Mesoscale processes regulating the upper layer dynamics of Andaman waters during winter monsoon

The characteristics of cold core eddies and their effects on the hydrodynamics and biological production in the Andaman seas were studied using in-situ and satellite observations. The existence and spatial extent of the eddy are revealed by the precise structure and patterns of the temperature-salinity (T–S) profiles, nutrients, and chl a. A good illustration of this is the anomaly in sea surface height (SSHA). The Cyclonic Eddies (CE1 and CE2) are monitored using the Okubo-Weiss parameter (− 2 × 10–11/s2) calculated from the satellite SSHA and a geostrophic current centered at 8°N and 92°E and 13°N and 93°E, respectively. At the eastern flank of CE1, measurements are taken in-situ along 8°N and 92.5°–93.5°E. Vertical currents measured with Acoustic Doppler Current Profiles (ADCP) at 0.3 m/s show northward flow along the track. A significant northward flow (0.3 m/s) can be seen in the vertical currents recorded with ADCP, yet a weak southern flow can be seen over the western margin. In addition, the SSHA acquired from altimetry demonstrates the spatial extent, supporting the occurrence of cyclonic eddies. The vertical shear in the horizontal flow is the main contributor to baroclinic instability (Ri 0.0001) in the water column, according to an analysis of the factors that could result in the formation of an eddy. The T–S profiles show that the area contains Bay of Bengal (BoB) water, and there are semi-annual Rossby waves there as well. However, the wind stress curl did not provide a reliable indication of divergence in the region. The region's biological production (chl a) and nutrient distribution (NO2, NO3, PO4, and SiO4) were impacted by the eddy. CE2 is connected to convective mixing processes that take place along the northwest coast of the Andaman Islands as a result of the primarily cold, dry continental air flowing from the north east.


Introduction
The monsoon is reversing, bringing dry continental air from the northeast during winter and wet summer winds over the sea near the Andaman and Nicobar Island chain (Potemra et al. 1991).The region receives a lot of runoff and suspended materials from the Ayeyarwady-Salween River system, which greatly affects the oceanography and hydrodynamics (Robinson et al. 2007).Strong stratification in the area limits vertical mixing, depletes nutrients in the top layers, and ultimately promotes oligotrophy.Although experienced in the summer and winter, seasonal winds of any strength are not seen to exert any divergence or positive curl, and nutrient pumping to enhance biological production is least common in these waters.In comparison to the Arabian Sea and the Bay of Bengal, the sea is less productive.According to Sanjeevan et al. (2011), the average primary production during the fall inter-monsoon is 283.19 mg C/ m 2 /d, followed by the spring inter-monsoon (249 mg C/ m 2 /d), summer monsoon (238.98 mg C/m 2 /d), and winter monsoon (195.47 mg C/m 2 /d).The western section of the island chain exhibits typical BoB characteristics, the northeast is heavily influenced by the Ayeyarwady and Salween River system, and the southeast by different water qualities, according to earlier observations (Salini et al. 2010).The region is the least studied in terms of oceanic processes, and studies done so far to understand the biodiversity and the basin-scale environment related to living resources show that there are no significant or seasonal activities that lead to nutrient pumping to change production patterns.The knowledge of the dynamics of the upper layer has been improved, however, with the development of satellite technologies, particularly the Altimetry and Ocean Colour imageries on mesoscale to basin scale.Explanations on these important processes in the BoB have emerged, particularly in relation to the quantity of eddies and gyres as well as the impact of cyclones, which results in significant mixing in its path (Nuncio and Prasanna Kumar 2012).Eddies are common ocean features that occur in both clockwise and counter clockwise directions, resulting in convergence or divergence at the centre and are mesoscale phenomena.
