The decline in oxygen supply to the ocean associated with global warming is
expected to expand oxygen minimum zones (OMZs). This global trend can be
attenuated or amplified by regional processes. In the Arabian Sea, the
world's thickest OMZ is highly vulnerable to changes in the Indian monsoon
wind. Evidence from paleo-records and future climate projections indicates
strong variations of the Indian monsoon wind intensity over climatic
timescales. Yet, the response of the OMZ to these wind changes remains poorly
understood and its amplitude and timescale unexplored. Here, we investigate
the impacts of perturbations in Indian monsoon wind intensity (from
The combination of strong organic matter decomposition and
poor ventilation explains the presence of large oxygen minimum zones (OMZs)
in the intermediate ocean of the eastern tropical Pacific and Atlantic oceans
as well as in the northern Indian Ocean. At low oxygen concentrations,
hypoxia-sensitive marine species are subject to varying environmental
stresses that can affect their growth and reproductive success and ultimately
cause their death
Dissolved O
An example of such perturbations is regional wind changes
In the Arabian Sea, the summer monsoon southwesterly winds drive strong
upwelling off the coasts of Oman and Somalia, giving rise to one of most
productive coastal upwelling ecosystems in the world
While there is still no consensus on the magnitude and the drivers of the
recent and future Indian monsoon wind changes, evidence from paleo-climate
records overwhelmingly suggests a strong link between Northern Hemisphere
temperatures and the Indian monsoon wind intensity on timescales ranging from
decades to thousands of years
Despite these previous studies, the amplitude of the OMZ sensitivity to potential wind changes remains largely uncertain. Indeed, whether in the context of past climate fluctuations or under future climate change, the response of OMZ to upwelling-favorable wind intensification is difficult to predict as such a perturbation may increase both oxygen supply through enhanced ventilation and oxygen demand via increased biological productivity, thus leading to an uncertain net effect. In the Arabian Sea, the picture is made even more complicated by the seasonal reversal of winds and the potential importance of changes in winter monsoon mixing. Additionally, the large denitrification fluxes in the Arabian Sea combined with the potential feedback of denitrification on biological productivity, and hence on oxygen consumption, further add to the intricacy of the problem. Finally, the question of the OMZ response timescales is essential but remains unanswered. Here, we address these questions and explore the mechanisms by which the Arabian Sea ecosystem responds to monsoon wind changes using a regional eddy-resolving model. We examine how idealized changes in summer and winter monsoon wind intensity affect the productivity and the volumes of hypoxic and suboxic water in the Arabian Sea and explore the biogeochemical and ecological implications of these changes. We show that the productivity increases on a timescale of years while the OMZ expands and deepens on a timescale of decades. This response is essentially driven by summer monsoon wind intensification and results from an enhanced biological consumption of oxygen opposed by increased ventilation near the surface. The enhanced upper ocean ventilation leads to intermittent expansions of habitats in the epipelagic zone, while the OMZ intensification at depth increases denitrification, thus amplifying the depletion of bioavailable nitrogen in the Arabian Sea. We conclude that changes in the Indian monsoon can affect the large-scale nitrogen marine budget on decadal to centennial timescales, with a positive feedback on warming in the case of stronger winds.
We use the Regional Ocean Modeling System (ROMS)_AGRIF (documented at
The model domain extends in latitude from 5
The model is forced with a monthly climatology. The absence of interannual
variability in the atmospheric forcing enables us to quantify the role of
internal variability associated with mesoscale eddies. The temperature,
salinity and currents are initialized and laterally forced using the Simple
Ocean Data Assimilation (SODA) ocean reanalysis. Oxygen and nitrate initial
and boundary conditions are derived from Garcia et al. (2010a, b)
dataset. The atmospheric boundary conditions for heat and freshwater fluxes
are based on the Comprehensive Ocean–Atmosphere Data Set (COADS;
The model is first spun up for 12 years and then is run for an additional
50 years using nine different wind stress scenarios. In the control run, the wind
stress is left unperturbed. In a first set of four perturbed monsoon
simulations, the wind stress was uniformly increased or decreased by 20 and
50 %, respectively. This amounts to wind speed perturbations of around 10 and
20 %, respectively. Two additional runs were conducted where the
perturbations of the summer and winter monsoon winds are set to be
antagonistic, i.e., a 50 % increase (respectively, decrease) of the summer monsoon
wind stress that is concomitant with a 50 % decrease (respectively, increase) in the
winter monsoon wind stress. Finally, two simulations with gradual increase
(respectively, decrease) of wind stress at a rate
Although these runs explore different wind perturbation scenarios, they are highly idealized by nature and are not intended to mimic past conditions from paleo-climatic reconstructions or realistic future trajectories but rather aim at exploring the sensitivity of the Arabian Sea OMZ to monsoon wind intensity changes and improving our understanding of the key mechanisms that control the OMZ response and its timescales. For model evaluation, we use the last 10 years of the control run.
