Title: Addition of Alkalinity to Rivers: a new CO2 Removal Strategy

Full text

Posted on 11 May 2023 — The copyright holder is the author/funder. All rights reserved. No reuse without permission. — https://doi.org/10.22541/essoar.168380809.92137625/v1 — This a preprint and has not been peer reviewed. Data may be preliminary.
Shannon Sterling1,2, Edmund Halfyard2, Kristin Hart1, Benjamin Trueman2, G¨unther
Grill3, and Bernhard Lehner3,4
1Department of Earth and Environmental Sciences, Dalhousie University
2CarbonRun Carbon Dioxide Removal Ltd
3Confluvio Consulting Inc
4Department of Geography, McGill University
May 11, 2023
1

1
Title: Addition of Alkalinity to Rivers: a new CO2 Removal Strategy
Authors: Shannon M. Sterling1,2*, Edmund Halfyard2, Kristin Hart1, Benjamin Trueman2,
Günther Grill3, Bernhard Lehner3,4
5
Affiliations:
1Department of Earth and Environmental Sciences, Dalhousie University; Halifax, Canada.
2CarbonRun Carbon Dioxide Removal Ltd.; Halifax, Canada.
3Confluvio Consulting Inc.; Montréal, Canada
10
4Department of Geography, McGill University; Montreal, Canada.
*Corresponding author. Email: shannon.sterling@dal.ca
15
One-Sentence Summary: Addition of alkalinity to rivers is a previously unexplored but
promising new tool to aid our global mission to reduce serious risks from climate change while
restoring aquatic habitats.
20

2
Abstract
In addition to rapid reductions of greenhouse gas emissions, large scale carbon dioxide removal
(CDR) is needed to limit severe climate warming. While recent CDR efforts have focused on land
surfaces and oceans, rivers have not yet been explored as a CDR medium. Here, we examine the
CDR potential of the addition of alkaline materials to rivers (RiverCDR). We synthesize
25
knowledge gained from long-term studies of river liming that demonstrate significant
improvements to water quality, fish populations and overall ecosystem health and combine it with
regional and global modeling approaches. We find RiverCDR can permanently sequester carbon
through both increasing carbonation weathering reactions (enhanced weathering) and increasing
rivers’ and receiving ocean water’s capacity to retain dissolved inorganic carbon (increased
30
alkalinity). Our thermodynamic modelling demonstrates the potential to use carbonate rocks for
RiverCDR: a safe and inexpensive source of alkalinity. Our global scale modelling shows that
while staying within safe limits RiverCDR has the potential to permanently store hundreds of
MtCO2e yr-1, which amounts to ~10% of current conventional land-based CDR methods (mostly
reforestation and afforestation) and exceeds other novel CDR technologies such as biochar. To
35
further explore opportunities, we propose a simple system for monitoring reporting and
verification at the river mouth, using standardized hydrological approaches. We recommend that
RiverCDR be added to our current options for CDR, as it has the potential to drawdown carbon
while enhancing biodiversity and ecosystem functions, employment, and local livelihoods, if done
correctly, and in the right locations.
40

3
Main
Carbon dioxide removal (CDR) technologies that are effective, affordable, scalable, safe,
permanent and verifiable are urgently needed to reduce severe risks from climate change and reach
45
net zero1; however, our current available CDR options are far from sufficient to meet these goals.
We urgently need to develop new CDR methods, particularly those that minimize land
requirements and energy consumption and that can scale up rapidly and cheaply2. The landscape
of CDR options include land-based approaches such as the application of alkaline materials to land
(via Enhanced Rock Weathering (ERW))3 and ocean-based approaches, such as Ocean Alkalinity
50
Enhancement (OAE)4 (Figure 1). As of yet, no CDR technologies exist for rivers, although recent
calls have been made to integrate river management and climate change5,6. Here, we propose a
new CDR pathway, RiverCDR: a technology that artificially accelerates dissolution of alkaline
materials in rivers. We examine how it may achieve CDR and meet key net-zero goals such as
permanence, safety, societal acceptance, scalability, and we propose a framework for monitoring,
55
reporting and verification (MRV).
Rivers are a hotspot for carbon flows7. Rivers and inland waters have been estimated to deliver
approximately 0.96 PgCyr-1 of organic and inorganic carbon from land areas to oceans via the
land-to-ocean aquatic continuum (LOAC)8. Water draining from land surfaces delivers up to 18%
of terrestrial gross primary production to inland waters via shunting of CO2 from soils or organic
60
carbon that is later mineralized to CO29–11, generating high pCO2 (partial pressure of CO2) in some
reaches (> 10,000 µatm)11. Carbonate alkalinity (CAlk) is produced from carbonation weathering,
a set of well-known mineral dissolution reactions of carbonate and silicate rocks that increase CAlk
and promote a transfer of CO2 from the atmosphere4 (Equation 1). Large amounts (approximately
0.4 PgC yr-1) of CAlk (comprising primarily HCO3-) are leached from soils and bedrock and
65

