Journal of Marine Science | Volume 01 | Issue 01 | April 2019

Journal of Marine Science http://ojs.bilpublishing.com/index.php/jms

ARTICLE Study of the Effective Factors on Shelf Currents in the Southern Caspian

Maryam Shiea1* Caspian Climate Co, Mashhad, Iran

ARTICLE INFO ABSTRACT

Article history The current and wind records and the physical parameter structures such Received: 22 February 2019 as temperature and in the southwestern part of the Caspian Sea adjacent to Anzali Port were investigated from November 2004 up to the Accepted: 26 February 2019 end of January 2005. Results show that, despite the existence of relatively Published Online:30 April 2019 weak winds along the coast in the area, the measurements indicate strong long shore currents. However, when heavy wind tension is observed in the Keywords: area, then strong currents are also present, which - from the perspective Caspian Sea of direction – also have good coordination with the wind. The direction of dominated currents was parallel to the coast from the west to the east. In Shelf currents most cases, the flow rate was identical from the surface to the , and Wind in this condition - because the values of the measured temperature points Low-frequency motions were almost identical-barotropic currents were present. However, in the autumn at some region, a significant difference was observed between the Kelvin waves surface currents and subsurface currents due to temperature differences that affected the density and caused the creation of barocilinic currents. Due to the high velocity of currents compared to wind velocity, and the intense slope of the coast and low-frequency movements in the area, we can hypothesize the existence of motions such as Kelvin waves and conclude that the effect of the wind compared to the other factors of the coastal current in the area was weaker.

1. Introduction about 80% of the climatological mean river discharge of 250 km3yr-1 [4]. The general circulation in the Caspian he Caspian Sea is the largest inland body of wa- Sea is reported to be anti-clockwise (cyclonic) based on ter on the planet, with a surface area of 379000 both indirect estimates of currents [5]. The isolated water 2 3 [1] Tkm and a volume of 78000 km . The Sea can and its inland position of the Caspian Sea are responsible be divided into three distinct basins: a northern (covering for many of the important outer thermohydrodynamic 2 2 80000 km ), a middle (covering 138000 km ) and south- factors, specifically the heat and water fluxes through 2 [2,3] ern (covering 1648400 km ) . There are about 130 the sea surface, river runoff for the variability, rivers that are connected to the Caspian Sea, and most formation of its 3D thermohaline structure, and water of them have small discharge rates. The largest inflow circulation [6]. of fresh water belongs to the Volga river, accounting for Spatially and temporally variable currents in the Cas-

*Corresponding Author: Maryam Shiea, Caspian Climate Co, Mashhad, Iran; E-mail: [email protected]

18 Distributed under creative commons license 4.0 DOI: https://doi.org/10.30564/jmsr.v1i1.550 Journal of Marine Science | Volume 01 | Issue 01 | April 2019 pian Sea are a consequence of the elongated geometry the experimental equations (salinity and density) accord- and strong topography of the basin, influenced by vari- ing to the ionic composition of the Caspian Sea waters able wind forcing and baroclinic effects. Investigations from water samples taken previously from the Sea [14]. done from the end of the 19th century until the 1950s Peeter et al.’s surface water density computation - based show that the overall general circulation of the Sea is cy- on their equation of state – was similar to Millero and clonic, even though strong sea current variability exists Chetirkin’s computation[13]. Therefore, the salinity calcu- [5,7]. In the years 1935–1937 [8,9], six instrumental surveys lated from CTD data by UNESCO standard processes[12] were done on the western coast of the middle Caspian is less than the salinity taken from the chemical methods. basin. The surveys indicated that mostly southward cur- Hence, by using a modification factor of 1.1017, the rents flowed alongside the western coast of the middle researchers were able to arrive at an ideal agreement be- Caspian basin that were influenced by wind-driven cur- tween the amount of salinity (calculated from CTD data rents near the surface. and the UNESCO formula) and the salinity (calculated The aim of this study was to investigate the shelf from the chemical method): current features and the role of wind, density and other forces in formation of surface and subsurface currents. SalinityCas = 1.1017 Salsea (1) In this paper, data measurements were analyzed and con- Where in 1.1017 is the correction factor, Salsea is com- clusions are presented. puted salinity using the UNESCO equations, and Salini- 2. Materials and Methods tyCas is the corrected salinity for the Caspian Sea water. The equation is free of pressure, temperature, and salinity 2.1 Study Area ranges and is versatile enough to be used for different conditions and depths. From November 2004 up to the end of January 2005 the current – aligned vertically to the cost, up to a depth of 200 m was measured in the southwestern area of the Cas- pian Sea between the estuary of the Sefidrood river and Anzali port at the width of the . In order to measure the currents three stations were set up with three current meters (RCM9-MKII). The first mooring station (at 20 m depth) held two devises at 5 and 18 m depth; the second station (at 50 m depth) held two devises at 5 m and 47 m depth; and the third station (at 200 m depth) held three current measuring devises at 3 m, 58 m and 122 m (Fig. 1). In autumn (October) and winter (January) the research- ers also took CTD measurements following the major axis of the current measurements alongside and in the direction of the moorings at the 17 stations (which were at a dis- tance of 1000 m from each other).

