Carbon dioxide sources and sinks in the delta of the Paraíba do Sul River (Southeastern Brazil) modulated by carbonate thermodynamics, gas exchange and ecosystem metabolism during estuarine mixing

Highlights

• Biological processes dominated the carbonate system in the freshwaters.

• The mixing zone exhibited permanent CO₂ undersaturation close to the expected conservative mixing.

• Heterotrophy, gas exchange and thermal variability create deviation from conservativity of CO_2.

• The mangrove creek presented high pCO₂ values, and important production of TA and DIC.

• Tropical coastal deltas have different CO₂ system controls compared to temperate and boreal.

Abstract

Tropical estuarine deltas generally present poorly buffered waters in their freshwaters. Carbonate chemistry predicts that mixture of such warm freshwater with seawater will create rapid consumption of the freshwater carbon dioxide (CO₂) by the carbonate buffering capacity of the seawater. In this study, we used the Paraiba do Sul River as a laboratory to investigate how thermodynamics compare with biological processes, gas exchange, and tidal advection from mangrove. We conducted three spatial surveys covering the salinity gradient of the main channel and surrounding mangrove waters and one 24-h mooring in a mangrove creek. In the main channel, dissolved inorganic carbon (DIC) and total alkalinity (TA) showed closely conservative distribution along the salinity gradient, increasing seaward. The partial pressure of CO₂ (pCO₂) followed a bell-shaped curve predicted by carbonate chemistry for conservative mixing of river and ocean endmembers. During high river flow, pCO₂ sharply decreased between salinities 0 and 5 (1800 to 390 ppmv), a pCO₂ drawdown attributed to riverine outgassing and thermodynamics. Indeed, the mixing of TA-poor freshwater (363 ± 16 μmol kg⁻¹) with TA-rich seawater creates a deficit of dissolved CO₂ not related to biotic processes. During low river flow, the entire mixing zone was undersaturated in pCO₂ with an increasing trend seaward. However, observed pCO₂ values were slightly above those predicted by conservativity. Approximately half of this deviation was attributed to biological activity (net heterotrophy), and remaining deviation was assigned to the effects of gas exchange (18%) and water heating (36%). The effect of gas exchange was higher in fresh and low salinity waters, reflecting the higher outgassing/ingassing of CO₂, and lower buffering capacity. Water heating was more important in mid- to high-salinities as a result of diel patterns of solar irradiance. Heterotrophy was slight and not able to outcompete thermodynamics and generate outgassing during estuarine mixing. Consistently, stable isotopic signatures of DIC (δ¹³C-DIC) presented slight deviations below the conservative mixing, corroborating net heterotrophy in the main channel. Areas of CO₂ uptake due to phytoplankton activity were identified but restricted to the freshwater endmember during low river flow, with lowest pCO₂ (up to 41 ppmv) and the highest chlorophyll a (up to 21.3 μg L⁻¹). The estuary was a CO₂ sink during low river flow (−1.34 to −5.26 mmolC m⁻² d⁻¹) and a source during high river flow (5.71 to 19.37 mmolC m⁻² d⁻¹). In the mangrove creek, the pCO₂, DIC, δ¹³C-DIC and TA presented deviations from the conservativity, with slopes between TA and DIC demonstrating organic carbon degradation mediated by aerobic respiration and sulphate reduction. Mangrove creek waters were a CO₂ source (average of 134.81 mmolC m⁻² d⁻¹), exhibiting high values of pCO₂ (up to 21,720 ppmv). The results reveal that the low buffering capacity in the main channel of tropical estuarine deltas can be the predominant driver of pCO₂, generating CO₂ undersaturation along the mixing zone, a process overlooked in estuarine systems. Moreover, air-water CO₂ exchange, thermal variability, and biological activities contribute to deviation of the carbonate system from conservative mixing in specific estuarine areas, also modulating pCO₂ variability.

