Advertisement

Influence of ENSO Events on the Agulhas Leakage Region

  • Morgan L. ParisEmail author
  • Bulusu Subrahmanyam
  • Corinne B. Trott
  • V. S. N. Murty
Original Paper
  • 136 Downloads

Abstract

This study explores the subsurface oceanic response to the El Niño Southern Oscillation (ENSO) in the southern tropical Indian Ocean and Agulhas leakage region. The subsurface temperature and salinity responses of the southern tropical Indian Ocean and Agulhas leakage are studied using Argo float data. In addition, AVISO satellite-derived sea surface height (SSH) data is used to examine the propagation of an ENSO signal across the Indian Ocean to the Agulhas leakage region and its impact on eddy propagation within the leakage region. During the peak of an ENSO event, the Indian Ocean basin is anomalously warm up to 500 m depth in response to El Niño, but anomalously cool in response to La Niña. Sea surface salinity is anomalously fresh for both events in the eastern and western Indian Ocean basins, but the subsurface salinity signal during El Niño is anomalously saline to about 150 m in both basins. The subsurface signals last approximately 1 year (2 years) after the peak of ENSO in the eastern (western) Indian Ocean basin. The subsurface signal in the Agulhas leakage region is anomalously warm 2 years after El Niño and anomalously cool 2 years after La Niña. Westward propagation of SSH anomalies across the Indian Ocean basin is evident at 12°S and 25°S latitudes and it takes 2 years for the initial signal of Rossby waves to reach Agulhas leakage region after the peak of an ENSO event.

Keywords

ENSO Argo profiles Rossby waves Agulhas leakage 

1 Introduction

The El Niño Southern Oscillation (ENSO) originates in the equatorial Pacific Ocean, but its influence reaches far beyond the Pacific Ocean’s boundaries. Air-sea interactions involved in the formation of El Niño and La Niña conditions have been linked to temperature and salinity anomalies in the Indian Ocean [13, 17]. Fluctuations in east-west zonal winds (Walker circulation) over the equatorial Pacific during an ENSO event alter the wind variability over the eastern Indian Ocean, which in turn changes sea surface temperature (SST) in the Indian Ocean basin [17, 18, 32]. A dominant basin wide warming (cooling) response during El Niño (La Niña) is observed in the boreal winter-spring [3] due to the persistence of easterly (westerly) surface winds over the eastern Indian Ocean that suppress (enhance) convective activity increasing (decreasing) the underlying temperatures ultimately causing basin wide temperature patterns [16, 31]. At the same time, these anomalous easterly (westerly) winds promote upwelling (downwelling) along the Sumatra coast bringing cool (trapping warm) waters to the surface. These anomalous winds force a westward downwelling (upwelling) Rossby wave that deepens (shallows) the thermocline in the western Indian Ocean [35]. As a result, during the boreal autumn, basin-wide patterns of SST are replaced by anomalously cool (warm) waters in the eastern Indian Ocean basin and anomalously warm (cool) waters in the western Indian Ocean basin [3, 31]. A similar response in sea surface salinity (SSS) is seen in which during El Niño (La Niña), large freshening (salting) occurs in the southwestern Indian Ocean with a coincident salting (freshening) off the southern Sumatra coast [13].

The Indian Ocean is linked to the Pacific Ocean by two major physical pathways north and south of Australia via the Indonesian throughflow (ITF) into the southern tropical Indian Ocean and Tasman leakage (TL) into the Antarctic Ocean [9]. Both pathways are considered source waters, contributing about 12.6 Sv (1 Sv = 106 m3 s−1) to Agulhas leakage, including recirculation from frontal regions of the Southern Ocean. Nearly half of this contribution (about 6.1 Sv) comes from ITF. ITF is composed of relatively fresh surface waters and saltier intermediate waters with a wide temperature range [12] that feed into the South Equatorial Current (SEC). These waters are primarily concentrated in the top 600 m between the island of Java and 13°S with an estimated total transport range of 10.7 to 18.7 Sv into the southern tropical Indian Ocean [29]. Evidence by Le Bars et al. [19] shows a co-dependency between ITF transport and Agulhas leakage. An increase in Agulhas leakage, or an increased number of eddies released into the Atlantic Ocean, was linked to an increase in the ITF transport. Furthermore, the transport of ITF has been found to change in response to ENSO events. ITF transport increases in response to La Niña and decreases in response to El Niño [20] with ITF lagging the ENSO cycle by 8–9 months [10]. In comparison, TL contributes less than 1.4 Sv to Agulhas leakage. TL transports waters westward into the south Indian sub-tropical gyre at intermediate depths of 500 to 1000 m. This transport pathway is concentrated between Australia and 40°S and contributes about 16 Sv into the Indian Ocean [9].

Air-sea interactions and physical connections between the Pacific Ocean and the Indian Ocean change the properties of the Indian Ocean in response to ENSO. Rossby waves play a dominant role by linking the circulation of the south Indian Ocean to Agulhas leakage [33]. Schouten et al. [27] identified two bands of enhanced variability east of Madagascar at 12°S and 25°S marking the preferred routes along which ENSO generated anomalies propagate westward as baroclinic Rossby waves. As previously mentioned, easterly (westerly) wind anomalies produced during El Niño (La Niña) generate westward propagating Rossby waves [11] thus explaining the SSH variability at 12°S. The SSH variability at 25°S is explained by the impinging Rossby waves on the east coast of Australia moving equatorward as coastally trapped Kelvin waves where they are finally released into the Indian Ocean as the secondary set of ENSO associated Rossby waves at 25°S [5, 34].

