Confidential manuscript submitted to
Geophysical Research Letters
1
P/Ca in Carbonates as a Proxy for
A
lkalinity
and Phosphate Levels
Miquela Ingalls
1
,
†
,
Clara
L.
Bl
ä
ttler
2
,
John
A.
Higgins
3
,
John
S.
Magyar
1
,
John
M.
Eiler
1
,
and
Woodward
W.
Fischer
1
1
Division of Geological and Planetary Sciences, California Institute of Technology, Pasadena,
CA 91125 USA
.
2
Department of the Geophysical Sciences, The University of Chicago, Chicago,
IL 60637 USA
.
3
Department of Geosciences, Princeton University, Princeto
n, NJ 08544 USA
.
Correspond
ing
author:
Miquela Ingalls
(ingalls@psu.edu)
†
Current address: Department of Geosciences, The Pennsylvania State University, University
Park, PA 16802, USA
.
Key Points:
•
Calcium
-
to
-
alkalinity ratios (Ca:ALK)
in water
determine rates and mechanisms of
carbonate
salt
formation
•
Shoreline tufa
carbonate facies
is a
useful
indicator
of low
Ca:ALK
aqueous
chemistry
in
the
rock
record
•
Calcium isotope proxy for Ca:ALK
in these mat
erials
is complicated by rate and
mineralogical impacts on
carbonate
-
water fractionation
•
Carbonate P/Ca provides a
nother
means to
assess
Ca:ALK
via the
phosphate content of
ancient aqueous environments
ESSOAr | https://doi.org/10.1002/essoar.10503082.1 | Non-exclusive | First posted online: Sat, 16 May 2020 04:53:45 | This content has not been peer reviewed.
Confidential manuscript submitted to
Geophysical Research Letters
1
Abstract
1
Understanding
mechanisms, rates, and drivers of
past
carbonate formation
provides
insight into
2
the chemical evolution of Earth’s oceans and atmosphere.
W
e
pair
ed
geological observations
with
3
element
al
and isotope geochemistry to
test potential
prox
ies
for
calcium
-
to
-
alkalinity ratios
4
(Ca:ALK)
.
Across diverse
carbonate facies
from
Pleistocene closed
-
basin lakes
in
Owens Valley,
5
CA
,
w
e
observed
less
d
44/40
Ca
variation
than theoretically predicted
(
>
0.75‰)
for
the
very low
6
Ca:ALK
in these systems
. Carbonate clumped isotope
disequilibria implied rapid carbonate
7
growth
—
kinetic isotope effects
,
combined with
the
diverse carbonate mi
nerals
present,
8
complicate
d
the
interpretation
of
d
44/40
Ca
as a
paleo
-
alkalinity
proxy
.
In contrast,
we observed that
9
the
h
igh
phosphate
concentrations
are
recorded by
shoreline and lake
bottom carbonates
formed
10
in
eleven
Pleistocene lakes
at
orders of magnitude greater
concentrations
than
in
marine
11
carbonates
.
Because the
maximum
phosphate content of water
depends on
Ca:ALK, we propose
12
that
carbonate
P/Ca
can
inform
phosphate
levels
and
thereby
Ca:ALK
of
aqueous
environments in
13
the
carbonate
record.
14
15
Plain Language Summary
16
Carbonate minerals record information about the
local
environments in which they form
,
for
17
example along the margins of
a lake or
on the
sea
bed
, as well as the
aqueous and atmospheric
18
chemistry at the time of
mineralization
(e.g. pCO
2
, pH
).
This
information
is recorded
in both the
19
chemical signatures and textures carbonate rocks acquire
during
precipitation
.
Formation of
20
calcium carbonate requires both a so
urce of calcium (Ca
2+
) and carbonate (CO
3
2
-
, or, carbonate
21
alkalinity [ALK]). The ratio of calcium to alkalinity (Ca:ALK) in lake or ocean
water
influenc
es
22
carbonate saturation state,
the rate of carbonate formation, the textures
that
carbonate will
devel
op
,
23
and the chemical signatures recorded by the carbonate mineral.
In this study, we test
ed
several
24
approaches to
identify very low Ca:ALK chemistry in modern and ancient lakes. We
found
that
25
water with extremely low levels of calcium and high levels of
alkalinity form
s
carbonates with a
26
characteristic “tufa” texture
in stratigraphic positions tied to riverine and groundwater sources of
27
Ca
2+
. Moreover
,
the
phosphate concentrations
in these rocks were
orders of magnitude higher than
28
carbonates that precipi
tate under the high Ca:ALK conditions of modern oceans.
Together the
29
results illustrated that
identification of
carbonate
tufa
textures in the rock record and phosphate
30
measurements of carbonate rocks
can
be use
d to study
the ancient
environ
ments and mechanisms
31
by which carbonate rocks formed.
