1997) and a e equa ion 6 (Mc
B). Below each al i udinal
ange, mean alues a e gi en.
# UTM WGS84 El . α hoh DBH Ma
BMb
BMc
B
[m] [m a.s.l.] [o] [cm] [m] [cm] [g m−2] [g m−2] [g m−2]
1 714,150/9,558,405 2,531 38 12.1±5.7; n=2 4.0 12.57±7.16; n=4 974 3,200 1,875
2 714,162/9,558,420 2,524 30 - 5.0 8.48±7.04; n=8 1,139 1,953 2,130
3 714,133/9,558,397 2,514 28 17.0±7.2; n=3 3.5 11.94±8.77; n=6 1,166 4,129 2,218
4 714,089/9,558,513 2,505 56 33.7±2.3; n=3 3.5 5.37±1.82; n=35 1,494 2,915 2,615
5 714,089/9,558,513 2,505 36 33.7±2.3; n=3 3.0 4.97±1.60; n=18 574 1,421 989
6 714,090/9,558,399 2,492 18 34.3±11.7; n=3 4.0 12.14±6.05; n=8 1,821 5,777 3,494
7 714,069/9,558,390 2,492 24 54.5±38.1; n=3 3.0 10.69±7.23; n=7 953 3,469 1,778
8 713,747/9,558,033 2,478 40 68.8; n=1 12.0 15.01±7.72; n=6 5,793 7,902 12,032
9 713,724/9,558,063 2,469 35 58.7±14.3; n=2 7.0 12.22±8.79; n=5 1,961 3,687 3,877
10 714,047/9,558,399 2,466 40 40.5±0.0; n=2 4.0 7.27±3.85; n=6 517 930 940
11 714,065/9,558,476 2,440 36 42.6±7.1; n=3 6.0 6.53±3.36; n=19 1,963 2,229 3,608
12 713,999/9,558,377 2,436 49 21.0±4.6; n=2 7.0 8.55±2.91; n=8 1,595 2,010 3,038
13 713,970/9,558,355 2,404 47 33.2±3.4; n=2 3.0 4.77; n=1 30 78 51
14 713,922/9,558,338 2,382 51 24.3±0.0; n=2 5.5 7.51±3.16; n=12 1,486 2,039 2,757
Ø Range: 2,382 o 2,531 37.5 35.3±18.2; n=31 5.0 7.84±5.23; n=143 1,310 2,044 2,430
15 713,971/9,559,299 2,331 8 30.4±9.3; n=17 6.0 8.63±4.06; n=19 3,326 4,900 6,291
16 713,986/9,559,334 2,325 45 61.2±34.8; n=13 4.0 9.43±3.58; n=12 1,693 4,065 3,164
17 713,992/9,559,348 2,303 49 49.4±15.5; n=17 4.0 8.97±3.70; n=23 2,950 6,679 5,485
18 713,936/9,559,392 2,295 28 23.6±6.6; n=11 10.0 9.16±5.97; n=32 10,187 9,921 19,911
19 713,976/9,559,428 2,294 43 41.0±23.6; n=3 4.0 8.79±2.47; n=8 989 2,187 1,834
20 713,976/9,559,446 2,287 36 15.1±12.0; n=3 3.0 7.43±3.80; n=3 205 494 368
21 713,959/9,559,390 2,285 36 14.6±4.7; n=10 5.0 10.36±4.87; n=24 4,999 10,826 9,545
22 714,014/9,559,348 2,282 48 18.4±6.6; n=18 8.0 13.18±6.31; n=21 10,783 19,227 21,634
23 713,860/9,559,631 2,264 36 21.7±6.0; n=17 2.5 5.53±1.10; n=11 361 947 623
24 713,857/9,559,681 2,256 28 25.6±4.6; n=12 5.0 7.60±3.33; n=17 1,969 2,997 3,642
25 713,864/9,559,655 2,252 36 40.5±12.4; n=17 6.0 7.06±4.71; n=16 1,915 2,289 3,548
26 713,953/9,559,686 2,243 40 21.0±1.1; n=2 6.0 10.47±4.78; n=9 2,274 4,192 4,388
27 713,918/9,559,725 2,236 10 28.3±0.0; n=2 6.4 10.26±4.09; n=13 3,360 5,696 6,492
28 713,896/9,559,704 2,236 15 48.3±14.4; n=18 6.0 9.15±7.93; n=44 8,613 13,596 16,391
29 713,972/9,559,770 2,232 36 61.1±34.9; n=2 5.0 6.57±1.34; n=7 615 833 1,120
30 713,932/9,559,718 2,229 40 20.2±5.7; n=2 5.0 9.68±3.78; n=15 2,746 5,500 5,206
31 713,982/9,559,756 2,228 35 31.8±43.3; n=3 5.0 7.87±1.91; n=7 865 1,365 1,606
32 713,847/9,559,721 2,225 18 14.1±4.9; n=17 6.0 11.01±5.84; n=24 6,669 13,000 12,937
33 713,857/9,559,739 2,222 24 58.3±21.7; n=18 7.0 7.26±4.17; n=31 4,527 4,789 8,478
34 713,965/9,559,412 2,214 52 25.9±5.1; n=9 10.0 11.65±7.75; n=18 9,043 11,544 18,123
35 713,819/9,559,775 2,208 44 58.0±20.8; n=18 7.0 7.67±3.49; n=28 4,543 5,073 8,556
36 713,759/9,559,819 2,192 36 54.1±13.9; n=18 7.0 9.60±4.33; n=20 4,959 7,150 9,559
Ø Range: 2,192 o 2,331 33.8 37.2±22.2; n=247 5.8 9.15±5.29; n=402 3,459 5,627 6,571
37 713,234/9,560,163 1,915 36 32.4; n=1 5.0 6.87±2.92; n=37 3,533 4,913 6,464
38 713,248/9,560,451 1,843 54 19.8±10.7; n=13 7.0 10.79±4.97; n=18 5,570 9,176 10,870
39 713,179/9,560,469 1,809 45 25.5±8.1; n=14 6.0 8.94±3.82; n=22 4,121 6,331 7,824
Ø Range: 1,809 o 1,915 45.0 23.1±9.7; n=28 6.0 8.38±4.02; n=77 4,897 6,970 9,234
21

Chap e 4
Conclusion and Ou look
Below we discuss he main indings o his hesis along i s wo main hemes: i s he ole o
li e his o y ai s o species coexis ence, and hen hei ole in o es eco e y a e landslides.
Each sec ion is o ganized acco ding o he ques ions s a ed in he in oduc ion and closes wi h
po en ial s a ing poin s o u he esea ch a ising om ou indings. In he ou look, we ou line
u u e esea ch ac i i ies based on his hesis.
