Hydrodynamic comparison between the north and south of Mallorca Island

Journal of Marine Systems xxx (2014) xxx–xxx

MARSYS-02478; No of Pages 11

Contents lists available at ScienceDirect

Journal of Marine Systems

j ourna l homepage: www.e lsev ie r .com/ locate / jmarsys

Hydrodynamic comparison between the north and south ofMallorca Island

Angel Amores a,⁎, Sebastià Monserrat a,b

a Instituto Mediterráneo de Estudios Avanzados, IMEDEA (UIB-CSIC), Palma de Mallorca, Spainb Departament de Física, Universitat de les Illes Balears (UIB), Palma de Mallorca, Spain

⁎ Corresponding author at: IMEDEA, C/Miquel MaMallorca-Spain.

E-mail address: [email protected] (A. Amor

0924-7963/$ – see front matter © 2014 Elsevier B.V. All rhttp://dx.doi.org/10.1016/j.jmarsys.2014.01.005

Please cite this article as: Amores, A., Mons(2014), http://dx.doi.org/10.1016/j.jmarsys.2

a b s t r a c t
a r t i c l e i n f o
Article history:Received 17 June 2013Received in revised form 5 November 2013Accepted 14 January 2014Available online xxxx

Keywords:Western MediterraneanHydrodynamic comparisonBalearic SeaMesoscale eddiesBalearic currentMallorca Channel

A hydrodynamic comparison between two zones of fishing interest, one located to the north and the other to thesouth of Mallorca Island (Balearic Islands, Western Mediterranean) was done. The comparison was conductedusing the data from two moorings, one placed in the middle of the Balearic Current, in the Balearic subbasin(herein, Sóller) and the other in the Mallorca Channel, near the Algerian subbasin (called Cabrera). The instru-ments moored, continuously recorded the temperature, salinity and currents at different depths, for over15 months. The data analysis suggests that Sóller is hydrodynamically more active than Cabrera, at least duringthe time of recording the measurements. The mean currents were higher at Sóller than at Cabrera at all depths,also showing greater maximum speeds and variability. In addition, the presence of more mesoscale eddies inSóller became evident from the altimetry data. These eddies were not only significantly more energetic nearthe surface, they also generally reached to greater depths, affecting the velocities of the seabed currents. Subse-quent to each significant eddy episode, strong changes in temperature and/or salinity were observed, along theentire water column. Spectral analysis revealed the presence of high frequency oscillations with periods of a fewhours. One energy peak, with a period around 3.7 h, was observed at both locations, probably related to trappedwaves aroundMallorca or the Balearic Islands,while others (3 h and 2 h)were reflected only in Sóller, suggestingthey could be associated with some standing resonance waves between the Iberian Peninsula and Mallorca.

© 2014 Elsevier B.V. All rights reserved.

1. Introduction

The Balearic Islands, located in the Western Mediterranean, are thenatural limit between the Balearic and the Algerian subbasins (Fig. 1).These two areas are connected via channels between the islands. TheIbiza Channel (80 kmwide, with amaximumdepth of 800m) is locatedbetween Cape La Nao (Iberian Peninsula) and Ibiza Island; the channelbetween Ibiza and Mallorca is known as the Mallorca Channel (80 kmwide and 600m deep); meanwhile theMenorca Channel, the shallowest(100 m deep) and narrowest (35 km wide) is located between theMallorca and Menorca Islands (García et al., 1994).

The mean circulation in the Balearic subbasin, situated north of thearchipelago up to the Iberian Peninsula, is driven by a density gradientbetween the fresher coastal water and the relatively saltier water inthe center of the subbasin. Due to this density gradient, the NorthernCurrent (NC) flows southward along the Iberian Peninsula slope,following the isobaths.When the NC reaches the Ibiza Channel, two dif-ferent behaviors are possible, depending on the mesoscale situation(Pinot et al., 2002). Normally, a portion of the NCmay leave the Balearic

rquès, n°21, 07190 Esporles,

es).

ights reserved.

errat, S., Hydrodynamic comp014.01.005

subbasin via the Ibiza Channel andMallorca Channel while the rest getsreflected northward to form the Balearic Current (BC) which flowsalong the Islands' northern slope. The resident water between 200 mand 700 m is mostly the Levantine Intermediate Waters (LIW). Whenthe previous winter was particularly cold, the Western MediterraneanIntermediateWaters (WIW) generated in the Gulf of Lion and normallylocated between 100 m and 300 m, could have displaced the LIWforming the resident water at the intermediate depths (Mertens andSchott, 1998; Millot, 1999; Pinot et al., 2002). When these WIW reachthe Balearic Channels in late spring, the NC most often gets blockedand a different configuration of flow through the channels is observed.Most of theNC reflects northwards, reinforcing the BC, and the southwardflow through the channels is reduced (Monserrat et al., 2008). TheWestern Mediterranean Deep Waters (WMDW) are located in thedeeper part of the water column, below 700 m (refer Massutí et al., inthis issue; Millot, 1999; Pinot et al., 2002 for a better description ofthe circulation and the properties of these water masses).