32
1 Introduction
33
Geochemical signatures recorded by carbonate minerals are used to reconstruct
ancient
34
environmental and atmospheric conditions. However, the rates and mechanisms of carbonate
35
formation,
which are set by the geochemistry
(and often biology)
of the depositional environment,
36
impact th
o
se
materials
.
It is often challenging to determine the processes
and
environments
37
responsible for carbonate precipitation in
the geological record
—
particularly in strata devoid of
38
fossils, as
occurs
in Precambrian sedimentary basins
.
39
Broadly
,
carbonate
formation is determined by
a
balance of
cations (e.g.
Ca
2+
or Mg
2+
)
and
40
the
carbonate
an
ion (CO
3
2
-
), which scales with carbonate
alkalinity (ALK
=
H
CO
3
-
+
2CO
3
2
-
)
in
41
systems where
pCO
2
is largely fixed and
carbonate is the primary buffer
,
e.g. seawater and most
42
lakes.
C
arbonate formation
typically
occurs
between
two
end
-
member regimes
:
high alkalinity in
43
ESSOAr | https://doi.org/10.1002/essoar.10503082.1 | Non-exclusive | First posted online: Sat, 16 May 2020 04:53:45 | This content has not been peer reviewed.
Confidential manuscript submitted to
Geophysical Research Letters
2
excess of Ca
2+
(
ALK>>Ca,
as
in
alkaline
lakes),
and vice versa
(modern marine)
.
Processes that
44
source or consume
cation
s
(herein generalized as
Ca
2+
)
and
alkalinity set the
carbonate saturation
45
state
of a solution
(
W
=
[Ca
2+
][
ALK
]
/
Ksp*
,
where
K
sp
*
is the solubility product of
CaCO
3
at a given
46
temperature
,
pressure
,
and composition
)
,
and
determine whether
carbonate
is likely to precipitate
47
(
W
>
1
) or dissolve (
W
<
1
)
.
The
continuum of
calcium
-
to
-
alkalinity ratios (
Ca:ALK
)
that
produce
48
carbonate phases
in natural environments
var
ies
b
y orders of magnitude
.
Where a given
49
environment sits on that continuum impacts how
carbonates get made.
50
M
icrobial metabolism
s
can contribute to carbonate formation and dissolution through the
51
production
and consumption
of alkalinity
and
dissolved inorganic carbon
(
DIC
;
Vasconcelos &
52
McKenzie, 1997
;
Folk & Chafetz, 2000)
.
Photosynthesis and a
erobic
respiration
increase
and
53
decrease
W
,
respectively,
by consuming
and producing
DIC
(
∆
DIC)
with
little to
no change
to
54
ALK
(
∆
ALK;
Bergmann et al., 2013
; Higgins et al., 2009)
.
The
effect
of these processes
on
W
,
55
however,
depend
s
on the Ca:ALK of the environment.
In
high Ca:ALK
environments (
modern
56
marine
Ca:ALK=4.1)
,
W
depends
most
on
the production and consumption of
DIC by microbial
57
carbon cycling because Ca
2+
is not limiting
.
In contrast,
Ca
2+
-
starved
, hyperalkaline
environments
,
58
like
Mono Lake, California
(Ca:ALK=10
-
4
)
, are
less
sensitive to localized
∆
DIC because
∆
ALK
59
is
relatively
insignificant
compared
to total ALK.
Carbonate
production
from
very
low Ca:ALK
60
water is
controlled
by
processes that impact
Ca
2+
supply and cycling. For example, nearly all
61
carbonate
formed in
Mono Lake and other
alkaline,
closed
-
basin Pleistocene lakes can be
62
described
as shoreline
-
proximal
facies (e.g. tufa towers and shoreline crusts)
wherein carbonate
63
production is
tied to
external Ca
2+
input
from rivers and
emergent
groundwater
s
.
In these settings
,
64
distal
lake bottom sediments
tend to be comparatively
carbonate
-
poor
.
T
he shoreline
-
associated
65
carbonate facies
characteristic of low Ca:ALK systems
can be
useful
geological
indicators of
this
66
carbonate production
regime
, as long as they are preserved with sufficient
stratigraphic and
67
sedimentologic
detail
.
However,
diagnostic shoreline facies
are rarely exposed in sedimentary
68
basins
.
Due to the incompleteness of the record,
we aimed
to develop com
plementary geochemical
69
proxies
for Ca:ALK conditions
that
could
be
paired with field
observations to assess
paleo
-
70
alkalini
ty
in ancient carbonate
-
producing systems
.