4.1 Role o li e his o y ai s o species coexis ence
Which mechanisms enable coexis ence in species ich plan communi ies?
The ques ion o species coexis ence is o gene al na u e and applies o many ecosys ems – pa -
icula ly opical o es s. In he ligh o insu icien knowledge abou such di e se ecosys ems,
heo e ical in es iga ions o po en ial p ocesses ha d i e coexis ence o ees migh be a key
o a be e unde s anding o compe i ion. The e o e, in pape 1 we ook an idealized app oach
which ex ends beyond he speci ic s udy sys em ha mo i a ed ou ques ion, and de eloped a
heo e ical model o in es iga e he ques ion o species coexis ence in o es s. While many model
s udies in es iga ing species coexis ence ocus on single coexis ence mechanisms (e.g. Po ho
e al., 2006; P onk e al., 2007), he ques ion how di e en mechanisms in e ac emains open.
We he e o e analyzed ade-o s be ween di e en species ai s ( ee ec ui men , ee g ow h,
ee mo ali y), alone and in combina ion wi h addi ional mechanisms ha modi y local compe-
i ion (densi y-dependen mo ali y, ligh -dependen egene a ion) o hei po en ial o acili a e
species coexis ence. To his end, we de eloped a simple spa ially explici o es model. Ou main
indings om his s udy a e:
•The combina ion o li e his o y ai s is c ucial o species coexis ence: we ind e y na ow
coexis ence anges o a simula ion pe iod o 1000 yea s (Figu e 4 in pape 1). This means
ha ade-o s need o be ’ ine- uned’, i.e show a de ined ela ion, in o de o acili a e
coexis ence. Al eady small de ia ions om his ela ion esul in compe i i e exclusion.
Howe e , he conside ed ade-o s alone do no suppo long- e m coexis ence as hey
ha e no s abilizing e ec – hey ac only equalizing (Figu e 7 in pape 1).
•The na ow coexis ence anges o ade-o communi ies a e conside ably b oadened by he
inclusion o p ocesses ha modi y local compe i ion (densi y-dependen mo ali y, ligh -
91
4 Conclusion and Ou look
dependen egene a ion, c . Figu e 4 and 7 in pape 1). Such mechanisms ac s abilizing
and cons i u e an impo an con ibu ion o coexis ence in o es communi ies.
•The imescale on which di e en species ai s ope a e in compe i ion can di e consid-
e ably. A ade-o be ween one ai , which has a s onge e ec on compe i ion and a
second ai which has a weake e ec can esul in highly non-linea coexis ence a eas in
he ai space – hus he e migh be h eshold alues abo e o below which we canno
expec unc ioning ade-o s be ween s ong and weak ai s (Figu e 4 b,c in pape 1).
Ou in es iga ion e ealed ha he conside ed ade-o s alone a e insu icien o explaining
long- e m coexis ence and hus species di e si y. In ac , we we e su p ised o see how apidly
(wi hin ew gene a ions) compe i i e exclusion akes place i ade-o s a e no ’ ine- uned’, i.e.
show a de ined ela ion (c . Figu e 2a in pape 1). Consequen ly, in he absences o in e nal
(e.g. densi y-dependence, specia ion) o ex e nal (e.g. dis u bance, immig a ion o new species)
s abilizing mechanisms, species could no coexis in he long un. On he o he hand, simple
mechanisms ha modi y he local compe i ion (e.g. densi y-dependen mo ali y) could be su i-
cien o bu e apid exclusion. Only ew s udies examining ade-o s men ion he insu iciency
o ade-o s alone o suppo coexis ence (Chesson, 2000; Lischke and Lö le , 2006; Bani z e al.,
2008). In ac , one p ominen class o models in es iga ing he compe i ion-coloniza ion ade-o
(Tilman, 1994; Klausmeie and Tilman, 2002; Mulle -Landau, 2008) ha e seemingly showed he
opposi e – namely, ha ade-o s can acili a e coexis ence. This seeming con adic ion can
be se led when examining he mechanisms ac ing in he di e en models. Mos s udies o he
compe i ion-coloniza ion ade-o implemen an a p io i hie a chy o local compe i ion, which is
a s abilizing mechanism ha may lead o coexis ence o a po en ially unlimi ed numbe o species
(Tilman, 1994). In con as , he ade-o s in physiological ai s conside ed in pape 1 do no
bea a s abilizing componen , and i ness di e ences e ol e di ec ly om physiological species
ai s. Compa ed o he a he indi ec pa ame e s o compe i ion and coloniza ion s eng h o
he abo emen ioned models, ou p ocess-based app oach uses mo e di ec pa ame e s (g ow h,
mo ali y, dispe sal) which can mo e easily be obse ed in he ield.
In di e ence o many classical s udies on species coexis ence ha ocus on equilib ium s a es
o ecosys ems (e.g. Cha e e al., 2002) and o en use spa ially implici app oaches (e.g. Le ins
and Cul e , 1971; Du e and Le in, 1994; Tilman, 1994), ou model akes a non-equilib ium
pe spec i e and is spa ially explici . The eby, we ake in o accoun , i s , ha space plays an
impo an ole in o es ecosys ems (as hey a e no ’well-mixed’ sys ems) and second, ha hese
sys ems a e also a ec ed by bo h dis u bances and en i onmen al (e.g. clima ic) changes on di -
e en spa io empo al scales, and he e o e a e likely ne e in a eal equilib ium s a e (a medium
scales).
Implica ions o ou s udy sys em
How do he indings om pape 1 ela e o ou s udy sys em, he species- ich mon ane o es o
he T opical Andes? A la ge spa ial scales, a ious biogeog aphical hypo heses y o explain
he high species ichness in he opics compa ed o lowe species ichness in highe la i udes (e.g.