We use available satellite and in situ observations to assess the model
ability to reproduce observed physical and biogeochemical properties in the
Arabian Sea domain. To this end, we evaluate the model in terms of surface
currents and eddy kinetic energy (EKE), sea surface height (SSH) anomalies,
sea surface temperature (SST) and surface chlorophyll
The model simulates successfully the surface EKE as it
reproduces quite accurately the spatial distribution of the observed surface
eddy field (Fig.
Surface eddy kinetic energy. Surface eddy kinetic energy (in cm
The model captures the main patterns of the observed SST from the Advanced Very High Resolution Radiometer (AVHRR) satellite data
in winter and summer seasons (Appendix A, Fig.
However, an examination of the north–south vertical distribution of
temperature reveals that the model (i) tends to underestimate the subsurface
water temperature in the northern Arabian Sea particularly in winter and
(ii) slightly overestimates the mixed layer depth in both seasons (Appendix A,
Fig.
The model successfully simulates the high chlorophyll concentrations in the
northern and western Arabian Sea associated with the winter and summer blooms,
respectively (Fig.
Surface distribution of chlorophyll
We further compare simulated primary production and export fluxes with
estimates based on available field data from the US Joint Global Ocean Flux Study (JGOFS) Arabian Sea Process
Study
The simulated surface distribution of nitrate is consistent with observations
from the World Ocean Atlas dataset in both seasons (Fig.
Horizontal and vertical distributions of oxygen. Distribution of annual-mean oxygen (in mmol m
A more quantitative evaluation of model skill is performed using in situ
observations from the World Ocean Database (2013) that we binned into a
0.5
In summary, despite some local biases, the model generally shows reasonable skill in reproducing the large-scale features of the circulation in the Arabian Sea as well as the seasonal dynamics of phytoplankton blooms in the region. More importantly, it reproduces fairly well the location and structure of the Arabian Sea oxygen minimum zone. The discussion of the potential impact of the model limitations on our results will be addressed in detail in the discussion section.
Taylor diagram displaying statistical comparison of modeled and observed fields. Taylor (2001) diagram of simulated
To explore the sensitivity of the Arabian Sea ecosystem to changes in the
intensity of monsoon winds, we consider various scenarios of idealized wind
perturbations. A first set of scenarios consists in increasing (respectively,
decreasing) the wind stress over the whole domain and throughout the year by
20 and 50 %, respectively. In a second set of experiments, we increase the
wind stress by 50 % in summer (respectively, winter) and decrease it by 50 % in
winter (respectively, summer). This is to account for perturbation scenarios where
summer and winter monsoon winds evolve in opposite directions as suggested by
multiple paleo-records
The magnitude of the Arabian Sea net primary production (NPP) response is
proportional to the amplitude of the perturbation
(Fig.
Biogeochemical response to monsoon wind intensity changes. Relative changes in response to monsoon wind intensity perturbations in net primary production (green), denitrification (blue), suboxic volume (red) and hypoxic volume (orange). Open circles (respectively, squares) indicate the results from the simulation where summer monsoon wind stress is increased (respectively, decreased) by 50 % and winter monsoon wind stress is decreased (respectively, increased) by 50 %.
Under monsoon wind intensification (respectively, weakening), the OMZ (defined here
as hypoxic water with O
Response to monsoon wind increase as a function of time. Response to wind stress increase (
As denitrification develops only under suboxic conditions, its response to
monsoon changes is modulated by the response of the suboxic volume.
