Available via license: CC BY 4.0
Content may be subject to copyright.
SPECIAL ISSUE-CURRENT EVIDENCE
Stoichiometry of carbon, nitrogen, and phosphorus through the
freshwater pipe
Roxane Maranger ,
1
* Stuart E. Jones ,
2
James B. Cotner
3
1
Groupe de Recherche Interuniversitaire en Limnologie et en Environnement Aquatique (GRIL), D
epartement de Scien-
ces Biologiques, Universit
e de Montr
eal, Montr
eal, Quebec, Canada;
2
Department of Biological Sciences, University of
Notre Dame, Notre Dame, Indiana;
3
Department of Ecology, Evolution and Behavior, University of Minnesota-Twin
Cities, St. Paul, Minnesota
Abstract
The “freshwater pipe” concept has improved our understanding of freshwater carbon (C) cycling, however, it
has rarely been applied to macronutrients such as nitrogen (N) or phosphorus (P). Here, we synthesize knowl-
edge of the processing of C, N, and P together in freshwaters from land to the ocean. We compared flux esti-
mates into and out of the N and P “pipes” and showed the net removal rates of N and P by inland waters
were less than those for C. The C : N : P stoichiometry of inland water inputs vs. exports differed due to large
respiratory C and N losses, and efficient P burial in inland waters. Residence time plays a critical role in the
processing of these elements through the pipe, where higher water residence times from streams to lakes
results in substantial increases in C : N, C : P, and N : P ratios.
Stoichiometry of the freshwater pipe: The challenge
Our understanding of freshwater systems in the context
of the global carbon (C) cycle has benefited greatly in the
past decade from the development of the “pipe model”
(Cole et al. 2007; Tranvik et al. 2009) and a concerted
research effort to understand C biogeochemistry in inland
waters. The main objective of that model was to determine
whether internal C biogeochemical processing in freshwaters
*Correspondence: r.maranger@umontreal.ca
Author Contribution Statement: All authors contributed equally to this effort and order is presented in reverse alphabetical. Divergent opinions
between RM and JBC were successfully mediated by SEJ.
Data Availability Statement: Data and the R scripts to analyze the data are available at https://github.com/joneslabND/CNPfreshwaterpipes.
Additional Supporting Information may be found in the online version of this article.
This is an open access article under the terms of the Creative Commons Attribution License, which permits use, distribution and reproduction in any
medium, provided the original work is properly cited.
This article is an invited paper to the Special Issue: Carbon cycling in inland waters
Edited by: Emily Stanley and Paul del Giorgio
Scientific Significance Statement
Freshwaters play a critical role in changing the quantity, quality, and rates of carbon inputs from land to the ocean. This
role has been captured in the “freshwater pipe” concept, which has improved our understanding of freshwater carbon
cycling. However, the freshwater pipe concept has not been applied to macronutrients such as nitrogen and phosphorus,
and has not considered these nutrients jointly with carbon. Here, we synthesize the important drivers of freshwater
carbon-nutrient biogeochemistry within the freshwater pipe and develop a conceptual model for the coupled fate of car-
bon, nitrogen, and phosphorus from land to the ocean.
1
Limnology and Oceanography Letters 00, 2018, 00–00
V
C2018 The Author. Limnology and Oceanography Letters published by Wiley Periodicals, Inc.
on behalf of Association for the Sciences of Limnology and Oceanography
doi: 10.1002/lol2.10080
was relevant in a global context with the null model being
that freshwaters are simply passive pipes that carry all of the
allochthonous C loaded into them downstream to the
oceans. The important conclusion drawn from the original
synthesis, and subsequent studies, was that freshwaters are
far from a passive pipe and serve as active C reactors. Indeed
inorganic and organic C are extensively processed internally,
resulting in significant net losses of terrigenous organic
material along the way either through atmospheric flux or
sediment burial.
Most of these studies, however, paid little attention to the
accompanying fluxes and processing of other elements such
as nitrogen (N) and phosphorus (P), that critically influence
C biogeochemistry either as limiting nutrients to primary
production or in how they facilitate decomposition, includ-
ing anaerobic processing such as denitrification and metha-
nogenesis among others. Similar to C, the abundance and
flow of both of these elements have significantly increased
in the last 200 yr due to human activities, primarily through
land-use changes, human population growth, and the indus-
trialization of agriculture (Vitousek et al. 1997; Galloway
et al. 2004). The increased application of chemically synthe-
sized nitrogenous fertilizers and of mined phosphate rock to
agricultural lands particularly during the great acceleration
(Bennett et al. 2001; Galloway et al. 2003; Steffen et al.
2015) has dramatically impacted loading into freshwaters in
many regions of the world (Smith et al. 1999). Therefore, if
we think of the freshwater pipe globally, not only have the
absolute amounts of these elements entering aquatic net-
works changed, but their stoichiometric ratios, relative to C,
have likely also changed.
Although considering the stocks and fluxes of C, N, and P
are at the foundation of aquatic ecosystem science, explicit
consideration of the stoichiometric ratios of these elements
in freshwaters has almost exclusively been at community
and organismal scales. For example, we know that changes
in the total N : P availability influence phytoplankton com-
position and their biomass elemental stoichiometry (Watson
et al. 1997; Downing et al. 2001). For grazers, their stoichio-
metric composition is influenced both by the ratio of ele-
ments available in their prey as well as the relative excretion
rates of those elements (Sterner 1990). Despite the inherent
interlinking of C, N, and P through photosynthesis, how
these foundational elements are processed together and the
effects on stoichiometry have not been fully addressed at
ecosystem and broader spatial scales in inland waters.
Indeed, some studies have looked at the detailed stoichio-
metric changes of N and P using mass balance approaches at
ecosystem scales for reservoirs (Vanni et al. 2011; Grantz
et al. 2014), and assessed the influence of hydrology on
nutrient stoichiometries in streams (Green and Finlay 2009).
However, linking carbon processing to nutrient stoichiome-
try has been largely restricted to streams (Rosemond et al.
2015), with primarily a focus on the influence of net
metabolism on nutrient ratios (Schade et al. 2011). A frame-
work for characterizing changes in ecosystem level C : N : P
stoichiometry in terrestrial environments has been proposed
(Schade et al. 2005), and to some degree tested (Vitousek
2003), but doing so for the entirety of inland waters presents
a different challenge. Given that C, N, and P enter in differ-
ent forms predominantly from their watersheds and interact
via a multitude of physical, chemical, and biological mecha-
nisms within aquatic ecosystems, we argue that the inte-
grated stoichiometric changes along the freshwater pipe
have not yet received appropriate consideration.
In contrast to previous aquatic work that quantified the
stoichiometry of organisms, populations, and communities,
defining the C : N : P stoichiometry of an aquatic ecosystem
is more complex. The bulk C : N: P stoichiometry of an orga-
nism can be readily quantified but ecosystems contain par-
ticulate and dissolved as well as inorganic and organic forms
of C, N, and P. Often these elements are quantified at the
ecosystem scale individually with little work focused on bulk
stoichiometry. In the oceans, ecosystem stoichiometric
approaches are often compared to the Redfield ratio (molar
ratio of 106C : 16N : 1P). Although there are many papers
demonstrating deviations from this ratio in specific basins or
at different latitudes (Martiny et al. 2013), it remains a
benchmark in marine systems, likely due to the long resi-
dence times of the oceans (thousands of years), as well as
more constrained biomass and physiologies of plankton due
to low nutrient availability. In freshwaters, where contact
with the terrestrial landscape is more direct and residence
times are more variable and shorter (days to hundreds of
years), stoichiometric ratios are also much more variable
(Hecky et al. 1993). The challenge when considering how
stoichiometry is regulated at the ecosystem scale requires
untangling the way these elements enter, whether they are
processed together or independently, and what the avenues
for loss from the ecosystem are. Therefore, we propose a
framework that focuses on the relative importance of trans-
port, metabolism, evasion, and burial of C, N, and P along
the freshwater pipe in order to develop expectations for the
role of inland waters in regulating the stoichiometry
exported to global oceans.
Coupling and uncoupling the cycles
The importance of downstream transport of C, N, and P
echoes the key question raised by initial investigations of
inland waters in processing terrestrial carbon (Cole et al.
2007; Tranvik et al. 2009). We learned from these previous
studies that C is actively processed within the pipe, with a
large fraction of the organic C pool being metabolized to
CO
2
and lost to the atmosphere and a smaller fraction being
buried in sediments (Mendonc¸a et al. 2017) or lost as meth-
ane (Tranvik et al. 2009; Stanley et al. 2016). But, unless
organic N and P are processed very similarly to organic C,
Maranger et al. Stoichiometry of carbon, nitrogen, and phosphorus
2
we should expect that the stoichiometry of these elements
could change significantly, leading to very different export
ratiostotheoceanrelativetowhatisloadedintofreshwaters.