71
72
1.1
Geochemical proxies for
Ca:ALK in ancient environments
73
Based on a study of Ca isotope distillation in hyperalkaline Mono Lake
(
Nielsen &
74
DePaolo 2013)
,
Blättler
and
Higgins
(
2014)
presented
a
framework in which
Ca isotopes in
75
evaporite salts
could
be
used
to reconstruct relative alkalinity
.
They found
a notably large
Ca
76
isotope
range (
∆
d
44/40
Ca>
1.5
‰
)
preserved within CaSO
4
salts in
one
evaporite system
. From these
77
measurements and experiments, they
derived a relationship between
∆
d
44/40
Ca of
evaporit
e
78
minerals and cation
-
to
-
anion ratio of the aqueous solution to reconstruct Ca
2+
:SO
4
2
-
ratios in
79
ancient evaporite sequences.
By analogy to the Ca
-
SO
4
system and gypsum precipitation,
Blättler
80
et al.
(
2017)
applied the same
Ca
isotope
principles
to
evaluate how seawater carbonate chemistry
81
might have changed over Earth history.
Theoretically, w
hen ALK
greatly exceeds
calcium
82
concentration (Ca:ALK<
0.75),
like
in
Mono Lake,
distillation
driven by net evaporation
should
83
yield a
∆
d
44/40
Ca >
0.75
‰
within measurements from one evaporative carbonate sequence
.
T
he
84
carbonate record
, however,
has additional
mechanics that
complicate
∆
d
44/40
Ca
as a paleo
-
85
alkalinity proxy.
First
,
evaporitic
environments often precipitate multiple
c
arbonate
mineral
s
(e.g.
86
aragonite
,
vaterite,
calcite,
dolomite)
, and sometimes admixtures within co
-
occu
r
ring textures
,
87
each with unique
Ca isotope fractionation factor
s
(
e.g.
Gussone et al., 2005)
.
Second
, mechanisms
88
and rates
of carbonate precipitation are controlled by chemistry, hydrology,
and
biology
in
aqueous
89
ESSOAr | https://doi.org/10.1002/essoar.10503082.1 | Non-exclusive | First posted online: Sat, 16 May 2020 04:53:45 | This content has not been peer reviewed.
Confidential manuscript submitted to
Geophysical Research Letters
3
systems. The
rate
of carbonate precipitation
can
change
based on the relative contributions of these
90
factors
, thereby driving variable expression of kinetic isotope effects
(KIEs)
in
calcium
and other
91
isotope systems in the resultant carbonate record
.
92
There is a widely appreciated relationship be
tween dissolved phosphate concentrations and
93
DIC in modern alkaline lakes (e.g.
Toner & Catling, 2020)
. In typical marine environments,
94
phosphate abundance (
≤
1
μ
M
in surface waters) is set by biologic
al consumption, organic matter
95
decomposition, sedimentary burial and release of P scavenged by ferric iron oxides, and
96
formation/dissolution of fluorapatite
(Ruttenberg & Berner, 1993)
. With evolving atmospheric and
97
oceanic redox chemistry and evolution of microfauna, marine P concentrations are tho
ught to have
98
fluctuated throughout Earth history, potentially to higher concentrations due to limited P
99
scavenging under lower pO
2
conditions (e.g. Precambrian) or when expanses of deep ocean were
100
anoxic (e.g. Cretaceous).
However,
phosphate
concentration
s
can
reach
100 mM in
closed
-
basin
101
lakes
in
quasistatic
hydrological equilibrium (
inflow
approximately equal to
evaporation
).
102
P
hosphate is typically a limiting nutrient due to biological uptake
. However
,
the phosphate
-
103
concentrating mechanisms
in alkaline
lakewater greatly outpace biological consumption
even in
104
the
extremely productive Mono Lake
(10
9
-
10
10
cells
·
liter
-
1
[
Humayoun et al., 2003
]
versus
surface
105
seawater
, ~10
8
cells
·
liter
-
1
[
Whitman et al., 1998
]
)
. This is in part due to feedbacks on productivity
106
by fixed nitrogen species which are often limiting in these systems
(Herb
st, 1998; Jellison &
107
Melack, 2001)
.
Further,
in low Ca:ALK systems,
Ca
2+
is titrated by
carbonate mineral
ization
108
before low
-
solubility phosphate salts (e.g. apatite
, vivianite
)
precipitate
,
which allows
phosphate
109
to accumulate
in solution.
110
In this
stud
y
, we
hypothesized
that
phosphate
should
incorporate
more
-
or
-
less
proportional
111
to its aqueous concentration
in
the
carbonate phases
precipitated from Mono, Searles, and other
112
Pleistocene closed
-
basin
lakes
. We further proposed
that
carbonate P/Ca
could be used
to
113
determine
phosphate levels
and
relative
alkalinity
in the carbonate rock record
.