Mesoscale eddies are important for moving nutrients, heat, and salt through the ocean (Dong et al. 2014) and boosting local production in oligotrophic regions (Hyrenbach et al. 2006), which has an impact on the pattern of production at each trophic level (Bakun 2006).Numerous scholars have proposed various driving mechanisms for the creation of eddies, including Ekman pumping and remote forcing from the equatorial Kelvin wave reflecting off the eastern border as Rossby wave.According to Yu et al (1999), the remotely imposed Kelvin wave energised the westward propagating Rossby wave.According to Yu et al. (1999), the remotely driven Kelvin wave stimulates a westward propagating Rossby wave, which greatly contributes to the variability of the local ocean circulation.Potemra et al. (1991) described the coastal Kelvin wave, which originates at the equator and propagates across the entire western boundary of the area around the Andaman Sea and the Bay of Bengal.They did this using the multilayer model.Based on in situ hydrographic studies, mesoscale eddies are seen in the coastal seas of the Andaman and Nicobar Islands (Hacker et al. 1998and Chen et al. 2013). Burnaprathepart et al. (2010) discussed the Andaman Sea's eddies and how they contribute to increased primary productivity by combining vertical profiles of chl a, key nutrients, temperature, and salinity.However, no in-depth research has been done in this area to explain how eddies (cold and warm cores) affect the overall biological production in the Andaman waters.It is attempted to list various mesoscale processes in this context using SSHA imagery, geostrophic current pattern, and in situ evidences.This study's goals are to locate these processes in the basin, explain the forcing mechanism and how column dynamics and biogeochemistry are affected by it.

Data and methodology
Onboard FORV Sagar Sampada, measurements in situ were made between November 21 and December 14, 2011.Measurements made at stations in the east and west of the island chain have helped scientists understand the environmental characteristics.The transects were at depths of 50, 100, 200, and 1000 m, while the offshore station was at a depth greater than 1000 m.The study was carried out in 60 stations, out of the 60 stations four were eddy stations at 8°N.The Andaman water's hydrography, circulation, and related biological responses were all studied using these data sets.But getting a transect with 4 sites along the eddy (Fig. 1) was the main goal.Through the equipment and sensors mounted on the IRAWS's onboard, the meteorological parameters such as air temperature, air pressure, and humidity were also recorded at intervals of 15 min.Using a SeaBird 911 Plus CTD with Niskin water samplers and a deck unit for data collection, profiles of temperature, salinity, dissolved oxygen, and Sigma-t were collected.Datasets are processed using 1 m bin.Additionally, salinity is generated from water samples obtained using Niskin samplers and verified using a Guildline 8400A Autosal Salinometer.For assessments of dissolved oxygen and nutrients, water samples were taken from standard depths (surface, 10, 20, 30, 50, 75, 100, 120, 150, 200, 300, 500, 750, and 1000 m) using twelve different 10 L Niskin water samplers.Based on averaging (climatological) data from Levitus et al. (1994a, b), temperature-salinity profiles for water mass characteristics have been developed.The depth at which the seawater density (Sigma-t) is 0.2 kg/ Fig. 1 Station Location m 3 greater than the surface density is known as the mixed layer depth, or MLD, which is calculated using CTD profiles (Sprintall and Tomczak 1993).The depth at which the surface temperature drops by 1 °C from the sea surface temperature is known as the isothermal layer depth (ILD), which is the depth of the thermocline's top layer (Kara et al. 2000 andRao andSivakumar 2003).The difference between ILD to MLD is used to calculate the barrier layer's thickness (Lukas and Lindstrom 1991).
The OS II BB ADCP, installed on the ship's hull, operates at a frequency of 76.8 kHz, with a vertical range of operation between 16 and 400 m, to derive the vertical sections of currents.Along the coast, georeferencing is done using the bottom track option; in deeper areas, navigation mode is chosen.The most recent datasets were collected using VmDas in 8 m bins with a 2 s ensemble time.While gathering the raw data, the ship heading and navigational details are also captured.At a depth of 16 m, the first bin record of the current began.WinADCP was used to postprocess the data in earth coordinates for an ensemble duration of one minute.For the analysis, only processed data with a good percentage more than 80% is taken into account.To de-tide the currents using Python software, load the current data obtained from the OS II BB ADCP into a NumPy array.SciPy library was used to identify the tidal components in the data.Once the tidal components have been identified, remove them from the data using the harmonic analysis method by NumPy or SciPy libraries for this purpose.
