Chinese Journal of Oceanology and Limnology   2015, Vol. 33 Issue(5): 1164-1180     PDF       
http://dx.doi.org/10.1007/s00343-015-4125-7
Shanghai University
0

Article Information

ZHANG Lianxin (张连新), ZHANG Xuefeng (张学峰), HAN Guijun (韩桂军), WU Xinrong (吴新荣), CUI Xiaojian (崔晓健), SHAO Caixia (邵彩霞), SUN Chunjian (孙春健), ZHANG Xiaoshuang (张晓爽), WANG Xidong (王喜冬), FU Hongli (付红丽)
Impact of sea spray on upper ocean temperature during typhoon passage: simulation with a 1-D turbulent model
Chinese Journal of Oceanology and Limnology, 2015, 33(5): 1164-1180
http://dx.doi.org/10.1007/s00343-015-4125-7

Article History

Received May 23, 2014
accepted in principle Jul. 22, 2014;
accepted for publication Sep. 30, 2014
Impact of sea spray on upper ocean temperature during typhoon passage: simulation with a 1-D turbulent model
ZHANG Lianxin (张连新)1,2, ZHANG Xuefeng (张学峰)2, HAN Guijun (韩桂军)2 , WU Xinrong (吴新荣)2, CUI Xiaojian (崔晓健)2, SHAO Caixia (邵彩霞)2,3, SUN Chunjian (孙春健)2, ZHANG Xiaoshuang (张晓爽)2, WANG Xidong (王喜冬)2, FU Hongli (付红丽)2       
1 College of Physical and Environmental Oceanography, Ocean University of China, Qingdao 266100, China;
2 Key Laboratory of State Oceanic Administration for Marine Environmental Information Technology, National Marine Data and Information Service, State Oceanic Administration, Tianjin 300171, China;
3 National University of Defense Technology, Changsha 410073, China
ABSTRACT:At the interface between the lower atmosphere and sea surface, sea spray might significantly influence air-sea heat fluxes and subsequently, modulate upper ocean temperature during a typhoon passage.The effects of sea spray were introduced into the parameterization of sea surface roughness in a 1-D turbulent model, to investigate the effects of sea spray on upper ocean temperature in the Kuroshio Extension area, for the cases of two real typhoons from 2006, Yagi and Soulik.Model output was compared with data from the Kuroshio Extension Observatory (KEO), and Reynolds and AMSRE satellite remote sensing sea surface temperatures.The results indicate drag coefficients that include the spray effect are closer to observations than those without, and that sea spray can enhance the heat fluxes (especially latent heat flux) considerably during a typhoon passage.Consequently, the model results with heat fluxes enhanced by sea spray simulate better the cooling process of the SST and upper-layer temperature profiles.Additionally, results from the simulation of the passage of typhoon Soulik (that passed KEO quickly), which included the sea spray effect, were better than for the simulated passage of typhoon Yagi (that crossed KEO slowly).These promising 1-D results could provide insight into the application of sea spray in general circulation models for typhoon studies.
Keywordssea spray     upper layer temperature     typhoon     air-sea heat fluxes    
1 INTRODUCTION

As an energetic weather event that occurs over oceans, a typhoon is a powerful system that evolves via energy released by the ocean. The air-sea transports of momentum and enthalpy have been recognized as important elements in generating and maintaining typhoons. However, during high-wind conditions, e.g., the passage of a typhoon, large amounts of sea spray produced by the bursting of air bubbles in whitecaps and the tearing of spume from wave crests could significantly affect air-sea interactions.

Related studies have indicated that sea spray might also significantly affect the drag coefficient under conditions of high wind speed(Powell et al., 2003; Jarosz et al., 2007). Makin(2005)suggested that a thin boundary layer of air, adjacent to the sea surface, enters a regime of limited saturation because of the suspended sea droplets. This thin layer restrains the momentum transfer from the wind to the ocean, which explains the reduction of the drag coefficient observed by Powell et al.(2003). Furthermore, Makin(2005)derived a resistance law of the sea surface at high winds that makes it possible for us to study the effect of sea spray macroscopically.

Sea spray also has considerable influence on airsea heat fluxes in conditions with winds of typhoon strength, and Riehl(1954)revealed that sea spray evaporation produces a significant amount of heat. Fairall et al.(1994)simplified the timescale of the sea spray scheme(Andreas, 1989, 1992) and were the first to incorporate spray-based parameterization into a simple model of the tropical cyclone boundary layer. Their results showed that sea spray was a sink of sensible heat fluxes. Cloud microphysical equations(Pruppacher and Klett, 1978)have been used to study the thermal and moisture evolution of sea spray droplets(Andreas, 1989, 1990, 1992, 2003, 2004, 2010; Andreas and Emanuel, 2001; Andreas et al., 2008). The findings demonstrated that once wind speed is in excess of 11–13 m/s, the spray-induced sensible and latent heat fluxes become significant fractions of the corresponding air-sea interfacial fluxes. Although spray parameterization and its contribution to the heat fluxes have been studied for many years, the issue remains a challenging task because of the following two reasons. First, observations of sea spray are quite scarce for high winds and second, previous studies have considered the sea spray effect theoretically and microphysically, for which consistent and effective sea spray source functions are required. However, different sea spray source functions produce different results(Ling et al., 1980; Monahan, 1986; Bortkovskii, 1987; Andreas, 1992; Iida et al., 1992; Smith et al., 1993; Wu, 1993).

Because of the impact of sea spray on turbulent airsea fluxes, sea spray might subsequently influence the upper ocean temperature under conditions of high wind speed. However, most previous research(Fairall et al., 1994; Andreas and DeCosmo, 1999; Uang, 1999; Bao et al., 2000; Meirink and Makin, 2001; Wang et al., 2001; Perrie et al., 2005; Jeffrey and William, 2008)has focused on the effect of sea spray in simulations of typhoons. Little attention has been paid to the impact of sea spray on the upper layer sea temperature during a typhoon period using a numerical simulation approach.