2.2 Computation of Salinity

UNESCO has established empirical equations for the Figure 1. Caspian Sea and position of the study area computation of salinity and density from CTD data [10,11]. However, these equations are limited in their validity for 2.3 Principal Component Analysis only water that has an ionic composition of standard sea Principal component analysis (PCA) can be used to iden- water, so these equations cannot be used for the Caspian [12] tify data patterns and pinpoint their similarities and dif- Sea water . In laboratory experiments done by Millero [15] [13] ferences . Theoretically, PCA allows data rotation that and Chetirkin an empirical equation for the Caspian is placed onto a set of orthogonal axes with extremized Sea water was formulated by taking water samples from variation[16]. PCA is supposed to decrease the number of the southern part of the Caspian Sea adjacent to Iran. This variables of interest into smaller component-sets[16]. PCA equation was rendered valid only at surface pressure in a analyzes all the variances in the variables and restructures specific temperature range. Later Peeter et al.[12] adapted

Distributed under creative commons license 4.0 DOI: https://doi.org/10.30564/jmsr.v1i1.550 19 Journal of Marine Science | Volume 01 | Issue 01 | April 2019 them into a new set of components equivalent to the num- that they smooth the data and thus determine the current ber of original variables. trend. According to these figures, the existence of strong long shore currents that flow parallel to the shore can be 2.4 Spectral analysis observed in the absence of strong winds flow parallel to the Spectral analysis was used to partition the variance of a shore. Also, the wind flow along the vertical was stronger current time series as a function of frequency[17]. Spec- than the wind flow parallel to the shore, but the currents' tral formulas were derived to calculate the space-time velocity was much weaker in this direction. Considering (wave number frequency) spectra using time spectral Fig. 3 to 6, the surface currents perpendicular to the shore techniques such as the lag-correlation method, the direct in stations did not have that much linkage or coordination Fourier transform method and the maximum entropy and there was not much similarity between the long shore method [18-20]. To obtain the power spectrum of the mea- currents that flow parallel and perpendicular to the shore. sured currents and to distribute the total variance over Also, no relationship between the currents perpendicular to a range of frequencies the Fast Fourier Transform was the shore and the wind flow perpendicular to the shore has used. The highest frequency (i.e., the Nyquist frequency been observed. The direction of long shore currents that ( fN =1/2∆t ), which can be resolved with a sampling flow parallel to the shore in stations - most of the time - was interval of one hour, is 0.5 cph. The rotary spectra were in relatively high coordination with the wind, which can be formulated by Fofonoff [21]. The rotary spectral analysis observed in Fig.7. This figure also shows the results of the resolves a velocity vector into clockwise or negative principal component analysis (PCA). At the three stations sense (indicated by negative frequencies) and anticlock- for current measurement, in most cases the flow direction wise or positive sense (indicated by positive frequencies) of the long shore currents parallel to the shore are from the components. east to the west. According to the mentioned figures, in par- ticular times when the effect of winds parallel to the coast 3 Results and Discussion was stronger, the influence of the wind on the long shore currents is quite obvious. However, given the high velocity 3.1 Wind and Current variability rate of long shore currents parallel to the shore, relative to the wind speed values​(parallel to shore) and considering According to Fig. 2, the values of longshore parallel that typically the velocity of the currents' wind driving currents along the in the first and second force is about 3 percent of the wind velocity, strong correla- stations are very close and the difference between the sur- tion between wind and surface currents can not be observed face and deep currents are slight. However, the currents’ at the stations. velocity from the coast to the relatively deep areas has been rising significantly, which can be caused by friction of the U components of currents:"station 1" surface current bed in coastal areas. The currents' velocity value of the 100 depth current 50 surface, subsurface, and bottom (or seabed), with a 200 m 0