Winkler titration was used to measure dissolved oxygen.Using a Skalar Analyzer, analyses of nitrite, nitrate, phosphate, and silicate were carried out.With a spatial resolution of 25 km, the wind stress curl (daily) data utilised was obtained from ASCAT and processed by NOAA/ NESDIS; the chl a data was obtained from MODIS Aqua Level 3 with a spatial resolution of 4 km, acquired from the Ocean Color Website, and processed using SeaDas.SST data were downloaded from the Ocean Color website using MODIS Aqua Level 3 at a spatial resolution of 4 km, while SSHA data were collected from AVISO with a temporal resolution of 7 days covering the time period of January 2003 to January 2013.Geostrophic current images from https:// ocean watch.pifsc.noaa.gov was used by SSHA to identify the cold core eddy, which was found to be centred at 7° N and 90° E and to be travelling in a cyclonic manner.From http:// oaflux.whoi.edu, we acquired net heat flux, solar radiation, latent heat flow, and specific humidity.The OSCAR data utilised for the computation of vertical velocity are available at 5-day intervals with a 1/3° spatial resolution, obtained from www. oscar.noaa.gov (Table 1).
There are two techniques to locate the eddies.The first method makes use of SSHA contours and geostrophic currents, which are determined by the following geostrophic equations: where g is the gravitational acceleration, f is the Coriolis parameter, x and y are longitudinal and latitudinal coordinates, and h is the SSHA.Where u and v are the zonal and meridional components of geostrophic currents.
Using the Okubo-Weiss parameter, OW (Okubo, 1970 andWeiss, 1991), the second technique is as follows: where s n is the normal strain component, s s is the shear strain component and w is the relative vorticity .
The negative Okubo-Weiss are expected in the vortex core if vorticity dominates the vortex core.
By assuming a homogenous layer from the sea surface to 50 m below the surface, the vertical velocity at that depth is determined.The vertical velocity at 50 m depth would be because the layer is homogenous and divergence from the surface to the bottom of the homogeneous layer is constant (Pond and Pickard 1983).
An appropriate analytical method for studying multiscale, non-stationary phenomena that take place over a limited time and space domain is the wavelet transform.The wavelet analysis method was employed in this study to examine time series data of oceanographic parameters that contain non-stationary power at a variety of frequencies.Using a number of wavelet functions to convolution the original time series, this technique is used to break down time series into their frequency components and, if possible, identify the prominent modes of variability as well as how those modes change over time.It does this by expanding functions in terms of wavelets, which are produced by translating and stretching a fixed function known as the Mother Wavelet.In the current work, the wavelet is used to analyse the lifespan and frequency of the processes over the course of ten years and to explain the temporal change of SSHA in the eddy region.In an idealised numerical model of the Indian Ocean, Meyers et al. (1993) investigated the propagation of mixed Rossby-gravity waves using wavelet analysis.
The phase speed for long baroclinic Rossby wave is given by ( 7) where g is the reduced gravity term (taken as 0.04 m s ˗2 for the first baroclinic mode), H 0 is the thermocline depth (taken as an annual mean depth of 20 °C isotherm derived from Levitus and Boyer, 1994), f the Coriolis parameter and β = f ϕ , where ϕ is the latitude.

Physical characteristics of the Eddy region
The area is characterised by warm (27.6-28 °C), humid (72-77%), and north-easterly winds, which suggests the presence of northeast monsoon conditions with magnitudes in the range of 10-12 m/s, with comparatively lower speed (10 m/s) in the western part and higher speed (12 m/s) in the eastern part of the eddy (referred to hereafter as CE1).