Based on a 1-D turbulent model, this study introduces the spray effect into the parameterization of the sea surface aerodynamic roughness length to investigate its impact on the heat fluxes, drag coefficient, and upper ocean temperature. The sea spray effect is introduced using a macroscopic approach that avoids inconsistencies among different spray source functions under a microphysical framework. The remainder of this study is organized as follows. The 1-D turbulent model and the COARE model that computes heat fluxes are described in Section 2. The parameterization of sea spray used in this study is presented in Section 3. Section 4 describes the observations, typhoons, and experimental design of this study. The simulation results and conclusions are given in Sections 5 and 6, respectively.

2 MODEL DESCRIPTION2.1 GOTM Model

To address the problems raised in Section 1 and avoid the complexities of a general circulation model, a 1-D water-column model, i.e., the General Ocean Turbulence Model(denoted as GOTM; Burchard, 2006), which simulates vertical mixing and advection processes near the ocean surface, was employed in this study. The governing equations of temperature and salinity of the GOTM are formulated as follows:

where θ and S represent the mean potential temperature(units: °C) and salinity, respectively, v ' and v '' denote the molecular diffusivities of heat and salt, respectively, and C P denotes the heat capacity of seawater. The source terms of temperature and salinity, the terms on the right-h and sides of Eqs.1 and 2a, are the vertical divergence of solar radiation(I) and the sum of the turbulent and viscous transport terms Ds, where ρ 0 is a constant reference density resulting from the Boussinesq approximation. In Eq.2b, v t S and v S are the turbulent and molecular diffusivities of salinity, respectively, and denotes the non-local flux of salinity. The surface freshwater is given by means of the precipitation P−evaporation E.Ds= S(P–E) at the surface, with P–E given as a velocity(note that Ds is the flux in the direction of z and thus, it is positive for a loss of salinity). The mean vertical turbulent fluxes of momentum and heat appearing in Eqs.1 and 2 are computed according to: where K V denotes the vertical turbulent diffusivity of heat and is equal to c ' μ k1/2 l . The non-dimensional quantity c'μ is the function of the non-dimensional stability parameter that describes the influence of stratification on turbulent mixing. Parameter k denotes the turbulent kinetic energy and lk3/2ε -1 is the integrallength scale, computed here from the dissipation rate ε . We will use the k-epsilon model with second moment closure to simulate those turbulence parameters(Umlauf and Burchard, 2005).

2.2 COARE Model

The COARE version 2.6 bulk flux algorithm has been tuned to fit measurements made on the R/V Moana Wave during the three different cruiselegs undertaken as part of the Coupled Ocean-Atmosphere Experiment. To provide the forcing for the GOTM, COARE version 2.6(Fairall et al., 1996)was used to calculate the following turbulent fluxes based on the Monin-Obukhov Similarity Theory:

where H s, H l , and τ represent the turbulent fluxes of sensible and latent heat, and turbulent momentum flux, respectively; ρa is the air density; c pa is the specific heat of air at constant pressure; Le is the latent heat of vaporization of water; and T *, q *, and u * are the related Monin-Obukhov Similarity(MOS)scaling parameters. The parameters w ', θ ', and q ' denote the turbulent fluctuations of vertical wind, temperature, and water vapor mixing ratio, respectively; the over bar indicates the mean value.

The st and ard bulk expressions for the scalar fluxes and momentum flux are:

where C d, C h, and C k are the transfer coefficients for momentum flux, and sensible and latent heat, respectively; S is the average value of the wind speed relative to the fixed earth, each measured at some reference atmospheric height zr ; Ts is the sea surface interface temperature; usi is the surface current; and q s is the interfacial value of the water vapor mixing ratio. The parameter ui is one of the horizontal wind components relative to the sea surface at zr and averaged as in Eq.4. T is the air temperature and q is the specific humidity of air.

The transfer coefficients in Eq.5 are partitioned into individual profile components:

which are themselves functions of the fluxes in a manner described by the MOS surface-layer theory: Here, κ is the von Kármán constant, α accounts for the difference in the scalar and velocity von Kármán constants, ψ is the MOS profile function(assumed the same for temperature and humidity), and ξ = zr / L, where . The subscript n denotes the value in neutral conditions(i.e., ξ =0)where ψ =0. The neutral transfer coefficients are related to the roughness lengths(z0 for the velocity, z0t for temperature, and z0q for humidity), which are defined as the heights where the extrapolations of the log-z portion of the respective profiles(of u, t, or q)intersect the surface value:

Laboratory studies have proven that it is convenient to characterize the surface and flow regime by the roughness Reynolds number:

where v is the kinematic viscosity of air. and R t and R q, as functions of R r, are the roughnessReynolds number for temperature and moisture, respectively.

3 SURFACE DYNAMIC ROUGHNESS Z 0 PARAMETERIZATIONS

Based on the COARE model, this section introduces the effect of sea spray into the parameterization of z0, before which the traditional parameterization of z0 is first presented.

3.1 Traditional parameterization of surface dynamic roughness

In numerical weather prediction models and oceanographic applications, surface roughness is commonly computed by the following Charnock relation(Charnock, 1995):

where g represents gravitational acceleration and α is the Charnock parameter. According to COARE version 2.6, the Charnock parameter α is set to 0.011(Fairall et al., 1996). This Charnock relationship has been proven to work well in many applications for low-moderate wind speeds, and on a theoretical basis, proven accurate for treating interfacial turbulent fluxes in wind speeds of up to 10 m/s(Fairall et al., 1996). Previous work has demonstrated that the contribution of spray to the heat fluxes becomes significant when winds reach 11–13 m/s; therefore, it should still be accurate for interfacial turbulent fluxes when extrapolated to higher wind speeds.

