Max-Planck-Institute for
Chemistry Instituto Nacional de
Pesquisas da Amazônia
Flood-induced Endemism in Amazonian
Floodplain Trees
Florian Wittmann Ethan Householder, Jochen Schöngart, Maria T. F. Piedade, Rafael L. de Assis,
Pia Parolin & Wolfgang J. Junk
Monitoring of Amazonian
Wetlands - MAUA
_____________________________________________
_______Methods
Aquatic phase
Monomodal flood pulse
Terrestrial phase
February - July August - January
Me
an
am
pli
tud
e:
10
.2 m
!
Photographs:J. Schöngart
Jackson & Drew (2002): Ann Bot
Trees establish where annual
inundations average < 7.5 m (white-
water) or < 9.0 m (black water),
which correponds to a waterlogged
or submersed period of 230 and 270
days year-1 (forest border)
Inundation reduces oxygen
availability to trees by the
factor 104
Amazonian floodplain tree species combine several adaptive strategies
to tolerate the anaerobic site conditions:
• Morpho-anatomical adaptations: Increase of root surfaces, hypertrophic
lenticels, aerenquimatic tissues;
• Physiological adaptations: leaf shedding during high-water periods, reduction of
photosynthesys, switch to anaerobic respiration, elevated production of anti-
oxidant compounds = reduction of metabolism = cambial dormancy.
Lenticels
Adventitious roots Aerenquimatic tissue
Leaf shedding
Floristic inventories in várzea forests totaling >60 ha
across the Amazon basin
Total number of trees: 39.497
Total number of morphotypes: 1.900
Total number of identified species: 918
Total number of genera: 320
Total number of families: 73
Várzea forest are the most species-rich floodplain forests worldwide:
-20 x higher than in the European temperate zone (Schnitzler et al. 2005)
-10 x higher than in subtropical bottomland forests of N-America (Johnson & Little 1967, Clark & Benforado 1981)
-10 x higher than in neotropical savannas and SE-Asian floodplains (Junk et al. 2006, Campbell et al. 2006)
Tree species diversity in white-water floodplains
Wittmann et al. (2006): J Biogeography
0
20
40
60
80
100
120
140
160
180
Tre
e s
pecie
s h
a-1
7.0 6.5 6.0 5.5 5.0 4.5 4.0 3.5 3.0 2.5 1.5 1.0
Mean flood height (m)
n=83 plots totaling 5.24 ha; 2.631 individuals, 306 species
(Wittmann et al. 2002: J Trop Ecol)
n = 25, R2 = 0.4717
0
20
40
60
80
100
0 1 2 3 4 5 6
Mean flood height (m)
Alp
ha-c
oeffic
ient
n=44 plots totaling 62.3 ha; 39.497 individuals, 918 species
(Wittmann et al. 2006: J Biogeography)
Tree species richness and diversity along the flooding gradient
0
20
40
60
80
100
120
140
160
180
Tre
e s
pecie
s h
a-1
7.0 6.5 6.0 5.5 5.0 4.5 4.0 3.5 3.0 2.5 1.5 1.0
Mean flood height (m)
n=83 plots totaling 5.24 ha; 2.631 individuals, 306 species
(Wittmann et al. 2002: J Trop Ecol)
n = 25, R2 = 0.4717
0
20
40
60
80
100
0 1 2 3 4 5 6
Mean flood height (m)
Alp
ha-c
oeffic
ient
n=44 plots totaling 62.3 ha; 39.497 individuals, 918 species
(Wittmann et al. 2006: J Biogeography)
Tree species richness and diversity along the flooding gradient
Low várzea High várzea
Low várzea
High várzea
Floodplain data resumed in:
Wittmann et al. (2006): J Biogeography
Terra firme data resumed in:
Oliveira & Nelson (2001), Pitman et al. (2002), Ter Steege et al. (2006)
Floristic similarity between Amazonian várzea and upland forests
26
25-32
9.8 30.1
Terra firme
(non-flooded uplands)
Low várzea
> 3 m
High várzea
< 3 m
From where came the várzea tree species?
Are there endemic tree species?
1. Taxonomic-evolutionary hypothesys:
Floodplain genotypes originate from adjacent upland forests (Kubitzki 1989)
2. Physiological hypothesys:
Several Amazonian floodplain genera and species originate from ecosystems/biomes
with climatologically and/or edaphically aridity = neotropical savannas (Prance 1979, Worbes
et al. 1992).
3. Endemism:
Due to the exceptional high inundations, and the high number of adaptive strategies of
trees to flooding, Amazonian floodplains are rich in endemic tree species (Kubitzki 1989,
Junk 1989). This is in contrast to other neotropical wetlands, where endemic tree species
are rare, or even absent (Junk et al. 2006, Veneklaas et al. 2005, Wittmann et al. 2010)
Prance (1979): Brittonia; Junk (1989): Academic Press, London; Kubitzki (1989): Plant Syst Evol; Worbes et al. (1992): J Veg Sci;
Veneklaas et al. (2005): Ecography; Junk et al. (2006): Aquatic Sci; Wittmann et al. (2010): Springer , New York
Behling et al. (2001): Palaeogeo Plaeoclima
Palaeoeco
83-67 Ma
61-60 Ma Repeated marine ingressions in the pre-Andean
depression
12-10 Ma
43-40 Ma Formation of Lago Pozo (alluvial, Andean sediment)
20-12 Ma Formation of Lago Pebas (alluvial, Andean sediment)
13.5 Ma Fossil of fruit-feeding fish (Colossoma macropomum)
indicate no changes to recent diet
8 Ma Amazon starts to drain eastwards
< 1 Ma Pleistocene and Holocene climate changes strongly
influence the Sea level and thus the Amazon River
system.