Velocity (cm/s) Velocity 10/30/2004 12/19/2004 02/07/2005 depth, are significantly different and the currents' velocity Time value at bottom (or seabed) has reduced considerably. At V components of currents:"station 1" this station, the currents' velocity is stronger in fall compare 100 50 to winter. In the winter the velocity difference between the 0 Velocity (cm/s) Velocity 10/30/2004 12/19/2004 02/07/2005 layers, especially the surface and the subsurface layer, is Time reduced considerably, so that the currents’ velocity from the U components of currents:"station 2"

100 surface current surface to depth is approximately the same value. Figures 3 depth current and 4 show the relationship between daily wind speedmean 50 0

Velocity (cm/s) Velocity and surface current speed mean. There was a significant 10/30/2004 12/19/2004 02/07/2005 Time difference between the maximum mean value of daily long V components of currents:"station 2" shore parallel currents’ velocity at this period for stations 100 50

1 to 3 with the value of 65.4 cm/s, 84.6cm/s and 95.7cm/s, 0 Velocity (cm/s) Velocity respectively, compared to the maximum mean value of dai- 10/30/2004 12/19/2004 02/07/2005 Time ly currents’ velocity perpendicular to the shore for stations 1 to 3 with the value of 18.5 cm/s, 12.3 cm/s and 10.9cm/s Figure 2. Top: Current velocities at surfaceand subsurface respectively. (station 1), Middle:Current velocities at surfaceand sub- surface (station 2), Bottom: Current velocities at surface, The moving averages techniques were used in Fig. 5 subsurface and depth layers (station 3) and 6, and the main advantages of moving averages are

20 Distributed under creative commons license 4.0 DOI: https://doi.org/10.30564/jmsr.v1i1.550 Journal of Marine Science | Volume 01 | Issue 01 | April 2019

u components of current and wind "station 1" u components measurment current velocity: "Station1" 100 surface current wind 100 original mov-ave

50 5 50

0 0 0 -50 C u r r e n t ( c m / s ) w i n d v e l o c i t y ( m / s )

c u r r e n t v e l o c i t y ( c m / s ) c m / y ( t c i o l v e t n e r r c u -100 -50 -5 10/30/2004 12/19/2004 02/07/2005 Time 10/30/2004 12/19/2004 02/07/2005 Time

u components of current and wind "station 2" 100 surface current u components measurment current velocity: "Station2" wind

100 original mov-ave 50 5 50

0 0 0

-50 w i n d v e l o c i t y ( m / s ) c u r r e n t v e l o c i t y ( c m / s ) c m / y ( t c i o l v e t n e r r c u -50 -5 C u r r e n t ( c m / s ) 10/30/2004 12/19/2004 02/07/2005 Time -100

10/30/2004 12/19/2004 02/07/2005 Time u components of current and wind "station 3" 100 10 surface current wind u components measurment current velocity: "Station3"

50 5 100 original mov-ave

50 0 0 0 w i n d v e l o c i t y ( m / s )

c u r r e n t v e l o c i t y ( c m / s ) c m / y ( t c i o l v e t n e r r c u -50 -50 -5 10/30/2004 12/19/2004 02/07/2005 Time C u r r e n t ( c m / s ) -100

10/30/2004 12/19/2004 02/07/2005 Time Figure 3. Average daily longshore surface current velocity Figure 5. Longshore surface current and wind velocities and wind at surface layer (blue line: original wind, red line: original

v components of current and wind "station1" current,black line: moving average) 20 15 surface current wind 10 10 5 v components wind velocity 15 original 0 0 mov-ave 10 -5 -10 5 -10 w i n d v e l o c i t y ( m / s ) 0 c u r r e n t v e l o c i t y ( c m / s ) c m / y ( t c i o l v e t n e r r c u -20 -15 10/30/2004 12/19/2004 02/07/2005 -5

Time ) W i n d (/ s m -10

v components of current and wind "station2" 20 15 -15 surface current 10/30/2004 12/19/2004 02/07/2005 wind Time 10 10 v components measurment current velocity: "Station1" 5 30 original mov-ave 0 0 20

-5 10 -10 -10 0 w i n d v e l o c i t y ( m / s ) c ur r e n t v e l o c i t y ( c m / s ) c m / y ( t c i o l v e t n e r c ur -20 -15 -10 10/30/2004 12/19/2004 02/07/2005

