Air Sea Fluxes of Heat Freshwater and Momentum

Simon A. Josey

Abstract An overview of the air-sea fluxes of heat, freshwater and momentum is presented with the emphasis being on methods used to determine these fluxes and the role they play within the wider climate system. The equations used to determine the various heat flux components and the wind stress (which is equivalent to the momentum flux) are described in detail, together with the main spatial characteristics of the resulting global fields. This is followed by an overview of currently available flux datasets, including in situ, remotely sensed, atmospheric reanalysis and hybrid products. Methods for evaluation of these datasets are explored, including recent developments in the use of air-sea flux reference sites to discriminate between the different fields. Several topics that place surface fluxes in the context of global climate are then discussed including the ocean heat budget closure problem, climate change related trends in surface fluxes and impacts of extreme heat fluxes at high latitudes. Finally, some outstanding challenges are presented including the need for a better understanding of ocean-atmosphere interaction in the Southern Ocean and the potential for use of the integrated surface density flux to estimate variability in the Atlantic meridional overturning circulation.

6.1 Introduction

The exchanges of heat, freshwater and momentum between the oceans and the atmosphere play a pivotal role in the global climate system. In the tropics, there is a net input of heat to the ocean which is subsequently transported to mid-high latitudes and released back to the atmosphere, modifying the climate over land downstream (e.g. Rhines et al. 2008). At several high latitude sites, intense winter heat loss (together with the effects of net evaporation and brine rejection associated with ice formation) drives deep convection and dense water formation, supplying the deep limb of the global overturning circulation. The wind stress on the ocean, which

National Oceanography Centre, Southampton, UK e-mail: [email protected]

A. Schiller, G. B. Brassington (eds.), Operational Oceanography in the 21st Century, 155

DOI 10.1007/978-94-007-0332-2_6, © Springer Science+Business Media B.V. 2011

is equivalent to the momentum exchange, is the other major driver of the circulation, and regional wind forcing also plays a key role in dense water formation through preconditioning of water masses as a result of upwelling. The freshwater flux (evaporation-precipitation) has a major impact on the ocean surface salinity field which to a large extent reflects the pattern of surface net evaporation.

Despite their major role in the climate system, our level of knowledge regarding many aspects of ocean-atmosphere interaction remains at a basic level. Attempts to develop global datasets of these fluxes have been severely hampered by the lack of observations in many regions. The primary source of data has historically been merchant ship meteorological reports which tend to follow the main shipping routes leaving large areas of the ocean, particularly the Southern Ocean, extremely under-sampled. This situation has improved for some flux related variables (sea surface temperature, wind speed) with the advent of satellite observations but these are only available for the past two decades and do not as yet provide reliable estimates of all terms in the surface heat budget.

Anthropogenic climate change is widely expected to lead to changes in the fluxes of heat and freshwater as a result of global warming and strengthening of the hydro-logical cycle. There is compelling evidence that an increase in global ocean heat content has already happened (e.g. Levitus et al. 2009) and this implies an increase in the global mean net ocean heat gain. However, the expected change is small, only about 0.5 W m-2. This signal is too small to be detectable given the accuracy of currently available heat flux datasets and this situation is unlikely to change in the near future. A strengthening of the hydrological cycle will influence the ocean-atmosphere exchange of freshwater and potentially leave an imprint in ocean salinity. Due to problems with obtaining reliable precipitation measurements, the level of uncertainty in freshwater flux datasets is greater than that for heat flux and it is again difficult to detect anthropogenic climate change in this variable. However, there is some evidence that changes in the hydrological cycle have modified ocean salinity as this acts as an integrator of variations in the surface freshwater exchange (Stott et al. 2008).

