Andes to Amazon: What we think we know and what we don't

1
Mark Bush Florida Institute of Technology Mile
s Silman Wake Forest University
Andes to Amazonia What we think we know and
what we dont.
2
Some of the big questions before us To what
extent can Andean records inform us about
Amazonia? How much climatic variation was there
within ice age Amazonia? How robust are the data
for a 5oC cooling in the terrestrial
Neotropics? How dry was Amazonia during the last
ice-age?
3
Refugia

• Raised on modern biogeographic data but no
  reliable paeleocological data.
• Most recent manifestation is untestable as it has
  no time or habitat constraints (Haffer and Prance
  Amazoniana 2002).

Brown (1991)
4
Dry Forest Arcs

• Modern biogeographic data used to suggest the
  presence of a past habitat bridge of dry forests.
• Assumed to be ice-age
• But there is no basis for this assumption.
• Nor a real difference from modern required to
  allow movement of species.

Pennington et al. ( J. Biogeogr. 2000)
5
Different ideas about Arcs
Wüster et al. (Molec. Ecol. 2005)
Pennington et al. (J. Biogeogr. 2000)
6
Miles et al. (GEB, 2004)
7
Sites that well discuss
Amazon fan
Pata Verde
Maicuru
Negra
Carajas
Consuelo
Titicaca
8
TRMM data and paleo locations

• Not surprising that the different paleoecological
  records are providing a different story

9
Late Glacial and Holocene Vegetation from Lake
Titicaca
Paduano et al. (Palaeo3 2003)
10
The Glad 800 drill rig (NSF/ODP) was hauled to
Titicaca in 2001 and three cores were raised from
deep water under the supervision of Paul Baker,
Geoff Seltzer and Sheri Fritz.
11
Long record from core LTO1-2B, Titicaca

• Probably four, full glacial cycles.

MIS 5
Pollen concentration is a proxy for
productivity, and indicator of near Holocene
temperatures.
MIS 7
MIS 9
MIS 11
12
Long record from core LTO1-2B, Titicaca
Too cold for pollen or Isoetes

Isoetes may be important for glacial temperature.
MIS 5e
MIS 7
MIS 9
MIS 11
13
D-O cycles in Titicaca
Work in progress
14
Titicaca pollen concentration
D-O cycles in Titicaca
GISP
Work in progress
15
Lake level and insolation

• Lake Titicaca and the Salar de Uyuni, Bolivia,
  respond to DJF insolation.
• Lake level at Titicaca varies synchronously with
  Pata for the last 70,000 years.

20
40
K yr BP
Baker et al. (Science 2001)
16
Hill of Six Lakes
17
Lake level fluctuation in Amazonia

• Lakes Pata and Verde show continuous forest cover
  through last two glacial periods.
• Periods of low lake level correlate with periods
  of high K concentrations in sediment in Pata.
• Low lake level due to reduced precipitation.

18
Lake level and orbital forcing

• Peak low-lake level corresponds to 11 K BP and 33
  K BP with 22 K BP wet (matches Titicaca).
• Aligns to orbital rhythm of variation in
  insolation.
• Continuous forest cover suggests decrease in
  precipitation in the wet season rather than dry
  season.

Bush et al. (J. Paleolimn. 2002)
19
Bush et al. (Palaeo3 2004)
20
Lake Pata, Brazil
Cal BP
10, 000 18,000 22,000 34,000 48,000
Colinvaux et al. (Science 1996)
21
Fusing modern ecology and paleoecology
22

• Linking Amazon and Andean forests
• Single system
• Space and Time
• Fully vouchered 1 ha plots lowlands to treeline
  (17 to date)
• Series of lakes along transects
• Peru and Bolivia
• Andes Biodiversity Consortium

23
Modern Ecology 77 ha (Peru) -17
Montane, 60 Lowland w3tropicos database Paleoecolo
gy 24 Lakes
L
L
L
L
L
L
L
L
L
L
L
L
L
L
L
L
L
L
L
L
L
L
L
L
L
L
L
24

• Collection data for 4 genera

25
(No Transcript)
26

27
(No Transcript)
28
(No Transcript)
29
Representation of pollen taxa

• Continuous wet forest taxa for 50,000 yrs
• Gradual change in community composition

Bush et al. (Science 2004)
30
Consuelo Ordination

• Two basic states
• 30 kyr
• 11 kyr
• Evidence of cycles

31
DCA Axis 1 vs. Time

• Two major forest types
• Periodic changes
• ENSO, drought cycles, D-O events
• Gradual 8-10 ky transition
• Non-equilibrium?

32
Elevation for taxa found 21,000 years ago (500
cm pollen assemblage)
If taxa were found co-occurring today, most
likely elevation would be 2400m
Q What does time course look like?
33
(No Transcript)
34
Laguna Negra wavelets
Lake at 3300 m elev, 7oS.
Bush et al. (JQS In press)
35
Consuelo and Pata
Pata
36
Pata
37
Maicuru
15-30K
hiatus
Colinvaux et al. (Amazoniana 2001)
38
Carajas, eastern Amazonia
15-25K hiatus
15,000 25,000 31,000 gt50,000
Absy et al. (Compt Rendus 1991)
39
Amazon Fan data (S. Haberle 1999)
Modern Grass
Fan deposits reflect pollen from entire
basin Ice age does not have more grass
pollen than modern
Ice age Grass
40
Amazon fan reconsidered
41
Discharge source and Fan data
42
What is cooling?
All these sites provide evidence of a 5-7oC
cooling at the LGM
43
One way forward Improving pollen recognition
Using SEM and vector- based computer modeling we
aim to improve identification of key taxa to
species or species group. More precise taxonomy
more precise bioclimatic envelope modeling
Extract sub-region for training and recognition
44

• What we think we know
• Strong decline in species diversity with
  elevation.
• Species migrate individualistically in response
  to climate change.
• Strong cooling signature in all Pleistocene
  records with some species ranges typically
  moving 1000 m -1500 m downslope.
• Lowland taxa are adapted to Pleistocene cooling.
• Amazonian paleoclimates were temporally and
  spatially heterogeneous.
• Refugia and savanna corridors are not grounded in
  reality.

45
What we dont know Do we have to reconcile the
terrestrial record with the marine
record? Difference in temperature could be
absolute minima (NH) or mean temperature (SH)?
To what extent forest structure
changed. Fine-resolution climate change from the
lowlands. The full impact of Holocene human
activities.
46
The Coring Gang
Jen Hanselman
Alex Correa
Dunia Urrego
Andrea Gomez
Dr. Will Gosling
47
(No Transcript)
48
(No Transcript)
49
(No Transcript)