A critical process that couples all three biogeochemical
cycles is primary production. Although N and P are not
directly part of the biochemical process of C-fixation, these
essential macronutrients are required by primary producers
to generate biomolecules such as nucleotides, proteins, and
lipids. It is worth noting that past discussions of the role of
freshwaters in the global carbon cycle have essentially
ignored the role of primary production because they have
focused on the net effect of these systems on export of C to
the oceans. However, because in situ primary production is
important in coupling C to N and P, we have included it
here. Primary production is often constrained by the avail-
ability of P in freshwaters but can also be limited by N, and
frequently by both (Elser et al. 2007). Cyanobacteria and
bacteria capable of N-fixation can further facilitate coupling
of C, N, and P by accessing atmospheric N although this pro-
cess is energetically expensive and the ability to correct eco-
system deficits are not common (Paerl et al. 2016).
A number of catabolic processes could also act to shift C,
N, and P stoichiometry, particularly as they relate to atmos-
pheric losses. These biogeochemical processes can have
extreme stoichiometric signatures relative to the composi-
tion of organisms and therefore could significantly alter the
stoichiometry of inland waters at the ecosystem scale. The
two most important catabolic processes to consider along
these lines are C respiration and denitrification. Inorganic C
is produced through respiration and this process decreases
the organic C concentration within systems while increasing
the permanent loss of CO
2
to the atmosphere if concentra-
tions are saturating. N and P are also remineralized in the
process, primarily as ammonium and phosphate, which can
be recycled within the systems. However, the conversion of
ammonium to nitrate through nitrification in sediments can
have two fates. One is nitrate flux from the sediment to the
water column whereas the other is the conversion of nitrate
to gaseous N (N
2
or N
2
O) via denitrification (and to a lesser
extent anammox), which also results in permanent N loss to
the atmosphere. Nitrification and denitrification also influ-
ence C metabolism, through chemosynthesis and nitrate res-
piration, respectively (Zehr and Ward 2002). Contrary to C
and N, P biogeochemistry does not produce a gaseous end
product (Schlesinger and Bernhardt 2013). Previous work has
shown that the net losses of C and N gases from inland
waters to the atmosphere are meaningful emissions at the
global scale (Raymond et al. 2013; Soued et al. 2016). Fur-
thermore, the relatively independent loss of C and N via
these processes represents an additional mechanism by
which high inland water C : P and N : P ratios are decreased.
Using a framework that considers the relative importance
of transport and metabolism with subsequent burial and/or
atmospheric exchange, we attempted to couple global inland
water C, N, and P biogeochemical cycles by examining their
inputs, standing stocks, and outputs, as well as their stoichi-
ometry along the freshwater pipe. In this article, we address
three important questions: (1) Are inland waters an “active
pipe” stoichiometrically, meaning that they export C, N,
and P at ratios distinct from what is imported from the ter-
restrial landscape?; (2) What are the ranges of the C : N : P
stoichiometry of inland waters?; and (3) How have human
activities altered inland water C : N : P stoichiometry and
changed the role of inland waters in modifying the stoichi-
ometry of terrestrial exports as they are transported to the
ocean?
Question 1: Are inland waters a stoichiometrically
“active pipe” ?
To investigate the degree to which the biogeochemistry of
global inland waters alters the C : N : P stoichiometry of ter-
restrial inputs prior to their delivery to the ocean, we gath-
ered available information on inputs to freshwaters of C, N,
and P from their watersheds as well as outputs of C, N, and
P from inland waters to the global oceans. Our conceptual
framework described above suggests metabolism, burial and/or
atmospheric exchange should have a large impact on altering
the C : N : P ratios along the aquatic continuum. In contrast, if
we observe little freshwater-mediated change in stoichiometry,
it is likely that transport is the dominant process or that the
multitude of biogeochemical processes that make up inland
water metabolism, burial, and atmospheric exchange are bal-
anced and yield no net change in stoichiometry.
Terrestrial inputs vs. oceanic outputs
In order to examine the stoichiometry of inputs and out-
puts from the freshwater pipe and assess processing, we com-
piled current estimates of C, N, and P loads and exports
(Table 1; Fig. 1A–C). Terrestrial inputs of carbon into fresh-
waters have been estimated at about 2900 Tg yr
21
(Tranvik
et al. 2009). This includes contributions of dissolved organic
carbon (DOC), dissolved inorganic carbon (DIC), and partic-
ulate organic carbon (POC), of which the organic compo-
nents are less well constrained at the global scale. If we
consider the lake budgets presented by Tranvik et al. (2009)
as being potentially representative of different C forms, we
estimate that 55% of terrestrial carbon export is DIC,
4% is POC, and the remainder enters inland waters as DOC
(Fig. 1A). For the purposes of evaluating the relative inputs
and outputs of whole ecosystem stoichiometry of inland
waters, we focus on organic C inputs because the transforma-
tions of this pool to inorganic C vs. burial is important to
our understanding of the role of freshwaters in the global
carbon cycle and climate change. Indeed much of the terres-
trial DIC load that enters the pipe is quickly vented to the
atmosphere (McDonald et al. 2013). Furthermore, the most
abundant gaseous form of N (N
2
), which also enters at high
concentrations, is typically not accounted for in ecosystem
Maranger et al. Stoichiometry of carbon, nitrogen, and phosphorus
3
exchanges or stoichiometry. Similarly, we excluded particu-
late mineral P from our analysis, although it is a large part
of the P flux into freshwaters, it is largely unreactive and is
either buried in the sediments or transported to the ocean
without being biogeochemically transformed (Smil 2000;
Schlesinger and Bernhardt 2013).
Omitting DIC, we estimated that a total 1305 Tg C yr
21
is
loaded to inland waters as allochthonous DOC and POC.
Getting an accurate estimate for N was a bit more problem-
atic. The estimated amount of N loaded into inland water
varies widely, from 67 Tg N yr
21
(Beusen et al. 2016) to 129
Tg N yr
21
(Galloway et al. 2004). Both estimates were
derived from previous literature values as well as models to
determine the export of N to freshwaters and coastal ecosys-
tems. Rather than picking one or the other, we opted for the
mean of these divergent values and estimated total N input
at 98 Tg N yr
21
(Table 1). But it should be noted that given
these widely divergent values for N loading into freshwaters,
future work to refine our estimates of terrestrial N export to
freshwaters would be beneficial. There was more consensus
for the amount of P that entered freshwaters, at around 9 Tg
Pyr
21
(Smil 2000; Beusen et al. 2016) (Table 1). Therefore,
in terms of inputs, these values gave a molar terrestrial
export C : N : P ratio of about 375 : 24 : 1 suggesting an
export of C-rich and P-poor materials into freshwaters,
which is consistent with many observations of P-limited
freshwaters (Table 1) (Schindler 1977; Hecky and Kilham
1988, Elser et al. 2007).
For oceanic export, we used 900 Tg C yr
21
(Cole et al.
2007; Tranvik et al. 2009). However similar to the terrestrial
loads, we partitioned the carbon terms into DIC, DOC, and
POC (59%, 19%, and 22%, respectively) based upon Stets
and Striegl (2012), McDonald et al. (2013), and Seitzinger
et al. (2005), thus resulting in an estimated 369 Tg C yr
21
for combined DOC and POC export (Fig. 1A). Values used
for N and P delivery to the ocean from freshwaters were 43
Tg N yr
21
(Seitzinger et al. 2010) and 4 Tg P yr
21
(Beusen
et al. 2016), respectively (Fig. 1B,C). Thus, our final export
numbers yield a molar ratio of 238 : 24 : 1 representing a
36% decrease in the C : N and C : P ratios of export relative
to import (Table 1). Somewhat surprising was the fact that
the N : P molar ratio did not change relative to the input
ratio suggesting that these elements are processed similarly
in freshwaters. Although N and P losses through the pipe are
proportionately similar, the fate of those losses is very differ-
ent where 75% of N loss is as gas to the atmosphere (N
2
and
to a lesser extent N
2
O) while all P loss is via sediment burial
(see below).