4
delivered to oceans via the global inland water network8, driving the long-term carbonate-silicate
cycles, Earth’s natural regulator of atmospheric CO2 and global temperature over geologic time12
(Figure 1a). Upon arriving in the open water network, much of this dissolved inorganic carbon
(DIC, comprising CO2, HCO3- and CO32-) is evaded to the atmosphere as CO213,14, estimated as
between 1.9 and 2.3 PgC yr-1 15 and only a portion is exported via stable forms of CAlk directly to
70
the ocean for long-term storage8,11,14,15. Thus, rivers commonly have pCO2 values well above
atmospheric levels11,16. This lateral land-to-ocean carbon transport via inland waters and rivers has
only recently been conceptualized in detail and is not well represented in the current generation of
land carbon models7,17, nor in global carbon budget accounting18,19.
A prevalent view is that HCO3- behaves conservatively once it has entered the open water network
75
while transiting through rivers and inland waters8 and that HCO3- dominates DIC in most
freshwaters20. However, pH controls the speciation of DIC into carbonate anions or CO2. When a
volume of water containing HCO3- is mixed with a volume of low pH water (below approximately
pH 8.3), the low pH can convert stable HCO3- to CO2, and thus be subject to evasion to the
atmosphere (Figure 2). The range of natural river pH levels overlaps with the critical values pH
80
4.5-8.3 for carbonate speciation21. Thus, decreased pH, particularly around pH 5.5, where there is
a sharp transition between CO2 and HCO3-, can increase carbon losses along the LOAC through
increased CO2 evasion, lowering HCO3- exports to long-term storage in the oceans.
Human activities, such as fossil fuel burning and fertilizer application, have reduced base cation
concentrations in soils and lowered pH and alkalinity of inland waters. Currently many rivers
85
remain acidified with low alkalinity and pH22, despite reductions in acid emissions23,24. This lack
of recovery is due to both the continued exceedance of critical loads of acid deposition and the
slow recruitment of critical base cations in catchment soils25,26. To save aquatic species at risk,
such as Atlantic salmon (Salmo salar), from freshwater acidification, ground carbonate rocks have

5
been added to surface waters in river liming programs in North America and Europe for decades27–
90
30 (Supplementary Data). These programs use ‘lime dosers’ to obtain a cost-effective, precise and
continuous dispersal of alkaline materials directly to rivers. Lime dosing has been successful in
raising pH and base cation concentrations above threshold concentrations needed for ecosystem
health27 (Supplementary Data). However, the CDR potential of such alkaline additions to rivers
has not yet been explored.
95
Here we combine the concepts of river liming and CDR and identify the potential of a new
negative-emission technology, RiverCDR. We define RiverCDR as the direct addition of ground
alkaline materials to river waters using methods that favour carbonation weathering reactions
(Figure 3a). RiverCDR is thus a strategy to artificially accelerate the dissolution of alkaline
minerals to increase delivery of CAlk to the oceans via rivers for long-term storage. In RiverCDR,
100
carbon may be drawn down via two main pathways: carbonation weathering and increased
alkalinity along the treated water plume. The captured carbon is then transported by the river to
oceans for long-term storage (Figure 3a).
By mapping out the thermodynamic and stoichiometric limits of carbonation weathering, we find
that RiverCDR is possible with carbonate minerals as the alkaline amendment, but that not all
105
dosing/river chemistry conditions will generate CDR. Here we consider the case of the carbonate
mineral dolomite (Equation 1):
CaMg(CO3)2 + 2CO2 + 2H2O ßà Ca2+ + Mg2+ + 4HCO3- (1)
and examine chemical equilibria in response to application of dolomite to a range of natural river
water chemistries and prescribed pCO2 levels (reflecting the range of common fluvial values11
110
from Nova Scotia, Canada (see Methods)). As expected, the carbonation weathering efficiency of
the added dolomite increases with higher pCO2 of the receiving water for a given dose (Figure 4).

6
Carbon capture is not achieved at low doses, particularly when stronger (mineral/organic) acids
are present in higher concentrations (represented by a lowered y-intercept of the stoichiometric
limit) (Figure 4), setting a dose-limit below which CDR would not occur. Another limit is the 2:1
115
line here, representing the amount of carbon introduced into the water from the carbonate minerals.
Favourable geochemical conditions are those in which added alkaline material undergoes
carbonation weathering reactions that are spontaneously driven forward5.
For small amounts of alkalinity added, the initial pH must be above the pH0, the transition region
between mineral acidic and alkaline waters, but carbon capture can be achieved in mineral acidic
120
waters at higher rates of alkaline material addition21, suggesting that RiverCDR is possible in
acidified rivers, especially in locations with high pCO2. Thus, in rivers where pH < pH0 dosing
rates of carbonate minerals must be sufficient to avoid carbonate rocks becoming a CO2 source in
rivers, especially those with elevated strong acid concentrations21,31. Most river water chemistries,
if below saturation, can favour spontaneous forward reactions of carbonate weathering reactions
125
(e.g., Equation 1), even for carbonate minerals, given sufficient pCO2 and dosing rates.
We also identify physical characteristics that promote weathering in rivers. Rivers promote
chemical disequilibrium at point inputs of alkaline material through continuous refreshing of water
via advective import. And the hydraulics of running water systems with high surface roughness
favour dissolution reaction kinetics, as turbulence promotes disruption of chemically-equilibrated
130
films developing around the weathering front. Further, elevated shear stress along the riverbed
promotes suspension of particles in the water medium.
We identify another pathway for CO2 drawdown in RiverCDR via increased alkalinity in the
receiving river and ocean waters which increases their ability to store CO2 (Figure 3a). Increased
alkalinity increases pH, which promotes the conversion of CO2 to HCO3-, thereby reducing pCO2
135