This paper explores temperature and salinity changes in the water column of the southern tropical Indian Ocean basin in response to ENSO events and how changes at depth are further propagated across the Indian Ocean basin to influence the Agulhas leakage region dynamics. Agulhas leakage occurs at the southern tip of Africa where the Agulhas Current retroflects releasing warm, saline waters from the Indian Ocean into the south Atlantic in the form of Agulhas rings, eddies, and filaments [1]. Agulhas leakage is the primary connection between the Indian Ocean and the Atlantic Ocean and continues to feed the upper arm of Atlantic Meridional Overturning Circulation (AMOC). Understanding the leakage dynamics and the causes of fluctuations in Agulhas leakage is fundamental in understanding the impact to the strength of AMOC sequentially altering climate patterns.

2 Data and Methods

2.1 Satellite Data

We have used the AVISO Version 1 daily sea level anomalies from 1993 to 2017, which are Ssalto/Duacs altimeter products produced and distributed by the Copernicus Marine and Environment Monitoring Service (CMEMS) (http://www.marine.copernicus.eu) at a 0.25° × 0.25° resolution. The sea level anomalies represent variations of the SSH relative to the mean sea surface reference ellipsoid over a 20-year reference period from 1993 to 2012. Climatological data used in Fig. 1 are from the AVISO DUACS L4 product released in July 2018, which possesses improved variance over high-variability areas (such as western boundary currents) and reduced global mapping error.
Fig. 1

AVISO SSH anomalies averaged from 1993 to 2012 for January to March (top) and July to September (bottom; m). Line (a1) at [12°S, 50°–120°E], line (a2) at [25°S, 50°–115°E], box (b1) at [0°–20°S, 90°–105°E], box (b2) at [10°–30°S, 60°–80°E], box (c1) [37°–45°S, 10°–20°E], and line (c2) at [37°–45°S, 15°E] represent areas of interest in signal propagation corresponding to subsequent figures

2.2 In Situ Observations of Argo Temperature and Salinity

Temperature and salinity profile data are provided by Argo floats from the Asia Pacific Data Research Center (APDRC), University of Hawaii. This data is available at monthly intervals with 1° × 1° horizontal resolution from January 2005 through October 2017 at 29 depth levels up to a maximum depth of 2000 m (http://apdrc.soest.hawaii.edu/datadoc/Argo_iprc.php).

2.3 Oceanic Niño Index

ENSO events are determined using the Oceanic Niño Index (ONI) from the National Weather Service and Climate Prediction website (http://www.cpc.noaa.gov/products/analysis_monitoring/ensostuff/ensoyears.shtml). ONI calculates ERSST.v4 SST anomalies in the Niño 3.4 region (5°N–5°S, 120–170°W) with an applied 3-month running mean and is based on a centered 30-year base period updated every 5 years. El Niño events are classified as having an ONI value at or exceeding + 0.5° anomaly threshold for 3 consecutive months and a La Niña event as having an ONI value at or below − 0.5° anomaly for 3 consecutive months. This classification process is consistent with that used by the National Oceanographic and Atmospheric Administration’s (NOAA) Climate Prediction Center (http://www.cpc.ncep.noaa.gov/).

2.4 Methods

The monthly temperature and salinity interannual anomalies are obtained by subtracting the average monthly temperature and salinity data from the monthly temperature and salinity data for each year to remove the seasonal climatology from the monthly data. The average monthly temperature and salinity are computed from the full Argo data set (2005–2014) for each depth range.

We define the Agulhas leakage region based on the location in which prevalent transport of warm, saline waters are observed. This region spans from the tip of the African continental shelf to the oceanic subtropical front (37°S–45°S) [26] and has a western limit established by the Good Hope transect [30] and an eastern limit at the point of retroflection (10°E–20°E) [4]. Dencausse et al. [7] define the retroflection location as being largely skewed by a retroflection signal because of its highly variable nature and the average mitigates the influence of extreme values that may result from retroflection interaction. Furthermore, we also considered two box regions in the eastern and western Indian Ocean regions. The eastern Indian Ocean box spans from 90°–105°E to 0°–20°S while the western Indian Ocean box spans from 60°–80°E to 10°–30°S. These regions are shown in Fig. 1. The eastern box is placed at a position to capture the origin of ENSO signal response and its interaction with the Indonesian Throughflow. The position of the western box is designed to capture the propagating signal as it approaches the source currents while avoiding interaction with the East Madagascar Current retroflection. This box includes both the South Equatorial Current and region of Rossby wave propagation. The box averaged salinity and temperature of these three regions were obtained to create a time-series at depth from Argo data.

As SSH variability due to strong eddying correlates with large standard deviations, we examined the eddy leakage region in the Agulhas Current region via the mean standard deviation at 15°E from 37°S to 45°S [6]. The standard deviation of the 33 SSH points along c1 line (Fig. 1), selected to reflect the signal in the Agulhas leakage region, was found at each daily timestep over the study period and temporally analyzed (Fig. 7) and discussed. The 2-year lagged signal of the standard deviation as a response to El Niño events was compared to that of La Niña events using a paired sample t test [28]. Furthermore, the response of Agulhas leakage was analyzed at a 2-year lag based on a previous study by Paris and Subrahmanyam [21]. The paired sample t test evaluates whether the eddying response observed in the standard deviation of SSH between the two ENSO phase possesses equal means and variances. This statement is referred to as the null hypothesis, or the statement being tested. Our results of the t test find a rejection of the null hypothesis, indicating that the eddy variability as a response to the two different ENSO phases is different at the 99% confidence level (with a p value less than 0.01).

