Eddy-induced Track Reversal and Upper Ocean Physical-1 Biogeochemical Response of Tropical Cyclone Madi in the 2 Bay of Bengal 3

The life cycle of the tropical cyclone Madi in the southwestern Bay of Bengal (BoB) during 6 to 12 9 December 2013 was studied using a suite of ocean and atmospheric data. Madi formed as a depression on 6 10 December and intensified into a very severe cyclonic storm by 8 December. What was distinct about Madi was 11 its (1) swift weakening from very severe cyclone to a severe cyclone while moving towards north on 9, (2) 12 abrupt track reversal close to 180-degree in a southwestward direction on 10, and (3) rapid decay in the open 13 ocean by 12 December while still moving southwestward. Using both in situ and remote sensing data, we show 14 that oceanic cyclonic eddies played a leading role in the ensuing series of events that followed its genesis. The 15 sudden weakening of the cyclone before its track reversal was facilitated by an oceanic cyclonic (cold-core) 16 eddy, which reduced the ocean heat content and cooled the upper ocean through upward eddy-pumping of 17 subsurface waters. When Madi moved over the cyclonic eddy-core, its further northward movement was 18 arrested. Subsequently, the prevailing northeasterly winds assisted the slow moving system to change its track to 19 a southwesterly path. While travelling southwestwards, the system rapidly decayed when it passed over cyclonic 20 eddies near the western boundary of the BoB. Though Madi was a category-2 cyclone, it had a profound impact 21 on the physical and biogeochemical state of the upper ocean. Cyclone-induced enhancement in the chlorophyll a 22 ranged from 5 to 7-fold, while increase in the net primary productivity ranged from 2.5 to 8-fold. This 23 enhancement of chlorophyll a and net primary productivity was much higher than previous cyclones that 24 occurred in the BoB. Similarly, the CO2 out-gassing into the atmosphere showed a 3.7-fold increase compared 25 to the pre-cyclone values. Our study points to the crucial role oceanic eddies play in the life cycle of cyclones 26 and their combined impact on upper-ocean biogeochemical changes in the BoB. Eddies are ubiquitous and 27 tropical cyclones occur in the BoB; there is an urgent need to incorporate eddies in models for better prediction 28 of cyclone track and intensity. As cyclone and eddy co-exists in many parts of the world ocean our approach in 29 delineating the upper-ocean biogeochemical changes can be adapted elsewhere. 30


Introduction
The Bay of Bengal (BoB) (Fig. 1) is a tropical sea situated in the eastern part of the northern Indian Ocean.The two most important characteristic features of the BoB are the perennial presence of low salinity waters (30-34 psu) in the upper ocean and the seasonal reversal of atmospheric winds from northeasterly direction between Ocean Sci.Discuss., https://doi.org/10.5194/os-2018-133Manuscript under review for journal Ocean Sci. Discussion started: 28 January 2019 c Author(s) 2019.CC BY 4.0 License.
November and February (5 m/s, northeast or winter monsoon) to southwesterly during June to September (9 m/s, southwest or summer monsoon) (Narvekar and Prasanna Kumar, 2006).This perennial presence of low salinity waters enhances the stability of the upper water column through increased stratification and makes it one of the warmest regions in the Indian Ocean.The BoB is a site of tropical cyclones, which occur usually during pre-monsoon (April-May) and post-monsoon (October-November) periods.Though north Indian Ocean accounts for only 7% of the total number of tropical cyclones that occur worldwide, the frequency of occurrence of cyclones in the BoB is 4-times higher than that in the Arabian Sea (Dube et al., 1997).Each year 3-5 cyclones occur in the BoB, with a primary peak during post-monsoon and a secondary peak in pre-monsoon.
Though there have been several studies on the tropical cyclone-ocean interaction in the Pacific (typhoon) and the Atlantic (hurricane) that have advanced our understanding about the upper ocean response in terms of cooling of SST and enhancement of chlorophyll (e.g., Chang and Anthes, 1978;Price, 1981;Emanuel, 1999;Babin et al., 2004;Wada and Chan, 2008;Liu et al., 2009;Pun et al., 2011) and cyclone-eddy interaction (e.g., Shay et al., 2000;Jaimes and Shay, 2009;Lin et al., 2011;Yablonsky and Ginis, 2013;Sun et al., 2014), the depth-dependent temperature and chlorophyll response is still poorly understood.It is in this context that the present paper aims at understanding the (1) ocean-atmosphere condition associated with the evolution of cyclone Madi, a category-2 cyclone (Saffir-Simpson scale), during December 2013 in the BoB, its sudden weakening and close to180-degree track reversal before its dissipation, (2) time-evolution of depth-dependent temperature and chlorophyll profiles in the vicinity of cyclone Madi, and (3) cyclone-induced physical and biogeochemical response of the upper ocean.

