Australia: The Land Where Time Began

A biography of the Australian continent 

North Pacific Ocean – Wind vs Eddy-Forced Regional Sea Level Trends and Variability

Wind changes of low frequency are often considered to be the cause of induction of regional sea level trends. It has been demonstrated by this that the significant sea level trend and variability can also be generated by eddy momentum flux forcing resulting from time-varying instability of the background oceanic circulation. The sea level changes that are eddy-forced tend to have subgyre scales, compared to the broad-scale variability that is wind forced, and they are largely confined to the Kuroshio Extension region in the North Pacific (30^oN–40^oN, 140^oE-175^oE) as well as the subtropical countercurrent (STCC) region (18^oN-25^oN, 130^oE-175^oE. The relative importance of the sea level trends in the past 2 decades that were wind-forced and eddy-forced were quantified using a 2-layer primitive equation model driven by the ECMWF wind stress data and the eddy momentum fluxes specified by the AVISO sea surface height anomaly data. It was found that the increasing trend south of the Kuroshio Extension or decreasing trend north of the Kuroshio Extension is due to strengthening of the regional eddy-forcing over the past 2 decades. On the other hand, the decadal weakening of the eddy momentum flux forcing causes a decreasing sea level trend south of the STCC and an increasing sea level trend north of the STCC. These changes in the decadal eddy momentum flux that are caused by the background Kuroshio Extension and changes in the STCC in connection with the Pacific decadal oscillation (PDO) shifting of the wind pattern from a positive phase to a negative phase over the past 2 decades.

The globally averaged sea level has risen at an accelerated rate of 3.2 mm/year, and understanding of this rate, including the budget of the 2 contributing components, has increased as a result of the findings from many interdisciplinary studies (see Bindoff et al., 2007; Church et al., 2013, and references therein). Sea level trends can deviate from the global mean value significantly, however (e.g. Cazenave & Llovel, 2010; Stammer et al., 2010). The regional sea level trends can exceed by double the 3.2 mm/year value at various locations and signals over a wide range of length scales are contained in their spatial patterns. Regional sea level rises or falls on relatively broad scales are detected in tropical regions and in the central (eastern) extratropical regions, e.g. in the North Pacific Ocean. The regional sea level trend signals have, in contrast, meridional spatial scales in the western subtropical and subarctic regions that are generally smaller. It is the regional sea level variability that is of paramount importance for coastal regions that surround the Pacific Basin where human activities are concentrated and where most impact of landmass erosion and inundation can be exerted.

 To clarify the dynamical causes of the regional sea level trends in the tropical Pacific many analyses of data and numerical modelling studies have related to the sea level rise in the western basin and drop in the eastern basin to the strengthening of the trade winds across the tropical Pacific over the past 2 decades (Kӧhl & Stammer, 2008; Timmermann et al., 2010; Feng et al., 2010; Merrifield & Maltrud, 2011; Qiu & Chen, 2012). Regarding the regional sea level changes in the extratropical North Pacific, in the surface wind stress field the low frequency changes have also been considered to be a main driver in the determination of the regional sea level trends and variability in the western subarctic gyre, the Kuroshio Extension, and off the coast of the North American continent (e.g. Qiu, 2002; Cummings & Lagerloef, 2004; Qiu & Chen, 2005; Yasuda & Sakurai, 2006; Bromirski et al., 2011; Sasaki et al., 2014). Qiu et al. suggest it is worth emphasising that away from the coastal orography, there are generally large spatial scales in the surface wind stress field. As such, regional sea level variability that is wind-forced necessarily has broad spatial scales that are present in the extratropical central and eastern North Pacific Ocean. It was required to explore the physical mechanisms beyond the direct forcing by the wind in order to properly understand the smaller-scale regional sea level trends that have been detected in the western North Pacific Ocean.

Modulation of the strength of the ocean circulation gyres and its dynamical stability is an indirect mechanism through which changes in large-scale surface winds affect the regional sea level variability. Some parts of the wind-driven gyres have been found to be susceptible to dynamical instabilities because of the underlying ocean dynamics. Unstable regions of the extratropical North Pacific Ocean are confined largely to 2 bands of latitude.

1)    The location of the first band is along the 32^oN-38^oN, where the western boundary Current Kuroshio enters the open North Pacific Ocean basin, becoming a free, baroclinic jet, which is unstable, the Kuroshio Extension.

2)    The second band of high mesoscale eddy variability is present along the 18^oN-28^oN band between Taiwan and Hawaii. Along this band the high mesoscale eddy variability is due to the presence  of the Subtropical Countercurrent (STCC), which is wind-driven, the eastwards shear of which is relative to the underlying North Equatorial Current (NEC), which flows to the west and provides the energy source for the baroclinic instability.