3.2 Physical effect of sea spray on surface dynamic roughness

It is known that the large amount of sea spray produced by breaking surface waves forms a stable and limiting saturation suspension layer in the marine atmospheric surface layer. The use of Global Positioning System sondes to measure the profiles of strong winds in the marine boundary layer associated with tropical cyclones has shown that surface momentum fluxlevels off as wind speeds increase above typhoon force(Powell et al., 2003). Furthermore, when wind speed exceeds 25 m/s, the layer that includes the effect of sea spray affects the drag coefficient(Powell et al., 2003). Subsequently, Makin(2005)presented the following parameterization scheme of z0 at high wind speeds based on the observations of Powell et al.(2003):

where ω=min(1, acr/ku*) and cl=gh1/u*. Here, u* is the friction velocity, acr represents the terminal fall velocity of sea spray that is set to 0.64 m/s(Makin, 2005), κ is the von Kármán constant that is set to 0.4, and ω is the correction parameter indicating the impact of sea spray on the logarithmic wind profile. Note that the effect of sea spray is partly reflected in ω . The parameter h l is the height of the sea spray suspension, which is higher(lower)than that of the breaking(significant)wave height. Thus, h l is simply set to 1/10 of the significant wave height(Makin, 2005; Liu et al., 2012). The parameter c l is the nondimensionalized quantity of the height of the suspension layer of sea spray droplets. Therefore, we havewhere H denotes the significant wave height. According to the above description of the parameterization of z0 at high wind speeds, it can be seen that the effect of sea spray is implied in ω and c l .

For the parameterization of z0 at mid-to-low wind speeds, recent studies have indicated that z0 depends on the wind speed and wave states(Masuda and Kusaba, 1987; Donelan, 1990; Maat et al., 1991; Smith, 1992; Monbaliu, 1994; Vickers and Mahrt, 1997; Johnson et al., 1998). For open oceans, Donelan(1990)proposed a widely used parameterization of z0, which is the function of the wave state:

where c p is the peak wave phase velocity.

Finally, to provide the value of z0 with the sea spray effect, for the full wind speeds based on the open ocean observations, this study combined the sea surface dynamical roughness length for moderate(Eq.14) and high winds(Eq.12)using the 3/2 power law(Toba, 1972) and the relation between significant wave period and peak wave period:

where β * represents the wave age, which is computed by Toba’s(1972)3/2 power law and the relation between significant wave period and peak wave period as:

Here, Eq.15 includes the effect of the spray for the full wind condition. Apparently, for low wind conditions with ω =1, the full wind conditional roughness length(Eq.15)degrades the low wind conditional roughness length(Eq.14). Under high wind conditions(ω<1), the impact of sea spray on sea surface roughness is large, whichleads to a decrease of the roughness length and drag coefficient.

The COARE model and sea surface dynamical roughness length z0(Eq.15)are used to compute the total air-sea sensible and latent heat fluxes(denoted as H s, tot and H l, tot, respectively), and total momentum flux, which include the sea spray effect. In contrast, interfacial sensible heat flux, latent heat flux(H s and H l, respectively), and interfacial momentum flux are calculated through the COARE model and sea surface dynamical roughness length z0(Eq.11)without the spray effect.

4 OBSERVATIONS, TYPHOONS, AND EXPERIMENTAL DESIGN4.1 Observations

One primary observation data set used in this study was from the Kuroshio Extension Observatory(KEO)surface mooring located at(32.4°N, 144.6°E), which was first deployed in mid-June 2004. The KEO surface measurements include 3-m wind speed and direction, 3-m air temperature, 3-m relative humidity, rain rate, solar and longwave radiation, sea temperature profile, sea salinity profile, and sea surface temperature(SST) and salinity measurements at a 1-m depth. The temporal resolution of all measurements was 10 min with the exception of radiation(2 min). The observed layers of temperature and salinity were 1, 10, 15, 25, 50, 75, 100, 150, 200, 300, 400, 450, and 500 m and 1, 10, 15, 50, 75, and 400 m, respectively. The period of KEO data used in this study was from September 17 to October 29, 2006.

4.2 Typhoons

In this section, we introduce two cases of real typhoons(Yagi and Soulik), both of which passed the KEO station, but with differing speeds of movement. Based on these two typhoons, we studied the impact of sea spray on surface heat fluxes. The measurements recorded during these typhoons included the trajectory, central maximum wind speed, and time of the observation. The basic characteristics of the two typhoons are described in the following.

Case 1

Typhoon Yagi started as a tropical depressionlocated at about 19.7°N, 156.2°E on September 17, 2006. Figure 1 shows the track of Yagi and thetemporal variation of SST within the study area(25°–32°N, 140°–148°E)from September 22 to September27, 2006. Yagi was upgraded to a super typhoon onSeptember 22 and the SST exceeded 26°C in the area.Subsequently, Yagi moved northeast and passed theKEO station with a speed of movement and centralmaximum wind speed of 6.92 and 45 m/s onSeptember 23, 2006. During this period, there was abroad area of low SST around the right-h and side ofYagi(Liu et al., 2006) and although Yagi moved awayfrom the area on September 24, the area of lowtemperature persisted until September 27, 2006.

Case 2

Typhoon Soulik formed in the Northwest Pacific at12:00 UTC on October 9, 2006. Soulik reachedtyphoon status and approached the KEO station onOctober 15, 2006, at 14 m/s(Fig. 2). Soulik rapidlyintensified to category 1(i.e., wind speeds exceeding35 m/s). The SST on the right-h and side of Soulik wasless than 25°C and this persisted for about 4 days(September 16–19).