Data
In the present study, the information on cyclone Madi was taken from Indian Meteorological Department (IMD) (http://www.imd.gov.in), while the track information was taken from Unisys Weather (http://weather.unisys.com/hurricanes/search).The daily SST data was taken from Tropflux (Praveen Kumar et al., 2012)   (https://www.aviso.altimetry.fr/en/my-aviso.html).The zonal and meridional components of wind at 10 m height were taken from Advanced Scatterometer (ASCAT) level 3 product (Bentamy and Croize-Fillon, 2012) (https://opendap.jpl.nasa.gov/opendap/OceanWinds/ascat/preview/L2/metop_a/12km/contents.html).It is a daily product having a spatial resolution of 0.25 degree latitude by longitude.This has been further used for the calculation of wind stress curl and Ekman pumping velocity (Gill, 1982) as given below: where  ,  are the zonal and meridional wind stress components,  is the density of sea water with its value taken as 1026 k gm -3 , and f is the Coriolis parameter which varies with latitude.
The oceanic heat content (OHC) in the upper 300 m is calculated following Eq.( 3): where,  is the density of seawater,  is the specific heat capacity of sea water taken as 3.87 kJ kg -1 K -1 , h 1 and h 2 are the lower and upper water depths, and () is the temperature profile measured in Kelvin.
The relative humidity at 500 hpa was taken from NCEP 2 reanalysis daily data having 2.5 degree grid resolution (https://www.esrl.noaa.gov/psd/data/gridded/data.ncep.html)(Kalnay et al., 1996).The daily zonal and meridional components of wind at 850, 500 and 200 hpa having a spatial resolution of 0.5 degree were extracted from NCEP climate forecast system version 2 and used to compute vector wind.Winds at 850 and 200 hpa were used for the calculation of vertical wind shear (Saha et al., 2014) (http://www.ncep.noaa.gov).
In order to gain insight about the time evolution of temperature and chlorophyll in the upper water column in response to the passage of cyclone Madi, we have analyzed the trajectory of two Argo floats (WMO ID 2901288, 2901629) for temperature profiles that were in the vicinity of Track 2 and one Bio-Argo float (WMO ID 2902086) for chlorophyll profiles that was to the right of Track 1 as shown in Fig. 1.Argo/Bio-Argo data were downloaded from Argo CORIOLIS site (http://www.coriolis.eu.org/Data-Products/Data-Delivery/Dataselection).

Data Processing method of Chlorophyll and net primary production
The satellite-derived daily chlorophyll a (Chl-a) pigment concentration data and net primary production (NPP) estimated based on vertically generalized productivity model (VGPM) of (Behrenfeld and Falkoswski, 1997) were taken from Moderate Resolution Imaging Spectro-radiometer (MODIS) Aqua Ocean color (https://oceandata.sci.gsfc.nasa.gov/MODISA/).The Level 3 Chl-a dataset has a zonal and meridional resolution of 0.05 degree longitude by latitude.From the daily data, weekly composites were calculated.In order to determine the net CO 2 flux over southwestern BoB, before, during, and after the passage of the cyclone Madi, pCO 2 air data was taken from NOAA ESRL (ftp://aftp.cmd1.noaa.gov/product/trends/co2/co2_mm_g1.txt).Since the daily pCO 2 sea values are not available, the value of climatological air-sea difference in partial pressure of CO 2 was taken from (Takahashi et al., 2009) and the net flux was calculated using the following formula: where, k denotes the gas transfer velocity, a is the solubility of CO 2 in sea water which is dependent on sea surface temperature and salinity (Weiss, 1974) as per the following equations: The gas-transfer velocity "k" is calculated using wind speed following (Wanninkhof, 1992) by using the formula where  is the scaling factor and its value of 0.26 is taken from (Takashashi et al., 2009), while ∪ is the wind speed.S c is the Schimidt number (kinematic viscosity of water/diffusion coefficient of CO 2 in water), the value of which is 660 for CO 2 in seawater at 20 o C and is a function of temperature and is computed as: For the values of the constants A, B, C and D refer (Weiss, 1974;Wanninkhof, 1992).
We have divided the study region into Box A, Box B, Track 1, Track 2 and Box abcd.See Fig. 1 for the location of these sub-regions.