Based upon our global estimates of terrestrial loads and
freshwater exports of C, N, and P, inland waters clearly pro-
cess a considerable amount of the material they receive from
the terrestrial landscape, thus inland waters serve as an
active pipe for all three elements. One way to express the
effects of freshwaters on these biogeochemical cycles is to
assess their relative removal rates, the quantity of each ele-
ment that is either buried or released into the atmosphere in
relation to terrestrial loading (Finlay et al. 2013). Using this
approach, we estimate that 72% of C (organic C only), 56%
of N, and 56% of P is removed through the pipe (Table 1).
The differential removal rates of C relative to N and P high-
light why we see such dramatic differences in the C : N and
C : P stoichiometry of inland water loads and exports. Fur-
thermore, the biogeochemical mechanisms by which remov-
als occur are clearly different for each element and may
occur at different locations along the continuum, thus influ-
encing the stoichiometry among sub-components of inland
waters (i.e., streams vs. rivers vs. lakes vs. reservoirs).
Removal of C and N through the pipe is dominated by
atmospheric fluxes, but proportional losses of C via this
pathway are considerably higher than losses of N. We
Table 1. Estimates of C, N, and P terrestrial inputs, atmospheric inputs and losses and loss to burial through the freshwater pipe as
well as export to the coastal ocean. Molar ratios from different terms also provided; only organic carbon estimates are used in ratios.
See Supporting Information Material for details of data sources and calculations.
Tg yr
21
Molar ratio
DOC POC DIC N P C : N C : P N : P C : N : P
Watershed inputs 1189 116 1595 98 9 15.5 374.6 24.1 375 : 24 : 1
Atmospheric flux* Input — 2288 — 10 0.04 266.9 — — —
Output 1018 2056 — 49 0.04 73 — — —
Net 2786* 21064
†
239 — — — — — —
Total burial — 150 — 16 5 11 77.5 7 78 : 7 : 1
Coastal exports 171 198 531 43 4 10 238.3 23.8 238 : 24 : 1
Reservoirs processing
‡
— 60 38.6 6.5 1.3 10.8 119.2 11.1 119 : 11 : 1
*Calculated as the difference between gross primary production (atmospheric inputs to POC) minus respiration of DOC and POC (DOC and POC
atmospheric flux output).
†Calculated as the difference between watershed inputs of DIC and DIC coastal exports.
‡POC and P is burial; DIC is CO
2
respiratory losses; N is both denitrification and burial; C : N ratios are sum of losses to atmosphere and burial.
Maranger et al. Stoichiometry of carbon, nitrogen, and phosphorus
4
estimate that nearly 86% of total organic C inputs, including
gross primary production, are lost to the atmosphere
whereas only 45% of total N input is removed via this mech-
anism (Fig. 1B). Indeed the C : N of terrestrial inputs is 16
while the stoichiometry of exchange with the atmosphere is
much higher at 73 (Table 1). While gaseous loss of N and C
are both mediated by different microbial transformation
pathways, C loss can also be facilitated via photochemical
oxidation (Koehler et al. 2014), but that is usually less than
10% of total oxidation to CO
2
. Gaseous loss of N via denitri-
fication is much higher than inputs from in situ N fixation,
so that on average most of the N that is lost to the atmos-
phere enters from the watershed (around 90%), whereas
around 67% of the C exchanged can be derived from in situ
primary production (Table 1; Fig. 1A,B). Global N fixation in
freshwaters is estimated to be approximately 10 Tg N yr
21
(Cleveland et al. 1999) whereas denitrification is estimated at
49 Tg N yr
21
(Table 1), which is similar to a previously
accepted estimate of 66 Tg N yr
21
for denitrification pro-
posed by Seitzinger et al. (2006). Given the widely different
delta G’s between nitrogen fixation and denitrification, (i.e.,
Gibb’s free energy required to fuel the reaction), where deni-
trification is much more negative, the capacity of freshwaters
to “inhale” N
2
gas to achieve stoichiometric balance is lim-
ited energetically. Although N-fixation can be an important
component in the N-balance in some primarily unperturbed
N-limited systems (Smith 1990), there is an inherent ener-
getic bias toward N “exhalation” (Paerl et al. 2016) that is
evident at the global scale.
In contrast to C and N, essentially all P removal through
the pipe is lost via burial to the sediments (Table 1). Atmos-
pheric losses of P as phosphine gas (PH
3
) are not well under-
stood nor quantified, but compared to C and N, it is
negligible at less than 0.04 Tg P yr
21
globally (Schlesinger
and Bernhardt 2013). As a result, burial of P in freshwater
sediments is the only meaningful loss, representing 56% of
terrestrial P inputs into freshwaters (Fig. 1C; Table 1). Burial
of organic C in freshwaters has recently been revised to 150
Tg C yr
21
(Mendonc¸a et al. 2017) and is substantially lower
than the previous estimate of 600 Tg C yr
21
(Tranvik et al.
2009). Burial therefore represents a loss of only 4% of total
organic C inputs. For N, freshwaters appear to bury around
15% of terrestrial inputs (16 Tg N yr
21
, Table 1), however as
outlined above, this number is based on the difference of a
relatively imprecise loading estimate. Although nearly all P
losses are through burial, some of this buried P is bound to
Fe and Al hydroxydes that are solubilized when bottom
waters become anoxic and can result in a significant P input
term through internal loading (N€
urnberg 1984). Indeed, all
of the elements can be returned to the water column
through sediment processing but given that P does not have
an atmospheric loss term, and it often limits primary pro-
duction, internal P loading may have greater consequences
Fig. 1. Inputs, transformations, and outputs of different (A) carbon, (B)
nitrogen, and (C) phosphorus forms through the freshwater pipe. All val-
ues are in Tg yr
21
.
Maranger et al. Stoichiometry of carbon, nitrogen, and phosphorus
5
on internal ecosystem level stoichiometry than internal
fluxes of C and N from the sediment (discussed below).
The disproportionate loss of P via assimilation and/or
sorption followed by sedimentation relative to C and N is
clear when the C : N : P stoichiometry of terrestrial loads
(375 : 24 : 1), freshwater exports (238 : 24 : 1) and sediment
burial (78 : 7 : 1) are compared (Table 1). Presumably, the
limited burial of C is driven by high rates of decomposition/
respiration, and/or selective burial of low C : P and C : N
plankton biomass, whereas for N, limited burial and high
export ratios are the result of a combination of denitrifica-
tion losses rather than long term storage as well as an 8–10
orders of magnitude greater solubility of nitrate relative to
phosphate that favors its downstream export. Indeed,
although the proportion of export of N and P total inputs
are more or less the same at 44%, the dominant forms
exported to the oceans reflects the differential importance of
solubility of these two nutrients. Modeled particulate N and
P exported by rivers to the oceans differed substantially in
that only 31% of N was in particulate form while 77% of P
was in particulate form (Fig. 1B,C) (Seitzinger et al. 2010).
The processing and fates of the different elements can
occur in either the fluvial network (streams and rivers) or in
lentic waterbodies (lakes and reservoirs). The relative reten-
tion and permanent loss of C, N, and P is apparently spa-
tially distinct through the freshwater pipe. Burial of P occurs
primarily in lentic waterbodies (Sorrano et al. 2015), and res-
ervoirs are particularly effective at settling P (see below). By
comparison to P, N and C are more heavily processed in
streams and rivers. Denitrification losses throughout streams
and rivers account for over half of gaseous losses (Seitzinger
et al. 2006; Mulholland et al. 2008) where lower order
streams tend to remove proportionally more N than higher
order ones due to higher relative sediment contact. However,
in highly eutrophic systems, excess N can lead to saturation
of sediment transformation sites (Alexander et al. 2009)
resulting in reduced removal efficiency and greater down-
stream export. For C, it remains unclear where along the
continuum and what portion of the 2100 Tg C yr
21
lost to
the atmosphere (Raymond et al. 2013) originates from in
situ respiration or is a result of vented inputs from the water-
shed (McDonald et al. 2013; Raymond et al. 2013). Recent
work suggests that 28% of the DIC emitted from streams
and rivers is locally produced (Hotchkiss et al. 2015) and
that the relative proportion of in situ respiration increases
with stream size, reflecting reduced connectivity with the
terrestrial landscape. Small ponds, however, which also have
high relative sediment contact and are more terrestrially
connected, are hotspots of locally produced C emissions, at
583 Tg C yr
21
(Holgerson and Raymond 2016). Therefore,
much like N, approximately half of C that is internally cata-
bolized through the freshwater pipe is lost from fluvial net-
works and half from small lakes and ponds. Thus, shallow
freshwater systems, with their relatively high sediment
reactivity, are playing a crucial role in removing both C and
N permanently from the pipe whereas deeper lentic bodies
are where permanent P losses occur.
