ABSTRACT

The biogeochemical dynamics of carbon in the ocean is a subject of fundamental interest to environmental studies. In this context, we have implemented a ten year run of the Brazilian Earth System Coupled Ocean-Atmosphere Model (BESM-OA2.3) integrated with TOPAZ biogeochemical model for the Atlantic basin. The modeled ΔpCO₂ for the tropical Atlantic shows very clearly a high dominance of positive fluxes, that is, the CO₂ fluxes are sea-to-air throughout the tropical region and for both winter and summer periods. In the mid-latitudes regions negatives fluxes (air-to-sea) were observed for both seasons. An exception to this pattern is an extensive negative tongue on the latitude 10˚N. The occurrence of this negative ΔpCO₂ tongue region in the Tropical Atlantic is highly correlated to negative Evaporation-Precipitation values during this season. In the northern hemisphere (NH) summer the negative values of ΔpCO₂ in the tropical Atlantic region are concentrated in the adjacent zone of the Amazon river mouth due to the North Equatorial Counter Current intensification. This process favors the formation of a carbon sink in the adjacent region of the Amazon river mouth. Model results show lowest values of dissolved inorganic carbon (DIC) in a surface layer (100 - 150 m). Highest DIC values are observed in deeper layers and concentrated in an equatorial band. The chlorophyll bloom in equatorial zones was well represented by the model. These blooms are the result of equatorial upwelling that brings the high concentration tongues of DIC present in the equatorial band towards the euphotic zone. This is the first published paper about the BESM-OA2.3 integrated with TOPAZ. The presented results suggest that this modeling system is able to reproduce the main regional carbon dynamics features of the mid-latitude/tropical Atlantic.

1. Introduction

The biogeochemical dynamics of carbon in the ocean is a subject of fundamental interest to global environmental changes studies [1]. This interest has been motivated in the last decades mostly for the need to better understand the role of the oceans on the carbon dioxide dynamics, to predict the rate of increase of atmospheric carbon dioxide concentrations and its effects on the global climate [2].

The best estimates indicate that roughly one third of the carbon available in the atmosphere is sequestered by the ocean [1]. However, a great uncertainty still exists in the quantification of carbon concentration and about the seasonal dynamics of the current carbon present in the ocean. In this context, a better quantification of the ocean’s role in the carbon cycle is still one of the major scientific challenges for global change studies [3].

According to [4], the large uncertainty in the estimates of the net oceanic CO₂ balance is associated to a relatively high spatial and temporal variability of natural air-sea exchange of carbon dioxide. Although worldwide efforts for in situ measurements of air and water CO₂ concentrations and air-sea exchange are being made to better assess the oceanic uptake of this greenhouse gas, there is still a lack of a good representation of the spatial and temporal distribution of the carbon dioxide in oceanic waters.

In seawater, CO₂ is soluble and reacts to form bicarbonate and carbonate ions [5]. The sum of bicarbonate, carbonate ions and dissolved carbon dioxide corresponds to the dissolved inorganic carbon—DIC [6]. Due to the peculiar chemistry of seawater and its influence on dissolved salts, more than 99% of the dissolved inorganic carbon is in the form bicarbonate and carbonate ions [7]. The consequence of this conversion of atmospheric carbon dioxide into oceanic bicarbonate and carbonate ions is that the ocean holds fifty times as much carbon as the overlying atmosphere. This inorganic carbon in the ocean is about forty times larger than the amount held as organic carbon [7].

It is important to note that the DIC concentration controls the distribution and abundance of pelagic phytoplankton, net primary production, and the food web dynamics [8]. In this context, the use of this parameter is not restricted to climate studies, but it is also used to estimate rates of net community production, which can be used as an indicator of ecosystem functionality [9,10]. Marine phytoplankton is important for its central role at the base of the food chain that supports the majority of life in the ocean [11].

The Atlantic basin has a relevant ecological contribution for the global biogeochemical cycle, mostly due the high climate oscillations observed in the region, and which occur irregularly on subseasonal to decadal time scales [12,13]. Moreover, this region shows severe storms, atmospheric dust deposition, mesoscale ocean features and extended periods of elevated precipitation or evaporation and warmings [14]. All these characteristics can affect directly or indirectly the CO₂ balance and the concentration of DIC in the water column, with conesquences to the ecological and physical forcings at work in the coupled ocean-atmosphere system [15].

