Max Planck Institutefor Chemistry

National Institute for Amazon Research

Tree species distribution, endemism, and oligarchy in Amazonian floodplain forests:

A species colonization concept for Amazonian white-water floodplains

Florian Wittmann

Rio Negro

Rio Solimões

Manaus

Landsat TM, bds 5,4,3 29/11/2005

(INPE)50 km

Manaus

Floodplains: 1.000.000 km2

Várzea: 200.000 km2

Igapó: 100.000 km2

Junk (1997)

____________________________________________________Methods

aquatic phase

Monomodal `flood pulsem

asl

terrestrial phase

February - July August - January

Daily water-level records of the Manaus harbour

Mean

am

plitu

de:

10.2

m !

Trees establish where annual inundations average < 7.5 m,which correponds to a waterlogged or submersed period of 270 days year-1 (forest border)

Inundation reduces oxygen availability to trees by the factor 104

Wittmann et al. (2004): For Ecol Manag

Jackson & Colmer (2005): Ann Bot

Amazonian floodplain tree species combine several adaptive strategies to tolerate the anaerobic site conditions:

• Morpho-anatomical adaptations: Increase of root surfaces, hypertrophic lenticels, aerenquimatic tissue;

• Physiological adaptations: leaf shedding during high water periods, reduction of photosynthesys, switch to anaerobic respiration = reduction of metabolism = cambial dormancy.

Lenticels

Adventitious rootsAerenquimatic tissue

Leaf shedding

Amazonian várzea forest

Várzea forests are the most species-rich floodplain forest worldwide:

Inventories with an area of 62.34 ha scattered over the Amazon basin:

Total number of trees: 39.497Total number of morphotypes: 1.900Total number of identified species: 918Total number of genera: 320Total number of families: 73

-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 tropical SE-Asian floodplains (Campbell et al. 2006)

-3 x higher than in the Brazilian Pantanal (Junk et al. 2006)Wittmann et al. (2006): J Biogeography

0

20

40

60

80100

120

140

160

180

Tre

e sp

ecie

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)

0

20

40

60

80

100

0 1 2 3 4 5 6Mean flood height (m)

R2 = 0.68

n=83 plots totalling 5.24 ha; 2.631 individuals, 306 species (Wittmann et al. 2002: J Trop Ecol)

n=44 forest plots totalling 62.3 ha; 39.497 individuals, 918 species (Wittmann et al. 2006: J Biogeography)

Alp

ha

-co

eff

icie

nt

(Fis

he

r)

7

0 km 1000

Flooding gradient

R2 = 0.92

low várzea high várzea

Ter Steege et al. (2003, 2006) modeled tree species alpha-diversity across the Amazon basin and concluded that diversity tends to increase from eastern to western Amazonia ….

…. the model holds true for upland and high-várzea forests, but not for low-várzea forests, where alpha-diversity is nearly constant across the Amazon basin.

Ter Steege et al. (2003): Biodiv & Conservation

Ter Steege et al. (2006): Nature

Geographic gradient

LV LV LVHV HV HV

TF

TF TF

0

50

100

150

200

250

300

50-100 1500-2000 > 3000

Km W from the mouth

Tre

e s

pe

cie

s h

a-1

Floodplain data resumed in: Wittmann et al. (2006): J Biogeography

Terra firme data resumed in: Oliveira & Mori (1999), Ribeiro et al. (1999), Harms et al. (2001), Pitman et al. (2001), Duque et al. (2003), Ter Steege et al. (2003), Condit et al. (2004), Oliveira & Amaral (2004), Penaherrera & Asanza (2004), Valencia et al. (2004)

n = 48, R2 = 0.2253

0

20

40

60

80

0,1 1 10 100 1000 10000

IS(%

)

Distance (km)

Low várzea

n = 42, R2 = 0.2607

0

20

40

60

80

0,1 1 10 100 1000 10000

IS(%

)

Distance (km)

High várzea

24.7%

9.9%

Geographic gradient

Floodplain data resumed in: Wittmann et al. (2006): J Biogeography

Terra firme data resumed in: Oliveira & Mori (1999), Ribeiro et al. (1999), Harms et al. (2001), Pitman et al. (2001), Duque et al. (2003), Ter Steege et al. (2003), Condit et al. (2004), Oliveira & Amaral (2004), Penaherrera & Asanza (2004), Valencia et al. (2004)

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

Where do the várzea species came from?

1. Taxonomically-evolutionary explanation:Kubitzki (1989) stated that the floodplain genotypes originate from the surrounding uplands.

2. Physiological explanation:Prance (1979) and Worbes et al. (1992) stated that

floristic similarities are especially high between floodplains and neotropical biomes with climatically or edaphically induced aridity (Campinas, Cerrado, Caatinga, Llanos).