Time C u r r e n t ( c m / s ) -20

-30 v components of current and wind "station3" 10/30/2004 12/19/2004 02/07/2005 20 15 Time surface current wind v components measurment current velocity: "Station2" 10 30 10 original mov-ave 5 20

0 0 10

-5 0 -10 -10 -10 w i n d v e l o c i t y ( m / s ) C u r r e n t ( c m / s ) c u r r e n t v e l o c i t y ( c m / s ) c m / y ( t c i o l v e t n e r r c u -20 -15 -20 10/30/2004 12/19/2004 02/07/2005 Time -30 10/30/2004 12/19/2004 02/07/2005 Figure 4. Average daily cross-shore surface current velo- Time v components measurment current velocity: "Station3" 30 city and wind original mov-ave 20

u components wind velocity 10 15 original 0 mov-ave 10 -10 5 C u r r e n t ( c m / s ) -20 0

-30 -5 10/30/2004 12/19/2004 02/07/2005 W i n d ( m / s ) W i n d (/ s m Time -10

-15 Figure 6. Cross- shore surface current and wind velocities 10/30/2004 12/19/2004 02/07/2005 Time at surface layer (blue line: original wind, red line: original current, black line: moving average)

Distributed under creative commons license 4.0 DOI: https://doi.org/10.30564/jmsr.v1i1.550 21 Journal of Marine Science | Volume 01 | Issue 01 | April 2019

3.2 Structure of Temperature and Salinity Temperature Contour 0 22

20 20 (a) 18 Figures 8 and 9 show the vertical temperature and salinity 12 -50 18 structures of the data taken from the 17 stations in autumn 16

8 14 and winter. In autumn there was a thick surface layer due -100 12 Depth(m) to weather temperature decrements and the 10 -150 was around 30 to 40 m. In the width of the layer, tempera- 8 6 ture change was 7°C, with the upper layer at about 19°C, -200 4

0 2 4 6 8 10 12 and the lower level at about 11.5°C. In winter, field data Cross-shelf distance(km) showed (Fig. 9) no season thermocline, but there was a Salinity Contour

0 11.7 11.9 12.5 uniform temperature (9.5°C to 10.5°C) up to a depth of 12.3 100 m (Fig. 9). Little vertical gradient saline change oc- (a) -50 12.2 curred in this coastal region during the four seasons (Fig. 12.3

-100 8 and 9). Due to the structure of the vertical density, the 12 12.4 slight saline change was in agreement with temperature Depth(m) -150 change.

-200 Scatter diagram of surface currents:"Station1" 11.5 40 0 2 4 6 8 10 12 Currents Cross-shelf distance(km) PC 20 Figure 8. Vertical temperature (top) and salinity (down) structure in the study area in October (2004) 0

V (cm/s) Temperature Contour 0 10.5 22 -20 20 (b) -50 18 10 -40 16 -60 -40 -20 0 20 40 60 80 100 120 U (cm/s) 9.5 14 Scatter diagram of surface currents:"Station2" -100

40 8.5 12 Currents Depth(m) 10 PC -150 8 20 7.5 6 -200 4 0 0 2 4 6 8 10 12 Cross-shelf distance(km) V (cm/s)

-20 Salinity Contour

0 12.2 12 12.6 (b) -40 -60 -40 -20 0 20 40 60 80 100 120 12.4 U (cm/s) -50 12.2 Scatter diagram of surface currents:"station3" 40 12 Currents -100 PC 11.8 Depth(m) 20 11.6 -150 11.4 12.4 0 11.2 -200

V (cm/s) 11 0 2 4 6 8 10 12 Cross-shelf distance(km) -20 Figure 9. Vertical temperature (top) and salinity (down)

-40 -60 -40 -20 0 20 40 60 80 100 120 structure in the study area in Januaury (2005) U (cm/s) Scatter diagram of winds 10 Significant differences between autumn and winter ex- Wind 5 PC isted in the vertical transect of velocity. Current velocity

0 measured in this area 3 m from the surface and subsurface layer (58 m from the water surface) is shown in Fig. 2. -5

V (m/s) The depth of these points in autumn on the upper and -10 bottom layer is consistent with thermocline. We also saw -15 significant decreases in current velocity in the subsurface -20 -10 -5 0 5 10 15 20 25 30 layer when Significant differences between autumn and U (m/s) winter existed in the vertical transect of velocity. Current Figure 7. Scatter diagrams of surface current and wind (red velocity measured in this area 3 m from the surface and scatter: current, blue scatter: wind, green scatter: PC) subsurface layer (58 m from the water surface) is shown