While the surface salinity (33.00) and density (20.40 kg/ m 3 ) values are similar in coastal and offshore seas, the SST varies between 28.4 and 28.8 °C with lower temperatures along the coast than offshore.The region is occupied by BoB waters, which have temperatures between 28.0 and 28.5 °C, salinities between 33.2 and 33.8, and densities between 20.6 and 20.8 kg/m 3 .These properties of the regional water mass are revealed by temperature, salinity, and density profiles (Chandran et. al. 2018).Warm (over 28.5 °C) and thick isothermal layer (54 m) are present in the western half of the vertical temperature distribution along 8° N (Fig. 2b), which gradually decreases towards the east (20 m).The upsloping of the isothermal layer, which is evident in the subsurface From the surface to 90 m down, the horizontal current structure at 8°N and 92.5° E to 93.5° E displays an uneven current pattern (Fig. 3).The majority of the flow throughout the eastern portion of the 100 km transect is southward (30 km), with a short and weak northward flow adjacent to it.This is followed by a large southern drift of up to 40 m.However, the T-S profiles do not significantly reflect the response to this irregular pattern, therefore the eastern part of the transect (around 60 km) is not taken into account.The flow toward the north and then the flow toward the south on the western flank suggest cyclonic flow direction.The current measured at a depth of 16 m is used to determine the near-surface pattern, and this indicates the presence of a northern component with an eastward-directed magnitude of 0.3 m/s and insignificant speed in the west.However, at 40 m, the current direction changes from northeast to southwest, with a 0.1 m/s decrease in magnitude on the eastern flank and a 0.1 m/s rise on the western flank.Similar patterns are followed by the current at 88 m, while its speed varies from 0.5 m/s in the western part to 0.4 m/s in the eastern part.The T-S profile's upsloping at the same time as this verifies the feature's status as a subsurface cyclonic eddy.The flow is 0.3 m/s northward in the eastern flank and 0.5 m/s southward on the western flank.All 8 m cells of the data up to 88 m deep were examined, and it was discovered that the pattern of the data followed that of the near surface, but with a diminishing magnitude.Since the dataset was shown to have erroneous values below 88 m, hence it was discarded.

Eddy generation mechanism
The wind stress curl, topographic instability, shear flows, baroclinic instability, and the radiation of Rossby waves from poleward propagating coastal Kelvin waves are some of the potential physical mechanisms that could control the eddy (White 1977 andKessler 1990).To determine the local force that contributes to the formation and maintenance of the eddy, the daily wind stress curl is evaluated.As a result, the contribution from wind stress curl is disregarded.Curl of the eddy zone from ASCAT wind data indicates negative values in the range of − 5.6 × 10 -8 and − 8.24 × 10 -8 Pa/m, indicating convergence (Fig. 4).
Differential mixing with the nearby sea, especially by intake from the Malacca Strait, and freshwater ingress from rivers, which results in large density changes in the water column, are other potential eddy generation sources.This variance may lessen or increase the mechanical impacts in the area, such as eddy currents or meanders.Based on the estimated Richardson Number, this is measured (Ri).Miles (1961) asserted that the flow is steady if Ri > 0.25.
where N 2 is the Brunt Vaisala frequency (BV), and its formula is where g is the gravitational acceleration, ρ 0 represents the average sea water density, z represents depth, σ t is ρ˗1000 where ρ is the sea water density.The velocity gradient, which is the denominator term in Eq. ( 9), is a measure of the strength of mechanical generation determined from vertical current profiles obtained using ADCP.3.165 × 10 -5 s −1 ) and significant velocity gradient (avg.3.968 s −2 ).These cause the water column to become unstable and encourage eddy-like disruption in the area.
Shear flows or the mixing of various water masses are the two main causes of instability.Since the T-S profiles clearly show that BoB water is present in the eddy zone, mixing with other water masses can be ruled out.Another hypothesis is the presence of shear-causing instability caused by planetary waves, which has been thoroughly documented in this area by Schott et al. (2009) and Rao et al. (2010).Planetary waves affect the near-surface circulation through local and remote forcing.Altimeter data were used to analyse the contribution of such planetary wave influence on eddy production mechanisms, and mapping of planetary wave propagation was done to determine their contribution to regional circulation.According to Yu (2003), the planetary wave was tracked using the Hovmuller diagram of SSHA at 8° N along with 89° E to 94° E, and the results are presented (Fig. 6).From mid-November to mid-January, low SSHA in the eddy zone denotes the presence of an upwelling mode Rossby wave (Girishkumar et al. 2011).Rossby wave is rapidly propagating, as indicated by the almost horizontal negative SSHA.Negative SSHA displayed a higher slope to the west (closer to the eddy location), indicating a slower propagation.The phase velocity of the westward signal is 0.20 m/s, and it takes the westward propagating signal roughly 45 to 60 days to travel from the coast of the Nicobar Island chain to the core of the eddy region (Potemra et al. 1991).