114
2
Geologic setting, S
ample
s and Methods
115
2.1
Field geology,
facies descriptions
,
and mineralogy
116
Pleistocene Owens Valley, California,
hosted
a series of glacial lakes interconnected by
117
the Owens River
and
fed by str
eam
s
off of the Sierra Nevada range (Fig. 1). Today, these lakes
118
exist
either
as
restricted
basins (e.g. Mono Lake) or
as
predominantly dry lake beds with seasonal
119
wetting and drying (e.g. Searles Lake and Death Valley).
The Searles dry lake bed
and
the re
lict
120
shoreline facies
of both Searles
Lake
and
Pleistocene
Lake Russell (
proto
-
Mono
Lake
)
are
useful
121
testing grounds for
paleo
-
alkalinity proxies because
they comprise
geologically young
carbonate
s
122
ESSOAr | https://doi.org/10.1002/essoar.10503082.1 | Non-exclusive | First posted online: Sat, 16 May 2020 04:53:45 | This content has not been peer reviewed.
Confidential manuscript submitted to
Geophysical Research Letters
4
that
have
not yet been
impacted by
potential
geochemical
complications
for which
carbonate
rock
s
123
are notoriously sensitive
(
i.e. burial
-
related heating or
recrystallization
)
.
124
Figure 1.
Interconnected alkaline lakes of Owens Valley,
125
California
, adapted from
Smith
(
1979)
. The pale and dark
126
teal shading marks the paleo
-
and modern shorelines,
127
respectively. Line dashing denotes the water chemistry
128
of the Sierran drainages that feed Owens River.
Sampling
129
sites
are
marked with
stars.
130
131
132
133
134
135
Three field
expeditions
to Mono Lake
and Searles Valley were conducted during 2019
136
(Fig
s
.
S
1
, S2
; Table
S
1)
,
during which diverse carbonate facies were collect
e
d
(Fig. S3)
and a
137
stratigraphic section of the Pleistocene Wilson Creek Formation was measured (Fig.
S4
).
During
138
summer 2018, a one
-
m
eter sediment
gravity
core was collected
in collaboration with
the
139
International
Geobiology Course
from Mono Lake
at Station 6 (USGS), in the deepest part of the
140
lake
(
4
2.5
m)
.
This core
contains sediment deposited over
the past
~
150 y
ears
,
and
was
subsampled
141
every 5
cm.
142
Carbonate mineralogy was determined by x
-
ray diffraction (XRD) at the California
143
Institute of Technology
(
Text S1
)
. An aragonite
-
calcite calibration curve (Fig.
S
5
) was used to
144
quantify
samples
of mixed mineralogy.
Peak areas were
integrated at 2
q
of 25.6
°
to 26.7° for
145
aragonite and 28.7
°
to 30.0° for calcite.
146
2.1
Calcium isotopes
147
Approximately
5 mg of bulk
carbonate
were processed for calcium isotope analysis
148
following
the
methods
of
Blättler et al.
(
2018)
.
Carbonate minerals were dissolved in
0.1N buffered
149
acetic acid
,
and
centrifuged to separate insoluble material.
The
resulting s
upernatant
was diluted
150
to approximately 30 ppm
Ca
2+
,
and purified by
ion chromatography
using
an automated Thermo
-
151
Dionex ICS
-
5000+ with a fraction collector
.
The purified calcium solutions were
di
luted
in 2%
152
nitric acid to a concentration of 2 ppm calcium for mass spectrometry
(Text S2)
. Isotopic ratios of
153
calcium were analyzed at Princeton University with a Thermo Neptune Plus multi
-
collector
154
inductively coupled plasma mass spectrometer (ICP
-
MS) w
ith an ESI Apex
-
IR sample
155
introduction system.
Corrections for mass interferences and standardization procedures are
156
described in Text S2.
d
44/42
Ca
values
were calculated
relative to modern seawater
by s
ample
-
157
standard bracketing
. External precision on
d
44/
40
Ca values is
±
0.14
‰
(2
s
), derived from the long
-
158
40km
north
Sierra Nevadas
Sierra Nevada drainage
inflow water types:
Na - HCO
3
-SO
4
Cl - SO
4
Ca - Cl
White-Inyo Mtns
Owens River
Nevada
California
Present day playa or lake
Pleistocene lake
Volcanic areas
Pleistocene river
This study
Long
Valley
Caldera
Big Pine
Volcanic
Field
Lake Russell/
Mono Lake
Lake Manly
DeathValley
Panamint
Lake
Searles Lake
China
Lake
119°W
117°W
37°N
Owens
Lake
Station 6
[42.5m]
[Fig.S1]
[Fig.S2]
ESSOAr | https://doi.org/10.1002/essoar.10503082.1 | Non-exclusive | First posted online: Sat, 16 May 2020 04:53:45 | This content has not been peer reviewed.