TheAlgerian subbasin, the zone between the Balearic Islands and thenorth of Africa, is controlled by a completely different dynamics. Thecirculation of the Algerian subbasin is driven by the Algerian currentwhich interacts with the Alboran Sea eddies and spreads the Atlanticwater towards the Balearic Islands. The northern part of the subbasinhas no clear steady current, although its circulation is affected by thechanges in themesoscale structures in the southern part of the subbasin

arison between the north and south of Mallorca Island, J. Mar. Syst.

http://dx.doi.org/10.1016/j.jmarsys.2014.01.005

mailto:[email protected]


http://www.sciencedirect.com/science/journal/09247963


1oW 30’ 0o 30’ 1oE 30’ 2oE 30’ 3oE 30’ 4oE 30’ 5oE 38oN

30’

39oN

30’

40oN

30’

41oN

30’

42oN

Ebro Riverdelta

CapeLa Nao

IbizaChannel

MallorcaChannel

MenorcaChannel

Mallorca

Balearic

Subbasin

Algerian Subbasin

NC

BC

NC = Northern CurrentBC = Balearic Current

S C Mooring position

S

C

8oW 4oW 0o 4oE 8oE

36oN

38oN

40oN

42oN

44oN

Iberi

an P

en.

Balearic Islands

0 20 40 60 80 100

−1.5

−1

−0.5

0

km

CTDCTD + current meter


Mal

lorc

a

CabreraC

−1.5

−1

−0.5

0

Dep

th (

km)



Mal

lorc

a

SóllerS

Fig. 1.Map showing the main characteristics of the ocean circulation in the Balearic subbasin. The positions of themoorings are markedwith an S inside a star for Sóller mooring and a Cinside a square for Cabreramooring. Isobaths are plotted between 500m and 2500mwith a step of 500m. Black arrows indicate the permanent currents, while the gray ones indicate thetemporal features. The enclosed areas are 0.5° radius circles where the occurrence of eddies is checked (they seem to appear as ellipsoids due to the map projection). The bathymetricprofiles correspond to the red lines indicated in the map. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)

2 A. Amores, S. Monserrat / Journal of Marine Systems xxx (2014) xxx–xxx

and the entrance of Mediterranean water coming in from the Balearicsubbasin via the channels.

The main objective of this study, which is included in the IDEADOSproject (Massutí et al., in this issue), is to analyze the similarities anddifferences in the hydrodynamic conditions between two zones of fish-ing interest, one situated in the Balearic subbasin, to the north ofMallor-ca Island and the other in the northern part of the Algerian subbasin, tothe south of Mallorca. The results found are to be of significance forIDEADOS when assessing the manner in which the variability of thefishery resources in both areas might be driven by a different hydrody-namic frame. With this objective in mind, a mooring line was deployedin each zone to retrieve a set of hydrodynamic data to facilitate a com-parison. One instrument line was located in the Balearic subbasin,near Sóller, while a second one was placed in the Mallorca Channel, inthe northern part of the Algerian subbasin, close to Cabrera Island. Theterms Sóller and Cabrera are used to identify these two regions through-out the manuscript (Fig. 1).

In thiswork,we first present themooring lines, instrumentation andthe data setmeasured. Then, themethodology used is described and themost interesting results are shown. A detailed comparison between thetwo zones is elaborated. The last section includes the summary andconclusions drawn.

2. Data and methods

Hydrodynamic data were acquired using two identical mooringlines. The line from Sóller was placed on the northwest side of MallorcaIsland (39°49.682′ N–2°12.778′ E) and the line from Cabrera was lo-cated on the southwest side of the island, into the Mallorca Channel(38°59.484′ N–2°28.907′ E) (Fig. 1).

Themoorings were deployed at about 900m depth to a height of 600m above the seabed. Each mooring consisted of four CTD (conductivity,temperature and depth) Seabird 37 sensors placed at fixed depthsat around 300 m, 500 m, 700 m and 900 m. In addition, two NortekAquadopp current meters were installed in the middle levels (atabout 500 m) and near the bottom (around 900 m depth). A sedi-ment trap (not used in this study) was placed 30 m above the bot-tom. Observations of the thermohaline properties and currents were

Please cite this article as: Amores, A., Monserrat, S., Hydrodynamic comp(2014), http://dx.doi.org/10.1016/j.jmarsys.2014.01.005

collected at sampling rates of 10 min for the CTD and 30 min for thecurrent meters.

The moorings were continually recording data frommid-November2009 until mid-February 2011. During this period, two mainte-nances were required because of the selected sampling rate forthe sediment traps. These maintenances took place in mid-Marchand in mid-September 2010. The depth where the instrumentswere located did not change significantly after the maintenances.The instruments encountered no significant problems during thewhole period. The only exceptions were the CTDs at 500 m and900 m in Sóller and the one at 300 m in Cabrera which ran out ofbatteries around January 2011, about 1 month prior to the end of theexperiment.