In recent years, better forecasts of oceanic physical variables have been accomplished and operational systems have been established by different groups. These systems typically consist of coupled general circulation models (CGCM), whose results are used as input to force biogeochemical models at basin scale [16]. The establishment of such operational models have largely been made possible thanks to the recent large increase in supercomputing resources [11].

The biogeochemical models face an additional challenge compared to the pure oceanic models. Not only their numerous biological tracers make them computationally more costly, but also the large number of empirical parameters and the scarcity of data available for validation and tuning is a great difficulty [15]. Additionally, it is essential for the biogeochemical models to consider the atmospheric influence on the gases concentration in the ocean, reinforcing the need for coupled oceanic and atmospheric modeling [17].

For the present study we performed a ten-year simulation using the Brazilian Earth System Coupled OceanAtmosphere Model (BESM-OA2.3) [18,19] which runs associated with the TOPAZ ocean biogeochemistry model [21], hereafter referred to as BESM-OA2.3-TOPAZ. The main purpose of this work was to analyse the space/ time variability of DIC and CO₂ balance in the Atlantic basin for the analysed period. The relation between physical and biogeochemical (DIC and carbon fluxes) variables is herein discussed.

2. BESM-OA2.3 and TOPAZ Models

The focus of the work was the implementation of one global experiment using the CGCM Brazilian Earth System Coupled Ocean-Atmosphere Model (BESM-OA2.3) [18,19] integrated with TOPAZ biogeochemical model [21] for a ten-year run (2007-2016). The results presented herein were extracted for the Atlantic basin in a regular grid of the 1˚ × 1˚ for the longitudes 70˚W - 10˚E and latitudes 40˚S - 40˚N. The restart files of the experiment were updated every month. A more in depth description of BESM-OA2.3 can be found in [18].

In general terms, BESM-OA2.3 is a system including an Atmospheric General Circulation Model (AGCM) developed at the Brazilian Center for Weather Forecast and Climate Studies (CPTEC) [19] and the NOAA/ GFDL’s MOM4p1 OGCM. The coupling between these two models is done using the GFDL’s Flexible Modular System—FMS coupler [20]. CPTEC’s AGCM is a spectral model, with triangular truncation at wave number 62, equivalent to approximately 1.875˚ × 1.875˚ lat-lon horizontal resolution, and 28 sigma levels unevenly distributed in the troposphere between the surface and the lower stratosphere.

NOAA/GFDL’s MOM4p1 is a primitive equation model configured over the globe, with 1/4˚ × 1˚ lat-lon in the deep tropics (i.e., between 10˚S - 10˚N), spacing to 1˚ × 1˚ at higher latitudes of both hemispheres, and 50 levels in the vertical. The model uses an explicit free surface formulation and is fully coupled without flux correction each 2 hourly coupled time steps [20]. The AGCM component sends to coupler fresh water (liquid and solid precipitation), specific humidity, heat, and momentum (10 meters zonal and meridional wind) fluxes, surface pressure, and cosine of the zenith angle; the AGCM receives SST and ocean albedo from the CGCM flux coupler.

In our experiment the CGCM was integrated to the ocean biogeochemical and ecological component, namely Tracers of Ocean Phytoplankton with Allometric Zooplankton version 2.0 (TOPAZ2). Modifications described in the following were made in the TOPAZ2 code to calculate the CO₂ balance at the air-sea interface. TOPAZ2 includes 30 tracers to describe the main biogeochemical cycles. A technical description of TOPAZ is available in [21].

3. Carbon Dynamics in Seawater

In aqueous systems, carbon dioxide dissolves forming aqueous carbon dioxide and carbonic acid which dissociates into negatively charged bicarbonate and carbonate ions. The three following simplified equations show these chemical processes:

(1)

(2)

(3)

Since aqueous CO₂ and H₂CO₃ are difficult to distinguish it is customary to refer to their combined dissolved concentration, which we represent here by CO₂ dissolved.

