The world's largest coastal deoxygenation zone is not anthropogenically driven

Temporal changes between early, peak and late phases of the SM upwelling showed two significant features (figure 1): the gradual intensification of shelf oxygen deficiency from hypoxic during early SM to suboxic/anoxic by late SM and the other, a gradient in it, with the suboxia/anoxia (⩽5 µM) confined to the central shelf between 11º and 18º N, and hypoxia prevailing to the north and south of it. This reflects that almost half of the western Indian shelf in the central region is under acute stress. There are several factors that appear to unrelate the development of this coastal anoxia to the anthropogenic effect. First, the anoxic central zone is located away from receiving significant anthropogenic inputs unlike in the coastal waters off Kochi (10º N) in the south and Mumbai (19º N) in the north, the two largest coastal cities of west India, which receive substantial allochthonous inputs, yet remained at hypoxia during entire SM. The increased anthropogenic activities have not impaired these coastal waters (Balachandran 2001) as evident from their decadal changes. The Kochi estuary has been transformed from an autotrophic to a highly heterotrophic system between 1965 and 2005 (Gupta et al 2009) due to a 4–6 fold rise in anthropogenic nutrients (Martin et al 2010). Yet, its SM coastal hypoxia remains unchanged during this period (Gupta et al 2016) because the fast turnover rates of nitrogen within the flushing times of the estuary (Bhavya et al 2016) limits its export impact to the nearshore regions only (Gupta et al 2016, Bhavya et al 2017, 2018). This is also true with other monsoonal estuaries of west India which are acting as heavy sink zones and export only <10% of anthropogenic nutrients to the coastal sea (Krishna et al 2016).

Second, nutrient concentrations in the lower reaches of seven main estuaries along the west coast during early and late phases of SM showed more than two-fold lower concentrations in the estuaries of the central region, where coastal anoxia is confined, than those in the northern and southern regions (figure 2). Though not measured, the estuarine nutrients during peak SM can contribute to coastal deoxygenation. But the high nearshore surface salinity (upper 4 m) at the central shelf (Mangalore and Goa) during peak SM (34.27–35.36) compared to that of Kochi (24.26) shows weak runoff over the former shelf (figure 3), still, it maintained with nitrate replete (9.5–13 µM) and anoxic conditions (figure 1).

Zoom In Zoom Out Reset image size

Zoom In Zoom Out Reset image size

Third, we examined the data on the stable isotopic composition of particulate organic carbon (POC) and nitrogen (PN) during the SM (figure 3). High δ¹⁵N_PN (15.8 ± 1.6‰) and low δ¹³C_POC (−31.5 ± 1‰) were found in the sewage from a coastal outfall at Mumbai (Sarma et al 2019). These δ¹⁵N_PN are distinctly higher than our measured values at nearshore regions of Kochi (7.7‰–7.8‰) and the rest of the west coast (8‰–13‰) including those reported off Goa earlier (7.77 ± 1.57‰; Maya et al 2011), indicating an absence of anthropogenic influence due to intense mixing and rapid dilution of these waters once enter the coastal sea. Similarly, except nearshore waters of Kochi during late SM (−15.3‰), the δ¹³C_POC from rest of the coast, including the anoxic central zone (−20‰ to −25.5‰) and off Goa (−19.0 ± 0.67‰; Maya et al 2011), largely fell in the range reported for organic matter of marine origin (−18‰ to −24 ‰; Cloern et al 2002, Bouillon et al 2003). Since the nearshore waters of Kochi show high δ¹³C_POC with high salinity during late SM, it precludes the possibility of significant terrestrial influence as it is quite close to the values observed for sea grass or macroalgae (−16‰ to −19‰; Hondula and Pace 2014). Notably, δ¹³C_POC (−21.7‰ to −22.2‰) during the maximum runoff conditions at Kochi nearshore during peak SM (salinity 24.26–28.1) also did not deviate significantly from the typical marine signature, when it maintained with bottom hypoxia (17–31 µM). These are in accordance with the last three decadal regular monitoring of coastal waters around India, which shows that the coastal waters within ∼2 km from the vicinity of population centres, especially mega cities like Kochi and Mumbai, only felt with significant anthropogenic effect (Madeswaran et al 2018) due to their rapid dilution beyond this distance. Thus the observed weak or no anthropogenic signal even at the nearshore stations (8–13 km away from the coast) along the EAS is consistent with no terrestrial nutrients influence beyond ∼15 km from the coast in the Bay of Bengal (Sarma et al 2020b).