It should also be noted however, that efforts to understand
C processing have been heavily influenced by studies focused
primarily on boreal and temperate systems (Raymond et al.
2013). This may skew our overall interpretation of C processing
dynamics through the freshwater pipe. Indeed, some studies
suggest that the biogeochemical loading and processing in
tropical inland waters may be behaving quite differently than
temperate and boreal systems (Huszar et al. 2006). Regarding
N, there has been strong focus on denitrification N losses in
flowing waters, but the large majority of this work has been
conducted in temperate regions with the most concrete meas-
ures on low order streams (Mulholland et al. 2008; Schade
et al. 2011; Rosemond et al. 2015), again reflecting geographic
and ecosystem bias in our overall understanding of processing.
So, our current understanding of coupled biogeochemical
cycling through inland waters would profit from a broader geo-
graphical assessment, with a focus on tropical and subtropical
regions with rapid population growth and land-use develop-
ment changes.
Question 2: What is the C : N : P stoichiometry of
inland waters?
Because the stoichiometry of inputs to and outputs from
freshwaters differ from the Redfield ratio, we wanted to
know what expectations for inland water stoichiometry
should be. Although Elser et al. (2000) and Hecky et al.
(1993) examined the stoichiometry of particles in freshwater,
no studies to our knowledge have examined the stoichiome-
try of the combined dissolved and particulate components
(including C) at large scales. National surveys of streams, riv-
ers, lakes, and reservoirs across the United States (2007 U.S.
EPA National Lakes Assessment, U.S. Environmental Protec-
tion Agency 2010; 2007–2017 stream/river chemistry data,
U.S. Geological Survey 2016) provide a picture of inland
water stoichiometry for a geographically diverse set of sys-
tems (Table 2; see Supporting Information Material for meth-
ods). These observations corroborate a non-Redfield
stoichiometry in freshwaters that closely reflects the stoichi-
ometry of terrestrial source material. The median C : N : P
ratios in streams and rivers (167 : 25 : 1), reservoirs (417 : 38
: 1), and lakes (963 : 62 : 1) differed strongly from the Red-
field ratio and commonly observed ocean stoichiometry
(Martiny et al. 2013).
It should be noted, however, that there really should be
no expectation of Redfield-like stoichiometry in these meas-
urements for a number of reasons. First, the stoichiometry of
primary producers and organic matter on terrestrial land-
scapes is not particularly close to the Redfield ratio
(McGroddy et al. 2004) with C : N : P ratios of trees >1000 :
25 : 1 and soils at 186 : 13 : 1 (Cleveland and Liptzin 2007).
Maranger et al. Stoichiometry of carbon, nitrogen, and phosphorus
6
Furthermore, in contrast to the oceans where terrestrial
inputs are extremely low and the inorganic and plankton
nutrient pools mirror each other (Redfield 1958), freshwater
systems are open, with relatively short residence times of
days to years, which means that it is more likely that the dis-
solved and particulate stoichiometry diverges from each
other as we discussed above due to active processing (They
et al. 2017). This is reflected in our comparisons of the stoi-
chiometry of streams/rivers, reservoirs, and lakes.
Simple comparisons of streams/rivers, reservoirs, and lakes
indicate that residence time and the magnitude of interac-
tions with sediment surfaces appears to be important in dic-
tating inland water stoichiometry. As residence times
increase across inland water sub-components, the total C : P
and C : N stoichiometry also increases. Perhaps this reflects
increased burial of P through settling of particles with low C
: P and N : P ratios, likely due in part to enhanced water col-
umn primary production. It may also represent higher C
losses with increased sediment contact, as would be the case
in shallow streams, rivers, and reservoirs relative to deeper
lakes. The N : P stoichiometry also showed an increase with
residence time (25 : 1, 38 : 1, and 62 : 1 in streams/rivers,
reservoirs, and lakes, respectively). Lower ratios in streams
could be due to elevated rates of N loss through denitrifica-
tion whereas for reservoirs, it is likely a combination of both
denitrification as well as high P burial (see below).
Question 3: Anthropogenic alterations to inland
water stoichiometry and processing
Freshwater pipes: Changes in inputs and changes in
plumbing
Humans have exerted tremendous control on the supply of
N and P to inland waters. Urban and industrial point sources
have enhanced N and P supply, but P inputs from point sources
have been controlled in many regions of the developed world
(Carpenter et al. 1998), whereas treatments to remove point
source N lag behind (Conley et al. 2009). Municipal waste-
waters as point sources of pollution remain a major concern
particularly in the developing world due to rapid urbanization.
Despite increasing regulation, 40% of municipal wastewater
remains untreated globally (Mateo-Sagasta et al. 2015).
Agricultural non-point sources of N and P, however, are
pervasive and have far surpassed historical point-source
inputs in many areas of the globe (Carpenter et al. 1998).
For example, a global surplus of 25 Tg of P is created by the
mobilization of mined phosphate rock used as fertilizer for
agricultural activities each year (Cordell et al. 2009), and
between 20% and 100% of this surplus is thought to enter
surface waters via leaching and erosion (MacDonald et al.
2011). P fertilizer use has doubled the reactive P concentra-
tions in rivers worldwide (Smil 2000). In terms of N, massive
amounts of N fertilizer, currently synthesized through the
Haber-Bosch process, double the amount of bioavailable N
to the terrestrial biosphere annually (Vitousek et al. 2013).
This has resulted in what is termed the N cascade where N
fertilizer use is inefficient and most N is lost to various other
environmental compartments including surface and ground-
waters (Galloway et al. 2003). Over the last few decades,
NH
4
-NO
3
synthesis has been replaced by the chemical syn-
thesis of urea, making this dissolved organic nitrogen form
now the most widely used nitrogenous fertilizer (Glibert
et al. 2014). Furthermore, the N : P ratio of global fertilizer
use has increased, likely altering the N : P ratios of what is
entering surface waters (Glibert et al. 2014).
The role of sediments in both the processing and storage
of nutrients through the freshwater pipe will become
increasingly more important as more nutrients enter inland
waters. Indeed, the C : N, C : P, and N : P ratios for our esti-
mates of sediment burial are radically different from those
that enter from terrestrial landscapes (Table 1). Furthermore,
P that accumulates in sediments over time results in a legacy
P effect whereby sediments can retain and act as a buffer
against P loads but only to a certain point (Haygarth et al.
2014). Once this threshold of buffering has been met, accu-
mulated P will be more chronically available to the entire
system, or its downstream export will be promoted.
Increased P loading onto landscapes along with selective bur-
ial of P suggests that we are globally enhancing internal P
recycling through legacy P, potentially sustaining eutrophi-
cation for the long term (Carpenter 2005), as well as chang-
ing ecosystem level stoichiometry for many generations.
Humans have also directly altered the “piping” of inland
waters through channelization of streams, removal and crea-
tion of wetlands, and perhaps most dramatically through the
Table 2. Median and ranges in molar stoichiometric ratios of streams/rivers, reservoirs, and lakes in the United States. Values based
on total organic C, total N, and total P concentrations across systems. See Supporting Information Material for methods.
C:N
(25
th
,75
th
)
C:P
(25
th
,75
th
)
N:P
(25
th
,75
th
)
U.S. streams and rivers 2007–2017
(N57275)
7.1 (3.9, 12.4) 166.6 (82.6, 351.4) 24.7 (13.6, 44.6)
U.S. reservoirs 2007 (N5636) 11.4 (7.9, 15.5) 416.7 (191.0, 798.8) 38.0 (22.7, 59.1)
U.S. lakes 2007 (N5521) 13.0 (9.2, 18.2) 963.2 (372.2, 1854.8) 61.8 (35.4, 109.1)
Maranger et al. Stoichiometry of carbon, nitrogen, and phosphorus
7
construction of reservoirs. Reservoir surface area is currently
estimated at 11% of the global lentic surface (Lehner et al.
2011) and there is an anticipated growth of 3700 large dams
in the coming decades (Zarfl et al. 2015). Reservoirs and nat-
ural lakes are geomorphometrically and hydraulically quite
different from each other (Hayes et al. 2017). Compared to
natural lakes, reservoirs have much higher drainage ratios
(catchment area to surface area) and thus receive more material
from their catchments on an areal basis. Furthermore, reser-
voirs have higher settling velocities but also higher throughput
due to relatively shorter water residence times and process sig-
nificantly more material per unit area than natural lentic water-
bodies (Harrison et al. 2009; Clow et al. 2015). Thus, this
change in “piping” will likely influence the biogeochemistry of
C, N, and P as well as their stoichiometry. One obvious impli-
cation is that the global export C : P and N : P ratios should
increase due to increased residence times.