Simultaneous surface information was obtained from the AVISOservice available at http://www.aviso.oceanobs.com, which providesgridded Sea Surface Height (SSH) fields with a map sampling of 1/8° ×1/8°. The absolute dynamic topography is obtained as the sum of thesea level anomalies provided, measured by the altimetry satellites andthe mean dynamic topography based on 7 years of observations(1993–1999) (Rio et al., 2007). The regional sea level anomalies forthe Mediterranean Sea are a multi-mission product with up to 4 satel-lites at a given time, spanning the period from 1992 to the present.The data provided are corrected to include all standard geophysicalcorrections, including Dynamic Atmospheric Correction (DAC). Thislatter correction combines the high frequency of the Mog2D model(Carrère and Lyard, 2003) with the low frequency of the classicalinverted barometer correction.

The formation and evolution of the eddies in the region of interesthave been studied by applying an automated eddy detection scheme(Nencioli et al., 2010). This scheme is a flow geometry based scheme.The method fixes the center of the eddy as the local velocity minimuminto an area with rotating flow, and the eddy boundaries are defined asthe outermost closed streamline around the center, for which the veloc-ity continues to radially increase.

In order to improve the algorithm performance of the automaticmethod of detecting the eddies, the AVISO velocity fields are linearlyinterpolated from the 1/8° × 1/8° grid to 1/16° × 1/16° as performedby Liu et al. (2012). The number of grid points from a reference one



3A. Amores, S. Monserrat / Journal of Marine Systems xxx (2014) xxx–xxx

fromwhich the increase in themagnitude of the velocity is checked, hasbeen fixed to 3 (parameter a of the algorithm) and the number of gridpoints of the area used to find the velocity local minimum has been se-lected to be 2 (parameter b of the algorithm).

The equivalent radius of each eddy located has been calculated asthe radius of the circumferencewith the same areawhile the circulationis the vorticity integral within the eddy area.

The computed spectra have been performedwith a half-overlappingKaiser-Bessel window of 2048 points, the initial number of degrees offreedom being 14. These spectra have also been smoothed, averagingper frequency bands, which permitted an increase in the degrees offreedom towards the high frequencies. Wavelets used Morlet as themother function.

3. Results

The temperature and salinity time series measured by themooringsduring the time of deployment are shown in Figs. 2 and 3. The firstcolumn shows the results from the Sóller mooring, whereas the secondcolumn lists the measurements from Cabrera. Each row is assigned to a

13.300

13.160

13.020

12.880

12.740

12.600

13.300

13.250

13.200

13.150

13.100

13.050

13.180

13.149

13.118

13.087

13.056

13.025

13.120

13.102

13.084

13.066

13.048

13.030

Jan10 Apr10 Jul10 Oct10 Jan11

Sóller

1 2

a

c

e

g

3 4 5 6 7 8 9

Tem

pera

ture

(o C

)

Fig. 2. Temperature time series after having been filtered with a 24 h low pass running averagecolumn shows the data registered at Cabrera. Each row corresponds to an increasing depth (30ison with other diagrams. Black patches indicate when an eddy was detected. The gray patch i


particular depth in increasing order (300 m, 500 m, 700 m and 900 m).The time series colors denote time, starting with the purple color andmoving linearly to blue. This time color scale enables us to identify thetime of the relevant features in the TS and PVD diagrams shown lateron. This color scheme is followed throughout the manuscript.

A visual inspection of the temperature and salinity time series showsgreater amplitude oscillations in Sóller. This observation is quantified inFig. 4 where the ratios between the standard deviations of the seriesmeasured at the Sóller and Cabrera locations are shown. All the instru-ments (the only exception being salinity at 900 m) presented standarddeviations significantly greater in Sóller than in Cabrera, with the max-imum value at 500 m depth. This result suggests that Sóller is a regionwith greater variability than Cabrera.

Only eddiesmeeting the selection criteria as detected by the automat-ed scheme (Nencioli et al., 2010), are shown in Fig. 5. The detection re-quirements specify that eddies have a lifespan of more than 10 days andtheir centers remain for at least onewhole day in the selected area (ellip-soids in Fig. 1), similar to the criteria applied by Amores et al. (2013). Thepresence of the eddies is also shown in Figs. 2 and3, usingblack patches inthe lower parts of the frames. The eddies are also numbered.


900m700m

500m300m

b

d

f

h

1 2 3 4

Cabrera

filter. The first column shows the measurements from the Sóller mooring and the second0m, 500m, 700m and 900m). Colors indicate the time evolution to facilitate the compar-n the Sóller time series indicates the presence of two eddies at the same time.



