Alkalinity, pH, and pCO2 in the Laurentian Great Lakes: An initial view of seasonal and inter-annual trends

https://doi.org/10.1016/j.jglr.2022.01.005Get rights and content

Abstract

Ongoing human perturbations to the global inorganic carbon cycle can cause various changes in the pH and alkalinity of aquatic systems. Here seasonal and inter-annual trends in these variables were investigated in the five Laurentian Great Lakes using data from the U.S. EPA GLENDA database. These observations, along with temperature, were also used to predict the partial pressure of carbon dioxide in surface water (pCO2). There are strong seasonal differences in pH in all five lakes, with higher pH levels in summer than in spring. All lakes show significantly higher pCO2 values in spring than in summer. Michigan and Ontario show higher alkalinity values in spring; Huron shows lower spring values. Inter-annually, open-lake pH shows the highest values in all lakes around 2010, the time frame of lowest lake water levels, though only lakes Superior and Erie show statistically significant inflection points at that time. Inter-annual alkalinity trends differ considerably from those of pH. Superior’s alkalinity increases until ∼2008 and then begins dropping; Ontario’s alkalinity decreases until ∼2004 and then begins increasing, with the decrease coinciding with the introduction and establishment of Dreissenid mussels. The other lakes show much less clear inter-annual alkalinity trends. For pCO2, inter-annual trends in each lake show either increases from 1992 to 2019 (for Superior, Michigan, and Huron) or shifts from slightly decreasing values to increasing values for the other lakes. The timing of this shift is from 2008 to 2010.

Introduction

Inorganic carbon cycling is involved in key aspects of aquatic ecosystems. The weak acid-base combinations provided by carbonic acid, bicarbonate and carbonate act as the main buffer system controlling pH in most natural waters. The pH of these water systems affects the solubility of compounds such as calcium carbonate, phosphate and biologically important trace metals (Doney et al., 2009). The consumption of inorganic carbon by photosynthesis and its release by respiration link the inorganic buffer equilibria directly to biomass production and organic carbon remineralization (Raymond and Hamilton, 2018). Carbon dioxide deviations from atmospheric equilibrium can occur due to changes in the physical regime (upwelling, internal waves, weather-related atmospheric temperature changes, etc) or as a function of net autotrophy or net hetrotrophy in a given aquatic ecosystem for the time period being observed (Alin and Johnson, 2007, Atilla et al., 2011 Baehr and DeGrandpre, 2004; Tranvik et al., 2009). Carbon dioxide fluxes driven by these differences directly link aquatic systems to the atmosphere, with the aquatic systems thus acting as either source or sink for greenhouse gases (Cole et al., 2009, Tranvik et al., 2018, Tranvik et al., 2009).

Despite the importance of inorganic carbon cycling as a factor in moderating biological activity, affecting climate and, through pH moderation, affecting the solubility of key compounds, little is known concerning inorganic carbon parameters in the Laurentian Great Lakes, which contain ∼18% of earth’s surface freshwater (Waples et al., 2008). Recent literature on these lakes indicates much uncertainty in past inorganic carbon trends, predictions of future lake pH, and estimates of current pCO2 fluxes (Atilla et al., 2011; Karim et al., 2011; Minor et al., 2019, Phillips et al., 2015). The most comprehensive data set for inorganic carbon parameters across the Laurentian Great Lakes is that taken semiannually (generally April and August) by the US EPA through the Great Lakes Water Quality Monitoring Program, with data available from 1986 to 2019 for Lakes Erie, Huron, Michigan, and Ontario and from 1992 to 2019 for Lake Superior. The US EPA visits eight to twenty open-water sites per lake (Barbiero et al., 2018) and measures water quality parameters, including pH, alkalinity and temperature. These data are often investigated directly for trends, such as the increase in alkalinity seen over the past 2.5 decades in Lake Superior (Minor et al., 2019). Implicit in such investigations is that there has not been a strong phenological shift in stratification state and lake surface temperature so that April and August have been measuring the same approximate seasons within the lake over the entire data set. Also implicit within many of these investigations is that the average of April and August measurements can be used to investigate inter-annual trends. Biogeochemical modeling work brings this into question for the Great Lakes, indicating that there is insufficient temporal and analytical resolution within the pH measurements to determine if the lakes are currently exhibiting acidification (Phillips et al., 2015).