Relative to other surface waters, reservoirs disproportion-
ately serve as settling basins for all elements. They retain a
huge amount of sediment, so much so that they have
reduced the accretion of several major coastal deltas (Syvitski
et al. 2009). In terms of C burial, reservoirs bury 60 Tg C
yr
21
(Mendonc¸a et al. 2017) (Table 1), which is 40% of the
total global C burial while for P, they bury 1.3 Tg P yr
21
(Maavara et al. 2015) or 26% of overall burial, despite only
covering 11% of the global lentic surface. Per unit area, res-
ervoirs are considered to be much more efficient at removing
N than lakes (Harrison et al. 2009). The N retained in reser-
voirs is estimated at 6.5 Tg N yr
21
and is mostly lost to the
atmosphere through denitrification (Harrison et al. 2009;
Beusen et al. 2016). C from reservoirs is also lost to the
atmosphere as both CO
2
through oxic respiration and CH
4
through methanogenesis, estimated at 36.8 Tg C yr
21
and
13.3 Tg C yr
21
, respectively (Deemer et al. 2016). In order to
be consistent with other C respiratory loss terms, only CO
2
loss is represented in Table 1. Although CH
4
emissions con-
tribute greatly to the global warming potential from inland
waters, from a total C processing point of view, C lost
through CH
4
emissions from freshwaters is relatively minor
(Tranvik et al. 2009). For this reason, it was not included in
our overall output terms. However, it should be noted that
current estimates for methane emissions from reservoirs rep-
resent 12% of their total C loss, proportionally much higher
than inland waters overall.
Given these global alterations in inputs and plumbing, we
wanted to assess the relative influence of land use and water
residence time on changes in inland water stoichiometry. In
order to do so, we compared the distributions of C : N, C : P,
and N : P ratios of reference vs. impacted lakes and reservoirs
using the same set of lentic systems described above (Fig. 2;
U.S. Environmental Protection Agency 2010; see Supporting
Information Material for methods). For C : P ratios, impacted
reservoirs had lower ratios than reference reservoirs, which
had lower ratios than impacted lakes, with reference lakes
having the highest C : P ratio on average (Fig. 2A). There-
fore, land use loading apparently impacted stoichiometry in
both lakes and reservoirs, but residence time represented a
stronger relative shift given that ratios for reference reser-
voirs were lower than those of impacted lakes. One possible
explanation for this, in addition to residence time, is that
reservoirs have greater internal P load brought on by an
increased incidence of anoxia (M€
uller et al. 2012) combined
with higher inputs. Another possibility is that increased P
load enhances C loss in reservoirs. In fact, more eutrophic
reservoirs tend to produce more methane and emit more
CO
2
(Deemer et al. 2016; Harrison et al. 2017).
Fig. 2. Kernel density plots of stoichiometric (A)C:P,(B)C:N,and(C)
N : P molar ratios of reference lakes and reservoirs vs. impacted lakes and
reservoirs included in the 2007 U.S. Environmental Protection Agency
National Lakes Assessment (U.S. Environmental Protection Agency 2010).
Maranger et al. Stoichiometry of carbon, nitrogen, and phosphorus
8
In terms of C : N, the mean ratios in impacted lakes and
both categories of reservoirs were similar and lower than ref-
erence lakes (Fig. 2). It is possible that different mechanisms
are responsible for this similar outcome. For impacted lakes,
this lower ratio may simply be a function of higher N load
and a signature of eutrophication. Alternatively, lower C : N
ratios in reservoirs relative to reference lakes may be a func-
tion of relatively higher C loss due to burial and respiration
but is more likely due to less N loss as a function of shorter
residence times relative to lakes. In impacted reservoirs, how-
ever, one possible explanation for why land use did not
result in lower C : N ratios relative to reference reservoirs is
that relatively more N may be lost via denitrification as a
function of additional P (Finlay et al. 2013). For N : P,
although reservoir ratios were systematically lower than
lakes, it was surprising to see no distinction between refer-
ence and non-reference systems in reservoirs and lakes for
these ratios (Fig. 2C). This could potentially suggest that
water residence time overrides relatively higher inputs, how-
ever, we do not have information on the relative loads of
the different nutrients between reference and impacted sys-
tems. It is also possible that internal processing maintains
the relative ratios of nutrients between impacted and refer-
enced systems likely as a function of the strong coupling
through metabolism. Indeed, there is a strong coupling of N
and P within the sediments where processing time is rela-
tively longer, which tends to result in sediment ratios that
are similar even across land use gradients (Vanni et al.
2011). However, the absolute amounts stored among these
systems likely differ across input gradients.
Nonetheless, the C : N and N : P ratios of the water were
the same between reference and impacted reservoirs but the
C : P ratios were lower in impacted systems (Fig. 2). The dif-
ferential fate of N and P could explain this discrepancy.
Indeed, the reference reservoirs were almost 2 m deeper on
average than the impacted reservoirs (mean depth ref-
res 510.5 m vs. imp-res 58.7 m), which could settle more P
(Collins et al. 2017) and decrease resuspension. Furthermore,
the shallower, impacted reservoirs may eliminate more N
due to increased sediment to water contact resulting in
increased denitrification and similar ratios for nutrients for
both types of systems.
Summary
One of the important conclusions from this effort is the
realization that, not surprisingly, a great deal more effort has
been expended to understand the processing of C than
understanding the fates of N and P throughout aquatic net-
works since the seminal Cole et al. (2007) paper. Nonethe-
less, in this first attempt to consider coupled biogeochemical
cycles of inland waters and their ecosystem level stoichiome-
try at the global scale, we find that input and export stoichi-
ometry of these elements differ due to differences in
retention among the elements through the freshwater pipe.
Furthermore, the fate of the retained elements differs due to
their biogeochemistry. More C is retained relative to N and P
(Table 1; Fig. 3) due to tight coupling of primary production
and ecosystem respiration, which means that most of the
“retention” is actually lost from freshwaters primarily as CO
2
(Raymond et al. 2013). Conceptually, this suggests that
organic carbon pools are rapidly turning over in freshwater
systems with rapid production being compensated by simi-
larly rapid decomposition. P and N are retained at equivalent
proportions through the pipe, but with very different fates,
often in different locations along the aquatic continuum.
Most P retention ends up in the sediments of lentic water-
bodies, in a large part due to limited solubility of inorganic
P relative to C and N. However, most of the N retained ends
up in the atmosphere due to an imbalance between N-
fixation and denitrification with the latter process dominat-
ing, particularly in small order streams and shallow lentic
systems with high sediment contact.
Another important conclusion from this effort is that resi-
dence time plays a critical role in the processing of these ele-
ments through the pipe. There were profound differences in
the stoichiometry of streams, reservoirs, and lakes in the
USGS and NLA surveys. Moving from streams to lakes, there
are large increases in residence times with substantial
increases in C : N, C : P, and N : P ratios. Certainly, settling
of particles likely plays a role along this continuum, but
increased light availability may also be important to support
the coupled settling of elements through primary production
(Urabe et al. 2002; Elser et al. 2003). Once algal particles
reach the sediments, respiratory C losses of this autochtho-
nous organic matter and subsequent denitrification-
mediated N losses decouples C from N from P.
Knowledge gaps
We have identified three critical knowledge gaps for fresh-
water scientists to consider in order to better characterize
coupled elemental cycling along the aquatic continuum.
First, although we have synthesized the best available
Fig. 3. Global estimates of stoichiometry of the freshwater pipe. C : P,
N : P, and C : N molar ratios of estimated inputs from terrestrial sys-
tems, loss to burial and the atmosphere and export to the ocean. Figure
adapted from Cole et al. (2007, fig. 1).
Maranger et al. Stoichiometry of carbon, nitrogen, and phosphorus
9
information to date on various components of C, N, and P,
none of these biogeochemical cycles are well constrained at
the global scale. The best estimates for global fluxes come
from the Global NEWS Project and subsequent iterations
(Seitzinger et al. 2005; Mayorga et al. 2010; Beusen et al.
2016) but these syntheses were primarily focused on the
exports from freshwaters rather than quantifying the materi-
als loaded into them. Consequently, there is a huge gap in
our understanding of what actually is happening inside the
pipe because we do not have a good understanding of what
is entering it in the first place. For example, in the case of C,
recently revised estimates suggest a 75% increase in terres-
trial loading to the pipe, upscaling it to 5100 Tg C yr
21
(Drake et al. 2017). Our inability to constrain terrestrial
inputs results from those values most often being estimated
as the sum of atmospheric exchange, sediment burial, and
export to the ocean. Increased observations of terrestrial C,
N, and P inputs at watershed and larger scales, as well as
process-based modeling efforts should increase our ability to
constrain this important, but highly uncertain, aspect of
inland water biogeochemistry.
