ArticlePDF Available

Inception, history and development of peatlands in the Amazon Basin

Authors:
27 
PAGES news • Vol 18 • No 1 • April 2010
Science Highlights: Peatlands
References
Anshari, G., Kershaw, AP. and van der Kaars, S., 2001: A Late Pleistocene
and Holocene pollen and charcoal record from peat swamp for-
est, Lake Sentarum Wildlife Reserve, West Kalimantan, Indone-
sia, Palaeogeography, Palaeoclimatology, Palaeoecology, 171:
213-228.
Page, S.E., Siegert, F., Rieley, J.O., Boehm, H.-D.V., Jaya, A. and Limin, S.,
2002: The amount of carbon released from peat and forest fires in
Indonesia during 1997, Nature, 420: 61-65.
Page, S.E., Wüst, R.A.J., Weiss, D., Rieley, J.O., Shotyk, W. and Limin, S.H.,
2004: A record of Late Pleistocene and Holocene carbon accumu-
lation and climate change from an equatorial peat bog (Kaliman-
tan, Indonesia): implications for past, present and future carbon
dynamics, Journal of Quaternary Science, 19: 625-635.
Page, S.E., Rieley, J.O. and Wüst, R., 2006: Lowland tropical peatlands
of Southeast Asia. In: Martini, P., Martinez-Cortizas, A. and
Chesworth, W. (Eds) Peatlands: basin evolution and depository
of records on global environmental and climatic changes, Elsevier,
Amsterdam (D evelopments in Earth Surface Processes series),
pp. 145-172.
Wüst, R.A.J. and Bustin, R.M., 2004: Late Pleistocene and Holocene
development of the interior peat-accumulating basin of tropical
Tasek Bera, Peninsular Malaysia, Palaeogeography, Palaeoclima-
tology, Palaeoecology, 211: 241-270.
For full references please consult:
http://www.pages-igbp.org/products/newsletters/ref2010_1.html
Inception, history and development of peatlands in the
Amazon Basin
outi lähteenoJa1 and katherine h. roucouX2
1Department of Biology, University of Turku, Finland; outi.lahteenoja@utu.fi
2School of Geography, University of Leeds, UK
The existence of peatlands in the Amazonian lowlands has only recently been confirmed, owing to the
remoteness of the area. These peatlands are important for regional carbon cycling and habitat diversity, and
represent valuable potential resources for paleoecological research.
The Amazon’s oodplain
peatlands
Amazonia, the world’s largest continuous 
area  of  humid  tropical  lowland  rainfor-
est, is famous  for its dense river network, 
large seasonal variations in water level (on 
average 10 m  at  Manaus, Brazil),  and ex-
tensive oodplains and wetlands covered 
by Mauritiapalms, oodplain forest or sa-
vanna-like vegetation (Irmler, 1977; Junk, 
1983;  Junk  and  Piedade,  2005;  Keddy  et 
al., 2009). Despite the great extent of wet-
lands within  the  Amazon  Basin, the exis-
tence of tropical peatlands has rarely been 
considered  (but  see  Suszczynski,  1984; 
Schulman et al., 1999; Ruokolainen et al., 
2001;  Guzmán  Castillo,  2007).  Two  stud-
ies  carried  out  recently  in  Peruvian  low-
land  Amazonia  (Loreto  region,  Fig.  1)  by 
members  of  the  Amazon  Research  Team 
of the  University of Turku (Finland) reveal 
that  peat  deposits,  up  to  6  m  thick,  are 
widespread on oodplain wetlands of the 
Western Amazon Basin (Lähteenoja et  al., 
2009a, 2009b). Sixteen of seventeen stud-
ied wetland  sites contained some kind  of 
peat deposit. According to the very rough 
estimate of Schulman  et  al.  (1999)  based 
on local land-cover maps, satellite images, 
grey literature and sporadic eld observa-
tions, Amazonian peatlands may cover up 
to 150 000 km2, an area equivalent to half 
of Finland, and about 75 % of the area cov-
ered  by  the  better-known  tropical  peat-
lands of Indonesia (Rieley and Page, 2005; 
Page et al., this issue).
History and development
Since  their  late  Holocene  inception,  the 
peatlands identied in Peruvian Amazonia 
have accumulated peat and carbon at rel-
atively high rates (0.94 - 4.88 mm per year, 
and 26 - 195 g C m-2 per year, respectively) 
(Fig. 2) and  therefore  constitute  a  strong 
carbon  sink  (Lähteenoja  et  al.,  2009b). 
These accumulation rates are comparable 
to those  of  the  Indonesian tropical peat-
lands  (Page  et  al.,  2004)  and  are  higher 
than  those  of  the  boreal  peatlands  (To-
lonen and Turunen, 1996).
The basal ages of ve dated peat de-
posits varied  from 0.588 cal ka  BP (at 164 
cm) to 2.945 cal ka BP (at 565 cm) (Lähtee-
noja et al., 2009b), which are considerably 
younger  than  basal  ages  determined  in 
peatlands  in  many  other  regions  of  the 
world (cf., Korhola et al., 2010). There  are 
several  possible  reasons  for  this.  A  pa-
leoecological  study  of  lake  sediments  in 
Peruvian Amazonia  suggests that the dry 
conditions  of  the  middle  Holocene  were 
followed  by  a  period  of  increasingly  wet 
conditions beginning some time between 
4.2 and 2.54 cal ka BP (Bush et al., 2007). 