Willig e al., 2003; K e and Je z, 2007). Clima e is in ol ed in di e en ways in such hypo he-
ses: on he one hand, con empo a y clima e (e.g. empe a u e, wa e a ailabili y) is assumed
92
4 Conclusion and Ou look
o be a d i e o la i udinal di e si y pa e ns (e.g. wa e -ene gy hypo hesis1; Hawkins e al.,
2003; F ancis and Cu ie, 2003). On he o he hand, he clima e-s abili y hypo hesis assumes
ha his o ical dynamics in clima e egimes (glacial-in e glacial cycles) go e n di e si y pa e ns
(McGlone, 1996; Dynesius and Jansson, 2000; Jablonski e al., 2006). Na u ally, hese explana-
ions could apply o ou opical s udy sys em. Addi ionally o hese la ge scale explana ions
ha gene ally apply o low la i udes, moun ain ecosys ems, due o hei complex opog aphy,
a e cha ac e ized by high le els o en i onmen al he e ogenei y and a b oad ange o clima ic
condi ions compa ed o lowland ecosys ems. This en i onmen al he e ogenei y mani es s in a
di e si y o habi a ypes ha is assumed o posi i ely a ec species di e si y (e.g. Hus on,
1994; Rosenzweig, 1995; Du ou e al., 2006). Mo eo e , ou s udy si e is loca ed in a unique
a ea, he Amo ape-Huancabamba dep ession. This dep ession is a pa ial in e up ion o he
Andean moun ain chain wi h he lowes poin a 2145 m asl in no he n Pe u. The Amo ape-
Huancabamba dep ession has been sugges ed bo h o be a mig a ion co ido (Weigend, 2002;
Beck e al., 2008a) be ween he Amazon and he paci ic side and a mig a ion ba ie in no h-
sou h di ec ion (Bo chsenius, 1997; Kea ing, 2008). The ex ao dina y high species di e si y has
been inc easingly ecognized in he las yea s (Young and Reynel, 1997; Kea ing, 2008; Rich e
e al., 2009) and sugges s ha his egion is a mee ing poin o lowland and upland species (Beck
and Rich e , 2008)
The abo emen ioned la ge scale hypo heses explain a he he o ma ion o di e si y bu no
he main enance o species ichness. He e, he smalle scale o local plan communi ies, whe e
species in e ac ions like compe i ion, p eda ion and spa ial he e ogenei y a e impo an , come
in o play. The e o e, in ou s udy, we ocused on in e ac ions be ween indi iduals and hei
e ec on popula ion dynamics. In he we opics, s able and a ou able clima e bo h wi hin
he yea and his o ically o e long ime pe iods migh ha e allowed o a s ong specializa ion o
ai s, esul ing in na owe pa i ioning o niches compa ed o ou side he opics (MacA hu
(1972); May (1973), bu see also Vazquez and S e ens (2004)). Bu ou indings ( om he mo e
local scale) sugges ha addi ional ac o s migh be necessa y o explain he main enance o
di e si y in species- ich habi a s like ou s udy a ea. Such addi ional ac o s could include bio ic
mechanisms e ol ing om in e - o in a-speci ic in e ac ions (e.g. densi y-dependen mo ali y),
bu hey could also be connec ed o abio ic ac o s like opog aphy (spa ial he e ogenei y) o
dis u bances (e.g landslides).
Impo ance o unde s anding mechanisms ha main ain biodi e si y
Biodi e si y has bo h di ec and indi ec alues o humans. Di ec alues conce n economically
impo an en i onmen al se ices like imbe , medicinal and ood p oduc s. Indi ec alues a e
he main enance o biogeochemical cycles and ecosys em unc ions like ca bon seques a ion and
egula ion o egional clima e. O he indi ec alues a e o e hical (’in insic alue’ o biodi e -
si y) and aes he ical na u e (Eh lich and Eh lich, 1992). O e he pas decades, he implica ions
1The wa e -ene gy hypo hesis s a es ha "species ichness a highe la i udes is con olled by he
a ailabili y o ambien hea , whe eas, in he he mally sui able opics, wa e - and humidi y- ela ed
a iables a e he main d i ing ac o s"(K e and Je z, 2007). I is known ha plan species dis-
ibu ion and ichness is in luenced by he ac o s ene gy (e.g. empe a u e, insola ion, po en ial
e apo anspi a ion) and wa e , which s ongly in luence he p oduc i i y o a sys em (S ephenson,
1990; O’B ien, 1993). Po en ial mechanisms leading o high species ichness unde ambien ene gy
and wa e a e low ex inc ion a es and high specia ion a es (Hawkins e al., 2003).
93
4 Conclusion and Ou look
o biodi e si y o ecosys em unc ions, in pa icula he ole o biodi e si y o ecosys em p o-
duc i i y and s abili y, also wi h espec o clima e change, ecei ed g owing scien i ic a en ion
(e.g. Hoope e al., 2005; Nad owski e al., 2010; Paque e and Messie , 2011). Acknowledging
he impo ance o biodi e si y and i s apid loss due o human in e en ions (Bu cha e al.,
2010) gi e ise o he u gen need o concep s o educe biodi e si y loss (Noss, 2001; Helle
and Za ale a, 2009). Fo he de elopmen o such concep s, he unde s anding o mechanisms
main aining biodi e si y a e indispensable.
Whe e o go om he e?
One s eng h o he modelling app oach chosen in pape 1 is i s lexibili y – i can, o example,
easily be ex ended o in es iga e o he coexis ence mechanisms. Po en ial di ec ions o u he
explo a ions a e:
•Upscaling: om ew o mul iple species compe i ion. An ex ension o a mul i-species
communi y may p o ide u he insigh s in o how he mechanisms ha we iden i ied as
impo an o he wo-species sys em ansla e in o species- ich communi ies. A i s s ep
in his di ec ion has been unde aken in he end o pape 1 (c . Figu e 5).
•Downscaling: inco po a ing in a-speci ic di e ences. Recen s udies sugges in a-speci ic
a ia ion o be highly impo an o species coexis ence (Adle e al., 2007; Cla k, 2010;
Cla k e al., 2010). In ou s udy we ocused on di e ences be ween species, ye , in p inciple
ou indi idual-based app oach also allows inco po a ing di e ences be ween indi iduals o
he same species.
•F om heo e ical owa ds mo e applied models: inco po a ing en i onmen al he e ogenei ies.
One could inco po a e spa ial he e ogenei y (e.g. moun ains, landslides) o o he po en-
ially ele an g adien s (e.g. nu ien s) and inco po a e species-speci ic di e ences in e-
sponse o such he e ogenei ies. Such an app oach may p o ide insigh s in o u he mech-
anisms ha main ain biodi e si y and ha a e pa icula ly ele an o moun ainous sys-
ems like ou s udy sys em.
4.2 Fo es dynamics, landslide dis u bance and li e his o y
ai s
How do mon ane o es s eac on he dis u bance o landslides?
Landslides ha e ecei ed much less scien i ic a en ion han o he dis u bances in o es s such as
gap-building and i e. Ye , hey a e a common dis u bance in many mon ane o es s (Ga wood
e al., 1979; Res epo e al., 2009). In addi ion, he majo i y o landslide esea ch ocused on
local p ocesses on he le el o single landslides (e.g. Dalling, 1994; Chaudh y e al., 1996; Fe che
e al., 1996; Kessle , 1999; Ohl and Bussmann, 2004; Velazquez and Gomez-Sal, 2008). While
hese in es iga ions a e impo an o gain knowledge abou he eac ion o di e en species on
his pa icula dis u bance, he e ec o landslides om he landscape pe spec i e was so a
a he neglec ed (Res epo e al., 2009). Thus, ou s udy con ibu es o unde s anding he e -
ec s o landslides on o es s uc u e and dynamics on he local and landscape scale.