Introduction

The concentration of carbon dioxide (CO₂) in Earth's atmosphere has been increasing at fast rates, and the current CO₂ level is unprecedented over the past 3 million years (Willeit et al., 2019), changing approximately from 277 ppmv in 1750 (Joos and Spahni, 2008) to 414 ppmv in April 2019 (NOAA, 2019), as a result of fossil fuel burning and land-use changes. According to recent estimates, the oceans and the land absorb approximately 23 and 32% of these total emissions, respectively, whereas the remaining 45% accumulate in the atmosphere (Le Quéré et al., 2018). This CO₂ accumulation in the atmosphere is causing global warming, with CO₂ as the most important anthropogenic-derived greenhouse gas (IPCC, 2013). In addition, penetration of CO₂ in the ocean is causing acidification, increasing the surface ocean partial pressure of carbon dioxide (pCO₂), with a net decline of pH (Gattuso et al., 2015). While carbon sink in the open ocean is currently estimated with a relative high degree of confidence (2.5 ± 0.5 GtC yr⁻¹; Le Quéré et al., 2018), the air-water CO₂ fluxes and controlling processes remain uncertain in the coastal ocean. Indeed, coastal oceans have not been satisfactory included in global carbon budget calculations, despite their importance in terms of global carbon cycling (Chen and Borges, 2009; Cai, 2011; Bauer et al., 2013).

Estuaries are aquatic ecosystems characterized by the mixing of fluvial and marine waters, hosting a large diversity of interfaces and gradients, with distinct physical, geomorphological and biological features (Cai, 2011; Borges and Abril, 2011). Studies have shown that CO₂ emissions by estuarine systems and near-shore coastal waters are globally significant (Chen and Borges, 2009). However, in 10 years of global data compilation, CO₂ emission estimates have decreased by a factor of 6. The first global estimate of estuarine CO₂ emissions calculated a degassing of 0.6 PgC yr⁻¹ (Abril and Borges, 2004), whereas the two lasts estimates showed degassing of approximately 0.1 PgC yr⁻¹ (Chen et al., 2013; Laruelle et al., 2013). This decline has been attributed to various factors, mainly the disproportional and insufficient global sampling (poor coverage in tropical regions and the Southern Hemisphere), few studies with direct and continuous measurements of pCO₂, and low temporal resolution (semi-diurnal, diurnal, seasonal, and annual) (Borges and Abril, 2011; Cai, 2011; Chen et al., 2013).

In general, the upper estuarine regions with low-salinity waters are important sources of CO₂ to the atmosphere. This is normally attributed their net heterotrophy and, to a lesser extent, to inputs of CO₂-enriched freshwater and lateral inputs from intertidal areas such as salt marshes and mangroves (Cai et al., 1999; Cai et al., 2011; Borges and Abril, 2011). However, in some estuaries, the riverine contributions of CO₂ (allochthonous sources) can be greater than contribution of estuarine heterotrophy (autochthonous sources) at certain times of the year (Jiang et al., 2008; Joesoef et al., 2015; Van Dam et al., 2018a). Climatological and hydrological characteristics may change seasonally according to river water discharge, temperature, and formation of vertical stratification, which modulate physical mixing and biological production (Frankignoulle et al., 1998; Salisbury et al., 2008; Borges and Abril, 2011; Dinauer and Mucci, 2017). Downstream in the seaward direction, estuarine mixing zones are considered as moderate CO₂ sources, whereas the marine domain from the river mouth to the shelf break generally behaves as a CO₂ sink (Frankignoulle et al., 1998; Borges and Abril, 2011; Chen et al., 2013). In general, the pCO₂ values and CO₂ emissions are much higher in river-dominated estuaries than in marine-dominated estuaries (Jiang et al., 2008; Cotovicz Jr. et al., 2015). In addition to the marked and complex natural variability, the coastal ocean hosts approximately 37% of the human population (Cohen et al., 1997), creating human-induced modifications in the metabolism of aquatic ecosystems associated with nutrient enrichment and eutrophication (Frankignoulle et al., 1998; Zhai et al., 2007; Cotovicz Jr. et al., 2015; Kubo et al., 2017).