The baroclinicity of the Kuroshio Extension jet and the changes in the STCC, which affects the intensities of their instabilities and the energy level of the regional mesoscale eddies, as the large-scale surface wind stress field fluctuates over time. Significant changes in the level of the eddy kinetic energy have been detected over the past 2 decades along the Kuroshio Extension and the STCC regions of the North Pacific Ocean. According to Qiu et al. it is well known that dynamically, the eddy momentum flux forcing can feedback, and thereby modify the time-mean circulation and its geostrophically balanced sea level (e.g. Rhines & Holland, 1979; Greatbatch et al., 2010a, b; Qiu & Chen, 2010a; Penduff et al., 2011). It is natural to expect the mean circulation, which is eddy-forced, and the sea level would also change, as the energy level of mesoscale eddies alters over time.

According to Qiu et al. the first objective of this study was to quantify the extent to which the regional sea level trend signals in the extratropical North Pacific Ocean are affected by the time-varying eddy momentum flux forcing, such as that detected by satellite altimeters over the past 2 decades. Qiu et al. were interested, in particular, in the sea level variability that was eddy-forced in the context of the regional sea level trends forced directly by the time-varying wind stresses. The second objective of Qiu et al. was to clarify the processes that control the observed wind- and eddy-forced the sea level trends with the aid of a 2-layer ocean dynamical framework. The importance of the large-scale atmospheric forcing that relates to the Pacific decadal oscillation (PDO; Mantua et al., 1997) was emphasised.

Discussion and Summary

Eddy-forced variability in the presence of wind-driven circulation

Qiu et al. quantified the regional sea level trend signals due to the wind stress, the eddy momentum flux, and their combined forcing, respectively, in the previous section. Qiu et al. emphasise that the results of the combined wind- and eddy-forcing run [of the model] are  not simply the sum of wind and eddy forcing runs, though the wind stress and eddy momentum flux appear as additive forcing. This is the case because as the upper layer thickness is deformed by H₁, it modifies the depth of the equivalent layer H_e = H₁H₂/H₁+H₂) and affects how the barotropic and baroclinic sea levels adjust to the wind and eddy forcing forcings. As without wind forcing there would be no eddy variability, a different way to quantify mean sea level and trend that are eddy-induced is the subtraction of the results of the combined wind- and eddy-forcing run from those of the wind-forcing run only.

In order to quantify the relative importance of wind forcing verses eddy forcing Qiu et al. divided the modelled sea level signals into those that were forced by wind stresses and those forced by eddy fluxes (defined by the difference between the combined wind forcing and the eddy forcing, and those that are wind forcing runs). In different latitude bands the wind and eddy forcings contribute differently. The modelled sea level trend north of the Kuroshio Extension is determined exclusively by eddy forcing. If acting alone the wind forcing would have produced a positive sea level trend in this band. On the other hand to the south of the Kuroshio Extension the eddy forcing predominantly contributes to the regional positive trend and decadal sea level modulations. The wind forcings and eddy forcings generate sea level trends along the band of the STCC that are oppositely signed and the positive trend that results is dictated by the eddy forcing. On the other hand, in the band to the south of the STCC the wind forcings and the eddy forcings contribute constructively to the regional negative sea level trend. Their relative contributions to the total negative trend are 34% and 66%, respectively.

Processes responsible for regional sea level trends that are wind forced

The regional sea level trends that are forced by wind stresses have spatial patterns that are of a gyre-scale and are positive in the western subtropical and eastern tropical and subarctic gyres, respectively. The dominant large-scale variability over the North Pacific Ocean is that which is associated with the Pacific decadal oscillation (PDO) which is defined by the leading principal component of the monthly sea surface temperature anomalies poleward of 20^oN in the North Pacific (Mantua et al., 1997). Qiu et al. plotted the PDO index time series of the past 2 decades which showed that there is a clear transition from a positive to a negative phase. It is shown that during the 1993-2013 period, during the period of interest to Qiu et al., the spatial wind stress vector and curl pattern regressed to the PDO index. The wind stress curl spatial pattern is, by and large, consistent with regional sea level trend patterns. E.g. in the western tropical and midlatitudes band of 35^o-53^oN, decreased wind stress curl is responsible for the sea level rise in these regions and the reduced wind stress curl in the western subtropical band of 19^oN-35^oN and in the Alaska gyre is responsible for the sea level drop in these regions.  

Theoretical analysis that is presented in the appendix indicates that the sea level variability, that is wind stress curl-induced, tends to migrate to the west resulting from the planetary β effect, rather than a local 1-to-1 correspondence between the wind stress curl and regional sea level signals. Such a shift to the west between the wind curl pattern and the wind-forced regional sea level trend pattern is easily discernible by comparing Figs. 10b and 4b in Qiu et al., 2015. The determination of the importance of the propagation to the west of the wind-forced sea level changes has been emphasised by many existing studies for the tropical and extratropical North Pacific Ocean (e.g., Kessler, 1990; Qiu & Chen, 2005, 2010b, Taguchi et al., 2007; Ceballos et al., 2009). No further quantification, as such, was further pursued in this study.