It is possible that the signal shown in the plot is a Rossby wave formed on the west coast of the Nicobar Island chain because the theoretical phase speed of a Rossby wave at 8°N that propagates westward is determined to be 0.21 m/s.The wave's estimated speed is in line with the theoretical wave speed, and it also holds up well when compared to earlier findings by Yang et al. (1998), Yu (2003), and Girishkumar et al (2011).In conjunction with poleward propagating coastal Kelvin waves, radiation from the west coast of the Nicobar Island chain created the Rossby waves (Potemra et al. 1991).Meanders and eddies in BoB are caused by the baroclinic instability caused by the combination of westward propagating Rossby waves and local wind stress curl (Nuncio and Prasanna Kumar 2012).Kurien et al. (2010) similarly came to the conclusion that baroclinic instability is crucial to meander expansion and eddy formation in BoB using a computational model.According to Sreenivas et al. (2012), mesoscale eddy formation in BoB is primarily driven by coastal Kelvin waves and their radiating Rossby waves from the east.According to Chen et al. ( 2012)'s study of the interannual variability mechanism of the mesoscale eddies in BoB, the eddy activities are sensitive to the baroclinic instability of the background flow rather than directly linked to El Nino Southern Oscillation (ENSO) occurrences.
The data is once again subjected to continuous wavelet transforms with Morlet wave as mother wavelet after Torrence and Compo (1998) in order to determine the periodicity of SSHA.The information in Fig. 7 that semi-annual variability is the predominant form of variation.Due to the influence of the westward-propagating Rossby wave from the coastally confined Kelvin wave, the wave period is more variable in the Andaman waters (Vialard et al. 2009 andNienhaus et al. 2012).The main frequencies can be determined from the power and global wavelet spectra and are in the semi-annual and annual modes.Compared to the semiannual mode, the annual mode seems to have less intensity.

Chemical and biological response of the eddy
The vertical structure of dissolved oxygen (DO) also exhibits changes above 90 m depth concurrent with the thermohaline oscillations.From a depth of approximately 47 m (92.3°E) to 25-30 m at the eastern side of the eddy (93.3°E), the 4.22 ml/L DO contour shoaled.The top nitrate (NO 3 ) concentration exhibits a little upslope toward the eastern edge (93.3°E) and is within the measurable range (0.67-0.98 M).A noticeable amount of phosphate (PO 4 ) was also present in the upper water, with a modest upslope toward the eastern side (0.12 M at 92.3°E and 0.27 M at 93.3°E).Additionally, there was a modest upsloping in the vertical distribution of silicate (SiO 4 ) towards the eastern periphery (0.77 M at 92.3°E to 1.62 M at 93.3°E).As a result, the vertical distribution of nutrients also demonstrated oscillations in the upper water column along with the thermohaline features (Table 2 figure not shown).
The surface chl a distribution shows that the physical and chemical variables do have an impact on the local life.Chl a, which is generated from ocean colour imaging and is 0.5 mg/m 3 in the eddy region (Fig. 8) compared to neighboring regions (0.1 mg/m 3 ), can show the standing stock of the major consumers for the optical depth.The influence of churning caused by the eddy is explained by an increase in this within the eddy in relation to the nutrient values.And this highlights the importance of such mesoscale phenomena that have a crucial impact on production in the Andaman waters.

