Search Results
You are looking at 1 - 10 of 16 items for
- Author or Editor: Florian Sévellec x
- Refine by Access: All Content x
Abstract
It is typically assumed that the meridional density gradient in the North Atlantic is well and positively correlated with the Atlantic meridional overturning circulation (AMOC). In numerical “water-hosing” experiments, for example, imposing an anomalous freshwater flux in the Northern Hemisphere leads to a slowdown of the AMOC. However, on planetary scale, the first-order dynamics are linked to the geostrophic balance, relating the north–south pressure gradient to the zonal circulation. In this study, these two approaches are reconciled. At steady state and under geostrophic dynamics, an analytical expression is derived to relate the zonal and meridional pressure gradient. This solution is only valid where the meridional density gradient length scale is shorter than Earth’s curvature length scale, that is, north of 35°N. This theoretical expression links the north–south density gradient to the AMOC and can be used as a closure for zonally averaged ocean models. Assumptions and shortcomings of the approach are presented. Implications of these results for paleoclimate problems such as AMOC collapse and asymmetry in the meridional overturning circulation of the Atlantic and of the Pacific are discussed.
Abstract
It is typically assumed that the meridional density gradient in the North Atlantic is well and positively correlated with the Atlantic meridional overturning circulation (AMOC). In numerical “water-hosing” experiments, for example, imposing an anomalous freshwater flux in the Northern Hemisphere leads to a slowdown of the AMOC. However, on planetary scale, the first-order dynamics are linked to the geostrophic balance, relating the north–south pressure gradient to the zonal circulation. In this study, these two approaches are reconciled. At steady state and under geostrophic dynamics, an analytical expression is derived to relate the zonal and meridional pressure gradient. This solution is only valid where the meridional density gradient length scale is shorter than Earth’s curvature length scale, that is, north of 35°N. This theoretical expression links the north–south density gradient to the AMOC and can be used as a closure for zonally averaged ocean models. Assumptions and shortcomings of the approach are presented. Implications of these results for paleoclimate problems such as AMOC collapse and asymmetry in the meridional overturning circulation of the Atlantic and of the Pacific are discussed.
Abstract
A weakly damped mode of variability, corresponding to the oceanic signature of the Atlantic multidecadal oscillation (AMO) was found through the linear stability analysis of a realistic ocean general circulation model. A simple two-level model was proposed to rationalize both its period and damping rate. This model is extended here to three levels to investigate how the mode can draw energy from the mean flow, as found in various ocean and coupled models. A linear stability analysis in this three-level model shows that the positive growth rate of the oscillatory mode depends on the zonally averaged isopycnal slope. This mode corresponds to a westward propagation of density anomalies in the pycnocline, typical of large-scale baroclinic Rossby waves. The most unstable mode corresponds to the largest scale one (at least for low isopycnal slope). The mode can be described in four phases composing a full oscillation cycle: 1) basin-scale warming of the North Atlantic (AMO positive phase), 2) decrease in upper-ocean poleward transport [hence a reduction of the Atlantic meridional overturning circulation (AMOC)], 3) basin-scale cooling (negative AMO), and 4) AMOC intensification. A criterion is developed to test, in oceanic datasets or numerical models, whether this multidecadal oscillation is an unstable oceanic internal mode of variability or if it is stable and externally forced. Consistent with the classical theory of baroclinic instability, this criterion depends on the vertical structure of the mode. If the upper pycnocline signature is in advance of the deeper pycnocline signature with respect to the westward propagation, the mode is unstable and could be described as an oceanic internal mode of variability.
Abstract
A weakly damped mode of variability, corresponding to the oceanic signature of the Atlantic multidecadal oscillation (AMO) was found through the linear stability analysis of a realistic ocean general circulation model. A simple two-level model was proposed to rationalize both its period and damping rate. This model is extended here to three levels to investigate how the mode can draw energy from the mean flow, as found in various ocean and coupled models. A linear stability analysis in this three-level model shows that the positive growth rate of the oscillatory mode depends on the zonally averaged isopycnal slope. This mode corresponds to a westward propagation of density anomalies in the pycnocline, typical of large-scale baroclinic Rossby waves. The most unstable mode corresponds to the largest scale one (at least for low isopycnal slope). The mode can be described in four phases composing a full oscillation cycle: 1) basin-scale warming of the North Atlantic (AMO positive phase), 2) decrease in upper-ocean poleward transport [hence a reduction of the Atlantic meridional overturning circulation (AMOC)], 3) basin-scale cooling (negative AMO), and 4) AMOC intensification. A criterion is developed to test, in oceanic datasets or numerical models, whether this multidecadal oscillation is an unstable oceanic internal mode of variability or if it is stable and externally forced. Consistent with the classical theory of baroclinic instability, this criterion depends on the vertical structure of the mode. If the upper pycnocline signature is in advance of the deeper pycnocline signature with respect to the westward propagation, the mode is unstable and could be described as an oceanic internal mode of variability.