In this paper, I provide a short overview of the current state of ocean-atmosphere interaction research. A thorough review of all aspects of air-sea exchanges was carried out by the Working Group on Air-Sea Fluxes in the late 1990s (WGASF 2000) and this remains a major resource which the interested reader is recommended to consult. A further valuable point of reference from the perspective of the ocean observing system is the Plenary White Paper on air-sea fluxes prepared for Ocean Obs'09 (Gulev et al. 2009). Progress in understanding ocean-atmosphere interaction in the face of a fundamental sampling problem and uncertainty over significant elements of the underlying physics has been the result of dedicated efforts by a wide international research community. I have attempted to summarise some of the key results here from a personal perspective which stems from my own research developing and analysing in situ observation based fields and more recently studying the wider role of fluxes using coupled models. I began my research career studying a very different class of surface flux, the effects of the infalling flux of primordial gas (primarily neutral hydrogen) onto the discs of spiral galaxies (Josey and Tayler 1991; Josey and Arimoto 1992). This presents a very different set of research prob lems but provides an interesting alternative perspective on the effects that surface exchanges have on a system. I count myself lucky to have worked initially in this field and subsequently on the equally fascinating, and arguably more important, role of surface fluxes in the global climate system.

Following this introduction, an outline of the formulae used to estimate surface fluxes is given in Sect. 6.2 and an overview of the different flux datasets in Sect. 6.3. Flux evaluation methods are then considered in Sect. 6.4. Several issues related to the role of surface fluxes in the global climate system are discussed in Sect. 6.5, while the final Sect. 6.6 highlights several outstanding issues and potential future applications of the air-sea exchanges, particularly as regards estimates of variability in the ocean overturning circulation.

6.2 Surface Flux Theory

6.2.1 Flux Components and Spatial Variation

The net air-sea heat flux is the sum of four components: two turbulent heat flux terms (the latent and sensible heat fluxes) and two radiative terms (the shortwave and longwave fluxes). These are shown schematically in Fig. 6.1 together with their global mean values from a globally balanced air-sea heat flux dataset (Grist and Josey 2003).

Fig. 6.1 Schematic representation of the different components of the air-sea heat exchange with global annual mean values of the key terms from a balanced flux dataset. (Grist and Josey 2003)

NOC1.1a Latent Heat Flux

80 40 0

80 40

80 40 0

Fig. 6.2 Climatological annual mean fields of the different heat flux components and the net heat flux. (Source: National Oceanography Centre 1.1a (NOCl.la) flux climatology, units W m-2, Grist and Josey 2003)

NOC1.1a Latent Heat Flux

200

80

100

40

0

0

-100

-40

-200

200

80

100

40

0

0

-100

-40

-200

200

100

0

-100

-200

S. A.Josey NOCl.la Sensible Heat Flux

100 200

200 100 0

Climatological annual mean fields of the different components and net heat flux are shown in Fig. 6.2. The sign convention is for positive fluxes to represent heat gain by the ocean. For the turbulent heat flux components (i.e. the sensible and latent terms), the areas of strongest loss are over the Gulf Stream and Kuroshio with latent heat losses of order 200 W m-2. Enhanced latent heat loss is also seen in the South-East Indian Ocean where the trade winds are particularly strong. The sensible heat flux is typically much smaller in magnitude than the latent term, the strongest losses occur in regions where very cold air is advected over the ocean from neighbouring land masses particularly the Labrador and Norwegian Seas.

The global variation in the net longwave flux is relatively small, typical values ranging from 30-70 W m-2. However, within this range there is a degree of structure which reflects the balance between the sea-air temperature difference, the cloud cover and the amount of water vapour. The most noticeable feature is a band of reduced longwave loss under the Inter-Tropical Convergence Zone (ITCZ). In contrast, the shortwave field has a primarily meridional variation determined by the mean solar elevation with peak values of order 200 W m-2. The main departures from this variation occur under regions of increased cloud cover such as the ITCZ. Finally, the net heat flux field is seen to be dominated by the contributions from the shortwave and latent heat fluxes with shortwave driven ocean heat gain in the Tropics and latent heat driven ocean heat loss over the western boundary current regions. The processes controlling these exchange terms and methods for their estimation are discussed below, for a more detailed review see WGASF (2000).