7
(decreasing CO2 evasion / increasing CO2 invasion) in the treated water plume and facilitating the
retention and delivery of DIC to the ocean. Furthermore, as we discuss below, the addition of
alkalinity to rivers increases biological productivity in acidified rivers and estuaries, removing
additional CO2 from the atmosphere as organic carbon32.
Losses to CAlk generated through carbonation weathering may occur via three paths in rivers: 1)
140
reverse reactions in which a carbonate mineral is precipitated, 2) mixing with a minerally acidic
body of water causing HCO3- to convert to CO2, or 3) uptake of HCO3- via photosynthesis33 (Figure
2). Carbonate precipitation results when saturation limits are exceeded and would reverse the
alkalinity and CDR generated. A recent modelling study estimated that calcite saturation limits do
not pose a large constraint for CDR in rivers34.
145
In summary, the feasibility of RiverCDR depends on the distance of the addition point from the
ocean, the type of minerals applied, and the river hydraulic and chemical properties (pCO2, pH,
alkalinity, temperature, dissolved constituents, and saturation limits) at the addition point and
downstream that govern the likelihood of the reaction Equation 1 to take place or to be reversed.
Here, we review how RiverCDR may meet established CDR criteria35.
150
Permanence. Once CAlk (mostly in the form of HCO3-) reaches the ocean in river estuaries, the
permanence of the carbon capture depends on the length of time HCO3- remains in solution (i.e.,
that the reaction is not reversed), estimated as 10,000s of years8,36,37. Reverse reactions of
carbonate weathering are less likely in areas where more rapid mixing and dilution occurs36,38,
areas of lower salinity water21, and in cases where alkaline solutions are equilibrated with
155
atmospheric CO2 prior to addition to the oceans4. Freshwater plumes in the ocean are dilute,
equilibrated with the atmosphere, and well-mixed, favouring carbon capture permanence of
RiverCDR.

8
Verifiability. For a CDR technology to be able to be part of the carbon market it must provide
160
transparent MRV of the net amount of carbon captured. We have developed an MRV for
RiverCDR based on changes in carbonate alkalinity export rates calculated at a single point at the
river mouth (Equation 6). In doing so, we avoid having to use forward mass balance calculations
that include all the sources and sinks of carbonate alkalinity in rivers (Equation 2-5, Figure 2) that
would be highly uncertain due to difficulty in estimating variables such as gas transfer velocities
165
and short-term changes in carbonation weathering efficiencies. At-a-point river mouth flux
estimations integrate all upstream carbonate alkalinity inputs and outputs such as photosynthesis,
precipitation, and other unanticipated gains and losses (Figure 2), as rivers transport dissolved and
suspended constituents from the watershed to a single outlet point (i.e., the river mouth). Our
RiverCDR MRV draws on standardized and well-established measurement techniques8,18. This
170
RiverCDR MRV calculation (Equation 6) is conservative because it does not account for increased
organic carbon stores associated with increased primary production or biomass associated with
higher trophic levels and does not include carbon drawdown of CO2 by increased alkalinity in
receiving ocean waters.
175
DCAlk = W + A – L - G (2)
where
DCAlk = increase in carbonate alkalinity at the river mouth relative to background
and is equivalent to CDR in [Mass CO2eq] [Time]-1
W = weathering-based CDR = M *S * Ew * Et
180
(3)

9
where
M = mass alkaline material added
S = stoichiometric relation of CDR vs tonnes alkaline material added
Ew = carbonation weathering efficiency
185
Et = CAlk transport efficiency to ocean, as a function of %R
A = alkalinity-based CDR = D + F (4)
D = reduced loss of carbon inputs to treated river segment and receiving
ocean waters = ICin * %R
190
ICin = inorganic carbon inputs into treated reach via groundwater, surface
water, or organic carbon mineralization
%R = the percentage increase in dissolved inorganic carbon (DIC) as
bicarbonate instead of CO as a function of river pH
F = reduced evasion of atmospheric CO2 by increased alkalinity = SW * S
195
* kw * ((pCO2 before - pCO2 after)water - (pCO2)atm))
(5)
SW = area of surface water in treated river and ocean plumes
S = solubility of CO2 in water for given temperature and salinity conditions
kw = gas exchange velocity
200
pCO2 = partial pressure of CO2
L = losses (Figure 3)

10
G = grey emissions produced by the supply chain of alkaline material (mining +
grinding + transport + dosing)
205
DCAlk = ([HCO3-]treatment – [HCO3-]control) * Q – C (6)
where
[HCO3-] = concentration of bicarbonate and carbonate anions, compared to
background levels in [Mass CO2e] [Length]-3
Q = river discharge in [Length]3 [Time]-1
210
C = the carbon added in any alkaline amendment (e.g., CaMg(CO3)2) in [Mass
CO2e] [Time]-1
To achieve net carbon removal, more dissolved CAlk must be delivered to the river mouth than
was added in the mineral amendments (Figure 4).
215
Safety. We identify potential risks of RiverCDR to aquatic ecosystems and identify that the safety
of RiverCDR depends on the type of material added, the nature of the receiving waters, the
dose/method of application, and the ecology of the selected river. We review numerous long-term
studies of lime dosing in acidified rivers and summarize how addition of carbonate minerals impact
aquatic ecosystems (Supplementary Data). While the impact of the addition of carbonate minerals
220
to rivers is generally considered safe and is well known (Supplementary Data), the impact of other
alkaline materials on aquatic ecosystems is not well-studied and poses added risk (Table 1).
Addition of strong bases, such as CaO and MgO, can cause rapid rises in pH leading to fish kills
and ecological harm40. Addition of silicate minerals, particularly of those which undergo