4.3 Experimental design

Based on the GOTM model, COARE 2.6 model, and two parameterizations of z0, four experiments(listed in Table 1)were designed to investigate the impact of sea spray on the upper ocean temperature during the passage of a typhoon. It should be noted that to use the KEO observations efficiently, the 1-D GOTM was fixed to the position of the KEO station with 200-m depth and 200 vertical layers with intervals of 1.0 m. Furthermore, to use the typhoon data, the simulation period was extended from September 17 to October 20, 2006. Equation 15(Eq.11)is the sea surface dynamical roughness length to include(exclude)the effect of sea spray. The COARE model and Eq.15 are used to compute the total air-sea H s, tot, H l, tot, and total momentum flux, which include the sea spray effect. In contrast, H s, H l, and interfacial momentum flux are calculated using the COARE model and Eq.11 without the spray effect. Here, the word “total” indicates that it comprises contributions from both the usual interfacial turbulent fluxes and spray-mediated fluxes. Thus, sprayinduced sensible heat flux Qs and latent heat flux Ql can be given by:

Tab. 1 Summary of the GOTM simulations used in this study

Based on the above descriptions, Test1 computed the air-sea momentum flux and surface heat fluxes without the effect of sea spray. Based on Test1, Test2(Test3)introduced the effect of sea spray to the computation of air-sea momentum flux(surface heat fluxes). Test4 introduced the effect of sea spray to both the air-sea momentum flux and surface heat fluxes. Note that the ocean was initialized with the same temperature and salinity profiles from the KEO station in all four cases.

5 RESULT

According to the experimental design, we first examine the impact of sea spray on the drag coefficient and heat fluxes and then investigate the impact of sea spray on sea temperature.

5.1 Impact on drag coefficients

To investigate the impact of sea spray on the drag coefficient, the values of surface roughness z0 with and without sea spray were firstly computed using Eqs.15 and 11, respectively. Then, the drag coefficient without sea spray was computed using the Charnock relationship, which is denoted by the red line in Fig. 3. For the drag coefficient with sea spray, because the drag coefficient is a function of the 10-m wind speed and the wave age, the drag coefficients were calculated with different 10-m wind speeds and wave ages(Fig. 3). To verify the validity of the drag coefficient with the effect of sea spray, values of drag coefficient derived from the observations of Jarosz et al.(2007) and based on different resistance coefficients, are also plotted in Fig. 3. The drag coefficient with the effect of sea spray increases(decreases)with wind speed under low(high)wind conditions for different wave ages, which means the increase of sea spray mass with wind speedleads to a reduction of the drag coefficient when the wind speed exceeds about 35 m/s.

Fig. 3 Variation of drag coefficient multiplied by 1000 with respect to 10-m wind speed under different wave ages(8, 10, 15, 20, 25, 30, and 35) Drag coefficient is computed by Eq.15, which includes effect of sea spray(black marker lines). To facilitate the comparison, the drag coefficientcomputed by the traditional Charnock relation(Eq.11; alpha=0.011)without sea spray effect(red line) and those derived by the observations of Jarosz et al.(2007)based on different resistance coefficients(0.1, 0.0505, 0.02, and 0.001 cm/s)are plotted(blue marker).

For high wind speed, a suspension layer is formed by the spray droplets produced by intensive wave breaking in the marine atmospheric surface layer. In the suspension layer, the heaviest droplets are closer to the sea surface and thus, the spray droplets over the ocean form a very stable and limiting saturation boundary layer, which restrains momentum transfer from the surface wind to the ocean(Makin, 2005). Therefore, spray droplets can influence the airflow dynamics at wind speeds exceeding 33 m/s. Furthermore, the drag coefficient will reduce with the increment of wind speed, which is in agreement with the “observed” drag coefficient. If the wind speed is fixed, the drag coefficient decreases with the increment of wave age. In contrast, C d without the sea spray effect does not vary with different wave age, but does increase monotonously with wind speed.

It is well known that the ratio of the enthalpy exchange coefficient C k to the drag coefficient C d might have considerable influence on typhoon simulations(Emanuel, 1995; Black et al., 2007). Here, we also examine the impact of sea spray on the C k / C d ratio, as illustrated in Fig. 4. Equation 15(Eq.11)is used to calculate C k to include(exclude)the effect of sea spray. It can be seen that the ratio without the effect of sea spray(blue line)decreases with increasing wind speed, and it cannot increase to the 0.75 threshold value for typhoon development, as proposed by Emanuel(1986, 1995), at high wind speeds. However, the ratio with the sea spray effect initially decreases with the increment of the 10-m wind and then increases to the threshold value of 0.75(thin line)at wind speeds of typhoon strength(black marker lines). The ratio can increase to the threshold of 0.75 more easily in young waves than in old wave ages. It is worth mentioning that although no observations of C k are available to verify the corrections of the ratio with the effect of sea spray, based on the above analysis the introduction of sea spray might provide the possibility of improvement in typhoon simulations.

Fig. 4 Variation of the ratio of enthalpy exchange coeffi cient to drag coeffi cient with respect to 10-m wind speed under different wave ages (8, 10, 15, 20, 25, 30, and 35) C k and C d are computed by Eq.15, which includes the effect of sea spray (black marker lines). Blue line is the ratio derived by the Charnock relationship(Eq.11; alpha=0.011). Thin line represents the 0.75 threshold value of the ratio for tropical cyclone development, as proposed by Emanuel, 1995.
5.2 Impact on heat fl uxes