Observational data on pCO2 concentrations in the Great Lakes come mainly from calculations of surface water pCO2 using the pH and alkalinity data from the US EPA and an inorganic carbon calculation program such as CO2SYS (Lewis and Wallace, 1998, Pierrot et al., 2006). Fluxes are then estimated using a local or regional measurement of atmospheric pCO2, and a parameterization of the gas transfer coefficient using wind speed as an input (e.g., Atilla et al., 2011, Lin and Guo, 2016a, Lin and Guo, 2016b, Urban and Desai, 2009). In addition to this EPA data, there is a suite of inorganic carbon data (4 to 5 sampling sites per lake, sampled in August 2013) for each of the Laurentian Great Lakes, where pH, dissolved inorganic C and alkalinity were directly measured and pCO2 was calculated from these measurements (Lin and Guo, 2016a, Lin and Guo, 2016b).

At the present time, Lake Superior is the most well studied Laurentian Great Lake in terms of inorganic carbon cycling, with additional measurements of water parameters (pH, alkalinity, and total inorganic carbon) and pCO2 available. Recently collected seasonal water column data from western Lake Superior were compared with inter-annual data from the EPA (Minor et al., 2019) to determine that outgassing occurred more strongly in spring, that alkalinity was increasing inter-annually and that there appeared to be little to no inter-annual trend in pH. Urban and Desai (2009) augmented their evaluation of EPA data from all five lakes with additional water samples from near the Keweenaw Peninsula in Lake Superior and measurement of air pCO2 from shipboard sampling in June 2017 in the same region. These researchers also report higher surface water pCO2 values for all five lake in spring than in summer and conclude from both spring and summer data that the lakes are net sources of carbon dioxide to the atmosphere (Urban and Desai, 2009). Surface water pCO2 variations in Lakes Superior and Michigan have also been investigated using coupled physical and biogeochemical models with the EPA derived data used for comparison with model output in the Lake Superior study (Bennington et al., 2012, Pilcher et al., 2015). Lake Superior was found to exhibit significant variability in pCO2 values on both spatial and seasonal scales (Bennington et al., 2012).

Yet even in Lake Superior, potential inter-annual and seasonal alkalinity and pH trends remain uncertain. Alkalinity over the past few decades has been reported as both relatively stable, based mainly upon data from the EPA and Environment Canada, but also including older historical published data (Chapra et al., 2012), and increasing (Minor et al., 2019), based upon the EPA data and including additional years of measurement. Lake Superior’s inter-annual pH over the same time frame has been reported as stable (Minor et al., 2019) or slightly increasing (Phillips et al., 2015), with the latter study also pointing out the difficulty in discerning clear trends with the existing data sets.

Lake Superior’s role in CO2 cycling is also unclear. Currently, there are only two published studies with directly measured surface-water pCO2 data for Lake Superior (Kelly et al., 2001, McManus et al., 2003). The earlier study does not report sampling locations in the lake and appears to consist of four sampling dates from June to October 1990. The second study is from the far western region of the lake; it consists of directly measured pCO2 from a mooring as well as pCO2 calculated from alkalinity and pH data from water column profiles collected in summer 2001. Kelly et al (2001) find Lake Superior to be a sink for atmospheric CO2 and thus net autotrophic. McManus et al find the lake’s surface water supersaturated with respect to CO2 and thus net heterotrophic, McManus et al further determine that this net heterotrophy is fueled by consumption of dissolved organic carbon.

This study extends detailed analysis of inorganic carbon parameters to all five Laurentian Great Lakes, using the same approach for each lake system. It is based upon EPA data through year 2019, which thus adds 7 to 10 years more data than have currently been explored for alkalinity in Lakes Erie, Ontario, Michigan, and Huron (Chapra et al., 2012). It applies statistical analyses of both seasonal (April vs August) and inter-annual trends in alkalinity, pH and pCO2 (as calculated from temperature, pH and alkalinity). It also examines CO2SYS calculations of pCO2 by over-constraining the inorganic carbon system when concurrent measurements of pCO2, pH, and dissolved inorganic carbon were obtained (n = 10). The over-constrained system is used to determine which Henry’s Law and carbonic acid dissociation constants among the suite available in CO2SYS are most applicable in Lake Superior, the headwater lake of the Laurentian Great Lakes system. While April and August, as snapshots of inorganic carbon parameters in the Great Lakes, do not fully assess overall seasonal variability, multi-decadal data allow a first approximation of possible climate related changes that can be compared with future, more complete data sets of spatial, diel, and seasonal variability in the Great Lakes systems.