38.550

1 2 3 4 5 6 7 8 9 1 2 3 4

a b

c d

e f

g h

Salin

ity38.470

38.390

38.310

38.230

38.150

38.540

38.528

38.516

38.504

38.492

38.480

38.510

38.504

38.498

38.492

38.486

38.480

38.493

38.490

38.488

38.485

38.483

38.480

Jan10 Apr10 Jul10 Oct10 Jan11 Jan10 Apr10 Jul10 Oct10 Jan11

900m700m

500m300m

Sóller Cabrera

Fig. 3. Similar to Fig. 2 but for the salinity time series.

0 0.5 1 2 3 4

300

500

700

900

σSóller

/σCabrera

Dep

th (

m)

TS

Fig. 4.Ratios between the Sóller and Cabrera standard deviations at different depths. Tem-perature standard deviation ratios are shown in blue and the salinity standard deviationratios in red. (For interpretation of the references to color in this figure legend, the readeris referred to the web version of this article.)



The total number of eddies detected is around twice the number inSóller (9) than in Cabrera (4) and almost all of them are anticyclonic.The mean of the absolute value of the circulation of the total numberof eddies (being the circulation sum of the vorticity inside the eddyarea during its entire lifetime) is around 1.5 times greater in Sóllerthan in Cabrera. The duration of the eddies remains almost the samein both areas (~19 days), although the mean equivalent radius is 5 kmgreater in Cabrera than in Sóller.

Observation of the temperature and salinity time series (Figs. 2and 3) recorded when the eddies remained in each area, clearly re-veals that the eddies in Sóller generally extend their effects downto the deeper levels than the eddies detected in Cabrera. One exam-ple of this fact is the significant drop in the temperature and salinityregistered in Sóller at 300 m between November and December2010, coinciding with the last eddy detected in this region (No. 9).This decrease at 300 m changed into an increase of both the variableswith depth, down to 900 m, which ensured the stability of the watercolumn (refer to Amores et al., 2013 for more details). The sameeffect, although of smaller magnitude, has also been observed inSóller, at least during the third and sixth eddies. Conversely, this be-havior has never been observed in the Cabrera time series of



Fig. 5. The first two panels show themean features (circulation, radius and duration) of all the eddies detected in each zone. Each following frame shows the features of every individualeddy for Cabrera (in red) and Sóller (in blue) with the same numbering as from Figs. 2 and 3. The trajectory from the starting point in green to the ending point in yellow is also indicated.The enclosed areas are the zones where the occurrence of eddies has been checked. Note that the sixth and seventh eddies in Sóller were simultaneous for some time but they have beenplotted in different frames. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)


temperature and salinity, where the eddy effects were not noticeablebelow 500 m.

The TS diagrams for the entire period of study are shown in Fig. 6.The Sóller measurements spread out across a wider area than forCabrera, which again reveals the greater variability present at Sóller.

The eddy episodes mentioned above are clearly reflected in Sóller at300m. The dark green portionmoving from the Levantine IntermediateWaters (LIW, the resident water at this depth) to theWestern Mediter-ranean Intermediate Waters (WIW) correspond to the eddy describedin Amores et al. (2013). The red–pink area, which indicates two differ-ent eddies (Nos. 2 and 3), are clearly separated into two parts. Thefirst part, which reveals a change in the salinity, however, maintainingthe temperature more or less unchanged, corresponds to the entranceof waters revealing Atlantic features. The second part, when changesin both salinity and temperature are seen, is related to the WIW.

Moving downwards in depth, from 500 m to 900 m, the characteris-tics of the water change gradually from LIW to Deep Waters (DW), anormal response for the resident waters of this area (Balbín et al., 2012;Pinot et al., 2002).

On the other hand, the Cabrera 300 m TS diagram reveals that thecharacteristics of the waters in this region were confined to a smallerarea (when compared with Sóller). Water properties deviated fromthe normal LIW (the resident water) toWIW only in the green portion,which corresponds to the presence of some eddies in this region.When


moving to the deeper levels, a slight displacement from LIW to DW isobserved. However, the water properties do not get as close to DW asthey do for the Sóller region.

A comparison of the Progressive Vector Diagrams (PVDs) and thespeed for each depth in which a current meter was deployed (500 mand 900 m) is presented in Figs. 7 and 8, respectively.

The current mean values are greater at Sóller (3.9 cm/s at 500m and5.2 cm/s at 900 m) than they are at Cabrera (2.3 cm/s at 500 m and2.2 cm/s at 900 m). Significantly greater speeds have been recorded at900 m than at 500 m in the Sóller region. Otherwise, at Cabrera, thespeeds are observed to be similar at both depths. Greater speeds inthe water column have been previously observed close to the bottom(Millot, 1994) although the exact reason for this phenomenon stillremains unclear (Pinot et al., 2002). A possible explanation for thesemeasurements could be attributed to the fact that the Sóller mooringhas been deployed along a very steep slope, whereas the Cabrera moor-ing is located in a relatively flat area (Fig. 1). The placement of the Sóllermooring, the steep slope, and the fact that the Balearic Current flows infrom the Northern Current, after experiencing a significant gyre just be-fore the Ibiza Channel, could explain the greater velocities measurednear the bottom.