Abstract
Ocean general circulation models (GCMs), as part of comprehensive climate models, are extensively used for experimental decadal climate prediction. Understanding the limits of decadal ocean predictability is critical for making progress in these efforts. However, when forced with observed fields at the surface, ocean models develop biases in temperature and salinity. Here, the authors ask two complementary questions related to both decadal prediction and model bias: 1) Can the bias be temporarily reduced and the prediction improved by perturbing the initial conditions? 2) How fast will such initial perturbations grow? To answer these questions, the authors use a realistic ocean GCM and compute temperature and salinity perturbations that reduce the model bias most efficiently during a given time interval. The authors find that to reduce this bias, especially pronounced in the upper ocean above 1000 m, initial perturbations should be imposed in the deep ocean (specifically, in the Southern Ocean). Over 14 yr, a 0.1-K perturbation in the deep ocean can induce a temperature anomaly of several kelvins in the upper ocean, partially reducing the bias. A corollary of these results is that small initialization errors in the deep ocean can produce large errors in the upper-ocean temperature on decadal time scales, which can be interpreted as a decadal predictability barrier associated with ocean dynamics. To study the mechanisms of the perturbation growth, the authors formulate an idealized model describing temperature anomalies in the Southern Ocean. The results indicate that the strong mean meridional temperature gradient in this region enhances the sensitivity of the upper ocean to deep-ocean perturbations through nonnormal dynamics generating pronounced stationary-wave patterns.
Abstract
Ocean general circulation models (GCMs), as part of comprehensive climate models, are extensively used for experimental decadal climate prediction. Understanding the limits of decadal ocean predictability is critical for making progress in these efforts. However, when forced with observed fields at the surface, ocean models develop biases in temperature and salinity. Here, the authors ask two complementary questions related to both decadal prediction and model bias: 1) Can the bias be temporarily reduced and the prediction improved by perturbing the initial conditions? 2) How fast will such initial perturbations grow? To answer these questions, the authors use a realistic ocean GCM and compute temperature and salinity perturbations that reduce the model bias most efficiently during a given time interval. The authors find that to reduce this bias, especially pronounced in the upper ocean above 1000 m, initial perturbations should be imposed in the deep ocean (specifically, in the Southern Ocean). Over 14 yr, a 0.1-K perturbation in the deep ocean can induce a temperature anomaly of several kelvins in the upper ocean, partially reducing the bias. A corollary of these results is that small initialization errors in the deep ocean can produce large errors in the upper-ocean temperature on decadal time scales, which can be interpreted as a decadal predictability barrier associated with ocean dynamics. To study the mechanisms of the perturbation growth, the authors formulate an idealized model describing temperature anomalies in the Southern Ocean. The results indicate that the strong mean meridional temperature gradient in this region enhances the sensitivity of the upper ocean to deep-ocean perturbations through nonnormal dynamics generating pronounced stationary-wave patterns.
Abstract
Variations in the strength of the Atlantic meridional overturning circulation (AMOC) are a major potential source of decadal and longer climate variability in the Atlantic. This study analyzes continuous integrations of tangent linear and adjoint versions of an ocean general circulation model [Océan Parallélisé (OPA)] and rigorously shows the existence of a weakly damped oscillatory eigenmode of the AMOC centered in the North Atlantic Ocean and controlled solely by linearized ocean dynamics. In this particular GCM, the mode period is roughly 24 years, its e-folding decay time scale is 40 years, and it is the least-damped oscillatory mode in the system. Its mechanism is related to the westward propagation of large-scale temperature anomalies in the northern Atlantic in the latitudinal band between 30° and 60°N. The westward propagation results from a competition among mean eastward zonal advection, equivalent anomalous westward advection caused by the mean meridional temperature gradient, and westward propagation typical of long baroclinic Rossby waves. The zonal structure of temperature anomalies alternates between a dipole (corresponding to an anomalous AMOC) and anomalies of one sign (yielding no changes in the AMOC). Further, it is shown that the system is nonnormal, which implies that the structure of the least-damped eigenmode of the tangent linear model is different from that of the adjoint model. The “adjoint” mode describes the sensitivity of the system (i.e., it gives the most efficient patterns for exciting the leading eigenmode). An idealized model is formulated to highlight the role of the background meridional temperature gradient in the North Atlantic for the mode mechanism and the system nonnormality.