Dissolved carbon dioxide is the result of the chemical reaction that appears in consequence of the carbon dioxide dissolution in seawater, forming bicarbonate and carbonate ions, which collectively are referred to as dissolved inorganic carbon, DIC:

(4)

The square brackets in Equation (4) represent the solute concentrations in seawater normally defined per unit mass in μmol·kg⁻¹.

The air-sea exchange of carbon dioxide depends on the partial pressure, pCO₂, or concentration of CO₂ dissolved in the surface waters, while the global ocean storage of carbon is measured by DIC.

Changes in DIC are buffered with respect to changes in pCO₂. In other words, there are negative feedbacks in the carbonate system which dampens the impact of perturbations on the total dissolved inorganic carbon pool. On the other hand, this ocean feedback amplifies changes in atmospheric pCO₂ in response to changes in ocean DIC [22-24].

The general ratio between dissolved carbon dioxide and the effective partial pressure, pCO₂, is described by the solubility K_o [25], defined as:

(5)

The air-sea exchange of carbon dioxide depends on the partial pressure, pCO₂, or concentration of dissolved carbon dioxide in the surface waters, while the global ocean storage of carbon is measured by DIC. More details about air-sea partial pressure difference are explained in the next sub-topic.

To determine CO₂ balance at the air-sea interface, it is necessary to calculate the difference of the partial pressure of CO₂ (pCO₂) between the sea and atmosphere. To obtain the partial pressure of CO₂ in seawater (pCO₂(sw)) several biogeochemical variables derived by the coupled model were used, namely: dissolved inorganic carbon (DIC), total alkalinity (ALK), salinity (SSS) and sea surface temperature (SST). According to [1], pCO₂(sw) can be calculated by:

(6)

where, according to [25], K_o which represents the solubility of CO₂ in seawater, can be obtained by,

(7)

SST_k represents the absolute sea surface temperature (K). The constants are: a₁ = −58.0931, a₂ = 90.5069, a₃ = 22.2940, b₁ = 0.027766, b₂ = −0.025888 and b₃ = 0.00050578.

K₁ and K₂ terms in Equation (6) represent the dissociation constants of carbonic acid in seawater. These variables were described in [26] and their mathematical formulations are:

(8)

(9)

where the constants are: a₄ = 3670.7, a₅ = −62.008, a₆ = 9.7944, a₇ = 1394.7, a₈ = 4.777, b₄ = −0.0118, b₅ = 0.000116, b₆ = −0.0184 and b₇ = 0.000118.

Monthly atmospheric pCO₂ (pCO₂(atm)) was obtained by [28]:

(10)

where XCO₂ represents the molar fraction of CO₂, obtained from the Mauna Loa (155.6˚W - 19.5˚N) in situ station and available at the World Data Centre for Greenhouse Gases (WDCGG). A monthly mean time series of this variable from 1975 to 2012 is displayed in Figure 1.

The term p in Equation (10) is the sea-level pressure (in atm) [26] and VP_water is the water vapor pressure calculated according to Equation (11) [28].

Using the difference of the partial pressure of CO₂ (pCO₂) between the ocean and atmosphere it was possible to estimate the carbon balance in our study region. Positive values denote sea-to-air CO₂ flux, whereas negative values denote air-to-sea CO₂ flux.

(11)

4. Results and Discussions

The ten-year (2007-1016) mean values of ΔpCO₂ for the

NH winter and summer seasons are presented in Figure 2. In general terms, it is possible to observe the strong tendency in the tropical Atlantic to sea-to-air CO₂ fluxes in both seasons with regional maximum values up to 90 μatm near latitudes ±20˚. However, in mid-latitudes (poleward of ±30˚) this tendency is reversed, that is, the CO₂ fluxes are air-to-sea. Although in the detail, some seasonality is evident, similarly to [29] our results also support the view that throughout the year the tropical Atlantic is a CO₂ source to the atmosphere, while midlatitude regions are a carbon sink. Figure 2 shows that positive fluxes are concentrated in the Tropical South Atlantic in the winter and in the Tropical North Atlantic in the summer. Distributions of ΔpCO₂ calculated by [29] indicate that the mid-latitudes regions are very important for removing the atmospheric CO₂ by the world’s oceans. In our results, the strongest sinks were found in the latitudes above 30 degrees (N-S).