The PVDs at Sóller (Fig. 7a and b) clearly reflect the steady BalearicCurrent, flowing northeast along the islands slope. The dominant cur-rent is only disturbed by the presence of successive eddieswhich trigger



13.250 a

LIW

LIW

LIW

LIW LIW

LIW

DW DW

LIW

LIW

DWDW

DW DW

WIW WIW

b

c d

e f

g h

13.075

12.900

12.725

12.550

13.225

13.169

13.113

13.056

13.000

13.070

13.030

12.990

12.950

12.910

13.000

12.970

12.940

12.910

12.880

38.480 38.483 38.486 38.490 38.493 38.480 38.483 38.486 38.490 38.493

38.480 38.487 38.495 38.502 38.510 38.480 38.487 38.495 38.502 38.510

38.480 38.495 38.510 38.525 38.540 38.480 38.495 38.510 38.525 38.540

38.150 38.250 38.350 38.450 38.550 38.150 38.250 38.350 38.450 38.550

300m500m

700m900m

Sóller Cabrera

Fig. 6. Temperature–Salinity (TS) diagrams. Data have been smoothed by using a 24 h low pass running average filter. The colors are the same than those used in Figs. 2 and 3 and theyindicate the time evolution.


a change in the direction of the current (Fig. 7) as well as an increase inthe speed values in almost all the cases (Fig. 8). These gyres, associatedwith the eddies, are stronger at 500 m but still noticeable at 900 m, atleast for the third and ninth eddies. It should be stated that the dimen-sions of the Sóller PVD frame at 900 m are twice the size of the 500 mframe. This fact, which implies larger displacements at 900 m than at500 m, takes into account two different effects: first, the direction ofthe current is more constant at 900 m than at 500 m; and second, the


speed at 900m is also generally greater than at 500m. The same resultsare also observed in Fig. 8.

The PVDs from Cabrera, however, do not show a clear dominant di-rection for the currents, which is an intrinsic feature of the deploymentarea, due to the significant effect of the mesoscale events in this regionon the distribution of the currents (Pinot et al., 2002). Only a dominantnorthwards current is observed at 500 m during the initial period. Also,neither the PVDs (Fig. 7) nor the speed time series (Fig. 8) present a



−100 0 100 200 300 400 500 600 700 800

0

100

200

300

400

Dis

tanc

e (k

m)

Distance (km)

c

Sóller 500m

Cabrera 500m

Cabrera 900m

Sóller 900m

−50 0 50 100 150 200 250 300 350 400

−50

0

50

100 a

Sóller 500m

−30 −20 −10 0 10 20

−10

0

10

20

30 d

Cabrera 900m

−80 −60 −40 −20 0 20

0

20

40

60

80b

Cabrera 500m

Fig. 7. Progressive Vector Diagrams (PVDs) for the current meters of both moorings. Data have been smoothed by using a 24 h low pass running average filter. The scale relationshipbetween the PVD scales is visible inside the Sóller 900 m PVD. The colors are the same than those used in Figs. 2 and 3 and indicate the time evolution.


clear signal of the eddies observed in this region on the surface data.Conversely to Sóller, the PVD frame from 500 m is double the sizethan that of the 900 m PVD frame, although in this case it is due onlyto the previously mentioned dominant northwards direction of thecurrent during the first part of the measurements, as the speeds showsimilar values at both depths.

Finally, a comparison between the Sóller and Cabrera PVD framesizes reveals, once more, that the current speeds are significantly greaterin Sóller, and also that amore constant current direction is present,whichis related to the steady Balearic Current.

The differences between the current distributions in both zones be-come even clearer on analyzing the number of times a given direction ismeasured and the mean speed for this orientation (Fig. 9). In Sóller,there is a predominant direction, around 70° (taking the zero at thenorth and increasing clockwise) at both the 500 m and 900 m depths.This direction coincides with the orientation of the slope of the islandand, consequently, with the path of the Balearic Current. This


0.05

0.10

0.15

0.20

0.25

Spee

d (m

/s)

b

0

0.05

0.10

0.15

0.20

0.25

1 2 3 4 5 67 8 9a

Sóller

Fig. 8. Speed time series for the current meters of bothmoorings (panels a and c are from 500mpass running averagefilter. The colors are the same as those used in Figs. 2 and 3 and indicate thetime series indicates the presence of two eddies at the same time.


predominant direction shows a higher frequency at 900 m than at500 m, which indicates a greater variability in the upper layers. More-over, there is a second maximum in the opposite direction (250°–260°) which takes into account the reversal of the currents during theeddy events. Regarding the velocity distribution, the same two maxi-mums have been observed for the very same orientations, with almostthe same mean speed values (between 5 and 6 cm/s at 500 m and 6and 7 cm/s at 900 m).

The distribution of the speeds and directions is totally different atCabrera. There is no predominant direction and the mean speed is al-most the same in all directions. Values are also similar at both depths(between 2 and 3 cm/s).