22 Distributed under creative commons license 4.0 DOI: https://doi.org/10.30564/jmsr.v1i1.550 Journal of Marine Science | Volume 01 | Issue 01 | April 2019 in Fig. 2. The depth of these points in autumn on the upper nental shelf off Feridoonkenar Bay[19]. and bottom layer is consistent with thermocline. We also 10 spectral of analysis:"station 1" 10 saw significant decreases in current velocity in the subsur- surface face layer when compared to the surface layer (Fig. 2-bot- depth 2.5 day 8 10 day tom).In winter the thermocline layer vanished and surface 10 and subsurface currents were about equal, concluding that

6 this alteration in the current from surface to depth signals 10 temperature change, and due to the vanishing of the win- Rotary spectra(cm/s)2/cph

4 ter thermocline layer, the alteration in temperature in the 10 -0.1 -0.05 0 0.05 0.1 water column was roughly insignificant and consequently frequency(cph) current velocity change in this water column will be very

10 spectral of analysis:"station 2" slight. According to Fig. 8, in the autumn at station 1, the 10 surface measurements of temperature and salinity were conducted depth 10 day 8 2.5 day at a depth of5 mand18 m; both of the sea areas are located 10 in the surface layer (mixed layer). Therefore, it can be

6 concluded that almost homogeneous conditions existed 10 in these areas, and as a result the surface and subsurface Rotary spectra(cm/s)2/cph currents' velocities at this station are equal due to the exis- 4 10 -0.1 -0.05 0 0.05 0.1 tence of the barotropic current. frequency(cph) At the station 2, the measurements were conducted

10 spectral of analysis:"station 3" at a depth of 5 m and 47 m; one area was located in the 10 surface surface layer and the other one located at the thermocline sub surface 10 day depth 8 2.5 day layers. Thus the currents’ velocity difference at these two 10 areas were a little more natural due to the density differ- ences. 6 10 Considering the geometry of the seabed or substrate, Rotary spectra(cm/s)2/cph Rotary and a sharp slope of the study area (Fig. 8 and 9) that 4 10 -0.1 -0.05 0 0.05 0.1 created a pressure gradient, and the rate of the currents at frequency(cph) these three stations were more than what was expected of Figure 10. Top: Rotary spectrum of hourly currents for the currents' wind driving force, we can conclude that a station 1 (depth: 20 m). Middle: Rotary spectrum of hour- part of the strong coastal currents are caused by the costal ly currents for station 2 (depth: 50 m). Bottom: Rotary shape and its severe slope. spectrum of hourly currents for station 3 (depth: 200 m) 3.3 Low Frequency Motion 4. Conclusion The spectral analysis results for the u and v components Based on the researches conducted by Kynsh et al. [20], of the three stations are shown in Fig. 10. It shows that Ibrayev et al[21] and Shiea et al[22], one of the main factors on the west coast of the southern section of the Caspian of the general circulation model of the Caspian Sea is the Sea, the low-frequency motions have considerable energy wind force and then . But due to such that cyclonic and anticyclonic motions have been coastal currents in the area (the western part of the south- observed in a 10-day period. Also, there was a peak of en- ern Caspian Sea area), and the fact that the currents' veloc- ergy from the anticyclonic motion for a period of 2.5 days ity of the shoreline has much more impact than the effect in the area. The spectral analysis charts comparison of of winds on current wind driven force and that the cur- the three stations show an energy increase that can be ob- rents from the surface to the substrate or seabed, in most served from station 1 toward station 3. The existence of cases is bathymetric, we can conclude that other important motions such as Kelvin waves and similar to it in this area factors besides the density gradient and wind stress impact has been reinforced because the direction of the coastal coastal currents. Among the sea factors we can mention currents was from the west to the east (cyclonic) and the the severe costal slope that can cause a pressure gradient spectral analysis has indicated an energy peak for motions and the formation of the currents in the area. with a period of a few days. These motions in the low Also, one can hypothesize the existence of motions frequency that have considerable energy were observed such as Kelvin waves in the area because of the presence along the whole coast of the Caspian and also in the conti-