Case 1: Typhoon Yagi

Figure 5 shows the hourly time series of both the heat fluxes(Fig. 5d and e) and the air-sea bulk variables, including wind speed(blue curve in Fig. 5a), wind direction(dashed dots in Fig. 5a), sea-air temperature difference(blue curve in Fig. 5b), SST(black curve in Fig. 5b), and relative humidity(Fig. 5c)during the passage of typhoon Yagi. The heat fluxes were derived from the COARE model and the bulk variables obtained from KEO data from September 19–29, 2006. As shown in Fig. 5, prior to the passage of Yagi(September 19–22), wind speed over the KEO station wasless than 10 m/s and SST within the study area exceeded 28°C. Sea spray had almost no effect on the heat fluxes. When Yagi passed the KEO mooring, wind speed over the station reached two peak values(i.e., 32 and 34 m/s)on September 23. At 19:00 UTC September 23, the relative humidity reached a maximum, which coincided with a rapid decrease in SST and transition of wind direction. According to Fig. 5d and e, the spray enhanced the two peak values of heat fluxes, i.e., the total latent(sensible)heat flux was enhanced by up to 147.7(53.47)W/m 2, which is an increment of about 29%(26%)compared with the interfacial latent(sensible)heat flux. When typhoon Yagi was far from the mooring(September 24–29), sea surface mean wind speed was about 15 m/s(blue curve in Fig. 5a), but sea spray droplets still affected the air-sea heat fluxes. The spray-induced mean sensible and latent heat fluxes were about 4 and 12 W/m 2, respectively. It is shown that the ocean lost more heat to the air because of the sea spray effect during typhoon Yagi. Under such circumstances, the heat released by the ocean was more than that absorbed by the ocean during typhoon Yagi.

Fig. 5 Time series of (a) wind speed (blue curve, m/s, left-hand y -axis) and direction (dashed dots, in degrees, right-handy -axis) at 10-min intervals; (b) air-sea temperature difference (denoted as SST-Ta, blue curve in °C, left-hand y -axis)and sea surface temperature (denoted as SST, black curve in °C, right-hand y -axis); (c) relative humidity of the air;(d) total sensible heat flux (denoted as H s,tot , red bars), interfacial sensible heat fl ux (denoted as H s , blue bars), and spray-induced sensible heat flux (denoted as Q s , black bars) (W/m 2 ); and (e) total latent heat flux (denoted as H l,tot , red bars), interfacial latent heat flux (denoted as H l , blue bars), and spray-induced latent heat fl ux ( Q l , black bars) (W/ m 2 ) for typhoon Yagi The period is from 12:00 UTC September 19, 2006 to 12:00 UTC September 29, 2006.

Case 2: Typhoon Soulik

Similar to Fig. 5, Fig. 6 displays the related results for the passage of typhoon Soulik. Soulik arrived on October 10–14, 2006, when the sea surface wind speed over the KEO station increased to 15 m/s(blue curve in Fig. 6a). The sea spray firstly modulated the latent heat flux with the spray-induced latent heat flux being about 15 W/m 2 . Soulik crossed the KEO mooring on October 15, when the wind speed reached 31.7 m/s(blue line in Fig. 6a), corresponding to the maximum relative humidity(black curve in Fig. 6c) and a shift of wind direction(black curve in Fig. 6a). The total heat fluxes were affected consistently and significantly by sea spray. The spray-mediated latent(sensible)heat flux was enhanced to 107.05(35.6)W/m 2, which is an increment of about 26%(9.4%)compared with the interfacial latent(sensible heat flux). When Soulik was far from the KEO mooring(October 17–20), the wind speed wasless than 10 m/s. During this period, sea spray had no obvious influence on the heat fluxes.

Fig. 6 Same as in Fig. 5, but for typhoon Soulik(from 12:00 UTC Oct. 10, 2006 to 12:00 UTC Oct. 20, 2006)

As can be seen from Figs.5 and 6, the inclusion of sea spray can significantly enhance heat exchange at the air-sea interface, especially for latent heat exchange. These results are consistent with the microphysical processes of sea spray droplets. Spray droplets are ejected into the air-sea interfacial layer with the same temperature as the surface sea, whereupon they cool to the wet-bulb temperature. Thereby, the spray enhances the sensible heat transfer between the air and the ocean(Liu, 2007). The sprayinduced latent heat exchange begins after the sensible heat reaches the equilibrium between the spray droplets and the ambient air. Because the heat required for evaporation of the spray droplets is extracted from the air, the spray-mediated latent heat flux shows as a sink in the boundary condition for sensible heat(Andreas et al., 2008). Hence, the effect of sea spray on the latent heat flux is more prominent, which indicates that spray might resolve the problem of insufficient energy being supplied to the storm to maintain its development(Zhao, 2012).

Here, we also investigate the relationship betweenthe spray-induced heat fluxes and wind speed and wave state. Figure 7 shows the dependencies of thespray-mediated heat fluxes(Qs, Ql)on wind speedduring the passage of Yagi. The black curves in Figs.7 and 8 represent the trends fitted by 3-order cubicpolynomials. For wind speedless than 10 m/s, Qs and Ql are almost negligible, whereas for wind speedsover 20 m/s, Qs and Ql both increase rapidly as thewind speed increases. Thus, the nonlineardependencies of Qs and Ql on wind speed aresignificant for higher wind speeds.

Fig. 7 Variations of spray-induced sensible heat flux(Qs, a) and latent heat flux(Ql, b)with respect to 10-m wind speed for Yagi Black curves are the trends fitted by 3-order polynomial.

In addition to the dependencies of Qs and Ql on wind speed, recent studies of microscopic physics have revealed that the spray source function is also related to wave state(Zhao et al., 2006; Fairall et al., 2009). Here, we also examine the relation between Qs and Ql and the wave state from the macroscopic viewpoint, which in this study, denotes the wave state as the wave age(denoted as β *). Figure 8 shows the variations of Qs(Fig. 8a) and Ql(Fig. 8b)with respect to β * during the passage of Yagi. The results show that Qs and Ql decrease as β * increases, i.e., there is negative correlation between the spray-induced heat fluxes and wave age. Sea spray hardly affects the airsea heat fluxes if the wave age is greater than 20. For small values of β *, according to Eq.16, wind speed is much larger than wave speed, which causes a much rougher sea surface and greater concentration of sea spray. Under this situation, sea spray could significantly affect the total heat fluxes.