Abstract
Variations in the strength of the Atlantic meridional overturning circulation (AMOC) are a major potential source of decadal and longer climate variability in the Atlantic. This study analyzes continuous integrations of tangent linear and adjoint versions of an ocean general circulation model [Océan Parallélisé (OPA)] and rigorously shows the existence of a weakly damped oscillatory eigenmode of the AMOC centered in the North Atlantic Ocean and controlled solely by linearized ocean dynamics. In this particular GCM, the mode period is roughly 24 years, its e-folding decay time scale is 40 years, and it is the least-damped oscillatory mode in the system. Its mechanism is related to the westward propagation of large-scale temperature anomalies in the northern Atlantic in the latitudinal band between 30° and 60°N. The westward propagation results from a competition among mean eastward zonal advection, equivalent anomalous westward advection caused by the mean meridional temperature gradient, and westward propagation typical of long baroclinic Rossby waves. The zonal structure of temperature anomalies alternates between a dipole (corresponding to an anomalous AMOC) and anomalies of one sign (yielding no changes in the AMOC). Further, it is shown that the system is nonnormal, which implies that the structure of the least-damped eigenmode of the tangent linear model is different from that of the adjoint model. The “adjoint” mode describes the sensitivity of the system (i.e., it gives the most efficient patterns for exciting the leading eigenmode). An idealized model is formulated to highlight the role of the background meridional temperature gradient in the North Atlantic for the mode mechanism and the system nonnormality.
Abstract
This study investigates the excitation of decadal variability and predictability of the ocean climate state in the North Atlantic. Specifically, initial linear optimal perturbations (LOPs) in temperature and salinity that vary with depth, longitude, and latitude are computed, and the maximum impact on the ocean of these perturbations is evaluated in a realistic ocean general circulation model. The computations of the LOPs involve a maximization procedure based on Lagrange multipliers in a nonautonomous context. To assess the impact of these perturbations four different measures of the North Atlantic Ocean state are used: meridional volume and heat transports (MVT and MHT) and spatially averaged sea surface temperature (SST) and ocean heat content (OHC). It is shown that these metrics are dramatically different with regard to predictability. Whereas OHC and SST can be efficiently modified only by basin-scale anomalies, MVT and MHT are also strongly affected by smaller-scale perturbations. This suggests that instantaneous or even annual-mean values of MVT and MHT are less predictable than SST and OHC. Only when averaged over several decades do the former two metrics have predictability comparable to the latter two, which highlights the need for long-term observations of the Atlantic meridional overturning circulation in order to accumulate climatically relevant data. This study also suggests that initial errors in ocean temperature of a few millikelvins, encompassing both the upper and deep ocean, can lead to ~0.1-K errors in the predictions of North Atlantic sea surface temperature on interannual time scales. This transient error growth peaks for SST and OHC after about 6 and 10 years, respectively, implying a potential predictability barrier.
Abstract
This study investigates the excitation of decadal variability and predictability of the ocean climate state in the North Atlantic. Specifically, initial linear optimal perturbations (LOPs) in temperature and salinity that vary with depth, longitude, and latitude are computed, and the maximum impact on the ocean of these perturbations is evaluated in a realistic ocean general circulation model. The computations of the LOPs involve a maximization procedure based on Lagrange multipliers in a nonautonomous context. To assess the impact of these perturbations four different measures of the North Atlantic Ocean state are used: meridional volume and heat transports (MVT and MHT) and spatially averaged sea surface temperature (SST) and ocean heat content (OHC). It is shown that these metrics are dramatically different with regard to predictability. Whereas OHC and SST can be efficiently modified only by basin-scale anomalies, MVT and MHT are also strongly affected by smaller-scale perturbations. This suggests that instantaneous or even annual-mean values of MVT and MHT are less predictable than SST and OHC. Only when averaged over several decades do the former two metrics have predictability comparable to the latter two, which highlights the need for long-term observations of the Atlantic meridional overturning circulation in order to accumulate climatically relevant data. This study also suggests that initial errors in ocean temperature of a few millikelvins, encompassing both the upper and deep ocean, can lead to ~0.1-K errors in the predictions of North Atlantic sea surface temperature on interannual time scales. This transient error growth peaks for SST and OHC after about 6 and 10 years, respectively, implying a potential predictability barrier.