Although our oldest peat  initiation  dates 
coincide broadly with the onset of this wet 
interval,  some  of  the  peat  deposits  have 
much younger  basal  ages  (Lähteenoja  et 
al., 2009b), indicating that peat formation 
was  not  determined  purely  by  climate. 
Peat  initiation  may  be  controlled  by  the 
dynamic  lateral  migration  of  western 
Amazonian rivers, characterized by mean-
dering  and  avulsion  (Kalliola  et  al.,  1992; 
Neller et  al.,  1992;  Pärssinen  et  al.,  1996), 
which  have  the  potential  to  erode  and 
bury  peat  deposits.  Peat  accumulation 
probably began when an area with water-
logged  conditions  was  isolated  from  the 
immediate destructive inuence of rivers. 
Consequently, the Western Amazon Basin 
Figure 1: The location of the study sites (from Lähteenoja et al., 2009b, Fig. 1). The map is a mosaic of histogram-
equalized Landsat TM satellite images (www.glcf.umiacs.umd.edu/). Palm swamps and forested wetlands have a
reddish tone, more or less treeless open areas (like the open peatland Riñón) are blue-green, and other floodplain
forests are pinkish to white.
28 
PAGES news • Vol 18 • No 1 • April 2010
Science Highlights: Peatlands
may be too dynamic to allow very old peat 
deposits to form (Lähteenoja et al., 2009b), 
compared with, for example, those found 
in  the  more  geologically  stable  Indone-
sian tropical lowlands  (Page et al., this is-
sue).  This  reasoning  is  supported  by  the 
presence  of  several  buried  peat  deposits 
observed under the mineral subsoil of the 
peatlands (Lähteenoja et al., 2009b). Older 
peat deposits might well be found in geo-
logically  more  stable  peripheral  areas  of 
the  oodplains  close  to  the  non-ooded 
terra rme.
According  to  peat  nutrient  analyses 
and topographic measurements, some of 
the thickest, oldest and most stable peat-
lands  have  attained  ombrotrophic  (rain 
fed)  conditions,  despite  their  location  in 
the  middle  of  a  oodplain  environment 
(Lähteenoja  et  al.,  2009a).  The sur face  of 
these peatlands has risen above the maxi-
mum level of river oods because of their 
thick  peat  layer  and  convex  topography 
(Fig. 2). This change  of conditions, from a 
groundwater-fed system to a rain fed one, 
aects the ecosystem properties in a dras-
tic  way,  and,  consequently,  the  variation 
of ombrotrophic bogs and minerotrophic 
swamps in the Amazonian lowlands  con-
tributes to  the  regional  ecosystem diver-
sity (Fig. 3; Lähteenoja et al., 2009a).
Record of paleoclimate and
vegetation dynamics
The existence of peat deposits (especially 
ombrotrophic  ones)  within  the  Amazon 
Basin  potentially  oers  an  excellent  re-
source  for  studies  of  Holocene  climate 
variability,  paleohydrology and  rainforest 
vegetation  dynamics  in  the  Amazonian 
lowlands,  providing  histories  extending 
to the time of peat initiation (cf., Frost and 
Miller,  1987;  Ledru,  2001;  Hoorn,  2006). 
In  a  new  project,  due  to  begin  in  sum-
mer  2010,  we  will  apply  pollen,  charcoal 
and  sedimentological  analyses  to  three 
of the  peat  sequences identied by  Läh-
teenoja  et  al.  (2009b)  in  order  to  recon-
struct changes in forest composition over 
the past  3  ka,  focusing in  most detail  on 
the  last  1  ka,  the  interval  of  most  direct 
relevance  to  current  ecological  trends. 
Knowledge  of  tropical  forest  history  on 
this timescale is crucial to the interpreta-
tion  and  understanding  of  recent  eco-
logical changes taking place  in the forest 
(Malhi  et  al.,  2002).  For  example,  studies 
of a network of forest plots across Amazo-
nia (Malhi et al., 2002) show that  rates  of 
tree mortality  and recruitment (Phillips et 
al., 2004), growth rates (Lewis et al., 2004) 
and  overall  forest  biomass  (Baker  et  al., 
2004) have increased over the  past  three 
decades. Understanding the  mechanisms 
behind  these  changes  is  important  for 
predicting  their  consequences  for  forest 
biodiversity and  for  the  role  of  forests in 
the global  carbon cycle (Cox et al., 2008). 
One  possible  explanation  is  that  forests 
are  recovering  from  disturbance  events 
prior  to  the  start  of  monitoring  (Wright, 
2005).  We  intend  to  test  this  hypothesis 
by  applying  paleoecological  techniques 
to the newly discovered peat sequences. 
They are ideal for the purpose because: 1) 
they record an interval for which there are 
currently no  detailed,  high  resolution re-
cords in this region; 2) the peat accumulat-
ed rapidly so  should  yield  pollen  records 
with  decadal  scale  temporal  resolution; 
and 3) they are located in a region where 
forest ecology has been monitored in per-
manent census plots for the past three de-
cades (RAINFOR project: Malhi et al., 2002). 