This paper used a lead-lag correlation analysis of SST and SSS signals in the Agulhas leakage region and found that the lag between the peak ENSO event and SST and SSS response ranged from 20 to 26 months. Furthermore, circulation patterns in the Indian Ocean support this time frame because it takes a little more than ~ 1.3 years (~ 27 months) for the waters to cross the Indian Ocean basin and then ~ 6 months for eddies from the East Madagascar Current and Mozambique Channel to interact with the Agulhas Current [9].

3 Results and Discussion

3.1 Identified ENSO Events and Regions of Influence

Between January 2005 and October 2017, four El Niño and three La Niña events occurred. The peak of an ENSO event is known to occur during the December–March period [31]. Records from NOAA’s National Centers for Environmental Information State of the Climate (https://www.ncdc.noaa.gov/sotc/enso/200813) state that the 2004–2005 El Niño originally developed during July–August of 2004 and peak temperatures occurred as early as March 2005 but warming persisted until late April 2005. Another El Niño event developed in September–October 2006, peaking in December with a decline at the end of March 2007. The first La Niña episode developed during September 2007 but peaked in February 2008 and lasted until March 2008. It was followed by an El Niño episode, which developed in early May 2009 and reached its peak in late December 2009 before breaking down and returning to neutral conditions by April 2010. The following 2 years are defined as two separate La Niña events. The 2010–2011 La Niña event developed in July 2010 with a peak at the start of 2011 before it transitioned to neutral conditions around July 2011. These neutral conditions were quickly replaced by a secondary La Niña event in September 2011. This weak event dispersed by April 2012 and neutral conditions returned and persisted until an anomalous warming event occurred around May of 2014. This warming failed to produce an El Niño event, however, in early 2015, El Niño conditions were observed and continued to strengthen into a strong event by November 2015. This event persisted until nearly May 2016 before it transitioned back to neutral conditions.

All identified ENSO events from 2005 to 2017 and the propagation of the ENSO signal in the Indian Ocean were investigated via SSH, salinity, and temperature data. Furthermore, the SSH, salinity, and temperature conditions were analyzed 2 years after these ENSO events to examine whether these ENSO signals affected the Agulhas Current and ultimately Agulhas leakage.

Figure 1 represents the six regions of interest isolated by subsequent figures. These regions include line a1 and line a2 at 12°S and 25°S, respectively, where Rossby wave propagation is known to occur in response to ENSO events [5, 11]. Figures 2 and 3 represent the Hovmöller diagrams at these locations drawn to isolate Rossby waves with respect to the identified ENSO years. As described in Sect. 2, boxes b1, b2, and c2 correspond to Figs. 4, 5, and 6, respectively, and show the depth-time sections of temperature and salinity at these three locations. An additional line within the Agulhas leakage region, line c2, was also used to evaluate the standard deviation of SSH anomalies (Fig. 7). Figure 1 shows the climatological mean SSH anomalies for the period 1993–2012 for January–February–March (JFM, boreal winter) and July–August–September (JAS, boreal summer) seasons and reveals the presence of a seasonal zonal dipole structure across the southern tropical equatorial Indian Ocean. With respect to the 1993–2012 mean SSH, much of the Indian Ocean basin possesses anomalously high SSH, excluding the region off the Sumatra coast, which is anomalously low. During the boreal summertime, the opposite is seen, with low SSH off the Sumatra coast. This region is known for upwelling particularly during El Niño conditions. Two significant mechanisms for the redistribution of SSH anomalies in this region are variability induced by ENSO events and westward-propagating Rossby waves, both of which are explored in this paper. SSH anomalies within the Agulhas leakage region alternate between small regions of positive and negative SSH anomaly indicating eddy activity.
Fig. 2

Hovmöller diagrams from AVISO SSH deseasonalized anomalies along line (a1) in Fig. 1 at 12°S from January 2005 to December 2010 (left panel) continuing through January 2011 to October 2016 (right panel)

Fig. 3

Hovmöller diagrams from AVISO SSH deseasonalized anomalies along line (a2) in Fig. 1 at 25°S from January 2005 to December 2010 (left panel) continuing through January 2011 to October 2016 (right panel)

Fig. 4

Depth-time sections of box averaged Argo temperature anomalies (top panel) and Argo salinity anomalies (bottom panel) from January 2005 to October 2017 from box (b1) in Fig. 1. The boxes isolate the peak (December to March) of any El Niño (solid line) or La Niña (dashed line) events that occurred during this time series

Fig. 5

Same as Fig. 4, but in reference to box (b2) in Fig. 1

Fig. 6

Depth-time sections of box averaged Argo temperature anomalies (top panel) and Argo salinity anomalies (bottom panel) from January 2005 to October 2017 from box (c1) in Fig. 1. The boxes isolate 2 years after the peak (December to March) of any El Niño (solid line) or La Niña (dashed line) events that occurred during this time series