Fig. 8 Same as Fig. 7, but with respect to wave age(β *)
5.3 Impact on sea temperature

In this section, SSTs simulated by the GOTMduring the passages of the two typhoons are comparedwith the AMSRE and Reynolds satellite SSTobservations, as well as the KEO observed SSTs.Then, the simulated upper ocean temperature isexamined. As with Section 5.2, this section is dividedinto two parts.

Case 1: Typhoon Yagi

Figure 9 shows the temporal evolutions of the daily averaged SSTs simulated by Test1(red curves), Test2(brown curves), Test3(green curves), and Test4(black curves)during the passage of Yagi. For comparison, the observed SSTs from the KEO mooring(blue stars), AMSRE(dashed curve), and Reynold(dotted curve)are also plotted. The simulated SSTs of the four cases are consistent with the observed SSTs before Yagi crossed the KEO mooring(on September 19–22). Three days after Yagi hadleft the KEO station area(i.e., September 25), the simulated SSTs of Test2 and Test4 decrease to about 26.8°C, i.e., closer to the observations of the AMSRE and the Reynold SSTs than Test1 and Test3. On September 27, the results of Test2 and Test4 show a significant warming trend, which is almost the same as that observed by the KEO station. A comparison between the four experiments demonstrates that the improvement of the SSTs in Test4 is mainly because of the refinement of the drag coefficient, which considers the effect of sea spray rather than heat fluxes. In contrast, the drag coefficients in Test1 and Test3, without the impact of spray, cannot produce sufficient cooling of the SST. Generally, the simulated SSTs with the impact of spray outperform those without spray during the passage of Yagi.

Figure 10 displays the upper 60-m temperature profiles simulated by Test1(red curves), Test2(brown curves), Test3(green curves), and Test4(black curves)at 12:00Z on(a)September 22, (b)September 24, (c)September 25, (d)September 26, (e)September 27, and (f)September 28. The upper layer temperature profiles of the four cases are similar to the KEO observations on September 22. Therefore, sea spray does not affect the temperature profiles noticeably before typhoon Yagi reaches the KEO location. However, the temperature profiles at the station cooled significantly after the typhoon’s passage(September 24–26)for all four cases, especially Test2 and Test4 that include the effect of sea spray on C d . The temperature simulated by Test4 is closer to the observations than that simulated by Test1, which does not consider the effect of sea spray(temperature of water below 60 m becomes cold, so it is not shown). During September 27–29, Test4 and Test2 simulate the warming process in the upper layer commendably when Yagi was far from the KEO station.

The slow speed of movement of typhoon Yagi caused intense cooling of the upper ocean(Bender et al., 1993). Although Test4 greatly improves the quality of the simulated SST, it cannot completely capture the cooling process observed by the KEO mooring, which might be owing to the limitations of the 1-D ocean model(Yablonsky and Ginis, 2009).

Case 2: Typhoon Soulik

Similar to Fig. 9, Fig. 11 shows the results for typhoon Soulik. Before Soulik arrived at the KEO station(October 10–14, 2006), the simulated results of all four cases are almost the same as the KEO data, which further verifies that spray does not significantly affect SSTs before the passage of the typhoon. After Soulik hadleft the station area(October 16–18), Test4 and Test2 simulate clearly the reduction in SST, consistent with the KEO data. Because of the introduction of spray-induced heat fluxes, Test4(Test3)is slightly better than Test2(Test1). However, without considering the effect of spray in the drag coefficient, the amount of cooling simulated by Test1 and Test3 is much smaller than simulated by Test2 and Test4.

Fig. 9 Variation of daily averaged SST during passage of typhoon Yagi Red, brown, green, and black curves represent the results of Test1, Test2, Test3, and Test4, respectively, which are described in the experimental design. Blue asterisks, dashed line, and dotted line represent KEO, AMSRE, and Reynolds SSTs, respectively.

Fig. 10 Upper 60-m temperature profi les simulated by Test1 (red curves), Test2 (brown curves), Test3 (green curves), and Test4 (black curves) at (a) 12:00 UTC September 22, (b) 12:00 UTC September 24, (c) 12:00 UTC September 25, (d) 12:00 UTC September 26, (e) 12:00 UTC September 27, and (f) 12:00 UTC September 28, 2006 Blue asterisks indicate the KEO observed temperatures.

Fig. 11 Same as Fig.9, but for the passage of Soulik, but without the observations of AMSRE and Reynolds plotted

Figure 12 shows similar results to Fig. 10, but for the passage of Soulik. The results show that both Test4 and Test2 capture the reduction in the upper temperature after Soulik hadleft the KEO station area. Again, Test4(Test3)is better slightly than Test2(Test1)because of the inclusion of spray-induced heat fluxes. Without introducing sea spray into the drag coefficient and heat fluxes, the temperature profiles simulated by Test1(red curves)deviate considerably from the KEO observations during the passage of the typhoon. Therefore, sea spray significantly affects SST, primarily through the momentum flux. The heat fluxes, considering the effect of sea spray, do not remarkably affect oceanic cooling in the 1-D ocean model.

Fig. 12 Upper 75-m temperature profiles simulated by Test1(red curves), Test2(brown curves), Test3(green curves), and Test4(black curves)at(a)12:00 UTC September 14, (b)12:00 UTC September 16, (c)12:00 UTC September 17, and (d)12:00 UTZ September 18 For comparison, the KEO observed temperatures(blue asterisks)are also plotted.Fig. 11 Same as Fig. 9, but for the passage of Soulik, butwithout the observations of AMSRE and Reynoldsplotted

Additionally, because the speed of movement of Soulik was slower than that of Yagi when they crossed the KEO station, the cooling of the upper ocean temperature during Soulik’s passage isless intense. This might be the reason for the better simulation of temperatures in Test4 during the passage of Soulik compared with the passage of Yagi(comparing Fig. 12bd with Fig. 10bd).