The new records will greatly improve our 
understanding of the mechanisms driving 
Figure 2: Peat profile and accumulation rates from the San Jorge ombrotrophic bog (Loreto region, Peru). Brown
= peat, dark brown = clayey peat, gray = clay. Redrawn from Lähteenoja et al., 2009a). The core location and
peat accumulation rates (mm/a) are shown in red (from Lähteenoja et al., 2009b).
Figure 3: Four of the peatland sites indentified by Lähteenoja et al. (2009a, 2009b) in Peruvian Amazonia (Loreto): a)
Mauritia flexuosa peat swamp (Quistococha), b) forested peatland (San Jorge), c) savanna-like peatland (Riñón),
d) forested peatland (Fundo Junior).
29 
PAGES news • Vol 18 • No 1 • April 2010
Science Highlights: Peatlands
29 
current changes in tropical forest ecology 
and the  sensitivity  of  these forests to fu-
ture climatic change.
In  contrast  to  the  Southeast  Asian 
tropical peatlands, these Amazonian peat-
land sites do not appear to be currently di-
rectly aected by anthropogenic actions. 
Nevertheless,  climate  change,  deforesta-
tion,  large-scale  land-use  projects  (such 
as river damming, road  construction  and 
development of oil palm plantations) and 
extensive gas and oil exploration (Malhi et 
al.,  2008)  represent  an  indirect  threat  to 
the  peatlands  insofar  as  they  contribute 
to drying  of  the regional climate. Conse-
quently, there is an urgent need to inves-
tigate  further,  and  conserve,  these  little-
known Amazonian ecosystems.
References
Guzmán Castillo, W., 2007: Valor económico del manejo sostenible de
los ecosistemas de aguaje (Mauritia flexuosa). In: Feyen, J., et
al. (Eds), International Congress on Development, Environment
and Natural Resources: Multi-level and Multi-scale Sustainability,
Volume III, Publication of the Universidad Mayor San Simón, Co-
chabamba, Bolivia, 1513-1521.
Lähteenoja, O., Ruokolainen, K., Schulman, L. and Alvarez, J., 2009a:
Amazonian floodplains harbour minerotrophic and ombrotro-
phic peatlands, Catena, 79: 140-145.
Lähteenoja, O., Ruokolainen, K., Schulman, L. and Oinonen, M., 2009b:
Amazonian peatlands: an ignored C sink and potential source,
Global Change Biology, 15: 2311-2320.
Ruokolainen, K., Schulman, L. and Tuomisto, H., 2001: On Amazonian
peatlands, International Mire Conservation Group Newsletter,
2001(4): 8-10.
Schulman, L., Ruokolainen, K. and Tuomisto, H., 1999: Parameters for
global ecosystem models, Nature, 399: 535-536.
For full references please consult:
http://www.pages-igbp.org/products/newsletters/ref2010_1.html
Peatland exchanges of CO2 and CH4: The importance of
presence or absence of permafrost
torben r. chriStenSen1, m. maStepanov1, m. JohanSSon1 and d. charman2
1Department of Earth and Ecosystem Science, Lund University, Sweden; torben.christensen@nateko.lu.se
2School of Geography, University of Exeter, UK
The presence of permafrost is shown to have dramatic impacts on land-atmosphere exchanges of key
greenhouse gases.
Permafrost and the carbon cycle
Permafrost,  soil  that  stays  frozen for  two 
or more years in a row, is a hot topic that 
has  attracted  a  lot  of  attention  in  both 
the  scientic  and  popular  literature  in 
recent  years.  Permafrost  underlies  25% 
of  the  land  areas  in  the  Northern  Hemi-
sphere  including  substantial  areas  with 
peatlands. With a warming climate that is 
particularly pronounced at  high northern 
latitudes, where most permafrost is pres-
ent, many questions have been raised re-
garding  what  may  happen  to  peatlands 
and  their  functioning  when  permafrost 
thaws. In areas with infrastructure, such as 
towns in  northern  Siberia, or  oil  and  gas 
pipelines through areas underlain by per-
mafrost, the thawing represents a serious 
and possibly  very  expensive  issue. Thaw-
ing permafrost may, however, have global 
implications  through  changes  in  natural 
ecosystem greenhouse-gas emissions.
Permafrost  areas  in  the  circumpolar 
North  are  estimated  to  hold  more  than 
1600 Pg of organic carbon (C) including al-
most 300 Pg in the form of peat (McGuire 
et al., 2009; Tarnocai et al., 2009) most of 
which has accumulated since the last gla-
cial maximum. In terms of atmospheric ex-
change of  carbon, in the  form of CO2and 
CH4,  the  potential  for  additional  releases 
are  probably  greater  from  these  areas 
than anywhere else in the world. While the 
potential release  from the huge stocks of 
carbon is  signicant,  the  actual data and 
year-round  monitoring  of  atmospheric 
exchanges  remain  rare,  and  continuous 
ux  measurements  of  CO2  are  limited  to 
a  handful  of  sites.  Continuous  monitor-
ing of CH4 uxes is even rarer; the number 
of  operational  sites  is  less  than  ve.  Our 
empirically based understanding of what 
permafrost does to the dynamics and in-
terannual  variability  in  atmospheric  (and 
dissolved run-o) uxes of organic carbon 
is therefore still very poor. The longer-term 
dynamics on decadal to centennial times-
cales are even less well understood.