Similarly, it shows a slow response (on a timescale of several decades)
characterized by a strong internal variability
(Fig.
Under increased monsoon winds, hypoxia increases at depth but is reduced in
the upper ocean (Fig.
Changes in the Arabian Sea OMZ and denitrification as a function of depth.
In order to elucidate the factors driving the OMZ changes, we perform an
oxygen budget analysis in the OMZ volume. We first take into account the
whole water column and then carry out the analysis separately in three
layers: the epipelagic zone (0–200 m), the mesopelagic zone (200–1000 m) and
the bathypelagic zone (
This analysis shows that when considering the whole OMZ, the oxygen
accumulation is generally negative (corresponding to a net loss of O
OMZ response in different vertical layers as a function of time. Changes in hypoxic volume (in %) under wind
stress increase (
The oxygen budget in the upper (0–200 m) OMZ reveals that the O
Drivers of OMZ expansion under monsoon intensification. Annual O
The oxygen changes induced by monsoon wind perturbation have important ecological and biogeochemical implications. In particular, changes that affect the upper OMZ can have direct effects on marine habitats and may impact the ecosystem community structure. On the other hand, the changes in the OMZ intensity have the potential – via denitrification – to alter the marine nitrogen budget, and hence the efficiency of the biological pump of carbon and climate, on the longer timescales.
A 50 % increase in the monsoon wind stress leads to almost a doubling of the
surface eddy kinetic energy in the central and western Arabian Sea
(Fig.
Estimated biological productivity and nitrogen uptake in
Eddy kinetic energy as a function monsoon wind intensity.
Latitudinally averaged eddy kinetic energy (in cm
Denitrification increases by around 72 % in response to a 50 % increase in
wind stress. This strong increase of denitrification amplifies fixed nitrogen
removal. When integrated over the Arabian Sea domain, this represents an
additional loss of fixed nitrogen of around 9.5 Tmol N decade
The increase in the Arabian Sea denitrification and nitrification should also
lead to an increase in the N
In conclusion, reduced large-scale productivity under Indian monsoon (and
denitrification) intensification could reduce the efficiency of the
biological pump of carbon and together with enhanced N
Our study reveals a strong link between the strength of the monsoon and the
Arabian Sea productivity and denitrification. This validates the assumption
made in several paleo-climate studies that rapid variations in productivity
and denitrification in this region are tightly coupled to monsoon
fluctuations. Our study also confirms the potential for the Arabian Sea OMZ
to strongly impact the marine nitrogen budget at a larger scale in response
to Indian monsoon fluctuations, in agreement with conclusions of previous
paleo-studies. Besides monsoon strength, some studies have linked past OMZ
intensity changes to changes in the rate of formation and subduction of
oxygen enriched Subantarctic Mode Water (SAMW) and Antarctic Intermediate
Water (AAIW) in the Southern Ocean in association with Atlantic meridional
overturning circulation (AMOC) fluctuations
Oxygen variability as a function of monsoon wind intensity.
Our model represents explicitly a large fraction of ocean eddies thanks to
its relatively high-resolution of
Simulated response to monsoon wind changes as a function of model resolution. Responses of
The limitations of the model are among the study's main limitations. For
instance, our simulations are based on a simple biogeochemical model based on
nitrogen only with no representation of other limiting nutrients such as
iron, silicate and phosphate. We think this can lead to some local biases in
regions where other nutrients can limit productivity (e.g., off the Somali
coasts). However, at larger scales, the limitation by nitrogen has been shown
to dominate over other nutrient limitations in the Indian Ocean, and hence we
assume that this choice should not affect the main findings and conclusions
of the study
Another caveat of the study is the lack of a representation of the nitrogen
fixation in the model. Although denitrification is thought to largely
dominate over N
The primary focus of this study is the sensitivity of the Arabian Sea OMZ to
monsoon wind changes and its response timescale. This justifies the use of
highly idealized wind perturbations, as our simulations are not intended to
mimic realistic past or future changes but rather to deepen our understanding
of the key mechanisms at work and their potential implications. However, we
are aware that future and past monsoon changes generally come in complex
spatial and temporal patterns, which may affect the projected OMZ response.