Overall the isotopic data in conjunction with salinity during the study period (figure 3) suggests that the nutrient source of organic matter to the anoxic central zone is likely of marine origin. Higher δ¹⁵N_PN in the central region (>9‰) compared to Kochi (7.1‰–7.8‰), despite salinity at the former region was higher than the latter, suggest the isotopically enriched source of nitrogen during the formation of organic matter as these values are higher than typical δ¹⁵N for phytoplankton in marine waters (Kumar et al 2004). The enriched source of nitrogen for these waters could be sewage (15.8 ± 1.6‰; Sarma et al 2019) or denitrified upwelling (9‰–12‰; Rixen et al 2014) or sediment pore water. However, no significant drop in salinity at these locations indicates the limited contribution of freshwater mediated waste supply. Therefore, it is highly likely that the source of nutrients for the formation of organic matter at these locations is marine-derived natural processes. Moreover, even during peak SM, when significant freshwater was received at Kochi nearshore (surface salinity ∼24), no significant increase in δ¹⁵N_PN was observed implying rapid dilution of terrestrial inputs within few kilometres from the coast. Similar is the case with Mangalore where the low salinity front observed between stations 3 and 4 during peak SM shows δ¹⁵N_PN of denitrified marine signals (figure 3).

To further confirm the possible influence of anthropogenic sources on coastal anoxia, paleoceanographic records are examined. Organic carbon isotopic composition and variability in iron content in the sediment core were taken as proxies for the source of carbon and redox conditions in the water column, respectively. The sediment cores from the shallow stations of the central west coast of India, dated to ∼350 years, were found to have a narrowly varied isotopic composition of organic carbon (−23‰ to −21‰) indicating its marine origin (Fernandes et al 2020). The authigenic iron contents in these cores also did not show significant variability between the recent and past centuries. Similarly, almost uniform Fe/Al (0.78 ± 0.04) and Mg/Al (0.32 ± 0.02) ratios in the sediment core in the last ∼700 years infer no considerable change in nature and source of detritus (Agnihotri et al 2008). Dinoflagellate cysts are good indicators of ecosystem eutrophication by anthropogenic factors (Price et al 2018), and their tracking history in the sediment core from the west coast of India also did not indicate any sign of terrigenous derived organic matter (D'Silva et al 2012). Unlike the coastal systems of the Gulf of Mexico, Gulf of St. Lawrence, etc where paleo records show a significant impact of anthropogenically derived eutrophication (Rabalais et al 2007, Gilbert et al 2010, Price et al 2018), the studies on coastal system of west India have not shown such effect. These findings along with our earlier observation of unchanged hypoxic conditions compared to five decades ago over the southwest coast of India corroborate the view that the coastal deoxygenation is supported by the oceanic processes (Gupta et al 2016).

All these observations collectively strengthen the possibility that the anoxic conditions along the central west coast are not primarily related to the anthropogenic effect.

It is known that the upwelling waters of the EAS originate from the Arabian Sea OMZ (Banse 1959), however, it is unknown till now the depth and dissolved oxygen concentrations of source waters for the upwelling. Our results show that the extent of shelf deoxygenation follows the corresponding spatio-temporal changes in the distribution of dissolved oxygen concentrations in the offshore OMZ. To begin with, the upwelling of deoxygenated waters initiated in the south is progressed up to 15° N by early SM, and to the entire west coast by peak SM (figure 1). While doing so, the shelf exhibited significant deoxygenation gradients with anoxia confined to the central shelf between 11° and 18° N. The upper boundary of OMZ (20 µM oxygen), varied between ∼100 m in the south to ∼130 m in the north during early SM, shallows with the progression of the season to a peak to ∼70 m in the central region by late SM (figure 4), and accordingly forms the source for upwelling (figure 5). Further, the perennial OMZ has a southern boundary at ∼12° N (Naqvi 1991), its 0.2 ml l⁻¹ (∼9 µM) dissolved oxygen isoline slopes northward up to 18° N (figure 1). These lead to a condition wherein the upwelling waters over the central shelf are sourced from the core OMZ which are suboxic (⩽10 µM) while those of the south and north are from hypoxic waters (∼20 µM) outside the core OMZ (figure 4). These core OMZ denitrified waters have enriched δ¹⁵N_PN of 9‰–12‰ at 100–150 m (Rixen et al 2014), the upwelling of these waters contributed to higher δ¹⁵N_PN over the central shelf (>9‰), consistent with the values reported earlier (Maya et al 2011), compared to the southern shelf (<8‰) (figure 3).