Second, nitrogen remains the most poorly constrained of
the three elements, likely due to the mechanistic complexity
of the cycle, a bias at understanding P limitation in fresh-
waters, and the concerted effort to understand C cycling
since the seminal Cole et al. (2007) paper. There seems to be
good agreement on nitrogen exports to coasts but, as
described above, there is a wide discrepancy as to what is
believed to be going into the freshwater pipe (Galloway
et al. 2004; Beusen et al. 2016). Global efforts at estimating
exports are suggesting downscaling N inputs, however,
efforts at estimating N loading to freshwaters at more
regional scales suggest inputs are increasing at a rapid pace
(Howarth et al. 1996; Gao et al. 2015; Goyette et al. 2016).
Furthermore, although there has been a great amount of
work done to understand N processing in streams (Mulhol-
land et al. 2008; Schade et al. 2011; Rosemond et al. 2015),
recent denitrification and burial estimates modeled by Beu-
sen et al. (2016) were only 27 Tg N yr
21
whereas the previ-
ous estimate is around 66 Tg N yr
21
(Seitzinger et al. 2006).
Indeed simple mass balance budgets cannot capture the
more complex nature of N cycling through the pipe. Nitrifi-
cation may be more ubiquitous in freshwaters than previ-
ously thought (Botrel et al. 2017) and N-fixation remains
poorly constrained (Scott and McCarthy 2010). There is cer-
tainly a need improve our estimates of N loading to the pipe
well as a greater understanding of what influences N process-
ing in different aquatic compartments.
Finally, it is clear that where, when, and how terrestrial
inputs are processed along the freshwater continuum is rap-
idly changing due to human activities. Land use alterations,
the growing use of fertilizers, rapid urbanization, the lack of
wastewater treatment, and increased eutrophication are all
factors that have the potential to change freshwater
stoichiometry at regional and global scales. These alterations
of inputs are occurring in parallel with alterations to the
hydrological features of the network, including increased res-
ervoir construction as well as shifting precipitation and run-
off patterns as a function climate change. Future work must
consider the direct and interactive effects of hydrology on
human-driven alterations in C, N, and P inputs from catch-
ments, and how elements are processed together through
different aquatic sub-components. We must also expand our
understanding of coupled cycles in the freshwater ecosys-
tems of the rapidly developing regions of the global south
for a more complete portrait of the role of the freshwater
pipe.
References
Alexander, R. B., and others. 2009. Dynamic modeling of
nitrogen losses in river networks unravels the coupled
effects of hydrological and biogeochemical processes.
Biogeochemistry 93: 91–116. doi:10.1007/s10533-008-
9274-8
Bennett, E. M., S. R. Carpenter, and N. F. Caraco. 2001.
Human impact on erodable phosphorus and eutrophica-
tion: A global perspective: Increasing accumulation of
phosphorus in soil threatens rivers, lakes, and coastal
oceans with eutrophication. AIBS Bull. 51: 227–234. doi:
10.1641/0006-3568(2001)051[0227:HIOEPA]2.0.CO;2
Beusen, A. H. W., A. F. Bouwman, L. P. H. Van Beek, J. M.
Mogoll
on, and J. B. M. Middelburg. 2016. Global riverine
nitrogen (N) and phosphorus (P) input, retention and
export during the 20th century. Biogeosciences 13: 2441–
2451. doi:10.5194/bg-13-2441-2016
Botrel, M., L. A. Bristow, M. A. Altabet, I. Gregory-Eaves, and
R. Maranger. 2017. Assimilation and nitrification in
pelagic waters: Insights using dual nitrate stable isotopes
(d
15
N, d
18
O) in a shallow lake. Biogeochemistry 135: 221–
237. doi:10.1007/s10533-017-0369-y
Carpenter, S. R. 2005. Eutrophication of aquatic ecosystems:
bistability and soil phosphorus. Proceedings of the
National Academy of Sciences of the United States of
America 102: 10002–10005. doi:10.1073/pnas.0503959102
Carpenter, S. R., N. F. Caraco, D. L. Correll, R. W. Howarth,
A. N. Sharpley, and V. H. Smith. 1998. Nonpoint pollution
of surface waters with phosphorus and nitrogen. Ecol.
Appl. 8: 559–568. doi:10.1890/1051-0761(1998)008[0559:
NPOSWW]2.0.CO;2]
Cleveland, C. C., and others. 1999. Global patterns of terres-
trial biological nitrogen (N
2
) fixation in natural ecosys-
tems. Global Biogeochem. Cycles 13: 623–645. doi:
10.1029/1999GB900014
Cleveland, C. C., and D. Liptzin. 2007. C: N: P stoichiometry
in soil: Is there a “Redfield ratio” for the microbial bio-
mass? Biogeochemistry 85: 235–252. doi:10.1007/s10533-
007-9132-0
Maranger et al. Stoichiometry of carbon, nitrogen, and phosphorus
10
Clow, D. W., S. M. Stackpoole, K. L. Verdin, D. E. Butman,
Z. Zhu, D. P. Krabbenhoft, and R. G. Striegl. 2015.
Organic carbon burial in lakes and reservoirs of the con-
terminous United States. Environ. Sci. Technol. 49: 7614–
7622. doi:10.1021/acs.est.5b00373
Cole, J. J., and others. 2007. Plumbing the global carbon
cycle: Integrating inland waters into the terrestrial carbon
budget. Ecosystems 10: 172–185. doi:10.1007/s10021-006-
9013-8
Collins, S. M., S. K. Oliver, J.-F. Lapierre, E. H. Stanley, J. R.
Jones, T. Wagner, and P. A. Soranno. 2017. Lake nutrient
stoichiometry is less predictable than nutrient concentra-
tions at regional and sub-continental scales. Ecol. Appl.
27: 1529–1540. doi:10.1002/eap.1545
Conley,D.J.,H.W.Paerl,R.W.Howarth,D.F.Boesch,S.P.
Seitzinger, K. E. Havens, C. Lancelot, and G. E. Likens. 2009.
Controlling eutrophication: Nitrogen and phosphorus. Sci-
ence 323: 1014–1015. doi:10.1126/science.1167755
Cordell, D., J.-O. Drangert, and S. White 2009. The story of
phosphorus: Global food security and food for thought.
Global Environmental Change 19: 292–305. doi:10.1016/
j.gloenvcha.2008.10.009
Deemer, B. R., and others. 2016. Greenhouse gas emissions
from reservoir water surfaces: A new global synthesis. Bio-
Science 66: 949–964. doi:10.1093/biosci/biw117
Downing, J. A., S. B. Watson, and E. McCauley. 2001. Pre-
dicting cyanobacteria dominance in lakes. Can. J. Fish.
Aquat. Sci. 58: 1905–1908. doi:10.1139/f01-143
Drake, T. W., P. A. Raymond, and R. G. M. Spencer. 2017.
Terrestrial carbon inputs to inland waters: A current syn-
thesis of estimates and uncertainty. Limnol. Oceanogr.:
Lett. doi:10.1002/lol2.10055
Elser, J. J., and others. 2000. Nutritional constraints in terres-
trial and freshwater food webs. Nature 408: 578. doi:
10.1038/35046058
Elser, J. J., and others. 2003. Growth rate–stoichiometry cou-
plings in diverse biota. Ecol. Lett. 6: 936–943. doi:
10.1046/j.1461-0248.2003.00518.x
Elser, J. J., and others. 2007. Global analysis of nitrogen and
phosphorus limitation of primary producers in freshwater,
marine and terrestrial ecosystems. Ecol. Lett. 10: 1135–
1142. doi:10.1111/j.1461-0248.2007.01113.x
Finlay, J. C., G. E. Small, and R. W. Sterner. 2013. Human
influences on nitrogen removal in lakes. Science 342:
247–250. doi:10.1126/science.1242575
Galloway, J. N., J. D. Aber, J. A. N. W. Erisman, S. P.
Seitzinger, R. W. Howarth, E. B. Cowling, and B. J. Cosby.