6 CONCLUSION

At the interface between the lower atmosphere and the ocean surface during the passage of a typhoon, sea spray might significantly affect the air-sea interaction and subsequently, modulate the upper ocean temperature. To investigate this issue, we introduced the effect of sea spray into the parameterization of sea surface roughness for full wind speed conditions, which further affects the drag coefficient and heat flux through the bulk formula. The results showed that under conditions of high wind speeds, the unified parameterization of surface roughness performed much better than the traditional Charnock relation did. An area of the Kuroshio Extension, where the ocean and atmosphere interaction is intense, was selected to extend the investigation of this study. Two cases of real typhoons(i.e., Yagi and Soulik)that occurred in 2006 and crossed the study domain were chosen. Observations from the Kuroshio Extension Observatory(KEO)surface mooring, situated within the study area were used to explore the impact of sea spray on heat fluxes. The results demonstrated that the introduction of sea spray could considerably enhance the heat fluxes(especially latent heat flux)during the passage of a typhoon. Additionally, the spray-induced heat fluxes increased nonlinearly as the wind speeds increased under conditions of high wind speeds.

Fig. 10 Upper 60-m temperature profiles simulated by Test1(red curves), Test2(brown curves), Test3(green curves), and Test4(black curves)at(a)12:00 UTC September 22, (b)12:00 UTC September 24, (c)12:00 UTC September 25, (d)12:00 UTC September 26, (e)12:00 UTC September 27, and (f)12:00 UTC September 28, 2006Blue asterisks indicate the KEO observed temperatures.

Based on the GOTM 1-D oceanic turbulence model, the impact of sea spray on the upper ocean temperature was investigated using the results of the earlier experiments, AMSRE and Reynold SSTs, as well as KEO observations. The results showed that the introduction of sea sprayled to better simulation of the cooling process of the SST and the upper-layer temperature profiles than when sea spray was not considered. Additionally, the upper-layer temperature profiles simulated by the model with the sea spray effect during typhoon Soulik(that passed KEO quickly)were better than those for typhoon Yagi(that crossed KEO slowly). These findings might provide insight regarding the application of sea spray in general circulation models or ocean-atmosphere coupled models for typhoon studies. The principal limitation of this study was the use of the 1-D ocean model, which means that horizontal advection and other associated physical processes were ignored. Moreover, the conclusions derived from this study should be validated in other domains and for other typhoon passages. Future work should therefore introduce evaporation of sea spray into atmosphereocean coupled models to evaluate such effects during the passage of a typhoon. In the meantime, the microphysical processes of sea spray on air-sea heat and moisture fluxes will be considered further.