11
incongruent weathering where secondary minerals are formed, poses risks for release of potentially
225
toxic metals such as nickel and chromium to rivers and estuaries41, for production of clays that can
cloud river water or clog river substrates, and for silicosis from inhalation by humans during
RiverCDR operations (Table 1).
For all mineral amendment types, RiverCDR poses a risk if alkalinity were increased above
ecosystem health thresholds42, particularly for systems that have species at risk which are
230
acidophilic. RiverCDR would also pose a risk to aquatic health if the dosing method was not
continuously managed and produced large swings in pH/water chemistry. To reduce potential risks
to ecosystem health from RiverCDR dosing must be precisely calculated and executed, and river
biogeochemistry and ecology must be well-understood.
235
Scalability. Global-scale hydrological modelling shows that the potential of CDR is in the 100s
of Mt CO2e yr-1 (Methods Figure 1). We consider a scenario where RiverCDR is applied to low
alkalinity rivers ( <0.6 mM) across the globe, the amount of CAlk generated is limited to stay
within safe limits (0.6 mM) to avoid a shift in microbiological communities42. In this scenario,
RiverCDR has a potential of approximately 300 Mt CO2e yr-1.
240
Affordability. Affordability of RiverCDR is aided by its simplicity and its ability to use readily
available and inexpensive feedstock. Dosing technology is already well developed27. Existing
alkaline material dosers in Nova Scotia have a low land footprint (< 50 m2) and low energy costs
(< 600 kW·yr-1), which are provided by solar and wind renewable energy while river advection
245
transports the captured carbon downstream of the addition point to the river mouth. A range of
alkaline materials could achieve RiverCDR via carbonation weathering, such as silicate rock

12
minerals, and industrial minerals32,39. Alkaline minerals must be milled to small particle sizes (e.g.,
< 0.07 mm) so they can remain in suspension in the river until dissolution is complete. Thus,
carbon costs are incurred during transportation of alkaline material to the sites and during the
250
milling process. As with ERW and OAE, a potential for reusing waste materials exists, such as
waste dust in the aggregate industry, or in unused dolomite in the cement industry. A challenge
for RiverCDR, as with ERW and OAE, is the transport costs of feedstock to deployment sites.
RiverCDR provides an opportunity to use carbonate rocks as feedstock for significant capture and
storage of atmospheric CO2. Our finding that river water chemistry provides a thermodynamically
255
favourable medium for carbonate rocks to undergo spontaneous carbonation weathering opens up
new opportunities to lower risks and costs for CDR (Table 1). The ability to use carbonate
feedstock is attractive because carbonate minerals weather congruently (e.g., do not generate clay
minerals), contain fewer metal impurities than silicate rocks, and dissolve orders of magnitude
faster than silicate minerals under the same conditions32,34,44,45. Further, carbonate rock requires no
260
additional processing (e.g., does not require heating) and may require less energy for transport
because carbonate quarries are widely available globally46.
Uncertainties. Further research is needed on the impacts of alkaline additions on nitrogen cycling,
cyanobacterial blooms and dissolved humic substances in rivers, including possible increased
265
mineralization of dissolved organic carbon43 and effects of cation exchange processes during
humic flocculation in estuaries. The ecological impacts of the addition of alkalinity to
circumneutral rivers is yet unexplored. Information on regional or local impacts of estuarine
alkalinization is lacking38. Further research is needed on the ecosystem effects and general site
risks involved in the addition strong bases to rivers (Table 1); this uncertainty has caused concern
270

13
in ocean application of alkaline materials52, highlighting the importance of community
consultation and support.
Co-Benefits / Social acceptability. Social acceptability of RiverCDR is supported by its potential
to deliver co-benefits to the communities where it is deployed (Figure 3b). Long-term studies show
275
that addition of carbonate materials to these rivers can reverse the impact of freshwater
acidification and provide a range of ecosystem and societal benefits, depending on the site-specific
context and the method used (Supplementary Data). In fact, the addition of alkalinity to many
rivers is urgently needed to save species at risk from extirpation and increase aquatic ecosystem
productivity and biodiversity48. Increasing river alkaline exports may also deliver co-benefits to
280
coastal fishing communities through reduced acidification of the ocean receiving waters49.
RiverCDR directly delivers to estuaries, which are a hotspot for ocean acidification50,51. The co-
benefits at the local-watershed scale of CDR application may increase opportunities to improve
ecological justice of CDR and for community-based efforts and leadership by Indigenous peoples.
285
We find that RiverCDR can complement other CDR methods. RiverCDR may help increase the
efficacy of ERW that is dependent on the ability of rivers to transport the products of weathering
reactions to the ocean for storage5. In acidic rivers, small-scale ERW-generated CAlk may be lost
due to evasion21. Further, RiverCDR is well-suited for pairing with CO2 capture technologies
(DAC, BECCS, or flue gas) because RiverCDR provides a new way to sequester CO2 captured by
290
alkaline absorbants, and because the RiverCDR process is enhanced by increased pCO247.