Zoom In Zoom Out Reset image size

Zoom In Zoom Out Reset image size

Though the upwelling propagates from south to north, its onset in the south is evident at 100 m during January–February, peaks to the surface layers during the SM (June–August), and subsides abruptly thereafter (Gupta et al 2016). However, the present intra-seasonal data shows that while progressing to the north, the peak phase of upwelling over the southern and northern shelves end up at hypoxia while the upwelling over the central shelf is sustained till late SM but with suboxia/anoxia (figure 6). These changes in the coastal waters are also accompanied by similar changes in the offshore waters. The doming of the water column seen at 10° N during early SM shallows and strengthens in size up to 11° N by peak SM, and thereafter shifts to 14° N by late SM (figure 4). This shift, as shown by the weekly averaged satellite data on sea surface height anomalies, is governed by the formation of cold-core eddies in the south and central regions respectively during peak and late SM (figure 7). This leads to a shift in the upliftment of the water column/oxycline from outside core OMZ to within core OMZ during the progression of SM. Consequently, the core OMZ waters of the central region shoal from >100 m to ∼70 m between early and late SM, the corresponding change in oxygen regime in upwelling source waters from hypoxia to suboxia (figure 4) influences the intensification of deoxygenation over the central shelf (figures 5, 6 and 8). Such spatial and temporal shifting of cyclonic eddies, influencing the intensity and spatial spread of upwelling and, in turn, the patterns of distribution of dissolved oxygen over the shelf, as seen from the sea surface height anomalies from 2012 to 2017 (figures S1 and S2 (available online at stacks.iop.org/ERL/16/054009/mmedia)), is probably a recurrent feature.

Zoom In Zoom Out Reset image size

Zoom In Zoom Out Reset image size

Zoom In Zoom Out Reset image size

The poleward west India under-currents at 50–150 m, gain strength between peak and late SM, carry relatively ventilated waters (Naqvi et al 2006) up to 11–12° N (figures 4 and S3). Their progression also restricts the upwelling of suboxic (<10 µM) core OMZ waters to the north of 11° N and governs the southern boundary of shelf anoxia (figures 1 and 5).

The equatorward advection of Arabian Sea High Salinity Watermass (ASHSW >36) with WICC also has a role in maintaining the oxygen gradients between the north and central regions. The oxygenated ASHSW is about 150 m thick in the north and caps the cold (20 °C) and hypoxic (∼20 µM) waters (figure 4). The upwelling of these relatively aerated waters over the shelf north of 19° N maintains its hypoxic conditions (figures 5 and 8). But between 18° and 12° N, the ASHSW is gradually eroded to <70 m, enabling an intense upwelling of suboxic waters on to the central shelf. However, the ASHSW still acts as a barrier for upward movement of anoxic layers along the central west coast. Though coastal anoxia is confined to 11–18° N, north of 15° N (for example, off Goa—figure 8) where the pycnocline is thicker, it is less-spread whereas the shallow pycnocline south of 15° N leads to large parts of the water column becoming anoxic, thereby the severity of anoxic volume is higher at Mangalore than at Goa.

Though physical factors govern the spread of deoxygenated upwelling waters, shelf biogeochemistry and biology also play a crucial role in maintaining the observed gradients. In phase with the decreasing upwelling intensity towards the north, the phytoplankton biomass (column chlorophyll a) between the nearshore and mid-shelf regions during the SM months was higher in the south (77 ± 47 mg m⁻²) compared to the central (56 ± 26 mg m⁻²) and northern (56 ± 23 mg m⁻²) regions. But the corresponding bottom oxygen consumption rates (based on changes in apparent oxygen utilisation), following the decay of plankton, remained relatively higher in the central (1.91 ± 0.8 µM l⁻¹ d⁻¹) than in the southern (1.59 ± 0.6 µM l⁻¹ d⁻¹) and the northern (1.68 ± 0.4 µM l⁻¹ d⁻¹) regions. As the abundance of zooplankton and herbivorous fishes in the suboxic/anoxic (hypoxic) waters of the central (southern) shelf are expected to be lower (higher) (Stramma et al 2011, Gupta et al 2016, Roman et al 2019), the lower grazing loss of phytoplankton in the central shelf results in higher decomposition of sinking organic matter (figure S4) and intensifies oxygen depletion.