Carbon dynamics
Basic features of how ecosystems are func-
tioning with and without permafrost have 
recently  been  discovered.  At  a  central 
Alaskan site, Schuur et  al.  (2009)  demon-
strated that permafrost thawing is accom-
panied by respiration of previously frozen, 
ancient  organic  carbon.  In  Siberian  thaw 
lakes, methane has  been  observed  form-
ing from recently thawed Pleistocene or-
ganic deposits (Walter et al., 2007).
The  interannual  and  across-site  vari-
ability  of  CO2  exchange  in  continuous 
permafrost  ecosystems  are  driven  pri-
marily  by  growing-season  dynamics  and 
moisture conditions. Several studies have 
shown  that  growing-season  rates  of  CO2
uptake by these  ecosystems is closely re-
Figure 1: The Zackenberg valley in NE Greenland, an area underlain by continuous permafrost. The automatic
chambers were used for the studies of methane emission dynamics during freeze-in (Mastepanov et al., 2008).
Local inhabitants, the muskoxen, are present in the background. Photo by C. Sigsgaard, from Christensen et al.,
2009, reprinted with permission.
O. Lähteenoja and K.H. Roucoux
Baker, T.R., et al., 2004: Variation in wood density determines spatial patterns in Amazonian
forest biomass, Global Change Biology, 10: 545-562.
Bush, M.B., Silman, M.R. and Listopad, M.C.S., 2007: A regional study of Holocene climate
change and human occupation in Peruvian Amazonia, Journal of Biogeography, 34(8):
1342-1356.
Cox, P.M., Harris, P.P., Huntingford, C., Betts, R.A., Collins, M., Jones, C.D., Jupp, T.E.,
Marengo, J.A. and Nobre, C.A., 2008: Increasing risk of Amazonian drought due to
decreasing aerosol pollution, Nature, 453: 212-215.
Frost, I.G. and Miller, M.C., 1987: Late Holocene flooding in the Ecuadorian rainforest,
Freshwater Biology 18(3): 443-453.
Guzmán Castillo, W., 2007: Valor económico del manejo sostenible de los ecosistemas de
aguaje (Mauritia flexuosa). In: Feyen, J., et al. (Eds), International Congress on
Development, Environment and Natural Resources: Multi-level and Multi-scale
Sustainability, Volume III, Publication of the Universidad Mayor San Simón,
Cochabamba, Bolivia, 1513-1521.
Hoorn, C., 2006: Mangrove Forests and Marine Incursions in Neogene Amazonia (Lower
Apaporis River, Colombia), Palaios, 21: 197-209.
Irmler, U., 1977: Inundation forest types in the vicinity of Manaus, Biogeographica 8: 17-
29.
Junk, W.J., 1983: Ecology of swamps on the middle Amazon. In: Gore, A.J.P. (Ed), Mires:
swamp, bog, fen and moor, regional studies; ecosystems of the World 4B, Elsevier,
Amsterdam, The Netherlands, 269-294.
Junk, W.J. and Piedade, M.T.F., 2005: The Amazon River basin. In: Fraser, L.H., et al. (Eds),
The World’s Largest Wetlands: Ecology and Conservation, Cambridge University Press,
63-117.
Kalliola, R., Salo, J., Puhakka, M., Rajasilta, M., Häme, T., Neller, R.J., Räsänen, M.E. and
Danjoy, Arias W.A., 1992: Upper Amazon channel migration: Implications for
vegetation perturbance and succession using bitemporal Landsat MSS images,
Naturwissenschaften, 79: 75-79.
Keddy, P.A., Fraser, L.H., Solomeshch, A.I., Junk, W.J., Campbell, D.R., Arroyo, M.T.K.
and Alho, C.J.R., 2009: Wet and Wonderful: The World's Largest Wetlands Are
Conservation Priorities, BioScience, 59(1): 39-51.
Korhola, A., Ruppel, M., Seppä, H., Väliranta, M., Virtanen, T. and Weckström, J., in press:
The importance of northern peatland expansion to the late-Holocene rise of atmospheric
methane, Quaternary Science Reviews, 29(5-6): 611-617.
Lähteenoja, O., Ruokolainen, K., Schulman, L. and Alvarez, J., 2009a: Amazonian
floodplains harbour minerotrophic and ombrotrophic peatlands, Catena, 79: 140-145.
Lähteenoja, O., Ruokolainen, K., Schulman, L. and Oinonen, M., 2009b: Amazonian
peatlands: an ignored C sink and potential source, Global Change Biology, 15: 2311-
2320.
Ledru, M.P., 2001: Late Holocene rainforest disturbance in French Guiana, Review of
Palaeobotany and Palynology, 115: 161-170.
Lewis, S.L., et al., 2004: Are global change agents causing widespread changes in tropical
forest dynamics? Evidence from 50 South American long-term monitoring plots,
Philosophical Transactions of the Royal Society of London, Series B, 359: 421-436.
Malhi, Y., et al., 2002: An international network to understand the biomass and dynamics of
Amazonian forests (RAINFOR), Journal of Vegetation Science, 13: 439-450.
