!"#$$%&'!#()'(*$$+,$$-+.+/01'!"#$%&'()%*#%+#"#&,(%-.(+#/$'(')/"0&#/-

Henry Hooghiemstra (1), Vladimir Torres (1,2), Raul Giovanni Bogotá-Angel (1,3), Mirella Groot (1), Lucas Lourens (4),

Juan-Carlos Berrio (1, 5) & Project members #

(1) Institute for Biodiversity and Ecosystem Dynamics, University of Amsterdam, Netherlands(2) current address Exxon Mobil, Houston, USA

(3) current address Universidad Distrital Francisco José de Caldas, Bogotá, Colombia(4) Geosciences, University of Utrecht, Netherlands

(5) current address Dept. of Geography, University of Leicester, UK

DEBI

Long Records of Environmental and Climate Change

The 2,250,000-yr long record from the Bogotá basin

Age model development

Basin development

Biome evolution

are crucial to document the series of ice ages that characterise the Pleistocene. While in the northern hemisphere ice sheets expanded and shrunk periodically, in the tropical mountains the altitudinal vegetation distribution shifted downslope during relatively cold glacials and stadials and upslope during relatively warm interglacials and interstadials. This poster focus on development age models for the sediments in the basins of Bogotá and Fúquene (Fig. 1), the climate history of the northern Andes and the evolution of its high elevation biota.

In the Bogotá basin (4°N, 2550 m elevation) 586 m of sediments have been cored. The 2200-sample record of grain size distributions shows changes in sedimentary environments (1). The pollen record starts at 540 m core depth and samples at 20-cm increments along the core provide a record of 2100-samples (1). The record of aquatic vegetation re!ects changes in water depth. The record of trees, shrubs and herbs shows changes in the vegetation that covered the mountains around the lake.

The 357-m deep Funza-1 core (2) and the 586-m deep Funza-2 core (1) have been coupled at the "rst appearance event of Alnus (Fig. 2). Changing percentages of arboreal pollen (AP%) in the composite record Funza09 (540-1.6 m) show altitudinal shifts of the upper forest line (UFL) (3). Using a lapse rate of 0.6°C/100 m UFL displacement and a mean annual temperature (MAT) at the UFL of ~9.5°C paleo-temperatures have been calculated.During warm interglacial conditions the tree Alnus mainly forms swamp forest around the lake re!ected as high peaks in the AP% record. The sequence of Alnus-based interglacials have been matched with the marine record of glacial-interglacial cycles as shown by the LR04 benthic #18O stack record (Fig. 3) (4). The immigration event of Alnus has been dated 1.018 Ma. We used

cyclostratigraphy to develop the age model for the lower part of the Funza09 record. Frequency analysis in the depth domain shows peaks of 9.5-m and 7.6-m (Fig. 4). Wavelet analysis shows a stable presence of the 9-m frequency band which could be robustly linked obliquity as the main driver of climate change. The age model shows the Funza09 record re!ects the period from 2,250,000 to 27,000 years BP (Fig. 5).

The AP%-based temperature record re!ects marine isotope stages (MIS) 85 to 3 (Fig. 6 lower panel). The record of aquatics shows deep-water vs. shallow-water conditions re!ecting climate change superimposed by e$ects of basin evolution (Fig. 6 top panel). In the period of 2.25-1.47 Ma wetlands, !uvial channels and swamps prevailed. The basin !oor subsided rapidly between 1.47 and 1.23 Ma and a lake developed. Between 1.23 and 0.86 Ma lacustrine sediments accumulated in waters up to 50 m deep. Water depth was lower during the last 860,000 years but !uctuated in conjunction with the 100-kyr dominated glacial-interglacial cycles of the middle and late Pleistocene (3). Around 27,000 BP the lake disappeared probably because the Bogotá basin became over"lled with sediments.