Historically, studies addressing estuaries have identified biological processes as predominant drivers on pCO₂ distribution and CO₂ fluxes along salinity gradients (Frankignoulle et al., 1998; Cai et al., 1999; Borges and Abril, 2011). However, the theory of carbonate chemistry predicts drastic abiotic changes in proportions of dissolved CO₂, bicarbonate (HCO₃⁻), and carbonate (CO₃²⁻) during the mixing of freshwater with seawater. The influence of thermodynamics in the CO₂ system in estuarine waters was first described by the pioneer studies of Mook and Koene (1975) and Whitfield and Turner (1986). These authors showed that modification of the carbonate equilibria in estuarine waters during the mixing of freshwater with seawater results in changes in the pCO₂ and, in turn, in the CO₂ fluxes at the air-water interface (Whitfield and Turner, 1986; Hu and Cai, 2013; Cai et al., 2013). In particular, the displacement of acid-base equilibrium during the mixing of weakly buffered freshwater (low TA concentration) with well buffered seawater (high TA concentration) can generate pCO₂ values far below that of the atmosphere, with no need for biological uptake (Whitfield and Turner, 1986). In this regard, thermodynamic processes would be potentially significant compared with biological processes in oligotrophic/mesotrophic estuaries with short residence times as, for example, estuarine deltas. Thermodynamic changes as the main driver of carbonate chemistry changes along the salinity gradient in estuaries were have been demonstrated in theoretical studies (Hu and Cai, 2013; Salisbury, 2008; Cai et al., 2013), but rarely validated using field data. In the Amazon River plume, pCO₂ undersaturation has been attributed mainly to phytoplankton productivity (Ternon et al., 2000; Körtzinger, 2003); however, recent findings showed that, from a salinity value of approximately 10, surface water becomes undersaturated as a result of the mixing effect (Lefèvre et al., 2017).

The geographical position also exerts influence over CO₂ concentrations and emissions across different latitudes. Estuaries located at low and intermediate latitudes present the largest flux per unit area, whereas those located in regions north of 50° N and south of 50° S present the lowest flux per unit area (Chen et al., 2013). These differences are attributed to local/regional characteristics of the coastal zone, including the mixing of freshwaters (presenting variable levels of pCO₂) with low-pCO₂ shelf waters, water temperature, and biogeochemistry complexity (Chen et al., 2013). However, it should be highlighted that studies on carbon cycling conducted in tropical coastal regions are overlooked compared with those in temperate and boreal regions. In addition, tropical rivers present lower TA concentrations than temperate/boreal rivers (Cai et al., 2008), thus the impact of thermodynamic processes during river-ocean mixing can be especially important in tropical estuaries. Estuarine deltas can locally/regionally be the main estuarine type through which significant terrestrial carbon is transported to the ocean (Laruelle et al., 2013). Deltas are mostly located in tropical and sub-tropical regions (Laruelle et al., 2013).

The present study aims at investigating the carbonate chemistry and air-water CO₂ exchange in the Paraiba do Sul River estuary (PSRE), southeastern Brazil. This estuary is a tropical mesotrophic ecosystem that has been suffering with increasing eutrophication and construction of dams in its watershed. This ecosystem has a short residence time and low TA concentrations in freshwater, potentially generating pCO₂ undersaturation conditions during mixing with seawater as a result of thermodynamic changes - a fact that is still disregarded globally. Based on high resolution monitoring of carbonate chemistry parameters along the entire salinity gradient, we could differentiate between the biotic and abiotic contributions to water pCO₂ distributions and CO₂ fluxes at the air-water interface; we also analyzed the metabolism of the main channel using pCO₂ distributions and stable isotope data.