Processes responsible for eddy-forced sea level trends

Eddy-forced sea level trends of large amplitude are confined to the Kuroshio Extension and STCC regions, where there are intense mesoscale eddy signals (see Fig. 3a in Qiu et al., 2015), in comparison with the wind-forced sea level trends. The dipolar negative trends to the north of the Kuroshio Extension and dipolar positive trends to the south of the Kuroshio Extension versus the dipolar positive trends north of the STCC and dipolar negative trends to the south of the STCC are the noticeable features of the eddy-driven sea level trends. The increasing eddy kinetic energy (EKE) in the Kuroshio Extension region of 150^oE-170^oE is due to the dipolar trends across the Kuroshio Extension. It has been indicated by many theoretical and modelling studies in the past that a zonal jet that is not stable, such as the Kuroshio Extension, is capable of driving a low sea level cyclonic recirculation gyre northwest of the zonal jet and a high sea level anticyclonic recirculation gyre southwest of the zonal jet (e.g. Jayne et al., 1996; Qiu et al, 2008; Taguchi et al., 2010; Waterman & Jayne, 2011). As the strength of the recirculation gyre is proportional to the intensity of the eddy forcing, the increased EKE level in the downstream Kuroshio Extension region should lead to Intensification of the northern recirculation gyre and the southern recirculation gyre, respectively, or the regional sea level decrease north of the Kuroshio Extension jet and increase southwest of the Kuroshio Extension jet.

Along the STCC band of 20⁰N-24^oN that flows to the east in the western subtropical North Pacific there is a decreasing trend in the level in the eddy kinetic energy. The eddy-forced regional sea level trend should increase northwest of the STCC and decrease southwest of the STCC, following the same theoretical argument. The PDO wind strength forcing induces the EKE trend signals that are detected in the Kuroshio Extension and STCC regions. It is indicated by several recent analyses that the regional EKE signals are related positively to the PDO-related surface wind and heat forcing (Yoshida et al., 2011; Qiu & Chen, 2013). In particular, the enhanced midlatitude westerlies along 29^oN-35^oN and the enhanced tropical easterlies along 15^oN-23^oN intensify the meridional Ekman temperature flux convergence and strengthen the STCC which flows to the east in the western North Pacific. Baroclinic instability along the STCC is enhanced, by increasing the vertical shear of the STCC-NEC system, which results in a higher level of regional EKE. When the PDO switches to its negative phase the opposite is true.

On the other hand, along the Kuroshio Extension latitude band in the positive phase of the PDO the wind stress curl forcing tends to induce negative sea level anomalies in the eastern North Pacific basin as a result of surface Ekman flux diversion. These wind-forced negative sea level anomalies propagate to the west in the Kuroshio Extension following a delay of about 3 years of more, they work to shift the Kuroshio Extension jet to the south, which causes the EKE level to increase in the upstream Kuroshio Extension region of 140^oE-153^oE and a decrease in the downstream region of 153^oE-170^oE (Qiu & Chen, 2005, 2011, and references therein). The opposite processes are at work during the negative phase and the EKE level in the downstream Kuroshio Extension tends to increase following a delay of 3-4 years. At the present it is not clear what the exact mechanism is for the downstream EKE in the Kuroshio Extension region to decrease once the PDO-induced negative sea level anomalies, or to increase when the PDO-induced positive sea level anomalies reach the Kuroshio Extension jet. According to Qiu et al. there is a possibility that the zonal Kuroshio Extension jet undergoes large latitudinal migration following the PDO wind forcing and that its downstream EKE level is controlled by the interaction between the migrating Kuroshio Extension jet and the underlying bathymetry, which is meridionally aligned, such as the Shatsky Rise along approximately 159^oE and the Emperor Seamounts along approximately 175^oE. Future studies are required to clarify the processes that determine the EKE levels in the upstream and the downstream Kuroshio Extension regions, in order to gain a better understanding and prediction of the regional sea level variability.

Several recent studies for the variability of the decadal Kuroshio Extension have pointed to the importance of wind stress forcing that is associated with the North Pacific Gyre Oscillation (NPGO) index, which is defined by the 2^nd principal component of the monthly sea level anomalies in the Northeastern Pacific regions 25^oN-62^oN, 180^oW-110^oW (e.g. Ceballos et al., 2009; Pierini, 2014). While the PDO and the NPGO indices are to a large extent independent for the long period 1950-2004 (Di Lorenzo et al., 2008), during the last 2 decades of interest to this study they are correlated. E.g., the negative NPGO index time series from 1993-2013. When compared to the PDO index the 2 indices are correlated with a coefficient r = -0.41 between the original monthly time series and r = -0.65 between their annually averaged time series. The above discussions in connection with the PDO forcing can be effectively replaced by the negative NPGO forcing as far as wind- and eddy-forced regional sea level changes in the last 2 decades are concerned.

Sources & Further reading

1. Qiu, B., et al. (2015). "Wind- versus Eddy-Forced Regional Sea Level Trends and Variability in the North Pacific Ocean." Journal of Climate 28(4): 1561-1577.