Fig. 7

Time-series of standard deviation of daily AVISO SSH anomalies at line (c2) in Fig. 1 from January 2005 to December 2016. The boxes isolate 2 years after the peak (December to March) of any El Niño (solid line) or La Niña (dashed line) events that occurred during this time series

3.2 Signal Propagation

To better understand SSH anomaly variability across the Indian Ocean basin in the context of signal propagation, Hovmöller diagrams were constructed at 12°S and 25°S (Figs. 2 and 3, respectively) from 50°E to 110°E with time from January 2005 to October 2016 from the AVISO SSH product. Anomalous wind forcing over the Equatorial Indian Ocean initiated by ENSO events has been known to initiate Rossby waves, albeit at a variety of magnitudes and lags with respect to the peak of each event [14]. During the onset of an El Niño event, anomalously strong equatorial easterlies create two anticyclonic gyres (one to the north and one to the south of the equator) in the eastern Indian Ocean [14]. The northern gyre generally dissipates while the southern one strengthens, resulting in a downwelling effect to the southwest of the strong wind system [14, 33]. In each diagram, alternating bands of positive and negative SSH anomalies suggest the westward propagation of Rossby waves. Estimates of the theoretical speed of Rossby waves at 12°S and 25°S based on Killworth et al. [15] should be about 15 cm/s and 5 cm/s, respectively. Comparison of the steep diagonal bands of continuous negative or positive SSH anomalies spanning across the southern Indian Ocean at 25°S to their more shallow 12°S counterparts indicate faster speed at 12°S than at 25°S, as expected. Calculated (using the clearly visible westward propagations in the attached Hovmöller diagrams) propagation speeds at 12°S (Fig. 2) are 15.57 cm/s, comparable to the theoretical value of 15 cm/s. The propagation speed calculated at 25°S (Fig. 3) is 6.92 cm/s which is similar to the theoretical value of 5 cm/s. At 35°S, the westward propagation speed is estimated at 4.43 cm/s.

The Rossby wave SSH signals trigged by La Niña and El Niño can be isolated in Fig. 2 as well as Fig. 3 and is consistent between both figures. At both 12°S and 25°S, the slanting bands of low SSH anomalies across the Indian Ocean correspond to the months of the peak years of La Niña episodes (Nov 2007, Dec 2010, and Dec 2011). Similarly, the slanting bands of high SSH anomaly moving across the Indian Ocean correspond to the months of the peak year of an El Niño episode (Nov 2006, Nov 2009, and Oct 2015). In other words, the Hovmöller diagrams provide a visual representation of the transport of cold waters across the Indian Ocean basin during La Niña and warm waters during El Niño. The arrival of these low and opposite high SSH signals near the coast of Madagascar (50°E) occurs about 15 months after the initial development of La Niña and El Niño events in the eastern Indian Ocean at 12°S. At 25°S, the arrival of the corresponding low and high signals does not occur until nearly 21 months after those at 12°S. It may be noted that while additional regions of high and low SSH occur throughout these figures, this noise is expected because Figs. 2 and 3 represent regions of known SSH variability where Rossby waves constantly propagate regardless of ENSO events [27].

These results suggest that Rossby waves at 25°S moving at a phase speed of 6.92 cm/s initiate changes to Agulhas leakage as they carry the ENSO-induced SSH anomalous signal further southwestward with much reduced propagation speed (as Rossby wave speed decreases at higher latitudes) from Madagascar region to impact the properties of Agulhas source currents. The Indian Ocean response signal of the ENSO event continues to propagate westward beyond the south of Madagascar and thus reaches the Agulhas region and then into the Agulhas leakage in another 1 year 6 months (or sometimes 2 years) after the peak of an ENSO event. Our results are consistent with Putrasahan et al. [22], who note that the continued propagation of the temperature and salinity signal southwestward about the southern tip of Africa is attributed to Madagascar dipoles that cause anomalous volume transports by the southern branch of the East Madagascar Current into the Agulhas leakage region [25]. Southward and westward propagation of anticyclonic and cyclonic eddies from Madagascar into the Agulhas Current region has been well observed and provides a mechanism for the transport of temperature and salinity signals from the westernmost boundary of the Rossby waves to the Agulhas leakage region [8, 24]. The variability of such dipole events has been directly related to ENSO events [25].

3.3 Subsurface Variability in Salinity and Temperature Anomalies

Using Argo in situ observations, we are able to evaluate the temperature and salinity anomalies with depth in the upper 500 m to those signals identified at the surface by previous studies discussed in the introduction. Figures 4 and 5 correspond to box b1 and b2, respectively, (Fig. 1) to evaluate the response of both the eastern and western Indian Ocean basin. The large areas spanned by these boxes are designed to observe the large-scale signal from the ENSO events and reduce the impacts of smaller-scale variability within the box-averaged regions. Figure 6, corresponding to box c1 in Fig. 1, evaluates the same signal later seen in the Agulhas leakage region.