2003. The nitrogen cascade. BioScience 53: 341–356. doi:
10.1641/0006-3568(2003)053[0341:TNC]2.0.CO;2]
Galloway, J. N., and others. 2004. Nitrogen cycles: Past, pres-
ent, and future. Biogeochemistry 70: 153–226. doi:
10.1007/s10533-004-0370-0
Gao, W., R. W. Howarth, D. P. Swaney, B. Hong, and H. C.
Guo. 2015. Enhanced N input to Lake Dianchi Basin from
1980 to 2010: Drivers and consequences. Sci. Total Envi-
ron. 505: 376–384. doi:10.1016/j.scitotenv.2014.10.016
Glibert, P. M., R. Maranger, D. J. Sobota, and L. Bouwman.
2014. The Haber Bosch-harmful algal bloom (HB-HAB)
link. Environ. Res. Lett. 9: 105001. doi:10.1088/1748-
9326/9/10/105001
Goyette, J.-O., E. M. Bennett, R. W. Howarth, and R.
Maranger. 2016. Changes in anthropogenic nitrogen and
phosphorus inputs to the St. Lawrence sub-basin over 110
years and impacts on riverine export. Global Biogeochem.
Cycles 30: 1000–1014. doi:10.1002/2016GB005384
Grantz, E. M., B. E. Haggard, and J. T. Scott. 2014. Stoichio-
metric imbalance in rates of nitrogen and phosphorus
retention, storage, and recycling can perpetuate nitrogen
deficiency in highly-productive reservoirs. Limnol. Ocean-
ogr. 59: 2203–2216. doi:10.4319/lo.2014.59.6.2203
Green, M. B., and J. C. Finlay. 2009. Patterns of hydrology
controlling stream water total nitrogen to total phospho-
rus ratios, Biogeochemistry 99: 15–30. doi:10.1007/
s10533-009-9394-9
Harrison, J. A., and others. 2009. The regional and global sig-
nificance of nitrogen removal in lakes and reservoirs. Bio-
geochemistry 93: 143–157. doi:10.1007/s10533-008-9272-x
Harrison, J. A., B. R. Deemer, M. K. Birchfield, and M. T.
O’Malley. 2017. Reservoir water-level drawdowns acceler-
ate and amplify methane emission. Environ. Sci. Technol.
51: 1267–1277. doi:10.1021/acs.est.6b03185
Hayes, N. M., B. R. Deemer, J. R. Corman, R. Razavi, and K.
E. Strock. 2017. Key differences between lakes and reser-
voirs modify climate signals: A case for a new conceptual
model. Limnol. Oceanogr.: Lett. 2: 47–62. doi:10.1002/
lol2.10036
Haygarth, P. M., and others. 2014. Sustainable phosphorus
management and the need for a long-term perspective:
The legacy hypothesis. Environ. Sci. Technol. 48: 8417–
8419. doi:10.1021/es502852s
Hecky, R. E., and P. Kilham. 1988. Nutrient limitation of phyto-
plankton in freshwater and marine environments: A review
of recent evidence on the effects of enrichment. Limnol.
Oceanogr. 33:796–822.doi:10.4319/lo.1988.33.4part2.0796
Hecky, R. E., P. Campbell, and L. L. Hendzel. 1993. The stoi-
chiometry of carbon, nitrogen, and phosphorus in partic-
ulate matter of lakes and oceans. Limnol. Oceanogr. 38:
709–724. doi:10.4319/lo.1993.38.4.0709
Holgerson, M. A., and P. A. Raymond. 2016. Large contribution
to inland water CO
2
and CH
4
emissions from very small
ponds. Nat. Geosci. 9: 222–226. doi:10.1038/ngeo2654
Hotchkiss, E. R., R. O. Hall, Jr., R. A. Sponseller, D. Butman,
J. Klaminder, H. Laudon, M. Rosvall, and J. Karlsson.
2015. Sources of and processes controlling CO
2
emissions
change with the size of streams and rivers. Nat. Geosci. 8:
696–699. doi:10.1038/ngeo2507
Howarth, R. W., and others. 1996. Regional nitrogen budgets
and riverine N & P fluxes for the drainages to the North
Maranger et al. Stoichiometry of carbon, nitrogen, and phosphorus
11
Atlantic Ocean: Natural and human influences. Biogeo-
chemistry 35: 75–139. doi:10.1007/BF02179825
Huszar, V. L., N. F. Caraco, F. Roland, and J. J. Cole. 2006.
Nutrient–chlorophyll relationships in tropical–subtropical
lakes: Do temperate models fit? Biogeochemistry 79: 239–
250. doi:10.1007/s10533-006-9007-9
Koehler, B., T. Landelius, G. A. Weyhenmeyer, N. Machida,
and L. J. Tranvik. 2014. Sunlight-induced carbon dioxide
emissions from inland waters. Global Biogeochem. Cycles
28: 696–711. doi:10.1002/2014GB004850
Lehner, B., and others. 2011. High-resolution mapping of
the world’s reservoirs and dams for sustainable river-flow
management. Front. Ecol. Environ. 9: 494–502. doi:
10.1890/100125
Maavara, T., C. T. Parsons, C. Ridenour, S. Stojanovic, H. H.
D€
urr, H. R. Powley, and P. Van Cappellen. 2015. Global
phosphorus retention by river damming. Proc. Natl. Acad.
Sci. USA 112: 15603–15608. doi:10.1073/pnas.1511797112
MacDonald, G. K., E. M. Bennett, P. A. Potter, and N.
Ramankutty. 2011. Agronomic phosphorus imbalances
across the world’s croplands. Proc. Natl. Acad. Sci. USA
108: 3086–3091. doi:10.1073/pnas.1010808108
Martiny, A. C., C. T. A. Pham, F. W. Primeau, J. A. Vrugt, J.
K. Moore, S. A. Levin, and M. W. Lomas. 2013. Strong lat-
itudinal patterns in the elemental ratios of marine plank-
ton and organic matter. Nat. Geosci. 6: 279–283. doi:
10.1038/ngeo1757
Mateo-Sagasta, J., L. Raschid-Sally, and A Thebo. 2015.
Global wastewater and sludge production, treatment and
use, p. 15–38. In P. Drechsel, M. Qadir, and D. Wichelns
[eds.], Wastewater. Springer.
Mayorga, E., and others. 2010. Global Nutrient Export from
WaterSheds 2 (NEWS 2): Model development and imple-
mentation. Environ. Model. Softw. 25: 837–853. doi:
10.1016/j.envsoft.2010.01.007
McDonald, C. P., E. G. Stets, R. G. Striegl, and D. Butman.
2013. Inorganic carbon loading as a primary driver of dis-
solved carbon dioxide concentrations in the lakes and res-
ervoirs of the contiguous United States. Global
Biogeochem. Cycles 27: 285–295. doi:10.1002/gbc.20032
McGroddy, M. E., T. Daufresne, and L. O. Hedin. 2004. Scal-
ing of C: N: P stoichiometry in forests worldwide: Implica-
tions of terrestrial redfield-type ratios. Ecology 85: 2390–
2401. doi:10.1890/03-0351
Mendonc¸a, R., R. A. M€
uller, D. Clow, C. Verpoorter, P.
Raymond, L. J. Tranvik, and S. Sobek. 2017. Organic car-
bon burial in global lakes and reservoirs. Nat. Commun.
8: 1694. doi:10.1038/s41467-017-01789-6
Mulholland, P. J., and others. 2008. Stream denitrification
across biomes and its response to anthropogenic nitrate
loading. Nature 452: 202–205. doi:10.1038/nature06686
M€
uller, B., L. D. Bryant, A. Matzinger, and A. Wuest. 2012.
Hypolimnetic oxygen depletion in eutrophic lakes. Envi-
ron. Sci. Technol. 46: 9964–9971. doi:10.1021/es301422r
N€
urnberg, G. K. 1984. The prediction of internal phosphorus
load in lakes with anoxic hypolimnia. Limnol. Oceanogr.
29: 111–124. doi:10.4319/lo.1984.29.1.0111
Paerl, H. W., and others. 2016. It takes two to tango: When
and where dual nutrient (N & P) reductions are needed to
protect lakes and downstream ecosystems. Environ. Sci.
Technol. 50: 10805–10813. doi:10.1021/acs.est.6b02575
Raymond, P. A., and others. 2013. Global carbon dioxide
emissions from inland waters. Nature 503: 355–359. doi:
10.1038/nature12760
Redfield, A. C. 1958. The biological control of chemical fac-
tors in the environment. Am. Sci. 46: 205–221.
Rosemond, A. D., J. P. Benstead, P. M. Bumpers, V. Gulis, J.