Malhi, Y., Roberts, J.T., Betts, R.A., Killeen, T.J., Li, W. and Noble, C.A., 2008: Climate
Change, Deforestation and the Fate of the Amazon, Science, 319: 169-172.
Neller, R.J., Salo, J.S. and Räsänen, M.E., 1992: On the formation of blocked valley lakes by
channel avulsion in Upper Amazon foreland basins, Zeitschrift für Geomorphologie,
36(4): 401-411.
Page, S.E., Wüst, R.A.J., Weiss, D., Rieley, J.O., Shotyk, W. and Limin, S.H., 2004: A record
of Late Pleistocene and Holocene carbon accumulation and climate change from an
equatorial peat bog (Kalimantan, Indonesia): implications for past, present and future
carbon dynamics, Journal of Quaternary Science, 19: 625-635.
Page, S.E., Wüst R.A.J. and Banks, C., 2010: Past and present carbon accumulation and loss
in Southeast Asian peatlands, PAGES news, 18(1).
Pärssinen, M.H., Salo, J.S. and Räsänen, M.E., 1996: River floodplain relocations and the
abandonment of aborigine settlements in the Upper Amazon Basin: a historical case
study of San Miguel de Cunibos at the Middle Ucayali River, Geoarchaeology: An
International Journal, 11(4): 345-359.
Phillips, O., Baker, T.R., Arroyo, L., Higuchi, N., Killeen, T., Laurance, W.F., Lewis, S.L.,
Lloyd, J., Malhi, Y. and Monteagudo, A., 2004: Pattern and process in Amazon tree
turnover, 1976-2001, Philosophical Transactions of the Royal Society of London, Series
B, 359: 381-407.
Rieley, J.O. and Page, S.E., 2005: Wise use of tropical peatlands: Focus on Southeast Asia.
Alterra - Wageningen University and Research Center and the EU INCO-STRAPEAT
and RESTORPEAT Partnerships: 168.
Ruokolainen, K., Schulman, L. and Tuomisto, H., 2001: On Amazonian peatlands,
International Mire Conservation Group Newsletter, 2001(4): 8-10.
Schulman, L., Ruokolainen, K. and Tuomisto, H., 1999: Parameters for global ecosystem
models, Nature, 399: 535-536.
Suszczynski, E., 1984: The peat resources of Brazil. Proceedings of the 7th International
Peat Congress, Dublin 1: 468-492.
Tolonen, K. and Turunen, J., 1996: Accumulation rates of carbon in mires in Finland and
implications for climate change, The Holocene, 6(2): 171-178.
Wright, S.J., 2005: Tropical forests in a changing environment, Trends in Ecology &
Evolution, 20(10): 553-560.
... National maps of New Zealand peatlands were derived from the Fundamental Soil Layers (FSL) soil maps published at 1 : 50 000 scale by the New Zealand Land Resource Inventory (NZLRI; Landcare Research NZ Ltd, 2000). The polygons in the FSL maps were manually created from aerial photograph analysis with ground truthing. ...
Article
Full-text available
Peatlands store large amounts of soil carbon and freshwater, constituting an important component of the global carbon and hydrologic cycles. Accurate information on the global extent and distribution of peatlands is presently lacking but is needed by Earth system models (ESMs) to simulate the effects of climate change on the global carbon and hydrologic balance. Here, we present Peat-ML, a spatially continuous global map of peatland fractional coverage generated using machine learning (ML) techniques suitable for use as a prescribed geophysical field in an ESM. Inputs to our statistical model follow drivers of peatland formation and include spatially distributed climate, geomorphological and soil data, and remotely sensed vegetation indices. Available maps of peatland fractional coverage for 14 relatively extensive regions were used along with mapped ecoregions of non-peatland areas to train the statistical model. In addition to qualitative comparisons to other maps in the literature, we estimated model error in two ways. The first estimate used the training data in a blocked leave-one-out cross-validation strategy designed to minimize the influence of spatial autocorrelation. That approach yielded an average r2 of 0.73 with a root-mean-square error and mean bias error of 9.11 % and −0.36 %, respectively. Our second error estimate was generated by comparing Peat-ML against a high-quality, extensively ground-truthed map generated by Ducks Unlimited Canada for the Canadian Boreal Plains region. This comparison suggests our map to be of comparable quality to mapping products generated through more traditional approaches, at least for boreal peatlands.
... These wetland mapping products are, however, of limited utility for peatland modelling applications as they generally do not agree well amongst themselves (Melton et al., 2013), which is also the case for peatland mapping products (as is discussed later) and may exhibit biases depending on how they were generated (see discussion in Bohn et al., 2015). In addition, in the boreal zone and some areas of the tropics such as the Amazon (Lähteenoja and Roucoux, 2010), some peatlands are not inundated, and thus using hydrological characteristics alone can underestimate their extent (Matthews, 1989;Prigent et al., 2007). Other studies, such as Largeron et al. (2018) or Leifeld and Menichetti (2018), have used a global peatland distribution map derived from a paleontological perspective . ...
