The track and intensity of TC “Asani” from IBTrACS data was overlaid on the minimum ERA5 mean sea level pressure (Fig. 1a). Three boxes (2° × 2°) were identified corresponding to the genesis, maximum intensity, and landfall regions of the cyclone. All subsequent analysis were carried out based on these selected areas. We then attempted to identify suitable proxy measures for each of the four primary air-sea interaction mechanisms driving the ocean response: wind-stirred mixing, buoyancy flux driven mixing, open ocean upwelling and coastal upwelling.
The Monin-Obhukov similarity theory (Monin and Obhukov 1954) describes the comparison of the dynamic significance of buoyancy flux to wind-stirring in terms of their contribution to the turbulent kinetic energy (TKE) within the ocean boundary layer through the Monin-Obhukov length scale (L) (Zheng 2021), as:
$$L= \frac{{u}_{*}^{3}}{\kappa {B}_{0}}$$
1
where, \({u}_{*}\) is the friction velocity, \({B}_{0}\) is the surface buoyancy flux and \(\kappa =0.4\) is the Von Karman constant. The magnitude of L is an estimate of the depth at which buoyancy flux dominates over wind-stirring in TKE production. Based on this, we selected \({u}_{*}^{3}\) (which is proportional to the cube of surface wind speed) to be the proxy for wind-stirred production of TKE where,
$${u}_{*}=\sqrt{{\tau }_{0}/{\rho }_{a}}$$
2
\({\tau }_{0}\) being the surface windstress and \({\rho }_{a}\) being the surface air density. Different formulations of wind energy transfer at the surface use a similar dependence on the cube of surface wind speed; the total power dissipation is formulated by Emmanuel (2005) as a spatio-temporal integral involving the cube of surface wind speed. Niiler (1977) similarly quantified the wind-stirred TKE production within the ocean mixed layer as \({u}_{*}^{3}\) in the 1 dimensional bulk mixed layer theory.
The surface buoyancy flux (\({B}_{0}\)) was considered to be the proxy for buoyant TKE production. As the density of seawater depends on temperature and salinity, both air-sea heat and moisture fluxes could alter the surface water density making the water column more or less buoyant. Following Cronin (2009), the buoyancy flux (\({B}_{0}\)) could be expressed in terms of the net surface heat flux (\({Q}_{0}\)), rate of evaporation (\(E\)) and precipitation (\(P\)) as:
$${B}_{0}=-\frac{g\alpha }{\rho {c}_{p}}{Q}_{0}+g\beta {S}_{0}(E-P)$$
3
where, g is the acceleration due to gravity, \({S}_{0}\) is the surface salinity, \(\rho\) is the density of sea water, \({c}_{p}\) is its specific heat, \(\alpha\) (\(\alpha = -\frac{1}{\rho } \frac{\partial \rho }{\partial T}\)) is the thermal expansion coefficient, and \(\beta\) (\(\beta = \frac{1}{\rho }\frac{\partial \rho }{\partial S}\)) is the haline contraction coefficient. A positive buoyancy flux due to heat loss or excess evaporation makes the surface waters less buoyant (more dense), decreases convective stability of the water column with the resultant overturning leading to the entrainment of deeper waters to the surface.
The open ocean upwelling process is quantified in terms of the Ekman pumping velocity (\({w}_{E}\)) which is the vertical velocity produced at the base of the Ekman layer due to the divergence of surface currents resulting from a positive windstress curl at the surface. The Ekman pumping velocity is expressed as:
$${w}_{E}=\frac{\overrightarrow{{\nabla }_{z}} \times \overrightarrow{{\tau }_{0}}}{\rho f}$$
4
where, \({\tau }_{0}\) is the surface windstress, \(\rho\) is the density of water, and \(f\) is the Coriolis parameter (Smith 1968). In the vicinity of the coast, the alongshore windstress (AWS) could generate a divergence of surface currents along the coast, resulting in coastal upwelling (Smith 1968). This process is quantified in terms of the Ekman transport (M) which represents the horizontal flow of coastal waters in the cross-shore direction (Smith 1968; Varela et. al. 2015; Jayaram and Kumar 2018), given by:
|
\(M= \frac{{\tau }_{l}}{\rho f}\)
|
(5)
|
where,
|
\({\tau }_{l}= {\rho }_{a}{c}_{d}{v}_{l}\sqrt{{u}^{2}+{v}^{2}}\)
|
(6)
|
where,
|
\({v}_{l}= \pm (u\text{cos}\left(\theta -\frac{\pi }{2}\right)+v sin \left(\theta -\frac{\pi }{2}\right))\)
|
(7)
|
where, \({c}_{d}\) is the drag coefficient, \({\tau }_{l}\) is the alongshore windstress, \({v}_{l}\) is the alongshore wind speed, \(u\) is the zonal wind speed, \(v\) is the meridional wind speed, and \(\theta\) is the coastal angle, which is the angle subtended by the seaward normal to the coastline. The sign of \({v}_{l}\) is negative (positive) in the northern (southern) hemisphere.
Three 2°× 2° boxes (as in Fig. 1) were selected along the track of the cyclone covering three different stages. The ERA5 SST and the first three proxy parameters, \({u}_{*}^{3}\), \({B}_{0}\), and \({w}_{E}\) were computed within each box over the period from 0000 UTC of 3rd May, 2022 to 2300 UTC on 15th May, 2022. The first box extended between 9 °N – 11 °N and 89.35 °E – 91.35 °E and covered the location of genesis and the earlier part of the intensification of the cyclone. The second box which spanned the location just prior to the cyclone reaching its maximum intensity extended between 12 °N – 14 °N and 86 °E – 88 °E. The third box extends between 14 °N – 16 °N and 80.5 °E – 82.5 °E and stretched over the final part of the TC trajectory where it started weakening and finally made landfall. In addition to the above three, the Ekman transport (M) associated with coastal upwelling was also computed for the third box which spanned the coastal region. For this, five coastal points that lay within this box area were selected and the coastal angle was computed at each of those points. The zonal and meridional wind speed were sampled at each of these five points for every time-step using a Cressman window (Cressman 1959) of 50 km radius, and the mean Ekman transport (M) over the five points were obtained. Further, ERA5 SST was also inter-compared with OSTIA SST and REMSS SST over each of these boxes to ascertain the performance of the SST products.