Section snippets

Data sets

This study primarily used data from the U.S. Environmental Protection Agency (EPA) within the Great Lakes Environmental Database System (GLENDA). The data set currently is comprised of lake water-chemistry parameters from the Laurentian Great Lakes including pH, alkalinity, and water temperature. Existing data from 1983 to 2019 were used for seasonal analyses as described below. Data from 1992–2019 for Ontario and Michigan and 1993–2019 for Huron and Erie were used for investigating

Comparison of Henry’s law and acid dissociation constants

By calculating pCO2 (ppm) from measured DIC and pH (free scale) in discreet water samples and comparing the resulting calculated value in CO2SYS with measured pCO2 (ppm) collected from the ship’s underway system, the suite of available Henry’s law and acid dissociation constants available in CO2SYS was evaluated (Fig. 1). Surprisingly, the best match between calculated and measured values occurred when using the Cai and Wang constants (Cai and Wang, 1998) rather than using the freshwater

Conclusions

This study explored pH, alkalinity, and pCO2 seasonal and inter-annual trends in the Laurentian Great Lakes using existing data as well as Henry’s law and acid dissociation constants chosen by over-constraining the inorganic carbon system at selected Lake Superior sampling sites. The pH varied significantly in all lakes between spring and summer, with higher pH levels in the summer than in the spring months. Inter-annually, pH appears related to lake water levels, with higher pH values

Declaration of Competing Interest

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

Acknowledgements

Thanks to Prof. Mike DeGrandpre (University of Montana), the AE, and reviewers, whose comments considerably improved this paper. This work was supported by the Swenson Summer Undergraduate Research Program and the University of Minnesota Office of the Vice President for Research, Grant-in-Aid of Research, Artistry & Scholarship Program.

References (45)

  • A. Karim et al.

    Carbon and oxygen dynamics in the Laurentian Great Lakes: Implications for the CO2 flux from terrestrial aquatic systems to the atmosphere

    Chem. Geol.

    (2011)
  • T.J. Lueker et al.

    Ocean pCO2 calculated from dissolved inorganic carbon, alkalinity, and equations for K1 and K2: Validation based on laboratory measurements of CO2 in gas and seawater at equilibrium

    Mar. Chem.

    (2000)
  • F.J. Millero et al.

    Dissociation constants of carbonic acid in seawater as a function of salinity and temperature

    Mar. Chem.

    (2006)
  • F.J. Millero et al.

    Dissociation constants for carbonic acid determined from field measurements

    Deep. Res. Part I Oceanogr. Res. Pap.

    (2002)
  • E.C. Minor et al.

    The effect of a flood pulse on the water column of western Lake Superior, USA

    J. Great Lakes Res.

    (2014)
  • F.J. Mojica Prieto et al.

    The values of pK1 + pK2 for the dissociation of carbonic acid in seawater

    Geochim. Cosmochim. Acta

    (2002)
  • E.D. Reavie et al.

    Phytoplankton trends in the Great Lakes, 2001–2011

    J. Great Lakes Res.

    (2014)
  • R.N. Roy et al.

    to 45°C

    Mar. Chem.

    (1993)
  • S.R. Alin et al.

    Carbon cycling in large lakes of the world: a synthesis of production, burial, and lake-atmosphere exchange estimates

    Global Biogeochem. Cycles

    (2007)
  • N. Atilla et al.

    Observed variability of Lake Superior pCO2

    Limnol. Oceanogr.

    (2011)
  • M.M. Baehr et al.

    In situ pCO2 and O2 measurements in a lake during turnover and stratification: Observations and modeling

    Limnol. Oceanogr.

    (2004)
  • V. Bennington et al.

    Can spatial heterogeneity explain the perceived imbalance in Lake Superior’s carbon budget? A model study

    J. Geophys. Res. Biogeosciences

    (2012)
  • View full text