Figure 6. Seasonal fluctuations in sea surface temperature (SST in blue), instantaneous wind speed (U10n, ms-1 in grey), F CO2 density (pink), F CH4 density (green) and F N2O density (yellow) across the locations of Kunene, Walvis Bay, and Lüderitz. The coastal stations are positioned to the right of the plot. Note different scales on y-axis.
Lüderitz in the inner shelf, in contrast, shows less pronounced seasonal variations in F CO2 densities. The positive F CO2 densities imply a small yet consistent annual source of CO2 to the atmosphere (F CO2(Lüderitz-winter ): ± σM: 3.00 ± 0.12 mmol m-2d-1 and F CO2(Lüderitz-summer ): ± σM: 3.00 ± 0.17) (Figure 6).
The Kunene and Walvis Bay regions both exhibited distinct seasonality in sea surface CH4 concentrations and corresponding fluxes, whereas in Lüderitz, the seasonal fluctuations in sea surface CH4 concentrations and their corresponding fluxes are less pronounced, with flux densities consistently remaining low throughout the year (Figures 6 and S2). During our campaigns, we determined the maximal sea surface CH4 concentration at Walvis Bay during summer campaign with an average ofsummer ± σM: 72.45 ± 5.63 nmol L−1 and as high as 700 nmol L−1 in the inner shelf compared to Kunene and Lüderitz (summer-Kunene ± σM: 3.30 ± 0.02 nmol L−1 andsummer-Lüderitz ± σM: 3.17 ± 0.06 nmol L−1). Also, the inner shelf of Walvis Bay showed to be a source of CH4 to the atmosphere throughout the year with the maximum daily net F CH4; ± σM: 1319.34 ± 65.00 µmol m-2 d-1 and ± σM: 593.37 ± 22.12 µmol m-2d-1 in winter and summer, respectively (e.g. van der Plas et al., 2007; Sabbaghzadeh et al., 2021). The wider shelf with an extensive organic-rich diatomaceous mud-belt might fuel seasonal cycling of CH4 in the Walvis Bay region and with high wind speeds during winter may contribute to elevates sea-air fluxes (Calvert & Price, 1983; Mollenhauer et al., 2007; van der Plas et al., 2007; Sabbaghzadeh et al., 2021). The substantial F CH4density recorded during our surveys is not surprising since CH4 flux densities as high as 3000.00 μmol m−2 d−1 have been recorded previously in Walvis Bay in association with large sedimentary fluxes (Brüchert, et al., 2009; Sabbaghzadeh et al., 2021). Morgan et al. (2019) also reported the maximum F CH4 density of 3145.00 μmol m−2 d−1 with an average of 450.00 μmol m−2 d−1 in Walvis Bay – Lüderitz sections during upwelling events.
In general, the concentrations and flux densities of N2O demonstrate a closer mirroring relationship with CO2compared to CH4 (Figures 6 and S2). Sea surface N2O concentrations ranged between 5.00 and 35.00 nmol L-1, slightly exceeding the maximum of 31.00 nmol L-1 reported by Arévalo Martínez et al. (2019) in the region. However, these concentrations fall within the range of 5.00 and 51.00 nmol L-1 as reported for upwelling regions in the Eastern South Atlantic (Bange et al., 2009; Kock and Bange, 2015).
The computed flux densities consistently demonstrate the region as a perennial source of N2O to the atmosphere (Figure 6). Off Kunene, flux densities were significantly higher in winter compared to summer, with the computed flux being 200 times greater in winter relative to the summer campaign (F winter ± σM: 151.42 ± 1.85 µmol m-2d-1 and F summer ± σM: 0.60 ± 0.05 µmol m-2d-1). The daily average wind speed (U10n in m s-1) also demonstrated a marked difference, with winter ± σM: 12.34 ± 0.08 m s-1 in winter compared to summer ± σM: 1.22 ± 0.03 m s-1 in summer during our campaigns. The Kunene region has also been previously documented as exhibiting maximal sea surface N2O concentrations during the winter months, attributed to the presence of cold water parcels (13 – 16 °C) and seasonal upwelling (e.g. Arévalo Martínez et al., 2019). The shelf topography, incorporating a steep slope in the Kunene area, may facilitate upward advection, initiating rapid diffusive mixing into the surface ocean and establishing the Kunene as hotspots of gas efflux (Sabbaghzadeh et al., 2021). Frame et al. (2014) also reported a maximum F N2O density of 43.40 µmol m-2d-1 between November and December in the event of upwelling near Kunene. However, it is important to note that their data was collected from only one station within the Angola-Benguela frontal zone at coordinates 17.5°S and 11.25°E. Notably, the F N2O density reported by Frame et al. (2014) was considerably lower than the F N2O density observed across the entire inner shelf of Kunene during the course of this study in the winter campaign. Unlike this study, Frame et al. (2014) did not include stations within the Kunene cell. Moreover, in contrast to this study, the majority of their stations were positioned beyond the 250 km bathymetric contour. As a result, they missed capturing the inner shelf variability emphasized by our dataset.