On landslide si es, changed en i onmen al condi ions like educed nu ien con en s and soil
ins abili y a e likely o esul in slowe o es eco e y compa ed o o he well-s udied dis u -
bances such as ee- all gaps. Depending on he changes in ee li e his o y ai s due o al e ed
94
4 Conclusion and Ou look
o ees o p omo e economically aluable species and en ichmen plan ing o selec ed species
(Gün e e al., 2004; Agui e e al., 2006; Cab e a Cisne os e al., 2006; Mosandl and Gün e ,
2008). Ou dynamic o es model p o ides a sui able amewo k o compa e di e en managemen
op ions conce ning hei long- e m impac on o es s uc u e and composi ion (e.g. Hu h and
Di ze , 2001; Kammesheid e al., 2002; Rüge e al., 2007a). Combining ecological models like
ou o es model wi h economic models ha e alua e he economic bene i s o di e en land
use op ions will enable he de elopmen o en i onmen al policies ha balance ecological and
economic bene i s o land use.
101

Bibliog aphy
Adle PB, HilleRisLambe s J, Le ine JM, 2007. A niche o neu ali y. Ecol Le 10, 95–104.
Agui e N, Gün e S, Webe M, S imm B, 2006. En ichmen o Pinus pa ula plan a ions wi h
na i e species in sou he n Ecuado . Lyonia 10, 33–45.
Alcan a a-Ayala I, Es eban-Cha ez O, Pa o JF, 2006. Landsliding ela ed o land-co e change:
A diach onic analysis o hillslope ins abili y dis ibu ion in he Sie a No e, Puebla, Mexico.
Ca ena 65, 152–165.
Ash on PS, 1993. Conse a ion Biology. Chapman and Hall, New Yo k. chap e Species ichness
in plan communi ies. pp. 4–22.
A iwill PM, 1994. The dis u bance o o es ecosys ems - he ecological basis o conse a i e
managemen . Fo Ecol Manag 63, 247–300.
Bani z T, Hu h A, G imm V, Johs K, 2008. Clumped e sus sca e ed: how does he spa ial
co ela ion o dis u bance e en s a ec biodi e si y? Theo Ecol 1, 231–240.
Beck E, Ko ke I, Bendix J, Makeschin F, Mosandl R, 2008a. G adien s in a T opical Moun ain
Ecosys em o Ecuado . Sp inge , Be lin. olume 198 o Ecological S udies. chap e G adien s
in a T opical Moun ain Ecosys em - a Syn hesis. pp. 467–479.
Beck E, Mosandl R, Rich e M, Ko ke I, 2008b. G adien s in a T opical Moun ain Ecosys-
em o Ecuado . Sp inge , Be lin. olume 198 o Ecological S udies. chap e The in es iga ed
g adien s. pp. 55–62.
Beck E, Rich e M, 2008. The T opical Moun ain Fo es - Pa e ns and P ocesses in a Biodi e si y
Ho spo . Gö ingen Cen e o Biodi e si y and Ecology. olume 2 o Biodi e si y and Ecology
Se ies. chap e Ecological aspec s o a biodi e si y ho spo in he Andes o sou he n Ecuado .
pp. 195–217.
Bee C, Reichs ein M, Tomelle i E, Ciais P, Jung M, Ca alhais N, Rodenbeck C, A ain MA,
Baldocchi D, Bonan GB, Bondeau A, Cesca i A, Lasslop G, Lind o h A, Lomas M, Luyssae
S, Ma golis H, Oleson KW, Roupsa d O, Veenendaal E, Vio y N, Williams C, Woodwa d FI,
Papale D, 2010. Te es ial g oss ca bon dioxide up ake: Global dis ibu ion and co a ia ion
wi h clima e. Science 329, 834–838.
103
BIBLIOGRAPHY
Bendix J, Rollenbeck R, Rich e M, Fabian P, Emck P, 2008. G adien s in a T opical Moun ain
Ecosys em o Ecuado . Sp inge , Be lin. olume 198 o Ecological S udies. chap e Clima e.
pp. 63–73.
Bolke B, 2008. Ecological Models and Da a in R. P ince on Uni e si y P ess.
Bo chsenius F, 1997. Pa e ns o plan species endemism in ecuado . Biodi e s Conse 6,
379–399.
Bo ga M, Dalla Fon ana G, Da Ros D, Ma chi L, 1998. Shallow landslide haza d assessmen
using a physically based model and digi al ele a ion da a. En i on Geol 35, 81–88.
Bo ga M, Tonelli F, dalla Fon ana G, Cazo zi F, 2005. E alua ing he in luence o o es oads
on shallow landsliding. Ecol Model 187, 85–98.
Bossel H, K iege H, 1991. Simula ion-model o na u al opical o es dynamics. Ecol Model
59, 37–71.
B ehm G, Pi kin LM, Hil N, Fiedle K, 2005. Mon ane andean ain o es s a e a global di e si y
ho spo o geome id mo hs. J Biogeog 32, 1621–1627.
B ooks TM, Mi e meie RA, Mi e meie CG, da Fonseca GAB, Rylands AB, Kons an WR,
Flick P, Pilg im J, Old ield S, Magin G, Hil on-Taylo C, 2002. Habi a loss and ex inc ion in
he ho spo s o biodi e si y. Conse Biol 16, 909–923.
Bugmann H, 2001. A e iew o o es gap models. Clima e Change 51, 259–305.
Bussmann R, Wilcke W, Rich e M, 2008. G adien s in a T opical Moun ain Ecosys em o
Ecuado . Sp inge , Be lin. olume 198 o Ecological S udies. chap e Landslides as Impo an
Dis u bance Regimes - Causes and Regene a ion. pp. 319–330.
Bu cha SHM, Walpole M, Collen B, an S ien A, Scha lemann JPW, Almond REA, Baillie
JEM, Bomha d B, B own C, B uno J, Ca pen e KE, Ca GM, Chanson J, Chene y AM,
Csi ke J, Da idson NC, Den ene F, Fos e M, Galli A, Galloway JN, Geno esi P, G ego y
RD, Hockings M, Kapos V, Lama que JF, Le e ing on F, Loh J, McGeoch MA, McRae L,
Minasyan A, Mo cillo MH, Old ield TEE, Pauly D, Quade S, Re enga C, Saue JR, Skolnik
B, Spea D, S anwell-Smi h D, S ua SN, Symes A, Tie ney M, Ty ell TD, Vie JC, Wa son
R, 2010. Global biodi e si y: Indica o s o ecen declines. Science 328, 1164–1168.
Cab e a Cisne os O, Gün e S, Mosandl R, 2006. Dynamics in na u al and selec i ely logged
opical moun ain ain o es o Sou h Ecuado . Lyonia 10, 87–97.