The Bogotá basin record shows 5 stages in the evolution of montane forest and páramo vegetation (Fig. 6 middle panels) (3). (1) Period 2.25-2.02 Ma: vegetation above the UFL was poor in species (protopáramo). Aragoa occasionally reached high abundance. In the Andean forest Podocarpus increased in abundance. Proper lake conditions had not yet developed. In absence of Alnus, Morella (Myrica) formed swamp forest and thickets on the basin !oor.(2) From 2.02 Ma (MIS 75) onwards increasing proportions of Aragoa, Ericaceae and Hypericum indicate that shrub vegetation developed as a structural transition zone between upper montane forest and herbaceous páramo. Polylepis, originating from the southern Andes, started its presence 2.1 Ma (MIS 78).(3) From 1.58 Ma (MIS 54) onwards Borreria had developed as an element of the upper montane forest and was no longer an exclusive constituent of open vegetation. The páramo reached more closely its modern structure and taxonomic composition. In the montane forest the share of Podocarpus decreased.(4) After the closure of the Panamanian Isthmus several Holarctic trees immigrated into South America. Alnus immigrated 1.01 Ma (MIS 29). In the swamp forest around the lake Alnus replaced Morella (Myrica). On the north Andean slopes Alnus changed the composition of montane forests.(5) At 430,000 yr BP (MIS 12) Quercus (oak) arrived in the northern Andes and changed forest composition dramatically. At high elevations Quercus competed with Weinmannia and Podocarpus. Near the UFL Quercus

partially replaced Polylepis. However, Quercus expanded slowly and reached as late as 240,000 yr BP (MIS 7) signi"cant proportions. Most of the Pleistocene vegetation assemblages have no modern analogue. However, modern ecological constraints of suites of taxa allow a robust reconstruction of environmental and climate change.

Funza

Fúquene

Bogotá Basin

Fúquene Basin south

Fúquene Basin north

100

0

50

100

150

200

250

300

350

400

450

500

550

Alnus

Gap in Funza-2 core recovery

Funza-2

m

100

Alnus

Funza-1

Subandean forest

Andean forest

Subparamo

Páramo

Arboreal Pollen (AP)

Many gaps andlow pollen recovery

0

50

100

150

200

250

300

350

400

450

500

550

m0 20 40 60 80 100%0 20 40 60 80 100%

1800

200 2800

Funza09

0 20 40 60 80 100%0

50

100

150

200

250

300

350

400

450

500

m

Take

n fro

m F

unza

-2Ta

ken

from

Fun

za-1

0 200 400 600 800 1000 1200Age (ka)

5.0

4.5

4.0

3.5

3.0

OD

P Si

tes

846

& 84

9

0

40

80

120

160

Alnu

s (%

)

3

2

1

0

-1

-2

18O

Med

iterra

nean

0

40

80

120

160

Alnu

s (%

)

visual tiepoint

1

5.1

5.5

7.1 7.5 9.1

9.311

1315.1

15.517

19 21

23

25

2927

31

33

35

3

7.3

8.5

18.3

Alnus

Alnus

18O

excl

. Que

rcus

excl

. Que

rcus

O-tuning: Arboreal Pollen excluding Alnus and Quercus (detrended)

41ky

100ky

23ky

41ky

100ky

23ky

0 200 400 600 800 1000 1200 1400 1600 1800 2000 2200 2400

ODP Sites 846 & 849

5.2

4.8

4.4

4.0

3.6

3.2

2.8

18O

(per

mil)

-40

-20

0

20

40

AP (%

, det

rend

ed)

eccl

. Aln

us a

nd Q

uerc

us

4

168

3264

128256512

4

168

3264

128256512

0 200 400 600 800 1000 1200 1400 1600 1800 2000 2200

0 50 100 150 200 250 300 350 400 450 500Depth (m)

-40

-20

0

20

40

AP (%

, det

rend

ed)

eccl

. Aln

us a

nd Q

uerc

us

Funza: Arboreal Pollen excluding Alnus and Quercus (detrended)

1

42

8163264

128

0.5

~20m (100ky)

~8m (41ky)

Age (ka)

Perio

d (m

)Pe

riod

(ky)

Perio

d (k

y)

~4m (23ky)

0003000200010Age (ka)

0

200

400

600

Dep

th (m

)

10

20

30

40

50

60

Sed.

rate

(cm

/kyr

)

P-tuning

Hooghiemstra et al., 1993

O-tuning

O-tuning

P-tuning

-30

-20

-10

0

10

20

30

AP(%

, det

rend

ed) ex

cl. A

lnus

and

Que

rcus

0

40

80

120

160

Alnu

s &

Que

rcus

(%)

0 200 400 600 800 1000 1200 1400 1600 1800 2000 2200 2400Age (ka)

Quercus

Alnus

5

79 11

13

15

17 19 21 25 37

47

49

787064

-3.0

0.0

3.0

3.0

3.5

4.0

4.5

5.0

Medit

erra

nean

!18 O

ODP

Site

s 846

& 8

49!1

8 O

5

7

911

13

15 19

612

7793

81

82

87

64

91

78

72

637321

25

23

22

31

35

33

29

37

39

4749

5155

59

Extrapolation Visual match Astronomical tuning

40

80

120

40

80

120

Ande

an fo

rest

Subp

áram

o &

Pára

mo

PodocarpusMyrica

BorreriaAragoa

Hypericum Polylepis

Proper Páramo and Andean forestwith no Oak

Andean forest with no Alder and no Oak Development of shrub Páramo ProtopáramoDevelopment of Oak forest