Origin, evolution and decay of the cyclone Madi
On 30 th November 2013, as per the Indian Daily Weather Report of India Meteorological Department (IMD), a low pressure system was formed over the southwestern part of the BoB (Fig. 1) and slowly intensified into a depression (the classification of intensity of the system is based on IMD, http://imd.gov.in/section/nhac/termglossary.pdf) on 6 th December 2013 with its centre at 10 o N and 84 o E (Fig. 1).
The system intensified further into a deep depression (DD) on the same day with maximum sustained wind speed of 50-60 km/hr.Subsequently, when it turned into a cyclonic storm (CS) on 7 th December, the IMD named it as Madi.On further intensification into a severe cyclonic storm (SCS), the system started moving in a north/north-northeast direction with maximum sustained wind speed of 90-100 km/hr.Subsequently, on 8 th December, the system turned into a very severe cyclone (VSCS) with maximum sustained wind speed of 120-130 km/hr.where it remained stationary for a while.At that point the SCS deviated from its northward track, took a near 180 degree turn and veered southwestward (Fig. 1).During the course of its south-westward movement, the SCS weakened to CS with maximum sustained wind speed of 80-90 km/hr.On its further south-westward journey, the CS weakened to DD on 11 th December and further to a depression the same day with its centre at 12.9 o N and 82.7 o E. On 12 th December 2013 the depression further weakened to a well marked low pressure.

Results and Discussion
We start our analysis by examining the time-evolution of the spatial distribution of various oceanic and atmospheric parameters from 4 th to 15 th December 2013 to understand the thermo-dynamical and dynamical conditions that led to the formation and subsequent dissipation of cyclone Madi.

Thermodynamic conditions before, during and after the cyclone
Ocean heat content (OHC) plays an important role in the translation speed and intensification of cyclones over the BoB (Sadhuram et al., 2010).The time-evolution of the spatial distribution of the OHC on 4 th December 2013 showed large values ranging from 3.580 to 3.600 x 10 11 J/m 2 , except a meridionally-elongated region along the western boundary between 8 o and 20 o N, and another small patch in the central BoB centered at 13 o N, where the values were small (Fig. 2a, 2b).Note the meridional band of the large OHC, adjacent to the meridional band of small OHC hugging the western boundary, with three distinct patches of high values within them.The drastic decrease of OHC on 7 th December (Fig. 2c, 2d) indicated strong heat uptake by the cyclonic storm during the process of its intensification.As the system moves northward, passing over the region of high OHC it continues to take up heat from the upper ocean and intensifies further (Fig. 2e).Note that on 9 th December when the track of the system passes over a region of low OHC it weakens (Fig. 2f).On 10 th December, when the system it deviated from its northward track and took almost a 180-degree turn it was passing through low OHC (Fig. 2g).On its southward journey, the system passes over regions of lower OHC on 11 th and 12 th (Fig. 2h, 2i), when it dissipates into DD and to well-marked low pressure respectively.Once the system is dissipated, the spatial distribution of OHC showed a recovery in terms of heat gain by the upper ocean (Fig. 2j-l), especially in the region of the track of the cyclone.