Chaudh y S, Singh SP, Singh JS, 1996. Pe o mance o seedlings o a ious li e o ms on landslide-
damaged o es si es in Cen al Himalaya. J Appl Ecol 33, 109–117.
Cha e J, 1999. S udy o s uc u al, successional and spa ial pa e ns in opical ain o es s using
TROLL, a spa ially explici o es model. Ecol Model 124, 233–254.
Cha e J, Mulle -Landau HC, Le in SA, 2002. Compa ing classical communi y models: Theo e -
ical consequences o pa e ns o di e si y. Am Na 159, 1–23.
Chesson P, 2000. Mechanisms o main enance o species di e si y. Annu Re Ecol Sys 31,
343–366.
104
BIBLIOGRAPHY
Cla k J, Die ze M, Chak abo y S, Aga wal P, Ibanez I, LaDeau S, Wolosin M, 2007. Resol ing
he biodi e si y pa adox. Ecol Le 10, 647–659.
Cla k JS, 2010. Indi iduals and he a ia ion needed o high species di e si y in o es ees.
Science 327, 1129–1132.
Cla k JS, Bell D, Chu CJ, Cou baud B, Die ze M, He sh M, HilleRisLambe s J, Ibanez I, LaDeau
S, McMahon S, Me cal J, Mohan J, Mo an E, Pangle L, Pea son S, Salk C, Shen ZH, Valle
D, Wycko P, 2010. High-dimensional coexis ence based on indi idual a ia ion: a syn hesis
o e idence. Ecol Monog 80, 569–608.
Condi R, Hubbell SP, La ankie JV, Sukuma R, Manoka an N, Fos e RB, Ash on PS, 1996.
Species-a ea and species-indi idual ela ionships o opical ees: A compa ison o h ee
50-ha plo s. J Ecol 84, 549–562.
Connell J, 1971. Dynamics o popula ions. Cen e o Ag icul u al Publishing and Documen-
a ion, Wageningen, The Ne he lands. chap e On he ole o na u al enemies in p e en ing
compe i i e exclusion in some ma ine animals and in ain o es ees. pp. 298–313.
Connell JH, Sla ye RO, 1977. Mechanisms o succession in na u al communi ies and hei ole
in communi y s abili y and o ganiza ion. Am Na 111, 1119–1144.
Conse a ion In e na ional, . Biodi e si y ho spo s. www.biodi e si yho spo s.o g.
C awley M, 1997a. Plan Ecology. Blackwell Science, Ox o d. chap e S uc u e o Plan Com-
muni ies. pp. 475–531.
C awley M, 1997b. Plan Ecology. Blackwell Science, Ox o d. chap e Li e His o y and En i on-
men . pp. 73–131.
C uden D, 1991. A simple de ini ion o a landslide. Bulle in o he In e na ional Associa ion o
Enginee ing Geology 43, 27–29.
Dalling JW, 1994. Vege a ion coloniza ion o landslides in he Blue Moun ains, Jamaica. Bio op-
ica 26, 392–399.
Dalling JW, Tanne EVJ, 1995. An expe imen al s udy o egene a ion on landslides in mon ane
ain o es in Jamaica. J Ecol 83, 55–64.
Da win C, Wallace A, 1858. On he endency o species o o m a ie ies: and on he pe pe ua ion
o a ie ies and species by na u al means o selec ion. J Linnean Socie y (Zool) 3, 46–50.
Dhakal AS, Sidle RC, 2003. Long- e m modelling o landslides o di e en o es managemen
p ac ices. Ea h Su P oc Land 28, 853–868.
Di zo R, 2001. Global Biodi e si y in a Changing En i onmen . Sp inge , New Yo k. olume 152
o Ecological S udies. chap e T opical Fo es s. pp. 251–276.
Dislich C, Gün e S, Homeie J, Sch öde B, Hu h A, 2009. Simula ing o es dynamics o a
opical mon ane o es in Sou h Ecuado . E dkunde 63, 347–364.
Dislich C, Johs K, Hu h A, 2010. Wha enables coexis ence in plan communi ies? Weak e sus
s ong species ai s and he ole o local p ocesses. Ecol Model 221, 2227–2236.
105
BIBLIOGRAPHY
Dixon RK, B own S, Hough on RA, Solomon AM, T exle MC, Wisniewski J, 1994. Ca bon
pools and lux o global o es ecosys ems. Science 263, 185–190.
Dudek A, Ek A, 1980. A bibliog aphy o wo ldwide li e a u e on indi idual ee based s and
g ow h models. S a Pape Se ies, Dep. o Fo es Resou ces, Uni e si y o Minneso a 12, 1–34.
Du ou A, Gadallah F, Wagne HH, Guisan A, Bu le A, 2006. Plan species ichness and
en i onmen al he e ogenei y in a moun ain landscape: e ec s o a iabili y and spa ial con ig-
u a ion. Ecog aphy 29, 573–584.
Du e R, Le in S, 1994. The impo ance o being disc e e (and spa ial). Theo Popul Biol 46,
363–394.
Dynesius M, Jansson R, 2000. E olu iona y consequences o changes in species’ geog aphical
dis ibu ions d i en by Milanko i ch clima e oscilla ions. P oc Na l Acad Sci USA 97, 9115–
9120.
Eh lich PR, Eh lich AH, 1992. The alue o biodi e si y. Ambio 21, 219–226.
Ek A, Monse ud R, 1974. G ow h models o ee and s and simula ion. Royal College o Fo es y,
S ockholm, Sweden. olume 30 o Resea ch No es. chap e T ials wi h p og am FOREST:
G ow h and ep oduc ion simulal ion o mixed species e en- o une enaged o es s ands. pp.
56–73.
Elias RB, Dias E, 2009. E ec s o landslides on he moun ain ege a ion o Flo es Island, Azo es.
J Veg Sci 20, 706–717.
FAO, 2009. S a e o he Wo ld´s Fo es s 2009. Rome, I aly.
Fea nside PM, 2005. De o es a ion in B azilian Amazonia: His o y, a es, and consequences.
Conse Biol 19, 680–688.
Fe che N, Haines BL, Co de o RA, Lodge DJ, Walke LR, Fe nandez DS, Law ence WT, 1996.
Responses o opical plan s o nu ien s and ligh on a landslide in Pue o Rico. J Ecol 84,
331–341.
F ancis AP, Cu ie DJ, 2003. A globally consis en ichness-clima e ela ionship o angiospe ms.
Am Na 161, 523–536.
F olking S, Palace MW, Cla k DB, Chambe s JQ, Shuga HH, Hu GC, 2009. Fo es dis u -
bance and eco e y: A gene al e iew in he con ex o spacebo ne emo e sensing o impac s
on abo eg ound biomass and canopy s uc u e. J Geophys Res-Biogeosc 114, 1–27.