Evolution of Amazonian Wetlands
The existence of tropical forests within the Amazon basin is evident since the upper
Eocene (approx. 30 Ma B.P.) (Burnham & Johnson 2004: Phil Trans R Soc Lond)
Hura crepitans L. Couroupita subsessilis Pilg.
Occurrence and distribution of the 658 most important (abundant
and frequent) várzea tree species
● In Inventories, Checklists and Floras published in literature (up to700 ha and 280.000 individuals
across the entire Neotropics)
● In herbaria (e.g., MOBOT, NYBG, RBG Kew, Nationaal Herbarium Utrecht, Jardim Botânico
do Rio de Janeiro, herbário do INPA)
● In digital databases (e.g., Flora Brasiliensis, International Legume Database, etc.).
0 1.000 km
0°
10° S
60° W 70° W 50° W 80° W
10° N
North
Ecuador
Peru
Columbia
Venezuela
Brazil
RDSM
2 3
4
11
5
7
6
15
9
13
8
10
12
14 16
Bolivia
1
Pando
Manaus
Data base
● 22 own inventories in the central and western Brazilian Amazon
● Review of 30 inventories available in literature
Authors:
Black et al. (1950)
Pires & Koury (1959)
Balslev et al. 1987
Revilla (1991)
Campbell et al. (1992)
Worbes et al. (1992)
Ayres (1993)
Queiroz (1995)
Dallmeier et al. (1996)
Klinge et al. (1996)
Urrego (1997)
Nebel et al. (2001)
Bentes-Gama et al. (2002)
Cattanio et al. (2002)
Duque et al. (2002)
Wittmann et al. (2002)
Schöngart (2003)
Ferreira et al. (2005)
Carim et al. (2008)
Assis & Wittmann (2011)
J.C. Monteiro (unpubl.)
= 80 ha, 92.000
individuals
Species selection and definition of importance
● In each inventory, the Importance Value (IV) was derived:
IV = rAb + rDom + rF (Curtis & McIntosh 1951)
● For each species, the Overall (continental) Importance Value (OIV) was
calculated:
OIV = Σ 1-80 (rIVI) + rF (80 inventories) (Wittmann et al. 2006)
We only selected trees (by excluding vines, lianas, epiphytes), and species that
occurred at least in two distinct inventories, performed by distinct authors.
Singletons were excluded from the list.
= 658 tree species from the original 918
Most important biomes/ecosystems containing
várzea tree species
Caribbean Islands
Orinoco & Magdalena
Rivers
WA
terra firme CA
terra firme
EA
terra firme
Brazilian
Cerrado
Llanos
Bananal
Pantanal
Chaco
0 100 200 300 400 500
Terra firme (all)
Terra firme WA
Terra firme CA
Terra firme EA
Central America
Atlantic rainforest
Savanna
Caribbean Islands
Paleotropics
Orinoco
Igapó
Hyperseasonal savanna
Riparian savanna
Shared species with Amazonian várzea (out of 658)
Distribution of várzea tree species (90% occur in non-várzea environments)
We
tla
nd
s
No
n W
etl
an
ds
0
50
100
150
200
250
300
350
400
450
3000 1500 500 Todas as partes daAmazônia
No o
f tr
ee s
pecie
s
Distance to the mouth of the Amazon River (km)
Numbers of várzea tree species in Amazonian uplands
All parts of Amazonia
Western
Amazonia
Central
Amazonia
Eastern
Amazonia
www.geowissen.de
The várzea is a geochemical extension of the Andes (Irion 1984)
Irion (1984): Z. Naturwissenschaften
Indicator Species Analysis investigating the occurrence
of the 658 most important várzea tree species
140 várzea tree species (21.3%) with occurrence < 5% in non-várzea
environments = Habitat Specialists
68 várzea tree species (10.3%) restricted to Amazonian várzea
= Ecological Endemics
Cluster-Analysis (Ward) combined with association-tests
(Šidàk) of várzea tree species occurrence across Neotropical
biomes, indicating two centers of endemism
Central Amazonian
low várzea Western Amazonian
high várzea
Widespread
Amazonian
group
Widespread
Extra-Amazonian
group
restricted restricted
Wittmann et al. (subm. to Ecography)
0
5
10
15
20
25
30
Eastern Amazon Central Amazon Western Amazon
No
. o
f e
nd
em
ic tre
e s
pe
cie
s
LV HV LV & HV
Degrees of endemism vs. flooding gradient
0
5
10
15
20
25
30
Eastern Amazon Central Amazon Western Amazon
No
. o
f e
nd
em
ic tre
e s
pe
cie
s
LV HV LV & HV
Degrees of endemism vs. flooding gradient
Flood pulse: 10 m
Flood pulse: 6 m Flood pulse: 6 m
Immigration, species exchange,
radiation
Terra firme /
Andes High várzea
Low várzea
Time
Min. flood level
Max. flood level
„Point of no return“
(species dependent)
Flo
od level
ENDEMISM
Increasing selection pressure
Increasing species dominance Decreasing species richness
Formation
of pre-adaptations under
polymodal flood regimes
Tree species colonization concept in Amazonian
Várzea floodplains
Wittmann et al. (2010): Springer , New York
Amazonian lowland areas of endemism (developed on ranges of terrestrial organisms)
„For (terrestrial) animals, Rivers are important distribution barriers“ (Silva et al. 2005: Conserv Biol 19:689-964)
„For (terrestrial) tree species, Rivers are important areas of speciation,
being one of the most prevalent adaptive gradients in the Amazon basin“ (Wittmann et al. In press: Ecography)
THANK YOU !