When evaluating the peak of the 2007–2008 (and 2010–2011, 2011–2012) La Niña, obvious negative anomalies of temperature and salinity are noted for both the eastern and western basins (Figs. 4 and 5). The anomalously low temperature in the eastern box begins as early as mid-2007 (and 2010, 2011) reaching about 150 m depth and continues in both the surface and subsurface deepening to past 500 m depth until about mid-2008 (and 2011, 2012) when the surface signal is replaced by anomalously warm waters. The subsurface anomalously cold signal persists between 50 and 150 m depth range for much longer until about June of 2009 (nearly the end of 2014). Near the end of 2009 (2014), the surface warm signal dominates reaching a depth of about 100 m by March 2010 (start of 2015). The salinity during this same time period is consistently low in both the surface and up to 150 m depth. The downward propagation with time is more evident in the western Indian Ocean box (Fig. 5), but with a smaller magnitude of anomalies. This box is located within the broad westward-flowing South Equatorial Current Qu and Meyers [23] which contains the waters of ITF in the upper layers and forms the northern branch of the southern subtropical anticyclonic gyre. The SEC is stronger at the sea surface between 10°S-11°S and is a broad westward flow at surface depths in the latitudinal belt of 13°S-25°S Qu and Meyers [23]. A part of the SEC turns southward as the East Madagascar Current to form the subtropical gyral flow in the western basin south of 25°S. The core of the subtropical gyre is in the eastern basin off Western Australia and shifts southward at deeper depths Qu and Meyers [23]. Thus, the currents at subsurface depths in the western basin are weaker and allow the temperature and salinity anomalies to persist for longer duration, as observed in this study. Both the Ekman flow and westward propagating Rossby waves influence the circulation of this broad westward flow of the subtropical gyre.

The cold, fresh anomalies of the 2007–2008 (and 2010–2011, 2011–2012) La Niña event continued with depth for the next 2 years in the western Indian Ocean box (Fig. 6). The cool signal in the western Indian Ocean box persisted longer at the surface than the cool surface layer in the eastern Indian Ocean box. The surface layer in the western Indian Ocean box does not reverse to a warm signal until after the start of 2009 (2012). A similar pattern is evident between salinity and temperature in the eastern box. The only difference is that instead of a signal reversal at the start of 2009 (2012), a shallow region of high salinity only develops for approximately 3 months (6 months) in 2009 (2013) before transitioning back to anomalously fresh.

In the context of a La Niña event, the patterns in these figures suggest that the cold fresh signal response develops around the start of La Niña in the eastern Indian Ocean basin and propagates westward via Rossby waves. The signal is stronger at its origin in the eastern Indian Ocean basin. At the peak of the event, the cold fresh signal penetrates to a depth of 150 m in the eastern basin and over 500 m depth in the western basin. The difference in the depth of penetration between the eastern and western basins is due to the prevalence of weaker currents in the western basin compared to the stronger currents at subsurface to depths within the westward flowing northern branch of subtropical anticyclonic circulation gyre [23]. The current speed in the ITF in the eastern basin is stronger and weakens with depth as well as in its westward flow. Below the ITF, the westward flowing SEC is strong in the eastern basin and the core of the subtropical gyre shifts southward and deepens. This surface and subsurface signal persists for longer in the western basin than the eastern basin due to the slow movement of the westward propagating Rossby waves and the weaker mean currents at subsurface depths of the subtropical anticyclonic gyre in the western basin. Furthermore, basin-wide cold patterns are expected to transition to a dipole-like pattern in which the eastern basin is anomalously warm [3, 31] as represented by the warm surface waters introduced in mid-2008 (and 2011, 2012). It should be noted that this dipole response only influences the surface.

An evaluation of the peak of the 2004–2005 (and 2006–2007, 2009–2010, 2015–2016) El Niño is also performed (Figs. 4 and 5). A consistent positive anomaly trend is present in all solid boxes. In other words, the initial response during the peak of El Niño is anomalously warm and saline. This trend is expected to reach earlier than available Argo data in 2004 (and mid-2006, start of 2009, start of 2015) in the western Indian Ocean basin but is isolated to the top 100 m. Over the next year, the anomalously warm signal persists and deepens to past 500 m in the western basin. A few months into 2005 (and 2007, 2010), the warm signal is replaced at the surface by a cool signal but at depth, this anomalously warm temperature continues until 2007 in Fig. 5 (and 2008, 2012). Over the same time period in the box averaged eastern Indian Ocean region, a primary warm signal transitions to a secondary cool signal. The warm signal is greatest in the top 150 m but extends beyond 500 m. This signal is hypothesized to develop earlier than available Argo data in 2004 (and mid-2006, start of 2009, start of 2015) but only lasts until about January 2005 (and January 2007, April 2010, December 2017) when it is replaced in both the surface and subsurface by an anomalously cold signal. A similar pattern is observed in the salinity in the eastern basin but the saline signal lasts slightly longer. The salinity of the western Indian Ocean basin is high at the surface at the start of 2005 (and 2007, 2010, mid-2015) and similar to temperature propagates past a depth of 500 m throughout the year. Midway through 2006 (and 2007, 2012), the saline surface signal is replaced by a fresh signal at the surface but exists at a high magnitude below 150 m.

The timeline of signal propagation of the El Niño event in both the eastern and western Indian Ocean basin is similar to that of the previously evaluated La Niña. In particular, the warm, saline response develops during the start of the El Niño episode. At the peak of El Niño, this signal has penetrated to beyond 500 m depth in both the eastern and western basin. Once again the anomalous signal is greater in the eastern basin compared to the western basin but the subsurface signal persists longer in western basin.