S. Kominoski, D. W. P. Manning, K. Suberkropp, and J. B.
Wallace. 2015. Freshwater ecology. Experimental nutrient
additions accelerate terrestrial carbon loss from stream
ecosystems. Science 347: 1142–1145. doi:10.1126/science.
aaa1958
Schade, J. D., J. F. Espeleta, C. A. Klausmeier, M. E.
McGroddy, S. A. Thomas, and L. Zhang. 2005. A concep-
tual framework for ecosystem stoichiometry: Balancing
resource supply and demand. Oikos 109: 40–51. doi:
10.1111/j.0030-1299.2005.14050.x
Schade, J. D., and others. 2011. The stoichiometry of nitro-
gen and phosphorus spiralling in heterotrophic and auto-
trophic streams. Freshw. Biol. 56: 424–436. doi:10.1111/
j.1365-2427.2010.02509.x
Schindler, D. W. 1977. Evolution of phosphorus limitation
in lakes. Science 195: 260–262. doi:10.1126/science.195.
4275.260
Schlesinger, W. H., and E. S. Bernhardt. 2013. Biogeochemis-
try: An analysis of global change. Elsevier/Academic Press.
Scott, J. T., and M. J. McCarthy. 2010. Nitrogen fixation may
not balance the nitrogen pool in lakes over timescales rel-
evant to eutrophication management. Limnol. Oceanogr.
55: 1265–1270. doi:10.4319/lo.2010.55.3.1265
Seitzinger, S., J. A. Harrison, E. Dumont, A. H. W. Beusen,
and A. F. Bouwman. 2005. Sources and delivery of carbon,
nitrogen, and phosphorus to the coastal zone: An over-
view of Global Nutrient Export from Watersheds (NEWS)
models and their application. Global Biogeochem. Cycles
19: GB4S01. doi:10.1029/2005GB002606
Seitzinger, S., J. A. Harrison, J. K. B€
ohlke, A. F. Bouwman, R.
Lowrance, B. Peterson, C. Tobias, and G. Van Drecht.
2006. Denitrification across landscapes and waterscapes: A
synthesis. Ecol. Appl. 16: 2064–2090. doi:10.1890/1051-
0761(2006)016[2064:DALAWA]2.0.CO;2]
Seitzinger, S. P., and others. 2010. Global river nutrient
export: A scenario analysis of past and future trends.
Global Biogeochem. Cycles 24: GB0A08. doi:10.1029/
2009GB003587
Smil, V. 2000. Phosphorus in the environment: Natural flows
and human interferences. Annu. Rev. Energy Environ. 25:
53–88. doi:10.1146/annurev.energy.25.1.53
Maranger et al. Stoichiometry of carbon, nitrogen, and phosphorus
12
Smith, V. H. 1990. Nitrogen, phosphorus, and nitrogen fixa-
tion in lacustrine and estuarine ecosystems. Limnol. Oce-
anogr. 35: 1852–1859. doi:10.4319/lo.1990.35.8.1852
Smith, V. H., G. D. Tilman, and J. C. Nekola. 1999. Eutrophi-
cation: Impacts of excess nutrient inputs on freshwater,
marine, and terrestrial ecosystems. Environ. Pollut. 100:
179–196. doi:10.1016/S0269-7491(99)00091-3
Soranno, P. A., K. S. Cheruvelil, T. Wagner, K. E. Webster,
and M. T. Bremigan. 2015. Effects of land use on lake
nutrients: The importance of scale, hydrologic connectiv-
ity, and region. PLoS ONE 10: e0135454.
Soued, C., P. A. del Giorgio, and R. Maranger. 2016. Patterns
in N2O fluxes across boreal aquatic networks challenge
global emission models. Nat. Geosci. 9: 116–120. doi:
10.1038/ngeo2611
Stanley, E. H., N. J. Casson, S. T. Christel, J. T. Crawford, L.
C. Loken, and S. K. Oliver. 2016. The ecology of methane
in streams and rivers: Patterns, controls, and global signif-
icance. Ecol. Monogr. 86: 146–171. doi:10.1890/15-1027
Steffen, W., and others. 2015. Planetary boundaries: Guiding
human development on a changing planet. Science 347:
1259855. doi:10.1126/science.1259855
Sterner, R. W. 1990. The ratio of nitrogen to phosphorus
resupplied by herbivores: Zooplankton and the algal com-
petitive arena. Am. Nat. 136: 209–229. doi:10.1086/
285092
Stets, E. G., and R. G. Striegl. 2012. Carbon export by rivers
draining the conterminous United States. Inland Waters
2: 177–184. doi:10.5268/IW-2.4.510
Syvitski, J. P. M., and others. 2009. Sinking deltas due to
human activities. Nat. Geosci. 2: 681. doi:10.1038/
ngeo629
They, N. H., A. M. Amado, and J. B. Cotner. 2017. Redfield
ratios in inland waters: Higher biological control of C:N:P
ratios in tropical semi-arid high water residence time
lakes. Front. Microbiol. 8: 1505. doi:10.3389/fmicb.2017.
01505
Tranvik, L. J., and others. 2009. Lakes and reservoirs as regu-
lators of carbon cycling and climate. Limnol. Oceanogr.
54: 2298–2314. doi:10.4319/lo.2009.54.6_part_2.2298
Urabe, J., J. J. Elser, M. Kyle, T. Yoshida, T. Sekino, and Z.
Kawabata. 2002. Herbivorous animals can mitigate unfav-
ourable ratios of energy and material supplies by enhanc-
ing nutrient recycling. Ecol. Lett. 5: 177–185. doi:
10.1046/j.1461-0248.2002.00303.x
U.S. Environmental Protection Agency. 2010. National
aquatic resource surveys. National lakes assessment 2007
(data and metadata files); [accessed 2015 May 15]. Avail-
able from http://www.epa.gov/national-aquatic-resource-
surveys/data-national-aquatic-resource-surveys.
U.S. Geological Survey. 2016. National Water Information
System data available on the World Wide Web (USGS
Water Data for the Nation).
Vanni, M. J., W. H. Renwick, A. M. Bowling, M. J. Horgan,
and A. D. Christian. 2011. Nutrient stoichiometry of
linked catchment-lake systems along a gradient of land
use. Freshw. Biol. 56: 791–811. doi:10.1111/j.1365-
2427.2010.02436.x
Vitousek, P. M. 2003. Stoichiometry and flexibility in the
Hawaiian model system, p. 117–133. In J. M. Melillo, C.
B. Field, and B. Moldan [eds.], Interactions of the major
biogeochemical cycles. Island Press.
Vitousek, P. M., J. D. Aber, R. W. Howarth, G. E. Likens, P.
A. Matson, D. W. Schindler, W. H. Schlesinger, and D. G.
Tilman. 1997. Human alteration of the global nitrogen
cycle: Sources and consequences. Ecol. Appl. 7: 737–750.
doi:10.1890/1051-0761(1997)007[0737:HAOTGN]2.0.CO;2]
Vitousek, P. M., D. N. L. Menge, S. C. Reed, and C. C.
Cleveland. 2013. Biological nitrogen fixation: Rates, pat-
terns and ecological controls in terrestrial ecosystems.
Philos. Trans. R. Soc. Lond. B Biol. Sci. 368: 20130119.
doi:10.1098/rstb.2013.0119
Watson, S. B., E. McCauley, and J. A. Downing. 1997. Pat-
terns in phytoplankton taxonomic composition across
temperate lakes of differing nutrient status. Limnol. Oce-
anogr. 42: 487–495. doi:10.4319/lo.1997.42.3.0487
Zarfl, C., A. E. Lumsdon, J. Berlekamp, L. Tydecks, and K.
Tockner. 2015. A global boom in hydropower dam con-
struction. Aquat. Sci. 77: 161–170. doi:10.1007/s00027-
014-0377-0
Zehr, J. P., and B. B. Ward. 2002. Nitrogen cycling in the
ocean: New perspectives on processes and paradigms.
Appl. Environ. Microbiol. 68: 1015–1024. doi:10.1128/
AEM.68.3.1015-1024.2002
Acknowledgments
We would like to thank E. Stanley, P. Soranno, and two anonymous
reviewers for constructive comments on the manuscript. We thank Max-
ime Leclerc for designing Figs. 1, 3. R. M. was supported by NSERC Dis-
covery Grant, S. E. J. by NSF award DEB-1547866, and J. B. C. by NSF-
IOS 1257571.
Submitted 01 August 2017
Revised 11 January 2018
Accepted 01 March 2018
Maranger et al. Stoichiometry of carbon, nitrogen, and phosphorus
13