Dynamic conditions before, during and after the cyclone
The analysis of the time-evolution of the spatial distribution of sea level anomaly (SLA) over-laid with geostrophic current from 4 th to 15 th December revealed the presence of several meso-scale cyclonic (blue region with negative SLA) and anticyclonic (red regions with positive SLA) eddies (Fig. 3).The SLA and associated geostrophic current clearly indicated the presence of two cyclonic eddies along the western boundary and two in the offshore region (Fig. 3a, b).The region of occurrence of these cyclonic eddies coincided with the region of low OHC (Fig. 2).Note that the genesis of Madi in the form of a depression occurred on 6 th December in the region of positive SLA with an anticyclonic circulation (Fig. 3c), which was the same region that had high OHC.The intensification of Madi on 8 th December also occurred in a region of positive SLA with an anticyclonic circulation (Fig. 3d, 3e).On 9 th December when the system entered into a region of negative SLA with cyclonic circulation (Fig. 4f), which was also a region of low OHC it weakened as it was deprived of the thermal energy from the upper warm ocean that is essential to sustain the system.On 10 th December when the system moved further north entering towards the core of the cyclonic eddy (Fig. 3g) with low OHC (Fig. 2g) its further northward movement was arrested.It remained stationary for a while and changed its track to almost 180-degree in a southwestward direction.While doing so the cyclone Madi was moving further through the regions of strong cyclonic circulation/eddies (Fig. 3i), which rapidly reduced its strength and finally led to its dissipation on 12 th December.The passage of cyclone Madi modified the upper ocean circulation in the southwestern part of the BoB (Fig. 3j-l) into a large cyclonic gyre with strong southward western boundary current from 17 o to 10 o N along the west coast of India.The four cyclonic eddies were now prominently seen embedded in this large-scale gyre.
When a cyclone passes over the cyclonic eddy region, the colder temperature within the eddy could potentially reduce the translation speed of the cyclone as well as its intensity as it is unable to fuel the cyclone as effectively as in the case of the warm water region where it originates.In order to further ascertain the role of cyclonic eddy in weakening the strength of the cyclone before its track reversal, we calculated the translational speed of the system from its formation on 6 th to its dissipation on 12 th December and examined it along with its strength (Table 1).It is clear from Table 1 that on 9 th December when the cyclone entered the region of oceanic cyclonic eddy the translational speed of the cyclone decreased from 2.81 m/s to 1.96 m/s and the cyclone weakened from VSCS to SCS.Thereafter, subsequent to track reversal as the system moves south-westward out of the cyclonic eddy, the translation speed increases.
Though the weakening and the final dissipation of the cyclone Madi was easy to understand in the context of the prevailing oceanic cyclonic eddies, we examined the time-evolution of the spatial distribution of the atmospheric parameters such as wind at 850 hpa (Fig. 4), vertical wind shear between the 850 and 200 hPa (Fig. 5) and mid-troposheric (500hpa) relative humidity (Fig. 6) to understand the atmospheric condition.
The salient feature of the large-scale atmospheric circulation over the BoB, prior to the genesis of cyclone Madi, was the prevalence of an easterly zonal wind with speed between 5 and 15 m/s with an embedded cyclonic circulation located in the southwestern region (Fig. 4a, 4b).The wind speed associated with the cyclonic circulation was between 15 and 25 m/s.On 6 th December when the depression was formed, this broad cyclonic circulation becomes well organized with a small central region having lower wind speeds of 10 m/s, while the surrounding regions had higher wind speeds of 20-25 m/s (Fig. 4c).When the system developed into the CS (Fig. 4d) and intensified into a VSCS (Fig. 4e), the large-scale atmospheric circulation in the BoB showed a well defined "eye of the cyclone".Away from the cyclonic circulation, the winds in the northen part of the BoB were mostly southwestward.On 9 th December the weakening of the system was discernible as it moved northward (Fig. 4f).At this time the low vertical wind shear (10 to 15 m/s) (Fig. 5f) and high relative humidity (60-80 %) (Fig. 6f) were congenial for the system for further intensification or at least to sustain its intensity.In contrast the system weakened from VSCS to SCS.This indicated that the system evolution at this time was controlled by the oceanic cyclonic eddies rather than the atmospheric conditions.On 10 th when the system reached its northern most location (Fig. 4g), it was actually sitting right on the top of the cold-core of the cyclonic eddy (Fig. 3g).At this time the system became stationary and the prevailing easterly winds (Fig. 4g) were able to turn and move it towards southwesterly direction, a result which is consistent with that of (Bhattacharya et al., 2015).
Thus, our study showed that the weakening of cyclone on its northward journey was mediated by the oceanic cold-core cyclonic eddy while the change in the direction of the cyclone track when the system was stationary was brought about by the prevailing northeasterly winds.