The depth-time sections of temperature and salinity anomalies are shown for the box averaged Agulhas leakage region (Fig. 6). In this figure, dashed boxes correspond to 2 years after the peak of a La Niña event while solid boxes correspond to 2 years after the peak of an El Niño event. These depth-time sections are interesting because the Agulhas leakage region is anomalously cool and fresh from 2008 to mid-2010 for the entire depth profile. This timing is 2 years after the peak 2007–2008 La Niña event. The temperature shows a surface-warming signal briefly during early 2006 and 2007, 2 years after the peak of the 2004–2005 and 2006–2007 El Niño, respectively. Surface warming during mid-2010 also exists and penetrates to deeper depths of up to 500 m with time until mid-2014. Notice, this is also 2 years after the peak of 2009–2010 El Niño. The low salinity anomalies occupy the top 200 m at the beginning of 2008 and continue to propagate in time with depth lasting over 4 years until early 2012. After this time, the salinity is dominantly saline except for brief freshening 2 years after both the 2010–2011 and 2011–2012 La Niña events.

The surface and subsurface of the Indian Ocean basin was found to be anomalously cold in response to La Niña as opposed to El Niño where it was found to be anomalously warm. The observed trends support that this same response occurs in the Agulhas leakage region. It would also appear that the subsurface influence in the Agulhas leakage region is longer lasting than that of the surface. This would imply that changes could be longer lasting than originally assumed by studies of only the surface signal.

3.4 SSH Variability in the Agulhas Leakage Region

Figure 7 shows the standard deviation of AVISO SSH anomalies within the Agulhas leakage region (line c2 in Fig. 1). For each day, the standard deviation was found from the 33 points along line c2 in order to observe eddy-induced variability in the Agulhas leakage region, and then plotted for the entire study period. High standard deviations coincide with increased eddy activity [6]. In this figure, dashed boxes correspond to 2 years after the peak of a La Niña event while solid boxes correspond to 2 years after the peak of an El Niño event. In all instances, the higher standard deviation of SSH anomalies corresponds to 2 years after the peak of an El Niño event. Two years after the peak of a La Niña event, the standard deviation is lesser. The peaks in standard deviation 2 years after El Niño are most likely attributable to the passing of warm, saline eddies. Composite analysis of the eddy variability between the encapsulated El Niño and La Niña events reveals a statistically dissimilar trend between SSH anomaly standard deviations (when comparing El Niño and La Niña events) 2 years following the peak of each ENSO event. A paired sample t test between the 2-year signal of standard deviation between El Niño and La Niña events was conducted to statistically confirm the difference in the two signals [28]. Paired sample t tests evaluate the validity of a null hypothesis (the statement being tested) that the two signals (that of El Niño and that of La Niña) possess equal means and variances. Our analysis rejects the null hypothesis that the lagged signal of eddy variability possesses equal means and variances, confirming the difference in the strength of eddy variability in the Agulhas leakage region as a response to ENSO events (p value of 1.7044 × 10−7; [28]). A p value less than 0.01 signifies rejection of the null hypothesis at the 99% confidence level.

4 Conclusion

The results indicate that during the peak of an ENSO event, the southern tropical Indian Ocean basin is anomalously warm in response to El Niño and anomalously cool in response to La Niña, but with prevailing anomalously fresh waters at the surface during both the events. This is relatively consistent with previous studies [2, 21, 22] that looked at the response of the surface signal in the Indian Ocean to ENSO. The novel findings of this particular study are that this trend is also true with the subsurface signals in the eastern and western Indian Ocean basins. The eastern and western Indian Ocean basins respond to an ENSO event by exhibiting the respective anomalous signal and this signal penetrates to beyond 500 m depth by the peak of the event. After this time, the eastern basin transitions to the respective opposing signals. This transition is only at the surface for La Niña but also at depth for El Niño. In the western Indian Ocean basin, the subsurface signal of both ENSO events persists for nearly 2 years after the peak of the ENSO event. Furthermore, we hypothesized that Rossby waves triggered by ENSO-associated systems were responsible for the propagation of this signal. Hovmöller diagrams identified and also quantified this propagation. Hovmöller diagrams of SSH anomalies at 12°S indicate the propagation of high (low) SSH anomalies across the Indian Ocean basin from the start of the peak El Niño (La Niña) event, reaching the coast of Madagascar approximately a year later. The same signal is observed in the surface and depth profiles of Agulhas leakage 2 years after the peak of an ENSO event. This signal indicates that the intensity of Agulhas leakage is affected by ENSO events. Anomalously cold waters in the Agulhas leakage region 2 years after the La Niña event would suggest Agulhas leakage weakens in response to a La Niña episode. In contrast, anomalously warm waters in the Agulhas leakage region 2 years after the El Niño event would suggest Agulhas leakage strengthens in response to an El Niño episode. To further explore the dynamics of the Indian Ocean response to El Niño events, recommended future work is to isolate the signal from different types of El Niño events (in the Eastern Pacific) or El Niño-Modoki events (in the Central Pacific).

Notes

Acknowledgements

The author, VSN Murty is thankful to the Director, CSIR-NIO for his keen interest and support for this collaborative research work. Argo floats temperature and salinity data is courtesy of the Asia Pacific Data Research Center (http://apdrc.soest.hawaii.edu/datadoc/Argo_iprc.php). The authors would like to thank the following data centers for providing the datasets used to conduct this study: the IPRC APDRC (http://apdrc.soest.hawaii.edu/projects/argo/) and NASA JPL PO.DAAC for providing SST data. The authors would like to thank the anonymous reviewers and the editor, whose comments significantly contributed to the improvement of this paper.