Preprint
Peatlands store large amounts of soil carbon and freshwater, constituting an important component of the global carbon and hydrologic cycles. Accurate information on the global extent and distribution of peatlands is presently lacking but is needed by Earth System Models (ESMs) to simulate the effects of climate change on the global carbon and hydrologic balance. Here, we present Peat-ML, a spatially continuous global map of peatland fractional coverage generated using machine learning techniques suitable for use as a prescribed geophysical field in an ESM. Inputs to our statistical model follow drivers of peatland formation and include spatially distributed climate, geomorphological and soil data, along with remotely-sensed vegetation indices. Available maps of peatland fractional coverage for 14 relatively extensive regions were used along with mapped ecoregions of non-peatland areas to train the statistical model. In addition to qualitative comparisons to other maps in the literature, we estimated model error in two ways. The first estimate used the training data in a blocked leave-one-out cross-validation strategy designed to minimize the influence of spatial autocorrelation. That approach yielded an average r2 of 0.73 with a root mean squared error and mean bias error of 9.11 % and −0.36 %, respectively. Our second error estimate was generated by comparing Peat-ML against a high-quality, extensively ground-truthed map generated by Ducks Unlimited Canada for the Canadian Boreal Plains region. This comparison suggests our map to be of comparable quality to mapping products generated through more traditional approaches, at least for boreal peatlands.
... The controlling factor for development of peatlands and peats as proposed by Gore and Goodall (1983), Moore (1987), Moore and Bellamy (1974) and Shepard (2006) is hydrology from surface water flow to subsurface water or groundwater level. However, recent studies of Amazonian and Southeast Asian peatlands indicate that precipitation (e.g., rainfall) and surface waters or groundwater capillary rise are equally important factors (Anderson, 1961(Anderson, , 1964(Anderson, , 1983Cecil et al., 1985;Wüst et al., 2001Wüst et al., , 2007Wüst and Bustin, 2004;Moore et al., 1996aMoore et al., , 1996bMoore and Shearer, 2003;Lahteenoja et al., 2009;Lahteenoja and Roucox, 2010;Page et al., 2010). Amazonian floodplains, for example, form minerotrophic (low-lying) and ombrotropic (domed, convex topography) peatlands (Lahteenoja et al., 2009). ...
Article
Full-text available
Peat depositional environments, the sites where and conditions under which peat accumulates, significantly influence a resultant coal's physical properties, chemical composition, and coal utilization behavior. Recognition of peat depositional environments for coal is a challenging endeavor because coal's observed compositional properties not only result from a variety of geological processes operating during peat accumulation, but also reflect the influence of adjoining or external depositional sedimentary environments and alteration during later diagenesis and/or epigenesis. The maceral or microlithotype composition of any one layer of peat can be the product of years or decades of plant growth, death, decay, and post-burial infiltration by roots in addition to the symbiotic, mutualistic, parasitic, and saprophytic relationships with non-plant biota, such as arthropods, fungi, and bacteria. The overprint of increasing thermal maturation and fluid migration through time on the resulting coal can make these relationships difficult to recognize. Therefore, published models based on maceral composition alone must be used with great caution. Lipid compositions, even from lipid-poor low-rank coals, can provide important information about depositional environments and paleoclimate, especially if combined with the results of organic petrography and paleontological studies. Just as sulfur derived from seawater provides environmental clues, the ratios of two particularly relevant trace elements rather than a single trace element can be used to interpret peat depositional environments. Epigenetic minerals, as well as their corresponding chemical compositions should not be used for such a purpose; similarly, resistant terrigenous minerals deposited during peat accumulation in many cases should be used with considerable caution. The interactions of the biota present in the peat-forming ecosystem, often determined using palynological and geochemical proxies, and their interpretation in the context of geography and paleoclimate are important means for deciphering peat depositional environments. Overall, a combination of evidence from geochemistry, mineralogy, palynology, and petrology of coal and from stratigraphy, sedimentology, and sedimentary facies of related rocks is necessary for accurate and comprehensive determination of depositional environments. The need for interdisciplinary studies is underscored by peat compositional properties, which have been greatly affected by various processes during the syngenetic, diagenetic or epigenetic stages of coal formation.
... Esta nueva dinámica tectónica y uvial in uyó en la geomorfología de la región con extensas planicies aluviales, tierra rme y pantanos. Adicionalmente, el neotectonismo en las subcuencas del ante-arco impulsó la migración lateral de los ríos amazónicos convirtiéndose en una importante fuerza en la con guración del paisaje amazónico 46 . ...
... In contrast, another ecosystem type, peatland pole forests Page 2011, Draper et al. 2014) has only been loosely described (Lähteenoja and Roucoux 2010, García-Villacorta et al. 2011, Torres Montenegro et al. 2015, and its floristic composition remains largely unknown. These pole forests occur on raised ombrotrophic bogs (water and nutrients dominantly supplied by atmospheric deposition; Rydin and Jeglum 2006), normally on top of several meters of peat, and are characterised by their distinctive structure which consists of a dense, low stature canopy made up of many thin-stemmed trees (Draper et al. 2014). ...
Article
Western Amazonia is known to harbour some of Earth's most diverse forests, but previous floristic analyses have excluded peatland forests which are extensive in northern Peru and are among the most environmentally extreme ecosystems in the lowland tropics. Understanding patterns of tree species diversity in these ecosystems is important both for quantifying beta-diversity in this region, and for understanding determinants of diversity more generally in tropical forests. Here we explore patterns of tree diversity and composition in two peatland forest types – palm swamps and peatland pole forests – using 26 forest plots distributed over a large area of northern Peru. We place our results in a regional context by making comparisons with three other major forest types: terra firme forests (29 plots), white-sand forests (23 plots) and seasonally-flooded forests (11 plots).