Satellite evidence (SSHA based) for cyclonic eddies
For the winter monsoon, the distribution of favourable mesoscale production pockets is studied using monthly SSHA and geostrophic current pattern (Fig. 9a-d) (November-February, 2011).This demonstrates the existence of a single cyclonic eddy (CE), which was identified as CE1 during the in situ studies.Positive SSHA between 5 and 9° N with a core at 7°N suggested that CE1 was stronger and that it will propagate from 93 to 86 degrees East within a month (November to December).November and December experience higher eddy intensity, with a negative value of − 0.14 m.In December, CE1 is detected between 86 and 93° E and propagates westward to BoB.By January, it had been entirely replaced from the Andaman waters and displayed a favourable SSHA (0.18 m).However, even in February, BoB waters had a low SSHA that was centred around 86°E.The eddy has an oval form with an axis that faces east and west.In the aforementioned Sects.(3.3.1-3.3.3), the characteristics and generating process of the eddy CE1 are discussed using both in situ and satellite measurements.The SSHA maps also showed a low SSHA pocket with a negative anomaly of − 0.12 m in November, located between 13°N and 93°E.This has the CE2 marking.With an SSHA of 0.12 m in November, the negative anomaly is more pronounced, and its intensity drops in December with an SSHA of 0.10 m.Positive anomaly of 0.16 m replaces the negative anomaly in January.
In this work, the Okubo-Weiss (OW) parameter approach is also used to clearly identify eddies.Because vorticity predominates over strain components at the eddy core, eddies are characterised by a negative OW parameter, whereas strain predominates areas with a positive OW parameter.Isern-Fontanet et al. (2003) claim that the threshold value for defining eddies is represented by closed contours of OW with a value of − 2 × 10 -11 /s 2 .The threshold value was established to be the same as that used by Isern-Fontanet et al. (2003) to define eddies and identify areas where vorticity predominates.The confined contours of OW in Fig. 9 and the organisation of the cyclonic currents supported the existence of an intensified cyclonic eddy at 8°N and 93°E.
The eddy propagation is explained by the weekly ssha plot (Fig. 10).The in-situ data pocket for the first week of December (01-07) lies near the edge of the cold core eddy, which is situated at 8°N and 87-94°E.We sampled between December 6 and 7, 2011.After December 7th, the eddy began to spread to the west.Eddy centre was at 90°E for the first and second weeks before moving to 88°E.
Figure 11 shows the vertical velocity in the Andaman waters at a depth of 50 m, and the eddy region CE1 is distinguished by a higher positive vertical velocity of 0.5 to 1.5 × 10 -5 m/s.This shows that the upwelling process is developing and that the area's isolines are supporting an upward slope.Following the identification of eddies using the SSHA, OW, and geostrophic current maps, the presence of prevailing processes was further validated using SST and chlorophyll.At the centre of cyclonic eddies created by divergent forcing are waters that are rich in nutrients below the surface.These negative SSHA zones are distinguished by their high chlorophyll content and relative coolness.
SST is high in November, which represents the beginning of winter (Fig. 12a), with higher values throughout the Andaman waters region (28.2-28.8°C).In December, the temperatures rise to 27.6-28.8°C (Fig. 12b).Further, during January (Fig. 12c) and February (Fig. 12d), the basin wide temperature is in the range to 27-29 °C and 26-29 °C respectively.Despite the fact that the Andaman waters are generally warm, the cold core eddies found in this region exhibit relatively cool temperatures as a result of the predominate cyclonic flow associated with them.When the eddy approaches the Andaman waters, surface temperatures start to cool, with CE1 recording a temperature of 28.6° C in November.SST drops from 28.6 to 28.2 °C in December before dropping once more to 27.6 °C in January.However, the temperature in February is the same as it was in January.In November, CE2 shows a temperature of 28.6 °C; in December, it drops to 28.2 °C; then, in January and February (26.5 °C), it drops even further to 27 °C.According to Rama Raju et al. (1981) and  Tan et al. (2006), the influx of low salinity waters through the Malacca strait may be the cause of the increase in temperature along the eastern Andaman waters.