Funding Information

This work is supported by the ONR NASCar (Northern Arabian Sea Circulation-autonomous research) award #N00014-17-1-2468 awarded to BS.

References

  1. 1.
    Beal LM, De Ruijter WPM, Biastoch A, Zahn R (2011) On the role of the Agulhas system in ocean circulation and climate. Nat 472:429–436.  https://doi.org/10.1038/nature09983 CrossRefGoogle Scholar
  2. 2.
    Behera S, Yamagata T (2001) Subtropical SST dipole events in the southern Indian Ocean. Geophys Res Lett 28(2).  https://doi.org/10.1029/2000GL011451
  3. 3.
    Bernal A, Latif M, Legutke S (2001) On dipolelike variability of sea surface temperature in the tropical Indian Ocean. J Clim 15(11):1358–1368.  https://doi.org/10.1175/1520-0442(2002)015<1358:ODVOSS>2.0.CO;2 CrossRefGoogle Scholar
  4. 4.
    Biastoch A, Durgadoo JV, Morrison AK, Van Sebille E, Weijer W, Griffies SM (2015) Atlantic multi-decadal oscillation covaries with Agulhas leakage. Nat Commun 2015(6).  https://doi.org/10.1038/ncomms10082
  5. 5.
    Cai W, Meyers G, Shi G (2005) Transmission of ENSO signal to the Indian Ocean. Geophys Res Lett 32.  https://doi.org/10.1029/2004GL021736
  6. 6.
    Chelton DB, Schlax MG, Samelson RM, De Szoeke RA (2007) Global observations of large oceanic eddies. Geophys Res Lett 34(15).  https://doi.org/10.1029/2007GL030812
  7. 7.
    Dencausse G, Arhan M, Speich S (2010) Spatio-temporal characteristics of the Agulhas current retroflection. Deep-Sea Res 27:1393–1405.  https://doi.org/10.1016/j.dsr.2010.07.004 CrossRefGoogle Scholar
  8. 8.
    De Ruijter WP, Ridderinkhof H, Schouten MW (2005) Variability of the Southwest Indian Ocean. Philos Trans R Soc Lond 363(1826):63–76.  https://doi.org/10.1098/rsta.2004.1478 CrossRefGoogle Scholar
  9. 9.
    Durgadoo JV, Ruhs S, Biastoch A, Boning CWB (2017) Indian Ocean sources of Agulhas leakage. J Geophys Res Oceans 122(4):3481–3499.  https://doi.org/10.1002/2016jc012676 CrossRefGoogle Scholar
  10. 10.
    England MH, Huang F (2005) On the interannual variability of the Indonesian Throughflow and its linkage with ENSO. J Clim 18(9):1435–1444.  https://doi.org/10.1175/jcli3322.1 CrossRefGoogle Scholar
  11. 11.
    Feng M, Meyers G (2003) Interannual variability in the tropical Indian Ocean: a two-year time-scale of Indian Ocean dipole. Deep-Sea Res 50(12):2263–2284.  https://doi.org/10.1016/s0967-0645(03)00056-0 CrossRefGoogle Scholar
  12. 12.
    Gordon AL, Ma S, Olson DB, Hacker P, Field A, Talley LD, Baringer M (1997) Advection and diffusion of Indonesian throughflow water within the Indian Ocean south equatorial current. Geophys Res Lett 24(21):2573–2576.  https://doi.org/10.1029/97gl01061 CrossRefGoogle Scholar
  13. 13.
    Grunseich G, Subrahmanyam B, Murty VSN, Giese BS (2011) Sea surface salinity variability during the Indian Ocean dipole and ENSO events in the tropical Indian Ocean. J Geophys Res Oceans 116(C11).  https://doi.org/10.1029/2011jc007456
  14. 14.
    Jury MR, Huang B (2004) The Rossby wave as a key mechanism of Indian Ocean climate variability. Deep Sea Res I Oceanogr Res Pap 51(12):2123–2136.  https://doi.org/10.1016/j.dsr.2004.06.005 CrossRefGoogle Scholar
  15. 15.
    Killworth PD, Chelton DB, De Szoeke RA (1997) The speed of observed and theoretical long extratropical planetary waves. J Phys Oceanogr 27(9):1946–1966.  https://doi.org/10.1175/1520-0485(1997)027<1946:tsooat>2.0.co;2 CrossRefGoogle Scholar
  16. 16.
    Klein SA, Soden BJ, Lau NC (1999) Remote Sea surface temperature variations during ENSO: evidence for a tropical atmospheric bridge. J Clim 12:917–932.  https://doi.org/10.1175/1520-0442(1999)012<0917:rsstvd>2.0.co;2 CrossRefGoogle Scholar
  17. 17.
    Kug JS, Kang IS (2005) Interactive feedback between ENSO and the Indian Ocean. J Clim 19(9):1784–1801.  https://doi.org/10.1175/jcli3660.1 CrossRefGoogle Scholar
  18. 18.
    Lau NC, Nath MJ (2000) Impact of ENSO on the variability of the Asian–Australian monsoons as simulated in GCM experiments. J Clim 13(24):4287–4309.  https://doi.org/10.1175/1520-0442(2000)013%3C4287:ioeotv%3E2.0.co;2 CrossRefGoogle Scholar
  19. 19.