Abstract
A salient feature of paleorecords of the last glacial interval in the North Atlantic is pronounced millennial variability, commonly known as Dansgaard–Oeschger events. It is believed that these events are related to variations in the Atlantic meridional overturning circulation and heat transport. Here, the authors formulate a new low-order model, based on the Howard–Malkus loop representation of ocean circulation, capable of reproducing millennial variability and its chaotic dynamics realistically. It is shown that even in this chaotic model changes in the state of the meridional overturning circulation are predictable. Accordingly, the authors define two predictive indices which give accurate predictions for the time the circulation should remain in the on phase and then stay in the subsequent off phase. These indices depend mainly on ocean stratification and describe the linear growth of small perturbations in the system. Thus, monitoring particular indices of the ocean state could help predict a potential shutdown of the overturning circulation.
Abstract
A salient feature of paleorecords of the last glacial interval in the North Atlantic is pronounced millennial variability, commonly known as Dansgaard–Oeschger events. It is believed that these events are related to variations in the Atlantic meridional overturning circulation and heat transport. Here, the authors formulate a new low-order model, based on the Howard–Malkus loop representation of ocean circulation, capable of reproducing millennial variability and its chaotic dynamics realistically. It is shown that even in this chaotic model changes in the state of the meridional overturning circulation are predictable. Accordingly, the authors define two predictive indices which give accurate predictions for the time the circulation should remain in the on phase and then stay in the subsequent off phase. These indices depend mainly on ocean stratification and describe the linear growth of small perturbations in the system. Thus, monitoring particular indices of the ocean state could help predict a potential shutdown of the overturning circulation.
Abstract
We explore the mechanisms by which Arctic sea ice decline affects the Atlantic meridional overturning circulation (AMOC) in a suite of numerical experiments perturbing the Arctic sea ice radiative budget within a fully coupled climate model. The imposed perturbations act to increase the amount of heat available to melt ice, leading to a rapid Arctic sea ice retreat within 5 years after the perturbations are activated. In response, the AMOC gradually weakens over the next ~100 years. The AMOC changes can be explained by the accumulation in the Arctic and subsequent downstream propagation to the North Atlantic of buoyancy anomalies controlled by temperature and salinity. Initially, during the first decade or so, the Arctic sea ice loss results in anomalous positive heat and salinity fluxes in the subpolar North Atlantic, inducing positive temperature and salinity anomalies over the regions of oceanic deep convection. At first, these anomalies largely compensate one another, leading to a minimal change in upper ocean density and deep convection in the North Atlantic. Over the following years, however, more anomalous warm water accumulates in the Arctic and spreads to the North Atlantic. At the same time, freshwater that accumulates from seasonal sea ice melting over most of the upper Arctic Ocean also spreads southward, reaching as far as south of Iceland. These warm and fresh anomalies reduce upper ocean density and suppress oceanic deep convection. The thermal and haline contributions to these buoyancy anomalies, and therefore to the AMOC slowdown during this period, are found to have similar magnitudes. We also find that the related changes in horizontal wind-driven circulation could potentially push freshwater away from the deep convection areas and hence strengthen the AMOC, but this effect is overwhelmed by mean advection.
Abstract
We explore the mechanisms by which Arctic sea ice decline affects the Atlantic meridional overturning circulation (AMOC) in a suite of numerical experiments perturbing the Arctic sea ice radiative budget within a fully coupled climate model. The imposed perturbations act to increase the amount of heat available to melt ice, leading to a rapid Arctic sea ice retreat within 5 years after the perturbations are activated. In response, the AMOC gradually weakens over the next ~100 years. The AMOC changes can be explained by the accumulation in the Arctic and subsequent downstream propagation to the North Atlantic of buoyancy anomalies controlled by temperature and salinity. Initially, during the first decade or so, the Arctic sea ice loss results in anomalous positive heat and salinity fluxes in the subpolar North Atlantic, inducing positive temperature and salinity anomalies over the regions of oceanic deep convection. At first, these anomalies largely compensate one another, leading to a minimal change in upper ocean density and deep convection in the North Atlantic. Over the following years, however, more anomalous warm water accumulates in the Arctic and spreads to the North Atlantic. At the same time, freshwater that accumulates from seasonal sea ice melting over most of the upper Arctic Ocean also spreads southward, reaching as far as south of Iceland. These warm and fresh anomalies reduce upper ocean density and suppress oceanic deep convection. The thermal and haline contributions to these buoyancy anomalies, and therefore to the AMOC slowdown during this period, are found to have similar magnitudes. We also find that the related changes in horizontal wind-driven circulation could potentially push freshwater away from the deep convection areas and hence strengthen the AMOC, but this effect is overwhelmed by mean advection.