Due to the augmentation of nutrients at the surface, a high chlorophyll content is anticipated in the eddy zone.Because these cold core eddies are located in an area with significant biological activity and strong physical and biogeochemical coupling, which results in high chlorophyll concentration, they are significant.Typically, the Andaman seas have lower chlorophyll concentrations and are oligotrophic in nature (Vijayalakshmi et al. 2010).Chl a concentrations in the eddy zone are raised by the presence of cyclonic circulation.Increased chl a levels were seen in the CE1 and CE2 eddy areas as the cyclonic flow advanced.In November, CE1 reported 0.1 mg/m 3 , climbed to 0.8 mg/m 3 during December, and then declined yet again to 0.3 mg/m 3 in January (Fig. 12a-d).In February, the level of chlorophyll fell to 0.2 mg/m 3 .In November, CE2 showed a very low value (0.1 mg/m 3 ); in December, chl a concentrations in the eddy zone started to rise (0.4 mg/m 3 ); and in January, the pattern continued with a concentration of 0.4 mg/m 3 , which fell to 0.2 mg/m 3 in February.
Weekly progress in the wind stress curl (ASCAT) for the pockets was used to confirm the contribution of the wind stress curl to eddy induction.The signal's mode was − 1.47 × 10 -7 Pa/m at CE1, although the curl ranged from − 4.43 × 10 -7 to 1.28 × 10 -6 Pa/m.The wind curl at CE2 has values ranging from − 2.87 × 10 -7 to 2.09 × 10 -6 Pa/m, with -3.25 × 10 -8 Pa/m being the mode.The presence of maximum negative values, however, suggests that wind is not the primary cause of eddy production.
In comparison to other neighbouring places, the surface temperature at CE2 is lower (27-27.2°C), and the MLD is significantly deeper (> 70 m).Wind speed ranges from 4 to 7 m/s in the northeast.A range of 14 to 18 g/kg of specific humidity indicates dry continental air during the time.Between November and February, net heat flux ranges from − 98 to-134 W/m 2 .This results in cooling of the sea surface due to heat loss from evaporation (latent heat flux − 220 to-312 W/m 2 ).In the eddy region, solar radiation ranges from 114 to 170 W/m 2 .This low solar insolation lowers the SST, which causes the water to become denser.As a result, nutrient-rich water entrains from deeper depths while the surface water sinks.This shows that surface cooling is caused by atmospheric forcing, and the convective mixing that results carries nutrients into the upper layer and stimulates primary production (Prasanna Kumar andPrasad 1996, Madhupratap et al. 1996).According to Chatterjee et al. (2016), the equatorial signal of Kelvin flows southward along the east coast of the Andaman Islands before entering the BoB through Preparis Channel and entering the Andaman Sea through the Great Channel.In this situation, it is assumed that Kelvin is the process that generates CE2.Another explanation for the origin of CE2 is that the flow from the Ayeyarwady-Salween river basin causes instability.

Conclusion
Based on a suit of in situ and satellite datasets, the column dynamics, forcing mechanisms, chemical, and physiological responses of cyclonic eddies are discussed for the Andaman waters.Focusing on the threshold Okubo-Weiss parameter of − 2 × 10 -11 /s 2 , the eddy CE1 is stronger than CE2 when the eddies are tracked using the Okubo-Weiss parameter.The small-scale processes, which have a diameter of 100-350 km, are discovered to be caused by baroclinic instability, which develops as a result of the westward-moving Rossby wave.Kelvin may also be responsible for inducing the semi-annual mode, which has a phase speed of 0.20 m/s for CE1 and CE2, and the instability is brought on by the Ayeyarwady-Salween flow.While CE2 is linked to the convective mixing that takes place in the area as a result of the cold, dry, continental air coming in from the northeast.The study comes to the conclusion that convective mixing along the northwest coast of Andaman is significantly contributing to the biological production of Andaman waters, in addition to the mesoscale processes.Significant improvements in regional surface biological production show how these mechanisms complement one another in raising production quality in Andaman waters.For the first time, the dynamics of the Andaman water's convective mixing and eddies are explained.

Fig. 2 a
Fig. 2 a Sea surface height (cm-Aviso daily from Dec 7, 2011) and geostrophic current (cm/s), b vertical; section of temperature, c salinity, d density

Fig. 3 Fig. 4
Fig. 3 Horizontal current (m/s from ADCP) structure at different depth at 8°N

Fig. 12
Fig. 12 Overlap map of SST (℃-Monthly MODIS Aqua) and Chl a (mg/m 3 -monthly MODIS Aqua) during a November, b December, c January, d February