Off Walvis Bay, too, displays a comparable trend in flux densities between winter and summer seasons with flux being 100 times higher in winter compared to summer (F winter ± σM: 23.20 ± 0.90 µmol m-2d-1and F summer ± σM: 0.29 ± 0.01µmol m-2d-1), serving as a source of N2O emissions to the atmosphere, with a potentially greater contribution during the winter relative to summer. Continuous observations from a ground-based atmospheric observatory for greenhouse gases along the coastal shelf, spanning approximately from 22°S to 28°S, have documented an average F N2O density of 52.70 µmol m-2 d-1. This flux density is associated with specific upwelling events observed at coastal stations located between Walvis Bay and Lüderitz (Morgan et al., 2019). These findings are in good alignment with the recorded F- N2O density at coastal stations in Walvis Bay (F ± σM: 48.20 ± 2.01 µmol m-2d-1) during our winter campaign.
Lüderitz on the other hand, may serve as a less significant source of N2O emissions to the atmosphere compared to the other regions in the winter (F winter ± σM: 4.43 ± 0.08 µmol m-2d-1). Reduced N2O concentrations and a less pronounced gradient in the vicinity of the Lüderitz cell during the winter months relative to Walvis Bay have already been reported. This observed reduction aligns with several contributing factors, including diminished primary productivity and constrained organic matter supply to deeper waters (Arévalo Martínez et al., 2018). Our data further supports this notion by revealing relatively lower N2O concentrations in the Lüderitz region during the winter ( ± σM: 11.93 ± 0.05 nmol L-1), in contrast to the Kunene and Walvis Bay regions (Figure S2). Furthermore, the surfacing of low N2O waters were identified in Lüderitz, compared to Kunene and Walvis Bay during our winter campaign (Figures 3, 4 and 5). However, F N2O was considerably higher during summer campaign in Lüderitz rather than Kunene and Walvis Bay (F summer ± σM: 8.65 ± 0.35 µmol m-2 d-1). The daily average wind speed in the inner shelf of Lüderitz during the summer was also higher than in the other two regions (Kunene ± σM: 1.30 ± 0.07 m s-1,Walvis Bay ± σM: 5.70 ± 0.03 m s-1 and Lüderitz ± σM: 6.00 ± 0.06 m s-1) which may explain the higher recorded flux densities in summer in Lüderitz. However, it is notable that, in general, during our campaigns, we found that flux gradients were primarily associated with concentration fields than SST and wind speed (Figures 6 and S2).
Collectively, the positive sea surface gas levels and flux densities are generally observed across our studied areas indicate that the regions are sources of gas emissions throughout the year. However, the specific timing and magnitudes of these emissions exhibit variability, underscoring the intricate and context-dependent nature of gas dynamics in coastal environments.
We observed a noteworthy phenomenon during our winter campaign, noting depleted sea surface gas concentrations in the nearshore core of upwelled waters. This led to a subsequent decline in flux densities at multiple stations near the coasts of Kunene and Walvis Bay (see Figures 6 and S2). Importantly, despite no significant difference in mean temperature (p > 0.05) compared to adjacent stations located within 30 km offshore, the observed reductions were evident. These reductions are suggested to be linked to the upwelling of strongly hypoxic waters near the shelf. With respect to N2O, these waters are characterized by N2O-depleted conditions, a result of N2O consumption under denitrifying conditions. However, a comprehensive exploration of this phenomenon extends beyond the scope of our current paper and necessitates more extensive investigation.