... Moreover, Amazonian peatlands are thought to account for a substantial land area (i.e., up to 150 000 km 2 ) ( Schulman et al., 1999;Lahteenoja et al., 2012), and any differences in biogeochemistry among peat and mineral/organomineral soil wetlands may therefore have important implications for understanding and modeling the biogeochemical functioning of the Amazon basin as a whole. Since the identification of extensive peat forming wetlands in the north ( Lahteenoja et al., 2009a, b; Lahteenoja and Page 2011) and south ( Householder et al., 2012) of the Peruvian Amazon, several studies have been undertaken to better characterize these habitats, investigating vegetation composition and habitat diversity ( Draper et al., 2014;Kelly et al., 2014;Householder et al., 2012;Lahteenoja and Page, 2011), vegetation history ( Lähteenoja and Roucoux, 2010), C stocks ( Lahteenoja et al., 2012;Draper et al., 2014), hydrology ( Kelly et al., 2014), and peat chemistry ( Lahteenoja et al., 2009a, b). Most of the studies have focused on the PastazaMarañón foreland basin (PMFB), where one of the largest stretches of contiguous peatlands has been found ( Lahteenoja et al 2009a;Lahteenoja and Page, 2011;Kelly et al., 2014), covering an estimated area of 35 600 ± 2133 km 2 ( Draper et al., 2014). ...
Article
Full-text available
The Amazon plays a critical role in global atmospheric budgets of methane (CH4) and nitrous oxide (N2O). However, while we have a relatively good understanding of the continental-scale flux of these greenhouse gases (GHGs), one of the key gaps in knowledge is the specific contribution of peatland ecosystems to the regional budgets of these GHGs. Here we report CH4 and N2O fluxes from lowland tropical peatlands in the Pastaza–Marañón foreland basin (PMFB) in Peru, one of the largest peatland complexes in the Amazon basin. The goal of this research was to quantify the range and magnitude of CH4 and N2O fluxes from this region, assess seasonal trends in trace gas exchange, and determine the role of different environmental variables in driving GHG flux. Trace gas fluxes were determined from the most numerically dominant peatland vegetation types in the region: forested vegetation, forested (short pole) vegetation, Mauritia flexuosa-dominated palm swamp, and mixed palm swamp. Data were collected in both wet and dry seasons over the course of four field campaigns from 2012 to 2014. Diffusive CH4 emissions averaged 36.05 ± 3.09 mg CH4–C m⁻² day⁻¹ across the entire dataset, with diffusive CH4 flux varying significantly among vegetation types and between seasons. Net ebullition of CH4 averaged 973.3 ± 161.4 mg CH4–C m⁻² day⁻¹ and did not vary significantly among vegetation types or between seasons. Diffusive CH4 flux was greatest for mixed palm swamp (52.0 ± 16.0 mg CH4–C m⁻² day⁻¹), followed by M. flexuosa palm swamp (36.7 ± 3.9 mg CH4–C m⁻² day⁻¹), forested (short pole) vegetation (31.6 ± 6.6 mg CH4–C m−2 day⁻¹), and forested vegetation (29.8 ± 10.0 mg CH4–C m−2 day⁻¹). Diffusive CH4 flux also showed marked seasonality, with divergent seasonal patterns among ecosystems. Forested vegetation and mixed palm swamp showed significantly higher dry season (47.2 ± 5.4 mg CH4–C m⁻² day⁻¹ and 85.5 ± 26.4 mg CH4–C m⁻² day⁻¹, respectively) compared to wet season emissions (6.8 ± 1.0 mg CH4–C m⁻² day⁻¹ and 5.2 ± 2.7 mg CH4–C m⁻² day⁻¹, respectively). In contrast, forested (short pole) vegetation and M. flexuosa palm swamp showed the opposite trend, with dry season flux of 9.6 ± 2.6 and 25.5 ± 2.9 mg CH4–C m⁻² day⁻¹, respectively, versus wet season flux of 103.4 ± 13.6 and 53.4 ± 9.8 mg CH4–C m⁻² day⁻¹, respectively. These divergent seasonal trends may be linked to very high water tables (> 1 m) in forested vegetation and mixed palm swamp during the wet season, which may have constrained CH4 transport across the soil–atmosphere interface. Diffusive N2O flux was very low (0.70 ± 0.34 µg N2O–N m⁻² day⁻¹) and did not vary significantly among ecosystems or between seasons. We conclude that peatlands in the PMFB are large and regionally significant sources of atmospheric CH4 that need to be better accounted for in regional emissions inventories. In contrast, N2O flux was negligible, suggesting that this region does not make a significant contribution to regional atmospheric budgets of N2O. The divergent seasonal pattern in CH4 flux among vegetation types challenges our underlying assumptions of the controls on CH4 flux in tropical peatlands and emphasizes the need for more process-based measurements during periods of high water table.