Distributed under creative commons license 4.0 DOI: https://doi.org/10.30564/jmsr.v1i1.550 23 Journal of Marine Science | Volume 01 | Issue 01 | April 2019 of low-frequency oscillations with high energy that shows [12] Peeters F, Kipfer R, Achermann D, Hofer M, motions between 2 to 10 day periods in this area, and the Aeschbach-Hertig W, Beyerle U, et a: Analysis of direction of the prevailing currents in these stations, which deep-water exchange in the Caspian Sea based on have been from the west to the east. environmental tracers, Research Part I: Acknowledgment: We would like to thank the Depart- Oceanographic Research Papers. 2000, 47(4): 621– ment of Physical of the Iranian National 54. Institute for Oceanography and Atmospheric Science [13] Millero FJ, Chetirkin PV: The density of Caspian Sea (INIO) for providing current and temperature data. waters. Deep Sea Research Part A Oceanographic Research Papers. 1980, 27(3):265–71. References [14] IAEA: Research/Training on the Caspian Sea. Data Report 1995. International Atomic Energy Agency, [1] Froehlich K, Rozanski K, Povinec P, Oregioni B, Vienna. 1996. Gastaud J: Isotope studies in the Caspian Sea. Sci [15] Smith LI. A: Tutorial on Principal Components Anal- Total Environ. 1999, 237–238:419–27. ysis, 2002: 51, 52. [2] Aubrey DG, Glushko TA, Ivanov VA: North Caspian [16] Suz Tolwinski. Statistical Methods for the Geosci- Basin: Environmental status and oil and gas opera- ences and Beyond. University of Arizona, 2007. tional issues, Mobil Oil, 1994. [17] Emery W, E. Thomson R: Data Analysis Methods in [3] Aubrey DG: Conservation of biological diversity of : Third Edition, 2004: 638. the Caspian Sea and its coastal zone. A proposal of [18] Hayashi Y. A: Generalized Method of Resolving Dis- the Global Environment Facility, GEF, 1994. turbances into Progressive and Retrogressive Waves [4] Kosarev AN, Yablonskaya EA: The Caspian Sea. The by Space Fourier and Time Cross-Spectral Analyses. Hague: SPB Academic Pub, 1994. Journal of the Meteorological Society of Japan Ser [5] Terziev FS, Kosarev AN, KerimovAA,(Eds.): Hydro- II. 1971;49(2):125–8. meteorology and Hydrochemistry of . Caspian [19] Ghaffari P, Chegini V: Acoustic Doppler Current Sea, Hydrometeorological Conditions, 1992, vol. Vl, Profiler observations in the southern Caspian Sea: issue 1: 359. shelf currents and flow field off Freidoonkenar Bay, [6] Kostianoy AG, Kosarev AN: The Caspian Sea Envi- Science Discussions (OSD) ,Internet,. 2009, ronment. Springer Science & Business Media; 2005: http://agris.fao.org/agris-search/search.do?recor- 296. dID=DJ2012089181 [7] Bondarenko AL: Currents of the Caspian Sea and [20] Knysh VV, Ibrayev RA, Korotaev GK, Inyushina Formation of Salinity of the Waters of the North Part NV: Seasonal variability of climatic currents in the of the Caspian Sea , Nauka, Moscow (1). 1993: 122. Caspian Sea reconstructed by assimilation of climat- [8] Stockman, W: Investigation of current kinematics at ic temperature and salinity into the model of water the western shore of the central part of the Caspian circulation, Izv Atmos Ocean Phys, 2008;44(2): Sea., vol. 1; Transactions of the Azerbaidjan Scientif- 236–49. ic-Investigative Fishery station, 1938. [21] Ibrayev R, Özsoy E, Schrum C, İ. Sur H: Seasonal [9] Baidin SS, Kosarev AN: The Caspian Sea. Hydrolo- variability of the Caspian Sea three-dimensional gy and hydrochemistry, Nauka, Moscow. 1986. circulation, sea level and air-sea interaction. Ocean [10] Background Papers and Supporting Data on the In- Science (OS). 2010, 6. ternational Equation of State of 1980, Une- [22] Shiea M, Chegini V, Bidokhti AA: Impact of wind sco; 1981: 188. and thermal forcing on the seasonal variation of [11] Background papers and supporting data on the Prac- three- dimensional circulation in the Caspian Sea. tical Salinity Scale, 1978 - UNESCO Digital Library, IJMS,2016, 45(05); Internet. 2016, http://nopr.nis- Internet. cair.res.in/handle/123456789/35087

24 Distributed under creative commons license 4.0 DOI: https://doi.org/10.30564/jmsr.v1i1.550