Sele

cted

taxa

(%)

excl

. Aln

us a

nd Q

uerc

us

0

100%

20%40%60%80%

Low water stands. Mainly fluvio-lacustrine and swamp deposits.Development of proper lake

Highest lake levelsFluctuation of lake levels closely related to glacial/interglacial cycles

Ludwigia & Myriophyllum

Hydrocotyle

Cyperaceae

Isoëtes

Aqua

tic ta

xa in

dica

ting

lake

leve

l

Fig. 1

Fig. 2

Fig. 3

Fig. 4 Fig. 5

Fig. 6

DEBI

The 284,000-yr long record from the Fúquene basin

Age model development

Conclusions

Selected ReferencesAcknowledgements

Tropical montane forest and tropical alpine grassland (páramo) ecosystems are sensitive to variations in global ice volume and respond to global climate change with signi"cant altitudinal migrations. A one-to-one coupling between changes in tropical MAT and the North Atlantic climate variability at both orbital and millennial time scales is shown. Century-scale changes of 2-3.5°C during MIS 3-4 (Pleniglacial) and during previous glacial MIS 6, and extreme changes of up to 10±2°C at the major glacial Terminations are shown. Long continental records analysed at sub-centennial resolution show the dynamic history of terrestrial biomes and allow to assess current and projected future climate change in a Pleistocene framework. However, developing such records is still an academic challenge.

(1) Torres V et al. Palaeogeography Palaeoclimatology Palaeoecology 226, 127-148 (2005); (2) Hooghiemstra H. Dissertationes Botanicae 79 (1984); (3) Torres V et al. Quaternary Science Reviews 63, 59-72 (2013); (4) Lisiecki LE, Raymo ME, Paleoceanography 20 PA1003,doi:10.1029/2004PA001071 (2005); (5) Sarmiento G et al., Geomorphology 100, 563-575 (2008); (6) Bogotá-A RG et al. Ch. 5 in PhD thesis (2011b); (7) Groot MHM et al., Climate of the Past 7, 299-316 (2011); (8) Groot MHM et al., Quaternary Geochronology (2014). http://dx.doi.org/10.1016/j.quageo.2014.01.003; (9) Van Geel B, Van der Hammen T, Palaeogeography Palaeoclimatology Palaeoecology 14, 9-92 (1973); (10) GRIP-Members, Nature 364. 203-207 (1993); (11) Jouzel J et al., Science 317, 793-797 (2007); (12) Van der Hammen T, Hooghiemstra H, Global and Planetary Change 36, 181-199 (2003); (13) Mommersteeg HJPM, PhD thesis, University of Amsterdam (1998); (14) Vriend, M et al., Netherlands Journal of Geosciences 91, 199-214 (2012); (15) Bogotá-A RG et al. Quaternary Science Reviews 30, 3321-3337 (2011a).

We greatly acknowledge the contributions to the Funza and Fúquene Projects by A. Betancourt, L. Contreras, S. Gaviria, C. Giraldo, N. González, J.H.F. Jansen, M. Konert, D. Ortega, J.O. Rangel, C.A. Rodríguez-Fernandez, G. Sarmiento, E. Tuenter, P.C. Tzedakis, J. Vandenberghe, T. Van der Hammen, M. Van der Linden, J. Van der Plicht, B. van Geel, M. Vriend, S.L. Weber, W. Westerho$ and M. Ziegler. These Long Continental Record projects were "nancially supported by NWO (The Hague), WOTRO, (The Hague), AlBan (Madrid), NUFFIC (The Hague), IBED (University of Amsterdam), INGEOMINAS (Bogotá) and COLCIENCIAS (Bogotá).

This poster was designed by Jan van Arkel/IBED/University ofAmsterdam

We revisited the age models of pollen records Fq-3 (12, 15) and Fq-7C (13) with the new approach of cyclostratigraphy and we obtained a basin-wide biostratigraphy (15) constrained in time (Fig. 10). We recognised periods with distinct sediment compositions (14) and abundance of geochemical elements (6).