The frequency variability has been studied through wavelets andspectral analysis.

Wavelets for the velocitymodulus and thewhole set of available dataare shown in Fig. 10. The inertial frequency (f ~0.0534 h−1~18.74 h) isthe more energetic component during the whole period, being always


900m

d

1 2 3 4

500m

c

Cabrera

depth and b and d are from 900m depth). Data have been smoothed by using a 24 h lowtime evolution. Black patches reveal when an eddy is detected. The gray patch in the Sóller



Direction

Freq

uenc

y (%

)

900m

0º 45º 90º 135º 180º 225º 270º 315º 360º0

1

2

3

4

5

Sóller500m

0

1

2

3

4

5

900m

0º 45º 90º 135º 180º 225º 270º 315º 360º

Speed (cm/s)

0

1

2

3

4

5

6

7

Cabrera500m

0

1

2

3

4

5

6

7

Fig. 9. The number of times that a direction wasmeasured (black histogram) and the corresponding mean speed for this orientation (red line) for each location and depth. (For interpre-tation of the references to color in this figure legend, the reader is referred to the web version of this article.)


more energetic in Sóller than in Cabrera. Moreover, sudden energyspurts are observed in Sóller wavelets during the time when most ofthe eddies have been detected. At these times, the energy appears tospread outwards affecting the lower frequencies, these frequenciesbeing related to the characteristics of the eddies.

These energy spurts in the speed wavelets are also reflected in thepressure wavelets (Fig. 11). A clear energetic increase at the inertial fre-quency is observed during the most energetic eddies (third and ninth).These energy increases are related to the depressions of the instrumentsdue to the tilting of the mooring associated with the strong currents.This effect can be better visualized by performing a spectral analysis oftwo different periods of time: the first period (from November 18,2009 to March 14, 2010) includes the presence of two intense eddiesin Sóller (eddies 2 and 3) (Fig. 12a); meanwhile, the second period(from April 1 to July 26, 2010), does not show any particular feature,either in Sóller or Cabrera (Fig. 12b). Thus, to correctly interpret thepressure spectra in Sóller, it must be assumed that the currents

Fig. 10. Velocity wavelets from Sóller at 500 m (a) and 900 m (c) and for Cabrera at 500 m (bmooring maintenances. The mother function used in wavelet computations is Morlet.


associated with the eddies are strong enough to significantly tilt themooring. Therefore, the inertial frequency and its harmonics presentin the velocity field (Fig. 10), also becomeapparent in thepressure spec-tra (Fig. 12a). The energy observed at the inertial frequency and har-monics decreases as the depth of the instrument's location increases(i.e. as the instrument gets closer to the sea bottom). This occurs be-cause as the vertical distance to the bottom decreases, the vertical dis-placement due to the tilting becomes less. The inertial frequencypeaks almost disappear in the instrument located near the bottom(900 m). These depressions occur only at some specific moments witha maximum value of around 4 m (see Fig. 7f in Amores et al., 2013),which ensures that the water property (i.e. salinity and temperature)measurements are not significantly affected by these depressions. On an-alyzing a period without eddies, (Fig. 12b) no significant differences areobserved in the pressure spectra for the four available depths, as expect-ed. Besides the tidal components, three high frequency peaks can also beidentified (3.7 h, 3 h and 2 h). The origin of the high frequency peaks

) and 900 m (d) during the entire recording period. The vertical strips correspond to the



Fig. 11. Pressure wavelets from Sóller at 300 m (a) and 900 m (c) and for Cabrera at 300 m (b) and 900 m (d) during the entire recording period. The vertical strips correspond to themooring maintenances. The mother function used in wavelet computations is Morlet.


observed may be better interpreted when the Sóller and Cabrera spectraare plotted together. Spectra for the selected periodwith no eddy and forthe instrument located at 900m are shown in Fig. 13. Similar results (notshown) are obtained at different depths.

Although a suitable analysis should be performed using a harmonicanalysis, the computed spectra suggests that the tidal components aresimilar, although not identical, in the two zones. Semi-diurnal (SD)and quarter-diurnal (QD) components are clearly more energetic inSóller; however, the diurnal (D) and third-diurnal (TD) tidal compo-nents have the same energy content in both regions.

Regarding the high frequency energy contents, the peak at 3.7 h ap-pears in both the Cabrera and Sóller spectra. This has always been ob-served when other time periods are analyzed. Therefore, this surelycorresponds to some trapped wave surrounding Mallorca Island or thewhole Balearic archipelago. A numerical computation could help to con-firm or discard this hypothesis.

Fig. 12. Pressure spectra for Sóller, at all available depths, for a period when an eddy was prto 14/3/2010) and when no eddy was observed in the zone (b) (from1/4/2010 to 26/7/2010origin are indicated in red; the peaks related with the inertial frequency are indicated in greferred to the web version of this article.)