14
Carbon fluxes from land to rivers are projected to increase with climate change. And agents of
acidification that disrupt the land-to-ocean transport of carbon are projected to increase with
climate change, including increased intensity of precipitation53 and rising CO2 in the atmosphere54,
295
and possibly sulphur aerosol deposition from solar radiation management55. Direct human activity
and climate change have already increased the land-to-river flux of organic and inorganic carbon
by approximately 12%8,17, and this increasing trend is projected to continue to rise due to
reductions to the soil carbon pool, CO2 fertilization increasing terrestrial NPP, and via increased
metabolism in lakes11,56. Our findings suggest that RiverCDR may increase the ability of rivers to
300
maintain the delivery of land-based carbon to oceans, and therefore reduce CO2 evasion, under
these projected future stresses.
In conclusion, here we have identified a new negative-emission technology, RiverCDR, that may
add to the CDR strategies needed to achieve net zero. We show how RiverCDR can contribute to
climate action while providing demonstrated co-benefits such as increased freshwater ecosystem
305
health and alleviation of freshwater and ocean acidification acidification if applied correctly and
at the right location. Rivers are a promising arena for CDR as they are hotspots for carbon
exchange11 with elevated pCO2, and therefore large amounts of CO2 to capture. Rivers have
physical and chemical conditions that favour carbonation weathering reaction kinetics and reduced
likelihood of carbonate precipitation of alkalinity delivered to the ocean. Thermodynamic
310
modelling shows a potential use of carbonate rocks in rivers to remove CO2 from the atmosphere.
By being applied at a point and driven by gravity, RiverCDR minimizes land use and minimizes
energy consumption. Due to the history of alkalinity addition to rivers to reduce acidification
threats to species at risk, the technology is relatively mature and well-understood. RiverCDR thus
may meet leading CDR criteria, such as affordability, scalability, permanence, safety, and
315
verifiability (i.e., the ability to simply quantify the CO2 removed) placing it among current options

15
of promising CDR strategies. RiverCDR thus may provide an opportunity to restore water
resources and lower greenhouse gases to combat severe climate change, supporting progress
towards three Sustainable Development Goals (SDGs): SDG6 Clean Water and Sanitation,
SDG13, Climate Action, and SDG14, Life Below Water.
320

16
References
1. Lee, H. et al. Synthesis Report of the IPCC Sixth Assessment Report (AR6): Summary for
Policymakers. (2023).
2. Ho, D. T. Carbon dioxide removal is not a current climate solution — we need to change the
narrative. Nature 616, 9–9 (2023).
325
3. Beerling, D. J. et al. Farming with crops and rocks to address global climate, food and soil
security. Nature plants 4, 138–147 (2018).
4. Hartmann, J. et al. Stability of alkalinity in ocean alkalinity enhancement (OAE) approaches
– consequences for durability of CO2 storage. Biogeosciences 20, 781–802 (2023).
5. Knapp, W. J. & Tipper, E. T. The efficacy of enhancing carbonate weathering for carbon
330
dioxide sequestration. Frontiers in Climate 4, (2022).
6. Rahman, M. F. et al. As the UN meets, make water central to climate action. Nature 615,
582–585 (2023).
7. Ciais, P. et al. Empirical estimates of regional carbon budgets imply reduced global soil
heterotrophic respiration. National Science Review 8, nwaa145 (2021).
335
8. Regnier, P., Resplandy, L., Najjar, R. G. & Ciais, P. The land-to-ocean loops of the global
carbon cycle. Nature 603, 401–410 (2022).
9. Battin, T. J. et al. Biophysical controls on organic carbon fluxes in fluvial networks. Nature
Geosci 1, 95–100 (2008).
10. Johnson, M. S. et al. CO2 efflux from Amazonian headwater streams represents a significant
340
fate for deep soil respiration. Geophysical Research Letters 35, (2008).
11. Liu, S. et al. The importance of hydrology in routing terrestrial carbon to the atmosphere via
global streams and rivers. Proceedings of the National Academy of Sciences 119,
e2106322119 (2022).

17
12. Berner, R. A. & Kothavala, Z. Geocarb III: A Revised Model of Atmospheric CO2 over
345
Phanerozoic Time. American Journal of Science 301, 182–204 (2001).
13. Tranvik, L. J. et al. Lakes and reservoirs as regulators of carbon cycling and climate.
Limnology and Oceanography 54, 2298–2314 (2009).
14. Raymond, P. A. et al. Global carbon dioxide emissions from inland waters. Nature 503, 355–
359 (2013).
350
15. Battin, T. J. et al. River ecosystem metabolism and carbon biogeochemistry in a changing
world. Nature 613, 449–459 (2023).
16. Aufdenkampe, A. K. et al. Riverine coupling of biogeochemical cycles between land,
oceans, and atmosphere. Frontiers in Ecology and the Environment 9, 53–60 (2011).
17. Lauerwald, R., Regnier, P., Guenet, B., Friedlingstein, P. & Ciais, P. How Simulations of the
355
Land Carbon Sink Are Biased by Ignoring Fluvial Carbon Transfers: A Case Study for the
Amazon Basin. One Earth 3, 226–236 (2020).
18. Casas-Ruiz, J. P. et al. Integrating terrestrial and aquatic ecosystems to constrain estimates of
land-atmosphere carbon exchange. Nat Commun 14, 1571 (2023).
19. Friedlingstein, P. et al. Global Carbon Budget 2022. Earth System Science Data 14, 4811–
360
4900 (2022).
20. Schlesinger, W. H. & Bernhardt, E. S. Biogeochemistry (Fourth Edition). (Academic Press,
2020).
21. Bertagni, M. B. & Porporato, A. The Carbon-Capture Efficiency of Natural Water
Alkalinization: Implications For Enhanced weathering. Science of The Total Environment
365
838, 156524 (2022).