Article
Full-text available
El Perú tiene la cuarta mayor área de turberas de los trópicos. Su cobertura terrestre de turba más representativa es el pantano de palmeras (denominado a partir de ahora PP denso) dominado por la especie Mauritia flexuosa, que ha estado sometido a presión humana durante décadas debido a la alta demanda del fruto de M. flexuosa, que a menudo se recolecta cortando toda la palmera. La degradación de estos bosques densos en carbono puede afectar de manera sustancial las emisiones de gases de efecto invernadero y contribuir al cambio climático. El primer objetivo de esta investigación fue evaluar el impacto de la degradación de los PP densos sobre la estructura forestal y las reservas de carbono de la biomasa. El segundo fue explorar el potencial de mapear la distribución de los PP densos con diferentes niveles de degradación utilizando datos y métodos de teledetección. Las reservas de biomasa se midieron en parcelas de 0,25 ha establecidas en áreas de PP densos con degradación baja (n = 2 parcelas), media (n = 2) y alta (n = 4). Se combinaron datos de campo y de teledetección de los satélites Landsat TM y ALOS/PALSAR para diferenciar entre áreas que tipifican PP densos con degradación baja, media y alta y bosques de tierra firme, restinga y PP mixtos (no dominados por M. flexuosa). Para ello, se utilizó un algoritmo de clasificación de aprendizaje automático Random Forest. Los resultados sugieren un cambio en la composición forestal, de bosques dominados por palmeras a bosques dominados por árboles leñosos después de la degradación. También se encontró que la intervención humana en los PP densos se traduce en reducciones significativas en las reservas de carbono de los árboles, con disminuciones de las reservas de biomasa iniciales (aérea y subterránea) (135,4 ± 4,8 Mg C ha-1) de 11% y 17% luego de una degradación media y alta, respectivamente. El análisis de teledetección indica una alta separabilidad entre PP densos con degradación baja y todas las demás categorías. Los PP densos con degradación media y alta fueron altamente separables de la mayoría de las categorías con excepción de los bosques de restinga y los PP mixtos. Los resultados también mostraron que los datos de sensores de teledetección tanto activos como pasivos son importantes para el mapeo de la degradación de PP densos. La precisión general de la clasificación de la cobertura terrestre fue alta (91%). Los resultados de este análisis piloto son alentadores para seguir explorando el uso de datos y métodos de teledetección para el monitoreo de la degradación de PP densos a escalas más amplias en la Amazonia peruana. Brindar estimados precisos sobre la extensión espacial de la degradación de los PP densos y sobre las emisiones derivadas de la biomasa y la turba es necesario para evaluar las emisiones nacionales derivadas de la degradación forestal en el Perú y es esencial para apoyar iniciativas dirigidas a reducir las actividades de degradación.
Article
Full-text available
The Amazon basin is likely to be increasingly affected by environmental changes: higher temperatures, changes in precipitation, CO2 fertilization and habitat fragmentation. To examine the important ecological and biogeochemical consequences of these changes, we are developing an international network, RAINFOR, which aims to monitor forest biomass and dynamics across Amazonia in a co-ordinated fashion in order to understand their relationship to soil and climate. The network will focus on sample plots established by independent researchers, some providing data extending back several decades. We will also conduct rapid transect studies of poorly monitored regions. Field expeditions analysed local soil and plant properties in the first phase (2001- 2002). Initial results suggest that the network has the potential to reveal much information on the continental-scale relations between forest and environment. The network will also serve as a forum for discussion between researchers, with the aim of standardising sampling techniques and methodologies that will enable Amazonian forests to be monitored in a coherent manner in the coming decades.
Article
Full-text available
Wetlands perform many essential ecosystem services—carbon storage, flood control, maintenance of biodiversity, fish production, and aquifer recharge, among others—services that have increasingly important global consequences. Like biodiversity hotspots and frontier forests, the world's largest wetlands are now mapped and described by an international team of scientists, highlighting their conservation importance at the global scale. We explore current understanding of some ecosystem services wetlands provide. We selected four of these wetlands (the largest peatland, West Siberian Lowland; the largest floodplain, Amazon River Basin; the least-known wetland, Congo River Basin; and the most heavily developed wetland, Mississippi River Basin), and we illustrate their diversity, emphasizing values and lessons for thinking big in terms of conservation goals. Recognizing the global significance of these wetlands is an important first step toward forging global conservation solutions. Each of the world's largest wetlands requires a basinwide sustainable management strategy built on new institutional frameworks—at international, national, and regional levels—to ensure provision of their vital services.
Article
Blocked valley lakes in tectonically active upper Amazon foreland basins were investigated using satellite and radar imagery. The survey revealed that blocked valley lakes formed as a result of channel avulsions disrupting pre-existing drainage patterns, and that such lakes could be formed adjacent to either the newly formed or the abandoned floodplains. This contrasts with descriptions of blocked valley lake formation in Papua and in the middle Amazon basin, where such lakes form adjacent to relatively stable or confined meander plains. -Authors
Article
We agree that the Terrestrial Ecosystem Model analyses by Tianet al. of carbon fluxes of Amazonian ecosystems represent a methodological improvement compared with extrapolation from site-specific estimates, especially with regard to spatial resolution. However, the resolution of a model cannot be finer than that of the input data, and Tian et al. disregard one important group of ecosystems: peatlands. This is understandable, as the literature grossly underestimates the extent of peatlands in Amazonia. Our estimate is 150,000 km2, ten times more than previously reported.