References
Andreas E L, DeCosmo J.1999.Sea spray production and influence on air-sea heat and moisture fluxes over the open ocean.Air-Sea Exchange: Physics, Chemistry and Dynamics.In: Geernaert G L ed., Kluwer Academic.p.327-362.
Andreas E L, Emanuel K A.2001.Effects of sea spray on tropical cyclone intensity.J.Atmos.Sci, 58: 3 741-3 751.
Andreas E L, Persson P O G, Hare J E.2008.A bulk turbulent air-sea flux algorithm for high-wind spray conditions.J.Phys.Oceanogr., 38: 1 581-1 596.
Andreas E L.1989.Thermal and size evolution of sea spray droplets.CRREL Rep.89-11.U.S.Army Cold Regions.Research and Engineering Laboratory.37p.
Andreas E L.1990.Time constants for the evolution of sea spray droplets.Tellus B, 42: 481-497.
Andreas E L.1992.Sea spray and the turbulent air-sea heat fluxes.J.Geophys.Res., 97: 11 429-11 441.
Andreas E L.2003.An algorithm to predict the turbulent airsea fluxes in high-wind.spray conditions.Preprints.12th Conference on Interactions of the Sea and Atmosphere.Long Beach.CA, Amer Meteor Soc., 3.4.
Andreas E L.2004.A bulk air-sea flux algorithm for high-wind spray conditions.Version 2.0.Preprints.13th Conference on Interactions of the Sea and Atmosphere.Long Beach,CA, Amer.Meteor Soc., 1.5.
Andreas E L.2010.Spray-mediated enthalpy flux to the atmosphere and salt flux to the ocean in high winds.J.Phys.Oceanogr., 40: 608-619.
Bao J W, Wilczak J M, Choi J K et al.2000.Numerical simulations of air-sea interaction under high wind conditions using a couple model: a study of hurricane development.Mon.Weather.Rev., 128: 2 190-2 210.
Bender M A, Ginis I, Kurihara Y.1993.Numerical simulations of tropical cyclone-ocean interaction with a high-resolution coupled model.J.Geophys.Res., 98 (D12): 23 245-23 262.
Black P G, D'Asaro E A, Drennan W M et al.2007.Air-sea exchange in hurricanes: synthesis of observations from the coupled boundary layer air-sea transfer experiment.Bull.Am.Meteorol Soc., 88 (3): 357-374.
Bortkovskii R S.1987.Air-Sea Exchange of Heat and Moisture during Storms.D.Reidel.194p.
Burchard H, Bolding K, Kühn W, Meister A et al.2006.Description of a flexible and extendable physicalbiogeochemical model system for the water column.J.Mar.Sys., 61: 180-211.
Charnock H.1995.Wind stress on a water surface.Q.J.R.Meteorol.Soc., 81 (350): 639-640.
Donelan M A.1990.Air-Sea Interaction.The Sea: Ocean Engineering Science, 9.In: LeMéhauté B, Hanes D eds.John Wiley and Sons, Inc., New York.p.239-292.
Emanuel K A.1986.An air-sea interaction theory for tropical cyclones.Part I.Steady-state maintenance.J.Atmos.Sci., 43: 585-604.
Emanuel K A.1995.Sensitivity of tropical cyclone to surface exchange coefficient and a revised steady-state model incorporating eye dynamics.J.Atmos.Sci., 52: 3 969-3 976.
Fairall C W, Banner M L, Person W L et al.2009.Investigation of the physical scaling of sea spray spume droplet.J.Geophys Res., 114: C100001, http://dx.doi.org/10.1029/2008JC004918.;
Fairall C W, Bradley E F, Rogers D P, Edson J B et al.1996.Bulk parameterization of air-sea fluxes for Tropical Ocean-Global Atmosphere Coupled-Ocean-Atmosphere Experiment.J.Geophys.Res., 101: 3 747-3 764.
Fairall C W, Kepert J D, Holland G J.1994.The effect of sea spray on surface energy transports over the ocean.Global Atmos.Ocean Syst., 2: 121-142.
Iida N, Toba Y, Chaen M.1992.A new expression for the production rate of sea water droplets on the sea surface.J.Oceanogr., 48: 439-460.
Jarosz E, Mitchell D A, Wang D W et al.2007.Bottom-up determination of air-sea momentum exchange under a major tropical cyclone.Science, 315 (5819): 1 707-1 709, http://dx.doi.org/10.1126/science.1136466.
Jeffrey S G, William M F.2008.Effects of sea spray on tropical cyclones simulated under idealized conditions.Mon.Weather.Rev., 136: 1 686-1 705.
Johnson H K, Hojstrup J, Vested H J et al.1998.On the dependence of sea surface roughness on wind waves.J.Phys.Oceanog., 28: 1 702-1 716.
Ling S C, Kao T W, Saad A I.1980.Microdroplets and transport of moisture from ocean.Proc.ASCE J.Eng.Mech.Div., 106: 1 327-1 339.
Liu B, Guan C L, Lian X.2012.The wave state and sea spray related parameterization of wind stress applicable from low to extreme winds.J.Geophys.Res., 117: 1-10.
Liu B.2007.Physical Basis and Numerical Study of the Coupled Atmosphere-Wave Model.PhD Thesis, Ocean University of China.(in Chinese with English abstract)
Liu Z H, Xu J P, Zhu B K et al.2006.Upper ocean to tropical cyclones in northwestern Pacific during 2001-2004 by Argo data.Journal of Tropical Oceanography, 25: 1-8.(in Chinese with English abstract)
Maat N, Kraan C, Oost W A.1991.The roughness of wind waves.Bound Layer Meterol., 54: 89-103.
Makin V K.2005.A note on the drag of the sea surface at hurricane winds.Boundary Layer Meteorol., 115: 169-176.
Masuda A, Kusaba T.1987.On the equilibrium of winds and wind-waves in relation to surface drag.J.Oceanogr.Soc.Jpn, 43: 28-36.
Meirink J F, Makin V K.2001.The impact of sea spray evaporation in a numerical weather prediction model.J.Atmos.Sci., 58 (23): 3 626-3 638.
Monahan E C.1986.The ocean as a source for atmospheric particles.In: Buat-Menard P ed.The Role of Air-Sea Exchange in Geochemical Cycling, D.Reidel Publ.Co.,Dordrecht, Holland.p.129-163.
Monbaliu J.1994.On the use of the Donelan wave spectral parameter as a measure for the roughness of wind waves.Bound Layer Meterol., 67: 277-291.
Perrie W, Andreas E L, Zhang W et al.2005.Sea spray impacts on intensifying midlatitude cyclones.J.Atmos.Sci., 62: 1 867-1 883.
Powell M D, Vickery P J, Reinhold T A.2003.Reduced drag coefficient for high wind speeds in tropical cyclones.Nature, 422: 279-283.
Pruppacher H R, Klett J D.1978.Microphysics of Clouds and Precipitation.D.Reidel, Norwell, Mass.714p.
Riehl H.1954.Tropical Meteorology.McGraw-Hill.392p.
Smith J A.1992.Observed growth of Langmuir circulation.J.Geophys.Res., 97: 5 651-5 664.
Smith M H, Park P M, Consterdine I E.1993.Marine aerosol concentrations and estimated fluxes over the sea.Quart.J.Roy.Meteor.Soc., 119: 809-824.
Toba Y.1972.Local balance in the air-sea boundary process.I.On the growth process of wind waves.J.Oceanogr.Soc.Jpn., 28: 109-120.
Uang C L.1999.Impact of sea spray and oceanic on the development of tropical cyclones.Preprints, 23rd Conference on Hurricanes and Tropical Meteorology,Dallas, TX, Amer.Meteor.Soc.p.30-31.
Umlauf L, Burchard H.2005.Second-order turbulence closure models for geophysics boundary layers.A review of recent work.Cont.Shelf Res., 25: 795-827.
Vickers D, Mahrt L.1997.Fetch limited drag coefficients.Bound Layer Meteorol., 83: 53-79.
Wang Y M, Kepert J D, Holland G J.2001.The effect of sea spray evaporation tropical cyclone boundary layer structure and intensity.Mon.Weather.Rev., 129: 2 481-2 500.
Wu J.1993.Production of spume drops by the wind tearing of wave crests: the search for quantification.J.Geophys.Res., 98: 18 221-18 227.
Yablonsky R M, Ginis I.2009.Limitation of one-dimensional ocean models for coupled hurricane-ocean models forecasts.Mon.Weather Rev., 137: 4 410-4 419.
Zhao D L, Toba Y, Sugioka K et al.2006.New sea spray generation function for spume droplets.J.Geophys.Res., 111 (C2): C02007.1-C02007.11.
Zhao D L.2012.Progress in sea spray and its effects on Air-Sea interaction.Advances in Earth Sciences, 27: 624-632.(in Chinese with English abstract)