Ga wood NC, Janos DP, B okaw N, 1979. Ea hquake-caused landslides - a majo dis u bance
o opical o es s. Science 205, 997–999.
Gee sema M, Poja JJ, 2007. In luence o landslides on biophysical di e si y - A pe spec i e
om B i ish Columbia. Geomo phology 89, 55–69.
Gen y AH, 1988. Changes in plan communi y di e si y and lo is ic composi ion on en i on-
men al and geog aphical g adien s. Ann Mo Bo Ga d 75, 1–34.
Glade T, 2003. Landslide occu ence as a esponse o land use change: a e iew o e idence om
New Zealand. Ca ena 51, 297–314.
106

BIBLIOGRAPHY
Gün e S, S imm B, Webe M, 2004. Sil icul u al con ibu ions owa ds sus ainable managemen
and conse a ion o o es gene ic esou ces in sou he n ecuado . Lyonia 6 (1), 75–91.
G ace J, 2004. Unde s anding and managing he global ca bon cycle. J Ecol 922, 189–202.
G imm V, Railsback S, 2005. Indi idual-based Modeling and Ecology. P ince on Uni e si y
P ess.
G oene eld J, Al es LF, Be nacci LC, Ca ha ino ELM, Knogge C, Me zge JP, Pü z S, Hu h A,
2009. The impac o agmen a ion and densi y egula ion on o es succession in he A lan ic
ain o es . Ecol Model 220, 2450–2459.
G o e R, E ha d M, 1999. Simula ion o ee and s and de elopmen unde di e en en i on-
men al condi ions wi h a physiologically based model. Fo Ecol Manag 120, 59–76.
Gua igua a MR, 1990. Landslide dis u bance and o es egene a ion in he uppe Luquillo
Moun ains o Pue o Rico. J Ecol 78, 814–832.
Gu ie ez AG, A mes o JJ, A a ena JC, Ca mona M, Ca asco NV, Ch is ie DA, Pena MP, Pe ez
C, Hu h A, 2009. S uc u al and en i onmen al cha ac e iza ion o old-g ow h empe a e
ain o es s o no he n Chiloe Island, Chile: Regional and global ele ance. Fo Ecol Manag
258, 376–388.
Guzze i F, Ca a a A, Ca dinali M, Reichenbach P, 1999. Landslide haza d e alua ion: a e iew
o cu en echniques and hei applica ion in a mul i-scale s udy, Cen al I aly. Geomo phol-
ogy 31, 181–216.
Hansen MC, S ehman SV, Po apo PV, 2010. Quan i ica ion o global g oss o es co e loss.
P oc Na l Acad Sci USA 107, 8650–8655.
Ha ig F, Dislich C, Hu h A, in p epa a ion. App oxima e bayesian calib a ion by simula ed
pseudo-likelihoods o a s ochas ic model o opical o es s .
Has ings A, 1980. Dis u bance, coexis ence, his o y, and compe i ion o space. Theo Popul
Biol 18, 363–373.
Hawkins BA, Field R, Co nell HV, Cu ie DJ, Guegan JF, Kau man DM, Ke JT, Mi elbach
GG, Obe do T, O’B ien EM, Po e EE, Tu ne JRG, 2003. Ene gy, wa e , and b oad-scale
geog aphic pa e ns o species ichness. Ecology 84, 3105–3117.
Hec o A, Schmid B, Beie kuhnlein C, Caldei a MC, Dieme M, Dimi akopoulos PG, Finn JA,
F ei as H, Gille PS, Good J, Ha is R, Hogbe g P, Huss-Danell K, Joshi J, Jumpponen A,
Ko ne C, Leadley PW, Lo eau M, Minns A, Mulde CPH, O’Dono an G, O way SJ, Pe ei a
JS, P inz A, Read DJ, Sche e -Lo enzen M, Schulze ED, Siaman ziou as ASD, Spehn EM,
Te y AC, T oumbis AY, Woodwa d FI, Yachi S, Law on JH, 1999. Plan di e si y and
p oduc i i y expe imen s in Eu opean g asslands. Science 286, 1123–1127.
Helle NE, Za ale a ES, 2009. Biodi e si y managemen in he ace o clima e change: A e iew
o 22 yea s o ecommenda ions. Biol Conse 142, 14–32.
Homeie J, We ne FA, 2007. P o isional Checklis s o Flo a and Fauna o he San F ancisco
alley and i s su oundings (Es ación Cien í ica San F ancisco), Sou he n Ecuado . Socie y
o T opical Ecology. olume 4 o Eco opical Monog aphs. chap e Spe ma ophy a. Checklis
Rese a Biológica San F ancisco (P o . Zamo a-Chinchipe, S-Ecuado ). pp. 15–58.
107
BIBLIOGRAPHY
Homeie J, We ne FA, G ads ein SR, B eckle SW, Rich e M, 2008. G adien s in a T opical
Moun ain Ecosys em o Ecuado . Sp inge , Be lin. olume 198 o Ecological S udies. chap e
Po en ial ege a ion and lo is ic composi ion o Andean o es s in Sou h Ecuado , wi h a
ocus on he RBSF. pp. 87–100.
Hoope DU, Chapin FS, Ewel JJ, Hec o A, Inchaus i P, La o el S, Law on JH, Lodge DM,
Lo eau M, Naeem S, Schmid B, Se ala H, Syms ad AJ, Vande mee J, Wa dle DA, 2005.
E ec s o biodi e si y on ecosys em unc ioning: A consensus o cu en knowledge. Ecol
Monog 75, 3–35.
Hough on RA, Hall F, Goe z SJ, 2009. Impo ance o biomass in he global ca bon cycle. J
Geophys Res-Biogeosc 114, 1–13.
Hubbell S, 2001. The Uni ied Neu al Theo y o Biodi e si y and Biogeog aphy. Monog aphs in
Popula ion Biology, P ince on Uni e si y P ess.
Hus on M, 1994. Biological Di e si y - The coexis ence o species in changing landscapes. Cam-
b idge Uni e si y P ess.
Hu h A, 1999. Modellie ung des Wachs ums und de Nu zung on opischem Regenwald. Ha-
bili a ionssch i , Uni e si ä Kassel.
Hu h A, Di ze T, 2001. Long- e m impac s o logging in a opical ain o es – a simula ion
s udy. Fo Ecol Manag 142, 33–51.
Hu h A, D echsle M, Kohle P, 2005. Using mul ic i e ia decision analysis and a o es g ow h
model o assess impac s o ee ha es ing in dip e oca p lowland ain o es s. Fo Ecol Manag
207, 215–232.
Imaizumi F, Sidle RC, Kamei R, 2008. E ec s o o es ha es ing on he occu ence o landslides
and deb is lows in s eep e ain o cen al Japan. Ea h Su P oc Land 33, 827–840.