Abstract
At low-resolution, idealized ocean circulation models forced by prescribed differential surface heat fluxes show spontaneous multidecadal variability depending critically on eddy diffusivity coefficients. The existence of this critical threshold in the range of observational estimates legitimates some doubt on the relevance of such intrinsic oscillations in the real ocean. Through a series of numerical simulations with increasing resolution up to eddy-resolving ones (10 km) and various diapycnal diffusivities, this multidecadal variability proves a generic ubiquitous feature, at least in model versions with a flat bottom. The mean circulation largely changes in the process of refining the horizontal grid (along with the associated implicit viscosity and diffusivity), and the spatial structure of the variability is largely modified, but there is no clear influence of the resolution on the main oscillation period. The interdecadal variability appears even more robust to low vertical diffusivity and overturning when mesoscale eddies are resolved. The mechanism previously proposed for these oscillations, involving westward-propagating baroclinically unstable Rossby waves in the subpolar region and its feedback on the mean circulation, appears unaffected by mesoscale turbulence and is simply displaced following the polar front.
Abstract
At low-resolution, idealized ocean circulation models forced by prescribed differential surface heat fluxes show spontaneous multidecadal variability depending critically on eddy diffusivity coefficients. The existence of this critical threshold in the range of observational estimates legitimates some doubt on the relevance of such intrinsic oscillations in the real ocean. Through a series of numerical simulations with increasing resolution up to eddy-resolving ones (10 km) and various diapycnal diffusivities, this multidecadal variability proves a generic ubiquitous feature, at least in model versions with a flat bottom. The mean circulation largely changes in the process of refining the horizontal grid (along with the associated implicit viscosity and diffusivity), and the spatial structure of the variability is largely modified, but there is no clear influence of the resolution on the main oscillation period. The interdecadal variability appears even more robust to low vertical diffusivity and overturning when mesoscale eddies are resolved. The mechanism previously proposed for these oscillations, involving westward-propagating baroclinically unstable Rossby waves in the subpolar region and its feedback on the mean circulation, appears unaffected by mesoscale turbulence and is simply displaced following the polar front.
Abstract
Optimal surface salinity perturbations influencing the meridional overturning circulation maximum are exhibited and interpreted on a stable steady state of a 2D latitude–depth ocean thermohaline circulation model. Despite the stability of the steady state, the nonnormality of the dynamics is able to create some transient growth and variability through stimulation by optimal perturbations. Two different measures are compared to obtain the optimum—one associated with the departure from steady state in terms of density, and the other with the overturning circulation intensity. It is found that such optimal analysis is measure dependent; hence, the latter measure is chosen for studying the following physical mechanisms. The response to the optimal initial sea surface salinity perturbation involves a transient growth mechanism leading to a maximum modification of the circulation intensity after 67 yr; the amplification is linked to the most weakly damped linear eigenmode, oscillating on a 150-yr period. Optimal constant surface salinity flux perturbations are also obtained, and confirm that a decrease in the freshwater flux amplitude enhances the circulation intensity. At last, looking for the optimal stochastic surface salinity flux perturbation, it is established that the variance of the circulation intensity is controlled by the weakly damped 150-yr oscillation. Two approaches are tested to consider extending such studies in more realistic 3D models. Explicit solutions (versus eigenvalue problems) are found for the overturning circulation measure (except for the stochastic optimal); a truncation method on a few leading eigenmodes usually provides the optimal perturbations for analyses on long time scales.
Abstract
Optimal surface salinity perturbations influencing the meridional overturning circulation maximum are exhibited and interpreted on a stable steady state of a 2D latitude–depth ocean thermohaline circulation model. Despite the stability of the steady state, the nonnormality of the dynamics is able to create some transient growth and variability through stimulation by optimal perturbations. Two different measures are compared to obtain the optimum—one associated with the departure from steady state in terms of density, and the other with the overturning circulation intensity. It is found that such optimal analysis is measure dependent; hence, the latter measure is chosen for studying the following physical mechanisms. The response to the optimal initial sea surface salinity perturbation involves a transient growth mechanism leading to a maximum modification of the circulation intensity after 67 yr; the amplification is linked to the most weakly damped linear eigenmode, oscillating on a 150-yr period. Optimal constant surface salinity flux perturbations are also obtained, and confirm that a decrease in the freshwater flux amplitude enhances the circulation intensity. At last, looking for the optimal stochastic surface salinity flux perturbation, it is established that the variance of the circulation intensity is controlled by the weakly damped 150-yr oscillation. Two approaches are tested to consider extending such studies in more realistic 3D models. Explicit solutions (versus eigenvalue problems) are found for the overturning circulation measure (except for the stochastic optimal); a truncation method on a few leading eigenmodes usually provides the optimal perturbations for analyses on long time scales.