Cyclone-induced along track oceanic variability
In order to quantify the upper ocean response of the tropical cyclone Madi, we examined four oceanic parameters viz.SST, Ekman pumping velocity (EKV), SLA and OHC during the period 2 to 15 December 2013 at four locations : (1) Box A, the region of genesis of the depression which subsequently turned into cyclone Madi, (2) along Track 1, the northward path followed by the cyclone Madi during which time it intensified from CS to VSCS, (3) Box B, the region where the cyclone Madi weakened, remained stationary and eventually turned, and (4) along Track 2, the southwestward path of the cyclone which eventually dissipated.
The time-evolution of SST in Box A, showed a monotonic decline of 1.5 o C from 28.2 o C to 26.7 o C during the period from the genesis of the cyclone to its decay (Fig. 7).However, the rate of decrease during the entire period was not uniform.Even before the formation of the depression SST showed a weak decrease of 0.3 o C, however, during the period 6 th to 8 th December when the depression was formed within the Box A and turned into a cyclone, the SST decreased rapidly.Though the system was away from the region of Box A and was dissipating with time during 9 th to 11 th December, the SST within the Box A showed the most rapid decrease of 1.1 o C. The SLA, on the other hand, showed a continuous decrease, before the formation of depression and much after its dissipation.The SST showed a recovery/warming trend after 12 th December.The EKV showed a peak on 6 th December, at the time of formation of depression.This is expected, as under the action of cyclonic wind, the upward Ekman pumping will also increase in magnitude.What was unexpected was the temporal response of the OHC, which showed an initial decrease from 2 nd to 3 rd December followed by an increase reaching the highest value of 3.589 x 10 11 J/m 2 and a subsequent decrease.A secondary peak occurred on 6 th as the depression formed in the area of Box A. During 6 th to 8 th December when the system intensified and was located within the Box A, the OHC showed rapid decrease to a value of 3.574 x 10 11 J/m 2 .There after the values were closer to 3.576 x 10 11 J/m 2 , except on 13 th December when it once again peaked to 3.578 x 10 11 J/m 2 .
Though the response of all the four parameters along Track 1 (Fig. 8), Track 2 (Fig. 9) and at Box B (Fig. 10) were similar to that of Box A (Fig. 7), a closer similarity was noticed between Box A and Track 2, and between Track 1 and Box B. However, the magnitudes of response of each parameter and their times of occurrence were different depending on the position of the cyclone with respect to each of the four locations.For example, the OHC showed an inverse relationship with the Ekman pumping velocity along Track 1(Fig.8) and at Box B (Fig. 9), while at Box A (Fig. 7) and along Track 2 (Fig. 8) the OHC showed a double-peak structure.Along Track 1, the occurrence of highest value of EKV was consistent with the system intensifying into VSCS with a maximum sustained wind speed of 110-120 km/hr.Similarly, at Box B also the occurrence of the highest EKV coincided with the arrival of the cyclone at this location.The rapid decrease in SST occurred in all the four regions, in general, during 9 th to 11 th December, indicating a time-lag between the presence of the cyclonic storm and the peak of the upward EKV.Another noteworthy feature, common in all the four cases, was the co-variation of SST and SLA, both showing a monotonic decline, indicating the occurrence of colder waters associated with decreasing sea level, except along Track 2 (Fig. 9).Note that this lowered sea level and colder SST occurred well before the initiation of the upward Ekman pumping under the influence of the cyclone Madi.This pointed towards the pre-cyclone cooling of SST by oceanic cyclonic eddies, which was also evident from the time evolution of the spatial maps of daily SLA (Fig. 3).However, along Track 2, SLA showed a rapid increase from 2 nd to 5 th December followed by a slower increase until 8 th December, well before the passage of the cyclone through this region.This is primarily due to the fact that the location of Track 2 passes through an anticyclonic eddy.obtained from Argo float with ID-2901288 was 30m and temperature was 28.2 o C and after the passage of cyclone on 14 th December the MLD was 50 m and temperature was 26.5 o C (Fig. 11a).A similar change was also noticed in the vertical profiles of temperature obtained from Argo float with ID-2901629 (Fig. 11b).Thus, both the Argo floats captured the cyclone-induced mixed layer cooling and deepening.

Depth
The vertical profiles of Chl-a obtained from the Bio-Argo float (ID-2902086) showed low values prior to the cyclone (23 rd November to 3 rd December 2013) in the range of 0.10 to 0.15 mg/m 3 with constant value within the mixed layer and a subsurface chlorophyll maximum (SCM) located at about 50m (Fig. 11c).The vertical profiles of Chl-a showed a progressive increase during and after the cyclone in both the surface as well as the subsurface values reaching a maximum of 0.45 and 0.65 mg/m 3 respectively on 23 rd December 2013.Thereafter, it showed a decline on 28 th December 2013 when the value in the upper 60 m was 0.40 mg/m 3 with no perceptible SCM.Thus, the Chl-a profiles in the upper 60 m showed maximum impact due to the cyclone leading to an overall increase in the biomass.