18
22. Rotteveel, L., Heubach, F. & Sterling, S. M. The Surface Water Chemistry (SWatCh)
database: a standardized global database of water chemistry to facilitate large-sample
hydrological research. Earth System Science Data 14, 4667–4680 (2022).
23. Clair, T. A., Dennis, I. F., Scruton, D. A. & Gilliss, M. Freshwater acidification research in
370
Atlantic Canada: a review of results and predictions for the future. Env.Rev. 15, 153–167
(2007).
24. Monteith, D. T. et al. Dissolved organic carbon trends resulting from changes in atmospheric
deposition chemistry. Nature 450, 537 (2007).
25. Whitfield, C. J., Watmough, S. A., Aherne, J. & Dillon, P. J. A comparison of weathering
375
rates for acid-sensitive catchments in Nova Scotia, Canada and their impact on critical load
calculations. Geoderma 136, 899–911 (2006).
26. Lacoul, P., Freedman, B. & Clair, T. Effects of acidification on aquatic biota in Atlantic
Canada. Env.Rev. 19, 429–460 (2011).
27. Henrikson, L. & Brodin, Y. Liming of surface waters in Sweden — a synthesis. in Liming of
380
Acidified Surface Waters (eds. Henrikson, L. & Brodin, Y.) (Springer, 1995).
doi:10.1007/978-3-642-79309-7_1.
28. Clair, T. A. & Hindar, A. Liming for the mitigation of acid rain effects in freshwaters: a
review of recent results. Env.Rev. 13, 91–128 (2005).
29. Raymond, P. A. & Oh, N.-H. Long term changes of chemical weathering products in rivers
385
heavily impacted from acid mine drainage: Insights on the impact of coal mining on regional
and global carbon and sulfur budgets. Earth and Planetary Science Letters 284, 50–56
(2009).

19
30. Vesper, D. J., Moore, J. E. & Adams, J. P. Inorganic carbon dynamics and CO2 flux
associated with coal-mine drainage sites in Blythedale PA and Lambert WV, USA. Environ
390
Earth Sci 75, 340 (2016).
31. Taylor, L. L. et al. Increased carbon capture by a silicate-treated forested watershed
attenuated by acid deposition. Biogeosciences 18, 169–188 (2021).
32. Renforth, P. & Campbell, J. S. The role of soils in the regulation of ocean acidification.
Philosophical Transactions of the Royal Society B: Biological Sciences 376, 20200174
395
(2021).
33. Morel, F. M. M. et al. Zinc and carbon co-limitation of marine phytoplankton. Nature 369,
740–742 (1994).
34. Zhang, S. et al. River chemistry constraints on the carbon capture potential of surficial
enhanced rock weathering. Limnology and Oceanography 67, S148–S157 (2022).
400
35. Frontier Climate. Frontier Climate. https://frontierclimate.com/.
36. Rau, G. H. CO2 Mitigation via Capture and Chemical Conversion in Seawater. Environ. Sci.
Technol. 45, 1088–1092 (2011).
37. Archer, D. Fate of fossil fuel CO2 in geologic time. Journal of Geophysical Research:
Oceans 110, (2005).
405
38. Kirchner, J. S., Lettmann, K. A., Schnetger, B., Wolff, J.-O. & Brumsack, H.-J. Carbon
capture via accelerated weathering of limestone: Modeling local impacts on the carbonate
chemistry of the southern North Sea. International Journal of Greenhouse Gas Control 92,
102855 (2020).
39. Hartmann, J. et al. Enhanced chemical weathering as a geoengineering strategy to reduce
410
atmospheric carbon dioxide, supply nutrients, and mitigate ocean acidification. Rev.Geophys.
51, 113–149 (2013).

20
40. CBC, R. B. ‘Total devastation’: Dozens of fish killed in North, West Vancouver streams |
CBC News. CBC https://www.cbc.ca/news/canada/british-columbia/trout-kill-in-north-
shore-streams-1.4652572 (2018).
415
41. Haque, F., Chiang, Y. W. & Santos, R. M. Risk assessment of Ni, Cr, and Si release from
alkaline minerals during enhanced weathering. Open Agriculture 5, 166–175 (2020).
42. Iversen, L. L. et al. Catchment properties and the photosynthetic trait composition of
freshwater plant communities. Science 366, 878–881 (2019).
43. Hart, K. A., Sterling, S. M. & Halfyard, E. A. Application of dolomite to forested catchments
420
in Nova Scotia improves water quality (but more is needed to meet all targets). (In prep.).
44. Morse, J. W. & Arvidson, R. S. The dissolution kinetics of major sedimentary carbonate
minerals. Earth-Science Reviews 58, 51–84 (2002).
45. Kelemen, P. B. et al. Engineered carbon mineralization in ultramafic rocks for CO2 removal
from air: Review and new insights. Chemical Geology 550, 119628 (2020).
425
46. Hartmann, J. & Moosdorf, N. The new global lithological map database GLiM: A
representation of rock properties at the Earth surface. Geochemistry, Geophysics,
Geosystems 13, (2012).
47. Caldeira, K. & Rau, G. H. Accelerating carbonate dissolution to sequester carbon dioxide in
the ocean: Geochemical implications. Geophysical Research Letters 27, 225–228 (2000).
430
48. Sterling, S. M. et al. Ionic aluminium concentrations exceed thresholds for aquatic health in
Nova Scotian rivers, even during conditions of high dissolved organic carbon and low flow.
Hydrology and Earth System Sciences 24, 4763–4775 (2020).
49. Renforth, P. & Henderson, G. Assessing ocean alkalinity for carbon sequestration. Reviews
of Geophysics 55, 636–674 (2017).
435