For a comparison of temperature records we show Greenland #18O ice core record (10) for the past 180,000 years and the record from Antarctic core Epica Dome C (11) and our near equatorial record Fq-BC (Fig. 9) (7). During the last 130,000 years ice cores show 28 millennial-scale climate oscillations (Fig. 9, bottom panel), known as Dansgaard-Oeschger (DO) cycles. Most DO cycles are re!ected in equatorially located Fq-BC pollen record (Fig. 9, top panel) The clear signature of the Younger Dryas (constrained by 14C dates), and the interstadial DO cycles 8, 12, 14, 19 and 20 in the reconstructed MAT record suggest an unprecedented North Atlantic-equatorial link.

Sediments centrally located in the lake contain low percentages of carbon preventing robust 14C-dating (8). Cyclostratigraphy is a challenging alternative. Frequency analysis of the AP% record in the depth domain show peaks at 907 and 2265 cm. Wavelet power spectra show that the strengths of these frequencies are stable throughout the core in depth and time. These frequencies were explored and appear to be linked to the obliquity (41-kyr) and eccentricity (~100-kyr) cycles of orbital forcing of climate change (Fig. 8). Remarkable is the absence of a clear imprint of the precession-related 21-kyr cycle (7). Fig. 8A shows the Fq-9C AP% record as raw data and Fig. 8B as a detrended and interpolated depth series overlain by a ~9 m Gaussian "lter. Fig. 8C shows the LR04 benthic #18O stack record (4) overlain by a Gaussian "lter centered at the 41-kyr cycle. We correlated the "ltered 9-m signal in Fq-9C to the "ltered 41-kyr obliquity-related component of the LR04 benthic #18O stack record to develop the age model. The 4768-points Fq-9C record re!ects the interval from 284,000 to 27,000 yr BP. The Fq-9C record was extended to late Holocene time by adding part of the Fq-2 pollen record (9): the Fúquene Basin Composite (Fq-BC) record. The connection with modern instrumental data allows to calculate mean annual temperatures (MAT) throughout the record.

In Lake Fúquene (5°N, 2540 m), a colluvial damm blocked lake (5), the upper 60 m of sediments have been collected from a !oating raft in two parallel cores (Fig. 1). Downcore measurements at 1-cm increments revealed multiple 4768-points records. Pollen and carbon content show biome changes in the area and grain size distributions and XRF-based geochemistry show the abiotic changes in the basin. Downcore changes in lithology (5), Fe and Zr (6) in cores Fq-9 and Fq-10 allowed to develop the composite record Fq-9C (7). Geochemical records show changes in erosion and sediment transport in the basin. Downcore grain size distributions show changing sedimentary environments (Fig. 7). Downcore aquatic pollen show changes in water-levels (Fig. 7). Downcore regional pollen show temperature-driven changes in the altitudinal vegetation distribution.

&'&

&'!

%'(

%'$

$')

$'&

)*+,-./$%0

(

!'%

!'$

!'#

!'!

*!'#

*!'$

*!'%

)*+,-./$%0

$1$/)$&23$4567#3(

! "!! #!!! #"!! $!!! $"!! %!!! %"!! &!!! &"!! "!!! ""!! (!!!

8#97:$%;<(

*$!

*#!

!

#!

$!

!=$%>1$?#73#@?#?(

*(!

*&!

*$!

!

$!

&!

(!

!=$%>$1$A$<$4567#3(

)

B

C

/

D

E

F

*

A

).

))

)B

)

B

C

/

D

E

F A

).

))

)B*

GHIJ#@#$AK L

K

! "!! #!!! #"!! $!!! $"!! %!!! %"!! &!!! &"!! "!!! ""!! (!!!

8#97:$%;<(

!

$!

&!

(!

)!

#!!

!=$%>(

GHIJ#@#$AK !

,-./

!"#

!"$

!%#

!"#$%&'()*

&'($()("(&('

+,- .

/012'(34*

$ )$ "$ &$ '$ ($$ ()$ ("$ (&$ ('$

,5&'(67*

$

'

(&%89:0';

<$5=>&'0

?@AB.

.0

!9 !C!."

DE

.!.8

...1

.F

E#'0

.9 ."

G>&H&IJKL7M!0 !C

!."

.1

.F.9 ."

N,

O.

O0 O/ OFO8O!

<PQ=&I&'R7H$I'%32G3H$J&

!()

!'

!"

$

"

'

-&2G&>7J=>&'(34*

SG$%7'E32&'4

T?@AB

!0

,A+'"!. !0

("*+,-./0"(!123+4-5673-.689+:869.0

+E1!K.008U0

;/<83/0.-.689

Fig. 7

Fig. 10

Fig. 8

Fig. 9