For instance, a recent study by
Finally, we considered here the effects of monsoon changes in isolation. The response of the OMZ to such perturbations may however change when considered in combination with other potential perturbations such as surface warming or large-scale ventilation changes. This needs to be further investigated in a dedicated study.
A set of coupled physical biogeochemical simulations of the Arabian Sea ecosystem reveals a tight coupling between the intensity of the summer monsoon wind and the size and intensity of the Arabian Sea OMZ.
We find that the OMZ and ecosystem responses are largely determined by the perturbation of the summer SW monsoon, whereas the winter NE monsoon changes play a comparatively smaller role. We show that the intensification of monsoon winds strongly increases the ecosystem productivity, thereby amplifying the oxygen biological consumption and intensifying the OMZ at depth. Concurrently, increased monsoon winds also enhance the transport of oxygen to the OMZ. These opposing effects will lead in the upper ocean to a weakening of the OMZ as the supply of oxygen through enhanced ventilation exceeds the oxygen depletion resulting from increased remineralization there. In contrast, the OMZ intensifies and expands at depth as the increased biological consumption of oxygen overcompensates the effect of enhanced ventilation below the thermocline.
Our simulations indicate that the productivity responds to monsoon wind changes on a timescale of years, while the OMZ responds on a much longer timescale (i.e., several decades). This reflects the difference in the ocean circulation adjustment timescales between the surface and the intermediate ocean. The enhanced ventilation favors episodic injections of oxic waters in the lower epipelagic zone (100–200 m) of the western and central Arabian Sea, leading to intermittent expansions of habitats and a more frequent alternation of hypoxic and oxic conditions there.
The increased productivity and deepening of the OMZ also lead to a strong intensification of denitrification at depth, resulting in a substantial amplification of fixed nitrogen depletion in the Arabian Sea. We conclude that changes in the Indian monsoon can affect, on longer timescales, the large-scale biogeochemical cycles of nitrogen and carbon, with a positive feedback on climate change in the case of stronger winds.
While it has been suggested that OMZs may expand in the future due to
increased stratification (causing reduced ventilation), we show here that the
Arabian Sea OMZ can also expand as a consequence of increased upwelling
causing increased productivity and increased O
The model code can be accessed online at
The Simple Ocean Data Assimilation (SODA) reanalysis data can be downloaded from
Surface circulation. Surface circulation as simulated in the model
Sea surface height anomalies. Sea surface height seasonal anomalies (in meters) simulated
in ROMS
Sea surface temperature and salinity.
Vertical distributions of temperature and salinity in the upper ocean.
Biological productivity at mooring stations M1 to M5 on a transect extending 1500 km from the coast of Oman.
Export fluxes at mooring stations M1 to M5. Annual-mean export flux at the M1–M5 stations as estimated
from in situ observations (black) and simulated in the model (red) at 100
Surface distribution of nitrate. Surface distribution of nitrate (in mmol m
Response to monsoon wind weakening.
Drivers of OMZ contraction under monsoon weakening. Annual O
Meridional transport in O
OMZ response to gradual vs. abrupt wind changes. Changes in hypoxic volume (in %) under gradually increasing
or decreasing winds (thin lines) and instantly increased or decreased winds (thick lines). The smooth wind perturbation
corresponds to an increase or decrease of wind stress at a rate of 1 % yr
ZL conceived the study, performed the experiment and the analysis and wrote the manuscript. ML and SS contributed to the design of the study and participated in the interpretation of the results and the writing of the manuscript.
The authors declare that they have no conflict of interest.
Support for this research has come from the Center for Prototype Climate Modeling (CPCM) at New York University Abu Dhabi (NYUAD). This research was carried out on the high-performance computing (HPC) resources at NYUAD. We thank Benoit Marchand, Muataz Al Barwani and the whole NYUAD HPC team for technical support. We are thankful to Nicolas Gruber for allowing access to the biogeochemical model code. Edited by: Tina Treude Reviewed by: two anonymous referees