21
50. Azetsu-Scott, K. et al. Calcium carbonate saturation states in the waters of the Canadian
Arctic Archipelago and the Labrador Sea. Journal of Geophysical Research: Oceans 115,
(2010).
51. Bianucci, L. et al. Sensitivity of the regional ocean acidification and carbonate system in
Puget Sound to ocean and freshwater inputs. Elementa: Science of the Anthropocene 6, 22
440
(2018).
52. BBC. Protest calls for St Ives Bay chemical test delay. BBC News (2023).
53. Chemke, R., Ming, Y. & Yuval, J. The intensification of winter mid-latitude storm tracks in
the Southern Hemisphere. Nat. Clim. Chang. 12, 553–557 (2022).
54. Phillips, J. C. et al. The Potential for CO₂-Induced Acidification in Freshwater: A Great
445
Lakes Case Study. Oceanography 28, 136–145 (2015).
55. Ricke, K. Solar geoengineering is scary - that’s why we should research it. Nature 614, 391
(2023).
56. Hasler, C. T. et al. Biological consequences of weak acidification caused by elevated carbon
dioxide in freshwater ecosystems. Hydrobiologia 806, 1–12 (2018).
450
57. Charlton, S. R., Parkhurst, C. A. & Appelo, C. A. J. phreeqc: R Interface to Geochemical
Modeling Software. (2023).
58. Dunnington, D. tidyphreeqc: Tidy Geochemical Modeling Using PHREEQC. (2022).
59. Wickham, H. et al. Welcome to the Tidyverse. Journal of Open Source Software 4, 1686
(2019).
455
60. Dunnington, D. chemr: R Data Structures for Chemistry Data. (2022).
61. Vaughan, D. & Dancho, M. furrr: Apply Mapping Functions in Parallel using Futures.
(2022).
62. Wickham, H. & Bryan, J. readxl: Read Excel Files. (2023).

22
63. Lehner, B. & Grill, G. Global river hydrography and network routing: baseline data and new
460
approaches to study the world’s large river systems. Hydrological Processes 27, 2171–2186
(2013).
64. Lehner, B., Verdin, K. & Jarvis, A. New Global Hydrography Derived From Spaceborne
Elevation Data. Eos, Transactions American Geophysical Union 89, 93–94 (2008).
65. Döll, P., Kaspar, F. & Lehner, B. A global hydrological model for deriving water availability
465
indicators: model tuning and validation. Journal of Hydrology 270, 105–134 (2003).
66. Wallin, M. Evasion of CO2 from Streams: Quantifying a Carbon Component of the Aquatic
Conduit in the Boreal Landscape. (Swedish University of Agricultural Sciences, 2011).
67. Atekwana, E. A. & Krishnamurthy, R. V. Seasonal variations of dissolved inorganic carbon
and δ13C of surface waters: application of a modified gas evolution technique. Journal of
470
Hydrology 205, 265–278 (1998).
68. Lerman, A. & Mackenzie, F. T. Carbonate Minerals and the CO2-Carbonic Acid System. in
Encyclopedia of Geochemistry: A Comprehensive Reference Source on the Chemistry of the
Earth (ed. White, W. M.) 206–226 (Springer International Publishing, 2018).
doi:10.1007/978-3-319-39312-4_84.
475
480

23
TABLE 1. RiverCDR risks associated with type of alkaline material used. H indicates high, M
indicates medium, L indicates low, U indicates unknown.
Level of
Knowledge
Introduction
of heavy
metals
Increased
river
turbidity
Deposition
of fines on
riverbed
gravel
habitat
Risk of
rapid
increases
of river pH
Risk of
silicosis
when
inhaled
Calcite /
Dolomite
minerals
H
L
L
L
M
L
Pure
olivine
minerals
L
L
L
L
L
L
Alkali
metal
oxides
L
L
L
L
H
L
Alumino-
Silicate
minerals
L
M
H
H
L
H
Cement
Kiln Dust
L
H
L
L
H
U
485

24
a)
490
b)
Figure 1a) The carbonate silicate cycle12, and b) existing (grey boxes) and new (turquoise box)
CDR strategies that aim to increase alkalinity.

25
495
Figure 2. Inputs and outputs to carbonate alkalinity (CAlk) in rivers66,67. Here HCO3- represents
Calk, PS represents photosynthesis, W represents inputs due to carbonation weathering, R
represents respiration of organic matter, P represents precipitation of carbonate minerals, IC
500
represents all forms of inorganic carbon inputs to the stream, M represents mineralization of
dissolved organic carbon, and F represents CO2 invasion or evasion.

26
505
Figure 3. a) Mechanisms of carbon dioxide drawdown from addition of alkalinity to rivers. b) co-
benefits of addition of carbonate minerals to acidified rivers.

