A preview of this full-text is provided by Springer Nature.
Content available from Nature Climate Change
This content is subject to copyright. Terms and conditions apply.
Articles
https://doi.org/10.1038/s41558-018-0177-y
1Aquatic Ecology and Water Quality Management Group, Wageningen University, Wageningen, The Netherlands. 2Department of Environmental
Sciences, Copernicus Institute for Sustainable Development, Utrecht University, Utrecht, The Netherlands. 3Department of Physical Geography, Utrecht
University, Utrecht, The Netherlands. 4Resource Ecology Group, Wageningen University, Wageningen, The Netherlands. 5Biodiversity, Macroecology and
Biogeography Group, University of Goettingen, Göttingen, Germany. 6Earth System Analysis, Potsdam Institute for Climate Impact Research, Potsdam,
Germany. 7Faculty of Management, Science and Technology, Open University, Heerlen, The Netherlands. *e-mail: ariestaal@gmail.com
The Amazon rainforest and its wet climate are mutually depen-
dent1–3. Although their interactions have been a topic of
study for decades2,4–12, we still lack a deep understanding of
the positive effects between forests and rainfall across the Amazon
basin. Forests access groundwater in deep soil layers and release
it to the atmosphere by transpiration1,13. This transpired moisture
can precipitate and evapotranspire repeatedly over forests14,15, pro-
moting forest growth in a cascading way. The significance of such
cascades is poorly understood due to uncertainties regarding the
contribution of tree transpiration to total evapotranspiration, the
atmospheric path of this transpiration to rainfall, the number of
re-evapotranspiration cycles that water goes through and the effect
of the resulting rainfall increase on local forest stability3,8,16. Also,
the temporal variability of the forest’s self-stabilizing mechanism is
poorly understood8. In the wet tropics including the Amazon, tree
cover declines at higher seasonal and inter-annual rainfall variabil-
ity17. To understand the effects of extreme weather events and defor-
estation, a quantitative and temporally explicit assessment of the
forest-rainfall cascades18,19 is needed8. We can now track moisture
flows quantitatively using empirically derived atmospheric wind
patterns, evapotranspiration and rainfall2,20. Also, complex models
are being improved that can quantify the contribution of vegetation
to evapotranspiration on increasingly high spatial and temporal
resolutions21. Finally, remote sensing now provides tools to quan-
tify forest resilience22–25. Here, we capitalize on those technological
advances to provide an empirically derived quantification of the
spatial and temporal interactions between rainfall and tree cover in
tropical South America, focusing on the Amazon basin (Methods).
We use a Lagrangian moisture-tracking algorithm26–28 that calcu-
lates atmospheric water flows in time steps of 0.25 h. We use out-
put on a 0.25° grid (around 25 km × 25 km) and monthly basis for
2003–2014. We account for multiple re-evapotranspiration cycles of
this moisture15 and use a large-scale hydrological model to calcu-
late the evapotranspiration change from potential tree-cover loss for
each month21,29. Combined, these calculations allow us to estimate
the contribution of tree cover to rainfall and related forest resilience
in downwind areas. As tree transpiration is a source of atmospheric
moisture that can be maintained in periods when rainfall is absent30,
we focus on the role of transpiration-induced rainfall.
We find that trees within the Amazon have transpired 20% of all
rainfall in the basin at least once (Fig. 1). We call this contribution of
trees to rainfall the transpiration recycling ratio (TRR). Half of this
transpiration recycling occurs in a direct way, in which moisture falls
back as rainfall after having last entered the atmosphere through
transpiration. The other half is composed of cascading transpira-
tion recycling (see Methods) in which transpiration-induced rain-
fall goes through an additional evapotranspiration–rainfall cycle
at least once (Fig. 1 and Supplementary Fig. 10). Considering all
evapotranspiration (including transpiration), we find that 32% of
Amazonian rainfall originates from the basin, in good agreement
with 10 previous estimates based on different methodologies and
datasets (24‒41% (median 28%); see Methods and ref. 15, including
the references therein). Combining this evapotranspiration recy-
cling with our transpiration recycling estimates, we find that 64% of
all regionally recycled water has travelled through the pores of leaves
of trees in the Amazon. The cascading contribution of transpiration
to rainfall entails that its effects can be remote and, because a single
transpired water molecule can undergo multiple re-evapotranspira-
tion and rainfall cycles, could be larger than the transpired amount
of water itself. We estimate that the largest transpiration is from the
north-eastern, southern and south-western parts of the Amazon
basin (Fig. 2a and Supplementary Fig. 4), in agreement with other
global transpiration estimates31,32. Loss of tree cover in these regions
would thus result in the largest loss of moisture for the basin. The
small transpiration flux in the north-western Amazon is striking,
but could be explained by the high moisture interception by the for-
est canopy31,33 and lack of a pronounced dry season (Supplementary
Fig. 3). In addition, this region has relatively low estimated potential
evapotranspiration (not shown). Across the Amazon, we find little
spatial difference in the distances that transpired molecules travel
Forest-rainfall cascades buffer against drought
across the Amazon
Arie Staal 1*, Obbe A. Tuinenburg 2, Joyce H. C. Bosmans3, Milena Holmgren4, Egbert H. van Nes1,
Marten Scheffer1, Delphine Clara Zemp5,6 and Stefan C. Dekker2,7
Tree transpiration in the Amazon may enhance rainfall for downwind forests. Until now it has been unclear how this cascading
effect plays out across the basin. Here, we calculate local forest transpiration and the subsequent trajectories of transpired
water through the atmosphere in high spatial and temporal detail. We estimate that one-third of Amazon rainfall originates
within its own basin, of which two-thirds has been transpired. Forests in the southern half of the basin contribute most to the
stability of other forests in this way, whereas forests in the south-western Amazon are particularly dependent on transpired-
water subsidies. These forest-rainfall cascades buffer the effects of drought and reveal a mechanism by which deforestation
can compromise the resilience of the Amazon forest system in the face of future climatic extremes.
© 2018 Macmillan Publishers Limited, part of Springer Nature. All rights reserved.
NATURE CLIMATE CHANGE | VOL 8 | JUNE 2018 | 539–543 | www.nature.com/natureclimatechange 539
Content courtesy of Springer Nature, terms of use apply. Rights reserved