The peaks at 3 h and 2 h appear only in the Sóller spectrum, indi-cating they are probably related to some resonance between theBalearic Islands and the Iberian Peninsula. Again a numerical modelcould help to assess this hypothesis; however, by simply analyticallysolving the 1D shallow water equations we can see that the resonantmodes between Mallorca and the Iberian Peninsula are expected tobe of this order. Results for the 800 m constant depth and the200 km length profile (which are the mean dimensions of the zonebetween Mallorca Island and the Iberian Peninsula), the fundamen-tal period of resonance is found to be around 1.25 h. A betterapproximation, obtained numerically solving the shallow waterequations in one dimension using the real bathymetry of the zone(Fig. 1), gives as a result that the first mode of resonance is around2.18 h. This result shows that the 3 h and 2 h peaks found in the Sóllerspectrum are of the same order that the expected resonance modesbetween the Balearic Islands and the Iberian Peninsula. However this

esent in the area exerting a significant influence on the mooring (a) (from 18/11/2009). The main tide components are indicated in gray; the peaks with a possible resonantreen. (For interpretation of the references to color in this figure legend, the reader is



Fig. 13. Comparison between the 900 m depth pressure spectra from Sóller (red) andCabrera (blue) during the period between April 1 and July 26, 2010. Themain tide compo-nents are indicated in gray; the peaks with a possible resonant origin are indicated in red.(For interpretation of the references to color in this figure legend, the reader is referred tothe web version of this article.)


result is, at this stage, only speculative, and additional data and the useof numerical computations would be needed in order to confirm thesource of the peaks observed.

4. Summary and conclusions

Twomooring lines were deployed, one to the north and the other tothe south of Mallorca Island for over a year (November 2009–February2011) in order to study the hydrodynamic similarities and differencesbetween the two zones of fishing interest.

The Sóller mooring was deployed in the middle of a steep slope,dominated by the Balearic Current and located within the Balearicsubbasin. The Cabrera mooring, however, was installed in a flatterzone, near the Mallorca Channel, in the northern part of the Algeriansubbasin.

On analyzing the daily SSH images over the period that themooringswere deployed, several mesoscale eddies were identified, crossing bothregions. The total number of eddies in Sóller was found to be more thantwice the number of eddies detected in Cabrera. The fact that the Sóllermooring was deployed in the middle of the Balearic Current could ex-plain this difference, as the instabilities of the Balearic Current appearedto be the common source generating the eddies (Amores et al., 2013).

This resulted in the greater number of eddies being detected inSóller, as there were more intrusions of waters with properties verydifferent from that of the resident waters (LIW). Depending on theeddy intensity, these water intrusions revealed different ranges indepth, affecting only the surface layers on some occasions, but clearlypresent down to the bottom during the more intense episodes. Con-versely, the eddies in Cabrera appeared to affect the depths from thesurface to only around 300mdepth, as the effects of the eddies detectedat the surface and having some signal at 300 m did not disturb themeasurements at 500 m.

Data showed that the Sóller area was hydrodynamically muchmoreactive than the Cabrera region at least during the time of deployment ofthe moorings. The velocities observed were considerably higher, with


greater variability. This higher variability was also observed in thetemperature and salinity time series. Moreover, a greater number ofeddies was also detected in the Sóller area. Although the currents inthe Cabrera region changed direction continuously, their speed oscilla-tions were lower than those in Sóller.

The constant direction of the currents in Sóller, and the non-near-zero mean values of the currents, makes it difficult to fix a lowerlimit to the Balearic Current. Although it is true that the Balearic Frontis located from the surface up to 200 m (Ruiz et al., 2009) or 300 m(Font et al., 1988), its effects on the currents appear to extend deeper,down to the bottom. This fact becomes significant in two different as-pects. When calculating the dynamic height in this area, a referencelevel of zero velocity at about 500 m is usually taken. However, ourobservations show that this approach is not adequate, as the mean ve-locities around 4.5 cm/s have been measured from 500 m to 900 m.The non-zero speed values at the intermediate and deep layers arealso important while computing the transport in the Balearic Current.The transport is usually calculated by taking an area 35–50 km inwidth and 200–250 m in depth (coinciding with the Balearic Front)and giving them transport values ranging from 0.3 Sv up to 0.9 Sv(Font et al., 1988; García-Ladona et al., 1996; Pinot et al., 2002; Ruizet al., 2009). However, when computing the transport from 300 mto 900 m, with a width half of the width used in the upper layers(17.5–25 km, trying to take into consideration the presence of theslope) and 4.5 cm/s as the mean velocity, values from 0.48 Sv to0.68 Sv are obtained. These transport values obtained for the intermedi-ate and deep layers are comparable to the transport in the upper layersand they should possibly be taken into account for the following com-putations of the Balearic subbasin transport.