27
510
Figure 4. Thermodynamic predictions of Molal concentration of HCO3- as a function of dolomite
addition, over a range of partial pressures of CO2 (see Methods). The dotted line and annotated
linear equation, y=β1x+β0, represent the stoichiometric ratio of bicarbonate to dolomite (β1),
offset by the initial concentration of bicarbonate (β0). The 2:1 dotted line represents the amount of
inorganic carbon added from the dolomite. For CDR to occur, the dosing and HCO3- response
515
must occupy the space between the two lines.

28
Methods
Thermodynamic modelling of CDR potential in rivers from alkalinity
additions
520
Model solutions representing four neutral to acidic freshwater systems in Nova Scotia (the
Sackville, Roseway, and St Mary’s Rivers, Shortts Lake) and four alkaline systems in Ontario
(Arab, Big Clear, Big Salmon, and Birch Lake) were characterized by measured alkalinity (as
CaCO3), Al, Ca, Cl, Cu, Fe, K, Mn, Na, pH, SO4, and Zn. Alkalinity was below detection in
Roseway River, and so a value of zero was imputed. Each solution was equilibrated with (1)
525
dolomite in varying amounts and (2) a gas phase with a fixed CO2 partial pressure. All simulations
were run using PHREEQC (Version 3) in R with the PHREEQC database57,58. Several other
contributed R packages were used for data cleaning and visualization59–62.
In our simulations, dolomite was added to solutions in equilibrium via the range of pCO2 that can
be found in rivers and streams, from 400 µatm of CO2 (approximately the partial pressure of CO2
530
in the atmosphere) to greater than 12,000 µatm11.
Forecasting global potential of RiverCDR
The HydroSHEDS global hydrographic framework applied in this study includes a digital
elevation dataset with flow direction and flow accumulation grids at a spatial resolution of 15 arc-
535
seconds (approximately 500 m pixel resolution at the equator) that were derived from a digital
elevation dataset at 3 arc-second resolution (approximately 90 m resolution at the equator) using
automated and manual processing steps63,64. The database also includes an estimate of long-term
average “naturalized” runoff and discharge, derived by downscaling coarse resolution (0.5°) runoff
and discharge estimates of the global hydrological model WaterGAP (v2.2 as of 2014)65.
540

29
To generate a baseline scenario of HCO3-, a global map of present-day HCO3-concentrations42 was
resampled to the target spatial resolution of 15 arc-seconds. Smaller data gaps in the original layers
were filled using nearest-neighbor spatial allocation techniques. Next, we calculated the weighted
mean HCO3- concentration using mean annual river runoff for each watershed draining directly
into the ocean. The mass flux of HCO3- at the outlet of the watershed was then calculated using
545
long-term annual discharge from HydroSHEDS multiplied with the weighted mean HCO3-
concentration.
To estimate the global potential if RiverCDR were applied to low alkalinity rivers, we simulated
an increase of HCO3- concentration in low alkalinity areas (defined as less than 0.6 mMol) to a
threshold of 0.6mMol, recalculated the HCO3-export for each watershed, and finally calculated the
550
difference from the baseline scenario. The results at the watershed scale were summarized at
country, continental, and global scales (Methods Figure 1).
Methods Figure 1. Amount of CDR increase (Mt CO2e yr-1) by raising low alkalinity watersheds
(< 0.6 mmol/L) to threshold of 0.06 mmol/L.
555

30
Reporting summary
Code availability
Code and documentation for the PHREEQC code and global HCO3- flux data are publicly available
560
at (https://github.com/bentrueman/river-cdr.).
Author contributions:
Conceptualization: SMS (lead), EAH
Methodology: SMS (lead), EAH, GG, BL
565
Investigation: SMS (lead), EAH
Visualization: SMS (lead), BT (Figure 4), GG (Methods Figure 1).
Funding acquisition: SMS
Project administration: SMS
Supervision: SMS
570
Writing – original draft: SMS
Writing – review & editing: SMS, EAH, BT, GG, KH, BL
Supplementary Data: KH, SMS, EAH
Competing interests: Competing interests of the authors. SMS and EAH and sit on the board and
575
hold equity in a carbon dioxide removal company (CarbonRun). SMS and EAH have filed a patent
related to carbon dioxide removal in rivers.

31
Acknowledgments:
This research was funded by the Haynes Connell Foundation Grant (SMS), Meighen Family Grant
580
(SMS), NSERC Discovery Grant (RGPIN06958-19) (SMS). We acknowledge support from
CarbonRun Carbon Dioxide Removal Ltd. Giancarlo Otani assisted with graphical design.
Correspondence and request for materials should be addressed to S.M.S.
(Shannon.sterling@dal.ca).
585

32
Supplementary Information
Supplementary Data. Database of studies on the impact of river liming on water quality,
ecosystems, and society [see attached excel spreadsheet entitled Supplementary Data].
590
Supplementary Note: illustration of the sensitivity of CAlk to changes in pH commonly found in
rivers. The speciation of inorganic carbon as a function of pH68. To illustrate, we compare two
rivers each receiving 10 mg/L of DIC. A river with a pH of 7.5 holds 9.3 g of the DIC as HCO3-
anion. By contrast, a river with a pH of 4.5 holds only 0.13 mg/L of the DIC as HCO3- anion with
595
the remainder as CO2 and therefore subject to loss through evasion (following Figure 2).