Jablonski D, Roy K, Valen ine JW, 2006. Ou o he opics: E olu iona y dynamics o he
la i udinal di e si y g adien . Science 314, 102–106.
Jakob M, 2000. The impac s o logging on landslide ac i i y a Clayoquo Sound, B i ish
Columbia. Ca ena 38, 279–300.
Janzen DH, 1970. He bi o es and he numbe o ee species in opical o es s. Am Na 104,
501–528.
Jel sch F, Moloney K, Schu F, Kochy M, Schwage M, 2008. The s a e o plan popula ion
modelling in ligh o en i onmen al change. Pe spec Plan Ecol 9, 171–189.
Jel sch F, Wissel C, 1994. Modelling dieback phenomena in na u al o es s. Ecol Model 75-76,
111–121.
Johnson CM, Za in DJ, Johnson AH, 2000. Pos -dis u bance abo eg ound biomass accumula ion
in global seconda y o es s. Ecology 81, 1395–1401.
Johs K, Hu h A, 2005. Tes ing he in e media e dis u bance hypo hesis: When will he e be
wo peaks o di e si y? Di e s Dis ib 11, 11–20.
108
BIBLIOGRAPHY
Kammesheid L, Köhle P, Hu h A, 2002. Simula ing logging scena ios in seconda y o es
embedded in a agmen ed neo opical landscape. Fo Ecol Manag 170, 89–105.
Kea ing PL, 2008. The lo is ic composi ion and biogeog aphical signi icance o a megadi e se
pa amo si e in he sou he n Ecuado ian Andes. J To ey Bo Soc 135, 554–570.
Kessle M, 1999. Plan species ichness and endemism du ing na u el landslide succession in a
pe humid mon ane o es in he Boli ian Andes. Eco opica 5, 123–136.
Köhle P, 2000. Modelling an h opogenic impac s on he g ow h o opical ain o es s. Ph.D.
hesis. Uni e is ä Kassel.
Köhle P, Di ze T, Hu h A, 2000. Concep s o he agg ega ion o opical ee species in o
unc ional ypes and he applica ion on Sabah’s dip e oca p lowland ain o es s. J T op Ecol
16, 591–602.
Köhle P, Hu h A, 2010. Towa ds g ound- u hing o spacebo ne es ima es o abo e-g ound li e
biomass and lea a ea index in opical ain o es s. Biogeosciences 7, 2531–2543.
Klausmeie CA, Tilman D, 2002. Compe i ion and Coexis ence. Sp inge , Be lin. olume 161 o
Ecological S udies. chap e 3: Spa ial Models o Compe i ion. pp. 43–78.
Knei el JM, Chase JM, 2004. T ade-o s in communi y ecology: linking spa ial scales and species
coexis ence. Ecol Le 7, 69–80.
Knoke T, Cal as B, Agui e N, Román-Cues a RM, Gün e S, S imm B, Webe M, Mosandl R,
2009. Can opical a me s econcile subsis ence needs wi h o es conse a ion? F on Ecol
En i on 7, 548–554.
Kohyama T, 1992. Size-s uc u ed mul ispecies model o ain- o es ees. Func Ecol 6, 206–212.
K a N, Valencia R, Acke ly D, 2008. Func ional T ai s and Niche-Based T ee Communi y
Assembly in an Amazonian Fo es . Science 322, 580–582.
K e H, Je z W, 2007. Global pa e ns and de e minan s o ascula plan di e si y. P oc Na l
Acad Sci USA 104, 5925–5930.
Kü pick P, Kü pick U, Hu h A, 1997. The in luence o logging on a malaysian dip e oca p ain
o es : A s udy using a o es gap model. J Theo Biol 185, 47–54.
Lehne M, Kessle M, Salaza L, Na a e e H, G ads ein S, 2007. P o isional Checklis s o Flo a
and Fauna o he San F ancisco alley and i s su oundings (Es ación Cien í ica San F an-
cisco), Sou he n Ecuado . Socie y o T opical Ecology. olume 4 o Eco opical Monog aphs.
chap e P e idophy es. Checklis Rese a Biológica San F ancisco (P o . Zamo a-Chinchipe,
S-Ecuado ). pp. 59–68.
Le ins R, 1969. Some demog aphic and gene ic consequences o en i onmen al he e ogenei y o
biological con ol. Bull En omol Soc Am 15, 237–240.
Le ins R, Cul e D, 1971. Regional coexis ence o species and compe i ion be ween a e species
(ma hema ical model/habi able pa ches). P oc Na l Acad Sci USA 68, 1246–1248.
Liess M, Glase B, Huwe B, 2009. Digi al soil mapping in sou he n ecuado . E dkunde 63,
309–319.
109
BIBLIOGRAPHY
Lischke H, Lö le T, 2006. In a-speci ic densi y dependence is equi ed o main ain species
di e si y in spa io- empo al o es simula ions wi h ep oduc ion. Ecol Model 198, 341–361.
Liu JG, Ash on PS, 1998. Fo mosaic: An indi idual-based spa ially explici model o simula ing
o es dynamics in landscape mosaics. Ecol Model 106, 177–200.
Logo e DO, Lesnaya EV, 2000. The ma hema ics o Ma ko models: wha Ma ko chains can
eally p edic in o es successions. Ecol Model 126, 285–298.
MacA hu RH, 1972. Geog aphical ecology. P ince on Uni e si y P ess, P ince on, N. J.
Malhi Y, Baldocchi DD, Ja is PG, 1999. The ca bon balance o opical, empe a e and bo eal
o es s. Plan Cell En i on 22, 715–740.
Malhi Y, Robe s JT, Be s RA, Killeen TJ, Li WH, Nob e CA, 2008. Clima e change, de o -
es a ion, and he a e o he amazon. Science 319, 169–172.
May RM, 1973. S abili y and complexi y in model ecosys ems. P ince on Uni e si y P ess,
P ince on, N.J.
McGlone MS, 1996. When his o y ma e s: Scale, ime, clima e and ee di e si y. Global Ecol
Biogeog Le 5, 309–314.
Mi e meie RA, Robles Gil P, Ho mann M, Pilg im J, B ooks T, Goe sch Mi e meie C, Lam-
o eux J, Da Fonseca G, Alge K, Bol z F, B andon K, B une A, Da Sil a J, Ca e A, Ca al-
can i R, Ca e A, Chu ch R, Fos e M, Gascon C, Go en lo L, G a wicke B, Gue in-Mcmanus
M, Hannah L, Knox D, Kons an W, Lache T, Langhamme P, Lang and O, Lapham N,
Ma in D, Mye s N, Nask ecki P, Pa M, Pea son D, P icke G, Rice D, Rylands A, Sech es
W, Smi h M, S ua S, Thomson J, To en M, Wa d J, 2004. Ho spo s Re isi ed - Ea h’s
biologically iches and mos endange ed e es ial eco egions. CEMEX Books on Na u e.
chap e In oduc ion. pp. 19–68.