Differences in the frequency response in both zoneswere also found.As the eddies observed in Sóller were more intense than those inCabrera, the associated strong currents during at least two episodesclearly caused the tilting of the mooring. The changes in the depth ofthe instruments related to this tilt introduced the inertial frequency inthe pressure spectrum. Some differences between the TD and QD tidalcomponents were also found, showing Sóller having greater energythan Cabrera. Finally, high frequency peaks were found in both sites,which could be related to some kind of resonance in the subbasins.One peak (3.7 h) observed in both locations appears to be relatedto some trapped wave surrounding Mallorca Island or the wholeBalearic archipelago. The other two (3 h and 2 h) observed only atSóller probably reflect resonance modes between the mainland andthe island. More conclusive results could arise from a simulation ofthe 2D shallow water equations or after the use of a numerical modelin the region.

Acknowledgments

This research has been partially sponsored by the IDEADOS project(CMT2008-04489-C03-01 and CMT2008-04489-C03-03). The work ofA. Amores has been funded by a JAE-PreDoc grant from Consejo Superiorde Investigaciones Científicas (CSIC) and co-funded by ProgramaOperativo FSE 2007–2013. The authors also wish to express theirgratitude to the IDEADOS team for their collaboration in the data acqui-sition process and the crews of the R/V Odón de Buen and F/V Punta desVent for their valuable assistance during the mooring deployments.

References

Amores, A., Monserrat, S., Marcos, M., 2013. Vertical structure and temporal evolution ofan anticyclonic eddy in the Balearic Sea (Western Mediterranean). J. Geophys. Res.Oceans 118–4, 2097–2106.

Balbín, R., Flexas, M.M., López-Jurado, J.L., Peña, M., Amores, A., Alemany, F., 2012. Verticalvelocities and biological consequences at a front detected at the Balearic Sea. Cont.Shelf Res. 47, 28–41.

Carrère, L., Lyard, F., 2003. Modelling the barotropic response of the global ocean to atmo-spheric wind and pressure forcing — comparisons with observations. Geophys. Res.Lett. 30. http://dx.doi.org/10.1029/2002GL016473.




Font, J., Salat, J., Tintoré, J., 1988. Permanent features of the general circulation in theCatalan Sea. Oceanol. Acta 9, 51–57.

García, E., Tintoré, J., Pinot, J.M., Font, J., Manriquez, M., 1994. Surface circulation anddynamics of the Balearic Sea. Coastal Estuarine Stud. 46, 73–91.

García-Ladona, E., Castellón, A., Font, J., Tintoré, J., 1996. The Balearic Current and volumetransports in the Balearic basin. Oceanol. Acta 19, 489–497.

Liu, Y., Dong, C., Guan, Y., Chen, D., McWilliams, J., Nencioli, F., 2012. Eddy analysis in thesubtropical zonal band of the North Pacific Ocean. Deep Sea Res. Part I 68, 54–67.

Massutí, E., Olivar, P., Monserrat, S., Rueda, L., Oliver, P., 2014. Towards the Understandingof the Influence of the Environmental Conditions on the Demersal Resources andEcosystems in the Western Mediterranean: Motivations, Aims and Methods of theIDEADOS Project (in this issue).

Mertens, C., Schott, F., 1998. Interannual variability of deep-water formation in thenorthwestern Mediterranean. J. Phys. Oceanogr. 28, 1410–1424.

Millot, C., 1994. Models and Data: a Synergetic Approach in the Western MediterraneanSea. Kluwer Acad., Dordrecht, Netherlands 407–425.


Millot, C., 1999. Circulation in the western Mediterranean Sea. J. Mar. Syst. 20, 423–442.Monserrat, S., López-Jurado, J.L., Marcos, M., 2008. A mesoscale index to describe the

regional circulation around the Balearic Islands. J. Mar. Syst. 71, 413–420.Nencioli, F., Dong, C., Dickey, T., Washburn, L., McWilliams, J.C., 2010. A vector geometry-

based eddy detection algorithm and its application to a high-resolution numericalmodel product and high-frequency radar surface velocities in the Southern CaliforniaBight. J. Atmos. Ocean. Technol. 27, 564–579.

Pinot, J.M., López-Jurado, J.L., Riera, M., 2002. The CANALES experiment (1996–1998).Interannual, seasonal, and mesoscale variability of the circulation in the BalearicChannels. Prog. Oceanogr. 55 (3–4), 335–370.

Rio, M.H., Poulain, P.M., Pascual, A., Mauri, E., Larnicol, G., Santoleri, R., 2007. A mean dy-namic topography of the Mediterranean Sea computed from altimetric data, in-situmeasurements and a general circulation model. J. Mar. Syst. 65, 484–508.

Ruiz, S., Pascual, A., Garau, B., Yannice, F., Álvarez, A., Tintoré, J., 2009. Mesoscale dynamicsof the Balearic Front, integrating glider, ship and satellite data. J. Mar. Syst. 78,S3–S16.



Hydrodynamic comparison between the north and south of Mallorca Island

Documents

Transcript of Hydrodynamic comparison between the north and south of Mallorca Island