Cyclone-induced biogeochemical variability
It is well known that tropical cyclones bring about large changes in the upper ocean productivity as well as gasexchange between ocean and atmosphere.In order to understand and quantify the biogeochemical response due to the cyclone Madi, we examined along track variation of satellite-derived chlorophyll a pigment concentration (Chl-a), net primary production (NPP), and the net CO 2 flux.A major difficulty with the remotely sensed Chl-a pigment concentration is the lack of adequate cloud-free pixels along track on a daily time scale.In order to overcome this, we have used weekly composite data for Chl-a for the calculation of NPP from 30 th November to 28 th December in the four regions, viz.Box A and B and Track 1 and 2 (Fig. 1), while the net CO 2 flux was computed on daily time scale from 2 nd to 15 th December 2013.
The time variation of the weekly composite of Chl-a showed a pattern that was typical of the cyclone induced response (Fig. 12).Prior to the genesis of cyclone Madi, the Chl-a was in the range of 0.2 to 0.4 mg/m 3 , but the weekly composite values for the period 7 th to 14 th December, which includes the growth, decay and a couple of days after cyclone, showed several fold increase.The maximum increase of 2.7 mg/m 3 was in Box B, which was almost 7-times higher than the pre-cyclone period.The minimum increase of 1 mg/m 3 occurred along the Track 2, which was 5-times higher than the pre-cyclone period.In the Box A and along Track 1, the Chl-a values were 1.4 and 1.5 mg/m 3 respectively after the cyclone.It is pertinent to examine the chlorophyll enhancement by other cyclones in the BoB and compare with the present study.For example, the Orissa super cyclone in October 1999 produced a Chl-a enhancement in the rage of 0.38 to 0.97 mg/m 3 in the open ocean region (Madhu et al., 2002), while that near the land fall region was a maximum of 1.0 mg/m 3 (Patra et al., 2007).
( Vinayachandran et al., 2003) reported a value ranging from 0.5 to 2.0 mg/m 3 for the cyclones that occurred during November-December during the period 1996 to 2001.In the case of cyclone Sidr in 2007, (Maneesha et al., 2011) obtained an increase from 0.2 to 0.5 mg/m 3 .
Thus, the Chl-a enhancement by Madi was much greater than for previous cyclones that occurred in the BoB.
The obvious question would be why the Chl-a along both the tracks as well as the boxes showed an increase and why Box B showed the highest magnitude of Chl-a response to the cyclone.Recall that the EKV showed a rapid increase during the period when the cyclone was transiting these regions, while a concomitant rapid decrease of SST was also noticed.This indicated the upward transport of cold subsurface waters under the influence of the cyclonic winds.As the subsurface waters are nutrient rich, the increased Ekman pumping under the tropical cyclone would bring more nutrients to the upper oceans which will kick-start the photosynthesis (Subrahmanyam et al., 2002;Lin et al., 2003) resulting in the observed increase in the Chl-a biomass.Recall also that an oceanic cyclonic eddy was located in the region of Box B where the cyclone was stationary for a while.In the BoB, the nutricline is located just below the mixed layer, usually at a depth ranging from 20 to 40m (Prasanna Kumar et al., 2007), and the eddy-pumping (Falkowski et al., 1991) associated with oceanic cyclonic eddies is able to supply sub-surface nutrients to the surface waters (Prasanna Kumar et al., 2004).
Hence, we infer that under the combined effect of the oceanic eddy and the cyclone Madi, the upward Ekman pumping would have been stronger and more nutrients could be supplied to the upper ocean, which resulted in the observed 7-fold increase.The lowest response, 5-fold increase, was seen along track 2, which is to be expected as when the cyclone transited along this path it was decaying rapidly.Note that by the last week of The cyclone Madi triggered intense physical and biogeochemical response in the upper ocean.The weekly composite of satellite-derived Chl-a pigment concentration showed an enhancement that ranged from 5 to 7-fold with a maximum value of 2.7 mg/m 3 .A similar response was seen in the net primary productivity which showed a 2.5 to 8-fold increase, with a maximum value of 2500 mg C m -2 day -1 .The largest values of both Chl-a and NPP was greater than for previous cyclones in the BoB.Out study indicates that a combination of an oceanic cyclonic eddy along with cyclone Madi facilitated upward Ekman pumping of nutrient rich subsurface waters to the surface, thereby kick-starting the primary production and increasing the chlorophyll biomass.Consistent with this, the net CO 2 out-gassing to the atmosphere also was the greatest in this region amounting to 4.7 Tg carbon per day, which was 3.7-fold greater than the pre-cyclone values.Our study emphasizes the importance of eddy-cyclone interaction that led to the large increase in Chl-a, primary production and CO 2 out-gassing.Since cyclone and eddies co-occur in many parts of the world ocean our approach can be adopted in other regions to quantify the biogeochemical response.
One of the limitations of our study is the lack of modeling to quantify the eddy-cyclone interaction.Our study underscores the important role of oceanic eddies in understanding the life cycle of tropical cyclones in the BoB.
Since both cyclonic and anticyclonic eddies are ubiquitous in the BoB, they will impact both the translation speed and intensity of a tropical cyclone.Hence, for the accurate prediction of a cyclone track and its intensity, there is an urgent need to incorporate eddies into the predictive models; this action is still to be explored.This will be attempted in the near future.
all the support and encouragement for this research.The daily SST data are available from (http://www.incois.gov.in/tropflux_datasets/data/ daily/), daily SLA data along with zonal and meridional geostrophic velocity are from (https://www.aviso.altimetry.fr/en/my-aviso.html),surface wind data are from (https://opendap.jpl.nasa.gov/opendap/OceanWinds/ascat/preview/L2/metop_a/12km/contents.html), while the wind at 850 hpa is from(http://www.ncep.noaa.gov).The Argo data are from (http://www.coriolis.eu.org/Data-Products/Data-Delivery/Data-selection).The graphics were generated using MATLAB and the code used in this paper can be obtained from the first author.Riyanka Roy Chowdhury acknowledges Ministry of Human Resource Development for providing the research fellowship.NIO contribution number is XXXX.
Table 1 Translation speed of the cyclonic disturbance along with its category during the life cycle of cyclone Madi.L-low pressure, DD-deep depression, CS-cyclonic storm, SCS-severe cyclonic storm, VSCSvery severe cyclonic storm.
-dependent temperature and chlorophyll a response Having analyzed the cyclone-induced SST response along the track of the cyclone, it is pertinent to examine the vertical profiles of temperature before, during and after the passage of cyclone Madi.Hence, we examined the vertical profiles of temperature in the vicinity of Track 2 obtained by two Argo floats (ID-2901288 and ID-2901629) which transected the northern and southern parts of Track 2 during the period of study (see Fig. 1 for the location of Argo floats).The vertical profiles of temperature obtained from both the Argo floats (Fig. 11a, b) showed the presence of a thermal inversion (0.2 to 0.3 o C) located in the upper 40 m prior to the passage of the cyclone Madi, which disappears in the subsequent profiles.The most distinct change was in the mixed layer temperature and depth.On 4 th December prior to the formation of cyclone the mixed layer depth (MLD) Ocean Sci.Discuss., https://doi.org/10.5194/os-2018-133Manuscript under review for journal Ocean Sci. Discussion started: 28 January 2019 c Author(s) 2019.CC BY 4.0 License.