Kubitzki (1989): Plant Syst EvolPrance (1979): BrittWorbes et al. (1992): J Veg Sci

Couroupita subsessilis Pilg.Hura crepitans L.

Determining the occurrence and distribution of the 300 most common (= abundant & frequent) várzea tree species:

• In revised species lists and Floras published in literature (more than 700 ha containing up to 280.000 individuals across the Neotropis),

• In herbaria (e.g., Missouri Botanical Garden, New York Botanical Garden, Royal Botanical Gardens Kew, National Herbarium Utrecht, Botanical Garden Rio de Janeiro, INPA-Herbarium, Manaus)

• In digital databases (e.g., Flora Brasiliensis [Martius 1840-1906], International Legume Database, etc.)

Database-Own study sites: 16 permanent sample plots-Review of 28 forest inventories compiled by other

authors

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) Cattanio et al. (2002) Wittmann et al. (2002) Schöngart (2003) J.C. Monteiro (unpubl.)

62.3 ha, 39.497 individuals, 918 species

0°

10° S

60° W70° W 50° W80° W

10° N

North

Ecuador

Peru

Colombia

Brazil

RDSM

1234

8

56

7

910

Bolivia

Manaus

Pando

0 1.000 km

Defining the most common várzea tree species

• In each inventory, the Importance Value Index (IVI) was calculated:

IVI = rAb + rDom + rFr (Curtis & McIntosh 1951)

• Overall species importance across the Amazon basin was determinedby the Overall Importance Value (OIV):

OIV = Σ 1-44 (rIVI) + rF (44 inventories)

Species occurrence (% out of 300 species)

Endemism in Amazonian várzea

Species occurrence (% out of 300 species)

Occurrence in other Neotropical floodplains

Species occurrence (% out of 300 species)

Occurrence in non-flooded moist uplands

Species occurrence (% out of 300 species)

Occurrence in non-flooded semi-arid uplands

22

18

51

60

6

5

14

36

12

18

9

várzea

24

3

Shared species between várzea & other ecosystems

floodplains / riparian forest

moist uplands (non-flooded)

semi-arid uplands (non-flooded)

Shared species (%)

Low-várzea forests highly resemble each other, even when separated over large distances, but show low floristic similarity to the terra firme = many endemic species that occur over thousands of Km = lateral species exchange

High-várzea forests represent increasing floristic dissimilarity with increasing distance, but show generally high floristic resemblance to the neighbouring terra firme, with many singletons = vertical species exchange to adjacent uplands.

Conclusions

Low-várzea forests highly resemble each other, even when separated over large distances, but show low floristic similarity to the terra firme = many endemic species that occur over thousands of Km = lateral species exchange

High-várzea forests represent increasing floristic dissimilarity with increasing distance, but show generally high floristic resemblance to the neighbouring terra firme, with many singletons = vertical species exchange to adjacent uplands.

Conclusions

Expected reasons:

1) High connectivity of low várzea by means of hydrochoric and ichthyochoric dispersal (Gottsberger 1978, Goulding 1980, Ziburski 1991)

2) Low várzea was less affected by paleoclimatic induced water-level changes than the high várzea (Irion 1984, Junk 1989, Wittmann et al. 2006)

= even with postulated dryer conditions during glacial periods, várzea forests persisted (as linear habitats along the main-river channels)

Species immigration,radiation, and species exchange

terra firmehigh várzea

low várzea

Time

forest border

lowest flood level

highest flood level

Point of `no return´

(species dependent)

Alti

tude endemism

increasing selection pressure

increasing species dominance

decreasing species richness

pre-formation ofadaptations in

episodically flooded river margins

Tree species colonization concept in Amazonian floodplains

Wittmann et al. (in press): Phytogeography, species diversity, community structure, and dynamics of Amazonian floodplain forests. Springer Verlag, Berlin.

• The species colonization concept implies that the floodplains acted as linear refuges for moist-sensitive terra firme species during glacial periods (Sioli 1957, Irion 1984, Pires 1984, Sternberg 1986).

• As orographic or tectonical barriers within the Amazon basin are absent, the continuous disturbance regime imposed by the flood-pulse and the alluvial dynamism especially in the white-waters is the most probable factor influencing speciation processes in equatorial Amazonia, thus contributing to the elevated biodiversity also in upland forests.

Conclusions

www.geowissen.de

Paleovárzea

• The species colonization concept implies that the floodplains acted as linear refuges for moist-sensitive terra firme species during glacial periods (Sioli 1957, Irion 1984, Pires 1984, Sternberg 1986).

• As orographic or tectonical barriers within the Amazon basin are absent, the continuous disturbance regime imposed by the flood-pulse and the alluvial dynamism especially in the white-waters is therefore the most probable factor influencing speciation processes in equatorial Amazonia, thus contributing to elevated biodiversity also in equatorial upland forests.

Conclusions