4.1. Evaporation-Precipitation Balance Influences on ΔpCO₂

During the NH winter, the region between the latitudes 30˚S to 34˚N (red colors in left map—Figure 2) supply CO₂ to the atmosphere. An exception to this pattern is the extensive negative tongue (air-to-sea CO₂ fluxes) on the latitude 10˚N. The occurrence of this negative ΔpCO₂ tongue region in the Tropical Atlantic is highly correlated to negative Evaporation-Precipitation (E-P) values during this season in the region [30]. According to [31], the concentration of all chemical species present in seawater occurs in direct proportion to the salinity dilution. The opposite effect occurs if an excess of evaporation over precipitation leads to a net removal of freshwater from the surface layers. For example, if we take a water parcel with an initial pCO₂ of 300 μatm and a salinity of 35, a one unit increase of salinity, the final pCO₂ will be increased in 9 μatm [1]. The pCO₂ change associated with changes in salinity are determined by the influence of salinity on the dissociation constants (see Equations (8) and (9)). Variations in salinity in the surface ocean are mostly driven by changes in the balance between evaporation and precipitation [31,32]. The relation between ΔpCO₂ and E-P can be indirectly confirmed by an analysis of Figure 3, where the minimum values of salinity during NH winter correspond to the ΔpCO₂ negative tongue regions in the tropical Atlantic.

4.2. The Amazon River Influence on ΔpCO₂

In the summer season (Figure 2, right map) the negative values of ΔpCO₂ in the tropical Atlantic region are concentrated in the adjacent zone to the Amazon river mouth. According to [33] this signal is consequence of the meteorological seasonality regimes in the North Brazil Current (NBC), carrying the Amazon plume to different areas of the western tropical North Atlantic (WTNA) as a coherent mass. According to [34], the Amazon plume turns northward outside the river mouth into the North Atlantic. During NH winter, onshore wind stress is approximately normal to the Brazilian coast. The result is a northwesterly displacement of the Amazon plume in relation to the NBC, trapping the river plume closer to the coastline in a long, thin band [35]. During the NH winter, the North Equatorial Counter Current (NECC) is weakest [34], consequently the Amazon waters primarily flow to the Caribbean rather than veering offshore to the WTNA.

In NH summer, east-southeasterly winds carry the plume along the northern Brazilian coastal shelf with the North Brazil Current. The Coriolis acceleration acting on the wind-driven, faster, along-shelf NBC causes the plume to detach from the coast anticyclonically and retroflect northeastward into the WTNA [34]. The NBC retroflection occurs at 6˚N - 8˚N, feeding the NECC’s annual peak flow [36]. Figure 2 (right) shows the NECC’s signal identified by the negative ΔpCO₂. The influence of the Amazon fresh water discharge on the tropical Atlantic is visible in the minimum salinity values on Figure 3 (right).

Tropical Atlantic seawater normally releases CO₂ to the atmosphere because solubility-driven degassing exceeds biological production in lower latitudes [32]. In contrast, the Amazon river mouth region in the summer acts as a carbon sink, removing CO₂ from the atmosphere [37,38]. Previous investigations showed that the Amazon river discharge creates a 5 - 10 m deep plume of lowsalinity, low-inorganic carbon water atop the WTNA waters, forming a sink for atmospheric carbon, and covering approximately two million square kilometers [39, 40].

4.3. DIC Space-Time Variability

The air-sea exchange of carbon dioxide depends on the CO₂ sea-water partial pressure, pCO₂ (sw) or generically, the concentration of CO₂ dissolved in the surface waters. The global ocean storage of carbon is measured by DIC, which corresponds to the sum of dissolved carbon dioxide, bicarbonate and carbonate ions (see Equation (4)).

Surface water pCO₂ (sw) is controlled by SST, SSS, biological uptake of CO₂, remineralization of organic carbon back to CO₂, air-sea CO₂ exchange and mixing with CO₂ rich waters from depth [41].

Figure 4 shows the spatial distributions of DIC in surface waters in the study region for summer and winter. According to these results, the concentration of the dissolved inorganic carbon in the region varies between