Figure 1
Figure 1 Map showing the track of the tropical cyclone Madi (magenta filled circles inside the black circles) during 6-12 December 2013 in the Bay of Bengal obtained from UNISYS Weather.The shading is the sea level anomaly (m), while vectors are the wind (m/s) at 850 hpa, both are composite for the period 6-12 December 2013.Location of Box A, Track 1, Box B, Track 2, rectangular Box abcd, and Argo floats (ID-2901288 red plus & ID-2901629 yellow plus) near Track 2 are also shown in the map.The black hollow circles (seen as dark circles due to overlap) show the position of Bio-Argo float (ID2902086).

Figure 2
Figure 2 Spatial maps of oceanic heat content (x10 11 J/m 2 ) from 4 th (a) to 15 th (l) December 2013 with track of the cyclone overlaid.The black filled circles represent the position of the cyclone on a particular day, while the magenta filled circles indicate the track.

Figure 3
Figure 3 Spatial maps of sea level anomaly (m) from 4 th (a) to 15 th (l) December 2013 with track of the cyclone overlaid.The black filled circles represent the position of the cyclone on a particular day, while the magenta filled circles indicate the track.

Figure 4
Figure 4 Spatial maps of wind speed (shading, m/s) overlaid with wind vectors (thin arrow) at 850 hpa from 4 th (a) to 15 th (l) December 2013 with track of the cyclone overlaid.The black filled circles represent the position of the cyclone on a particular day, while the magenta filled circles indicate the track.

Figure 5
Figure 5 Spatial maps of vertical wind velocity difference between the 850 and 200hPa (shading, m/s) from 4 th (a) to 15 th (l) December 2013 with track of the cyclone overlaid.The black filled circles represent the position of the cyclone on a particular day, while the magenta filled circles indicate the track.

Figure 6
Figure 6 Spatial maps of relative humidity (%) overlaid with winds at mid-troposhere (500hpa) from 4 th (a) to 15 th (l) December 2013 with track of the cyclone overlaid.The black filled circles represent the position of the cyclone on a particular day, while the magenta filled circles indicate the track.

Figure 8
Figure 8 Along track variation of the sea surface temperature (SST, o C), Ekman pumping velocity (EKV, m/day, positive upward), oceanic heat content (OHC, x 10 11 J/m 2 ) and sea level anomaly (SLA, m) along Track 1 from 2-15 December 2013.These are daily averages along the track.

Figure 10
Figure 10 Along track variation of the sea surface temperature (SST, o C), Ekman pumping velocity (EKV, m/day, positive upward), oceanic heat content (OHC, x 10 11 J/m 2 ) and sea level anomaly (SLA, m) along Track 2 from 2-15 December 2013.These are daily averages along the track.