Moh en GMJ, Bu kha HE, 1994. Con as s be ween biologically-based p ocess models and
managemen -o ien ed g ow h and yield models - p e ace. Fo Ecol Manag 69, 1–5.
Mon gome y DR, Die ich WE, 1994. A physically-based model o he opog aphic con ol on
shallow landsliding. Wa e Resou Res 30, 1153–1171.
Mon gome y DR, Schmid KM, G eenbe g HM, Die ich WE, 2000. Fo es clea ing and egional
landsliding. Geology 28, 311–314.
Mosandl R, Gün e S, 2008. The T opical Moun ain Fo es - Pa e ns and P ocesses in a Biodi-
e si y Ho spo . Gö ingen Cen e o Biodi e si y and Ecology. olume 2 o Biodi e si y and
Ecology Se ies. chap e Sus ainable managemen o opical moun ain o es s in Ecuado . pp.
179–195.
Mose J, 1974. G ow h models o ee and s and simula ion. Royal College o Fo es y, S ock-
holm, Sweden. olume 30 o Resea ch No es. chap e A sys em o equa ions o he componen s
o o es g ow h. pp. 260–287.
Mulle -Landau HC, 2008. T opical Fo es Communi y Ecology. Blackwell Publishing; 2002.
chap e Coloniza ion- ela ed ade-o s in opical o es s and hei ole in he main anance o
plan species di e si y. pp. 182–195.
110
Danksagung
An diese S elle möch e mich bei allen Pe sonen bedanken, die zum Gelingen diese A bei beige-
agen haben.
Mein besonde e Dank gil dabei And eas Hu h, de mich in die Wel de ökologischen Mo-
dellie ung einge üh und mi dieses P ojek an e au ha . E ha meine A bei mi ielen
an egenden Diskussionen beglei e und mi iele F eihei en gelassen, wa abe auch imme da,
wenn ich F agen ha e. Ich eue mich schon au unse e wei e e Zusammena bei .
Deswei e en geh mein Dank an Bjö n Reineking, de die A bei on Uni e si ä ssei e aus be-
eu und mi hil eichen Kommen a en un e s ü z ha .
Fü ausdaue ndes Ko ek u lesen und Disku ie en on Teilen diese A bei möch e ich mich
seh bei Guy Pe’e ,Flo ian Ha ig und Jü gen G oene eld bedanken. Lucia Schobe , Vik o iia
Radchuk,Nadja Rüge ,Jule Schulze und Edua do Velázquez danke ich ü le z e Ko ek u en.
Fü die Übe se zung de Zusammen assung ins Spanische danke ich ganz he zlich Bea iz Vidal
Legaz.
Daß die Zei , die ich an diese A bei gesessen habe, eine schöne wa , ha o allem mi den
Menschen zu un, die in de OESA a bei en (und gea bei e haben). Sie machen in e essan e
A bei , on de ich iel le nen kann, und sind auch ein ach olle Menschen, mi denen ich ge ne
ausgedehn e (Mi ags-)pausen e b inge: Oli, Flo, Guy, Thomas, Sand o, Alex, Jü gen, Jula,
Bi gi , Tama a und Nadja, um einige zu nennen. Gabi Nagel, And eas Thiele, Michael Mülle
und Heike Reichel sei ü die Hil e bei ielen kleinen und g öße en o ganisa o ischen und ech-
nischen P oblemen sowie bei de Li e a u suche gedank .
Wäh end meine Au en hal e au de Fo schungss a ion in Ecuado habe ich Menschen ge o -
en, die mi iel En husiasmus in einem seh spannenden Ökosys em a bei en. An Jü gen
Homeie , S en Gün e und Flo ian We ne geh mein Dank o allem ü geduldiges Bean -
wo en iele F agen (eine Ma hema ike in) übe den Wald, sowie Hil e bei de Felda bei ,
besonde s beim Bes immen on A en. Mi Pe e Vo pahl, Ma eike Ließ, Ka in Wol und ie-
len ande en habe ich eine gu e Zei au de Es ación und im Feld e b ach , die ich nich missen
möch e.
Alexande Singe möch e ich da ü danken, dass wi unse Waldmodell je z in 3D und in
(beinahe) O iginalg öße sehen können. Ma kus Eichho n danke ich ü "e s e Hil e" in A cGis.
Bei Flo ian Ha ig bedanke ich mich ü die LaTeX-Vo lage und bei allen Au o en, die an de
Enwicklung on eie So wa e mi a bei en, insbesonde e: The R P ojec o S a is ical Com-
pu ing, LaTeX, Jab e und Sub e sion.
Dem Helmhol z-Zen um ü Umwel o schung (UFZ) gebüh mein Dank ü die Be ei s el-
lung eines seh gu ausges a e en A bei spla zes und eine he o agenden In as uk u . Fi-
nanziell wu de diese A bei e möglich du ch die DFG-Resea ch Uni 816 "Biodi e si y and
Sus ainable Managemen o a Megadi e se Moun ain Ecosys em in Sou h Ecuado ". Auße dem
wu de diese A bei eundliche weise un e s ü z du ch die Helmhol z In e disciplina y G adua e
School o En i onmen al Resea ch (HIGRADE).
Zu gu e Le z möch e ich mich bei meinen El e n bedanken, die mich seh un e s ü zen und
meine Wege (und Umwege) beglei en.
117

Hie mi e klä e ich an Eides s a , dass ich die o liegende Disse a ionssch i selbs ändig und
ohne emde Hil e e ass , keine ande en als die angegebenen Quellen und Hil smi el benu z
und die den benu z en We ken wö lich ode inhal lich en nommenen S ellen als solche kenn lich
gemach habe.
Hie mi e klä e ich an Eides s a , dass ich wede die o liegende noch eine gleicha ige Dok-
o p ü ung an eine ande en Hochschule endgül ig nich bes anden habe.
O , Da um Un e sch i
119
ISSN 1860-0387
PhD Disse a ion 01 / 2012
The ole o li e his o y ai s o coexis ence and o es
eco e y a e dis u bance – a modelling pe spec i e
Towa ds a be e unde s anding o species- ich o es s
Claudia Dislich
PhD Disse a ion 01 / 2012 I Claudia Dislich I The ole o li e his o y ai s o coexis ence and o es eco e y a e dis u bance ...
Helmhol z Cen e
o En i onmen al Resea ch – UFZ
Pe mose s aße 15 I 04318 Leipzig I Ge many
In e ne : www.u z.de