    Le Bars DLB, Dijkstra HA, DeRuijter WPM (2013) Impact of the Indonesian Throughflow on Agulhas leakage. Ocean Sci Discuss 10:353–391.  https://doi.org/10.5194/osd-10-353-2013 CrossRefGoogle Scholar
  20. 20.
    Meyers G (1996) Variation of Indonesian Throughflow and the El Niño-southern oscillation. J Geophys Res Oceans 101(C5):12255–12263.  https://doi.org/10.1029/95jc03729 CrossRefGoogle Scholar
  21. 21.
    Paris ML, Subrahmanyam B (2018) Role of El Niño southern oscillation (ENSO) events on temperature and salinity variability in the Agulhas leakage region. Remote Sens 10(1):127.  https://doi.org/10.3390/rs10010127 CrossRefGoogle Scholar
  22. 22.
    Putrasahan D, Kirtman BP, Beal LM (2016) Modulation of SST interannual variability in the Agulhas leakage region associated with ENSO. J Clim 29(19):7089–7102.  https://doi.org/10.1175/jcli-d-15-0172.1 CrossRefGoogle Scholar
  23. 23.
    Qu T, Meyers G (2005) Seasonal characteristics of circulation in the southeastern tropical Indian Ocean. J Phys Oceanogr 35(2):255–267.  https://doi.org/10.1175/jpo-2682.1 CrossRefGoogle Scholar
  24. 24.
    Quartly GD, Srokosz MA (2004) Eddies in the southern Mozambique Channel. Deep Sea Res II Top Stud Oceanogr 51(1–3):69–83.  https://doi.org/10.1016/j.dsr2.2003.03.001 CrossRefGoogle Scholar
  25. 25.
    Ridderinkhof W, Le Bars D, Heydt AS, Ruijter WPM (2013) Dipoles of the south East Madagascar current. Geophys Res Lett 40(3):558–562.  https://doi.org/10.1002/grl.50157 CrossRefGoogle Scholar
  26. 26.
    Ruijter WD, Biastoch A, Drijfhout SS, Lutjeharms JRE, Matano RP, Pichevin T, Weijer (1999) Indian-Atlantic interocean exchange: dynamics, estimation and impact. J Geophys Res Oceans 104:20885–20910.  https://doi.org/10.1029/1998jc900099 CrossRefGoogle Scholar
  27. 27.
    Schouten MW, De Ruijter WP, Van Leeuwen PJ (2002) Upstream control of Agulhas ring shedding. J Geophys Res Oceans 107(C8):3109.  https://doi.org/10.1029/2001jc000804 CrossRefGoogle Scholar
  28. 28.
    Snedecor GWC, William G (1989) Statistical methods. QA276 12:S6313Google Scholar
  29. 29.
    Sprintall J, Wijffels SE, Molcard R, Jaya I (2009) Direct estimates of the Indonesian Throughflow entering the Indian Ocean: 2004–2006. J Geophys Res Oceans 114(C7).  https://doi.org/10.1029/2008jc005257
  30. 30.
    Swart SS, Speich S, Ansorge GGJ, Gladysev S, Lutjeharms JRE (2008) Transport and variability of the Antarctic circumpolar current south of Africa. J Geophys Res Oceans 113(C09014).  https://doi.org/10.1029/2005gl023271
  31. 31.
    Tokinaga H, Tanimoto Y (2004) Seasonal transition of SST anomalies in the tropical Indian Ocean during El Niño and Indian Ocean dipole years. J Meteorol Soc Jpn 82(4):1007–1018.  https://doi.org/10.2151/jmsj.2004.1007 CrossRefGoogle Scholar
  32. 32.
    Wieners CE, De Ruijter WP, Ridderinkhof W, Von Der Heydt AS, Dijkstra HA (2016) Coherent tropical Indo-Pacific interannual climate variability. J Clim 29(11):4269–4291.  https://doi.org/10.11785/jcli-d-15-0262.1 CrossRefGoogle Scholar
  33. 33.
    Xie SP, Annamalai H, Schott FA, McCreary JP Jr (2002) Structure and mechanisms of South Indian Ocean climate variability. J Clim 15(8):864–878.  https://doi.org/10.1175/1520-0442(2002)015%3C0864:samosi%3E2.0.co;2 CrossRefGoogle Scholar
  34. 34.
    Xie SP, Hu K, Hafner J, Tokinaga H, Du Y, Huang G, Sampe T (2009) Indian Ocean capacitor effect on indo–western Pacific climate during the summer following El Niño. J Clim 22(3):730–747.  https://doi.org/10.1175/2008jcli2544.1 CrossRefGoogle Scholar
  35. 35.
    Yu W, Xiang B, Liu L, Liu N (2005) Understanding the origins of interannual thermocline variations in the tropical Indian Ocean. Geophys Res Lett 32(24).  https://doi.org/10.1029/2005gl024327

Copyright information

© Springer Nature Switzerland AG 2018

Authors and Affiliations

  1. 1.School of Fisheries, Aquaculture, and Aquatic SciencesAuburn UniversityFairhopeUSA
  2. 2.School of the Earth, Ocean, and EnvironmentUniversity of South CarolinaColumbiaUSA
  3. 3.Council of Scientific and Industrial Research (CSIR)-National Institute of Oceanography Regional CentreVisakhapatnamIndia

Personalised recommendations