Figure 12
Figure 12 Time variation of weekly composite of chlorophyll a pigment concentrations (Chl-a, mg/m 3 ) in the Box A (red) and B (blue) and along Track 1 (green) and 2 (black) from 30 November to 28 December 2013.The vertical lines are the standard deviations.

Figure 13
Figure 13 Time variation of weekly composite of net primary production (NPP, mg C m -2 day -1 ) in the Box A (red) and B (blue) and along Track 1 (green) and 2 (black) from 30 November to 28 December 2013.The vertical lines are the standard deviations.

Figure 14 Figure 1
Figure 14 Daily variation total CO2 flux (terra gram carbon per day) in the Box A (red) and B (blue) and along Track 1 (green) and 2 (black) from 2 to 15 December 2013.The vertical lines are the standard deviations.

Figure 2
Figure 2 Spatial maps of oceanic heat content (x10 11 J/m 2 ) from 4 th (a) to 15 th (l) December 2013 with track of the cyclone overlaid.The black filled circles represent the position of the cyclone on a particular day, while the magenta filled circles indicate the track.

Figure 3
Figure 3 Spatial maps of sea level anomaly (m) from 4 th (a) to 15 th (l) December 2013 with track of the cyclone overlaid.The black filled circles represent the position of the cyclone on a particular day, while the magenta filled circles indicate the track.

Figure 4
Figure 4 Spatial maps of wind speed (shading, m/s) overlaid with wind vectors (thin arrow) at 850 hpa from 4 th overlaid.The black filled circles represent the position of the cyclone on a particular day, while the magenta filled circles indicate the track.maps of wind speed (shading, m/s) overlaid with wind vectors (thin th (a) to 15 th (l) December 2013 with track of the cyclone .The black filled circles represent the position of the cyclone on a particular day, filled circles indicate the track.

Figure 5
Figure 5 Spatial maps of vertical wind (shading, m/s) from 4 th (a) to 15 The black filled circles represent the position of the cyclone on a particular day, while the magenta filled circles indicate the track.Spatial maps of vertical wind velocity difference between the 850 (a) to 15 th (l) December 2013 with track of the cyclone The black filled circles represent the position of the cyclone on a particular day, while the magenta filled circles indicate the track.between the 850 and 200hPa (l) December 2013 with track of the cyclone overlaid.The black filled circles represent the position of the cyclone on a particular day, while the

Figure 6
Figure 6 Spatial maps of relative humidity (%) overlaid with winds at (500 hpa) from 4 th (a) to 15 th black filled circles represent the position of the cyclone on a particular day, while the magenta filled circles indicate the track.

Figure 10
Figure 10 Along track variation of the sea surface temperature (SST, o C), Ekman pumping velocity (EKV, m/day, positive upward), oceanic heat content (OHC, x 10 11 J/m 2 ) and sea level anomaly (SLA, m) along Track 2 from 2-15 December 2013.These are daily averages along the track.

Figure 11
Figure 11 Time-series of the vertical profiles of temperature ( 2 obtained from (a) Argo float ID Argo float ID-2901629 for 2, 12 and 22 December 2013 and (c) chlorophyll the vicinity of Track 1 obtained from Bio 3, 8, 13, 18, 23 and 28 December 2013.

Figure 12
Figure 12 Time variation of weekly composite of chlorophyll (Chl-a, mg/m 3 ) in the Box A (red) and B (blue) and along Track 1 (green) and 2 (black) from 30 November to 28 December 2013.The vertical lines are the standard deviations.

Figure 13
Figure 13 Time variation of weekly composite of net primary production (NPP, mg C m day -1 ) in the Box A (red) and B (blue) and along Track 1 (green) and 2 November to 28 December 2013.The vertical lines are the standard deviations.

Figure 14
Figure 14 Daily variation total CO 2 flux (terra gram carbon per day) in the Box A (red) and B (blue) and along Track 1 (green) and 2 (black) from 2 to 15 December 2013.The vertical lines are the standard deviations.
The system moved further northward on 9 th December reaching the location 14.6 o N and 84.7 o E, when it weakened into SCS with maximum sustained wind speed of 110-120 km/hr.The system not only Ocean Sci.Discuss., https://doi.org/10.5194/os-2018-133Manuscript under review for journal Ocean Sci. Discussion started: 28 January 2019 c Author(s) 2019.CC BY 4.0 License.weakened but slowed down considerably while reaching the location 15.7 o N and 85.3 o E on 10 th December Ocean Sci.Discuss., https://doi.org/10.5194/os-2018-133Manuscript under review for journal Ocean Sci. Discussion started: 28 January 2019 c Author(s) 2019.CC BY 4.0 License.