6.2.2 Turbulent Flux Bulk Formulae

The latent and sensible heat fluxes are proportional to the products of the near surface wind speed with the sea-air humidity and sea-air temperature difference respectively. However, the detailed form of these relationships remains poorly known under certain conditions, in particular at high wind speeds and this provides a significant source of uncertainty in estimates of these fluxes.

The sensible and latent heat fluxes, QH and QE, are generally determined using the following bulk formulae:

where p is the density of air; cp, the specific heat capacity of air at constant pressure; L, the latent heat of vaporisation; Ch and Ce, the stability and height dependent transfer coefficients for sensible and latent heat respectively; u, the wind speed; Ts, the sea surface temperature; Ta, the surface air temperature with a correction for the adiabatic lapse rate, y, z, the height at which the air temperature was measured; qs, 98% of the saturation specific humidity at the sea surface temperature to allow for the salinity of sea water, and qa, the atmospheric specific humidity. A major amount of research has been devoted over the past few decades to accurately determining values for the transfer coefficients and their functional dependence on wind speed and near surface stability by means of direct flux measurements, in particular through the eddy correlation method. This work has lead to the development of the COARE flux algorithm (Fairall et al. 2003) which has greatly reduced uncertainty in the values of the transfer coefficients although questions still remain in several areas, particularly the high wind speed regime and inclusion of the effects of sea spray.

6.2.3 Radiative Flux Parameterisations

The shortwave flux is primarily a function of solar elevation and cloud amount with an additional dependence on ocean albedo. The longwave (infrared) flux is the difference between large downwelling and upwelling terms from the ocean and atmosphere respectively and depends on sea surface temperature, air temperature and humidity in addition to cloud amount. The longwave and shortwave flux components have been determined using a wide range of empirical formulae over the years

(e.g. Clark et al. 1974; Bignami et al. 1995; Josey et al. 2003). The performance of several bulk formula parameterisations for the net longwave flux has been assessed by comparison with radiometer measurements made at sea during a number of cruises (Josey et al. 1997). More recently Josey et al. (2003) carried out a detailed evaluation of both the Clark et al. (1974) and Bignami et al. (1995) formulae using measurements made on a long meridional research cruise from 20-63 °N at 20° W in the North Atlantic. This analysis made use of recent advances in understanding of various biases in the pyrgeometer instrument used to measure the longwave flux (Pascal and Josey 2000).

Neither formula was found to be capable of providing reliable estimates of the atmospheric longwave flux over the full range of latitudes. The Clark formula overestimated the cruise mean measured longwave flux of 341.1 W m-2 by 11.7 W m-2, while Bignami underestimated by 12.1 W m-2. Josey et al. (2003) developed an alternative formula which expresses the combined effects of cloud cover and other relevant parameters on the atmospheric longwave in terms of an adjustment to the measured air temperature. The net longwave flux, QL, across the ocean-atmosphere interface is given by:

where QLS is the emitted longwave radiation from the sea surface, QLA is the down-welling longwave radiation from the atmosphere, and the coefficient (1-aL), where aL is the longwave reflectivity, takes account of the component of the downwelling radiation reflected from the sea surface. They characterise the downwelling longwave radiation by an effective blackbody temperature, TEff, such that,

where oSB is the Stefan-Boltzmann constant (5.67 x 10-8 W m-2 K-4). Given that the observed variable is Ta instead of TEff, they write TEff as the sum of Ta and a temperature adjustment, ATa, which includes the effects of cloud cover, atmospheric humidity and other, as yet unknown, variables on the downwelling longwave, such that,

ATa is thus the difference between the measured air temperature and the effective temperature of a blackbody which emits a radiative flux equivalent to the atmospheric longwave. The problem of obtaining a reliable estimate for QLA then becomes one of parameterising the dependence of ATa on cloud cover, vapour pressure and any other relevant variables. The air temperature is adjusted by the amount necessary to obtain the effective temperature of a blackbody with a radiative flux equivalent to that from the atmosphere. A simple parameterisation of the temperature adjustment solely in terms of the total cloud amount leads to a net longwave flux formula which has an improved mean bias error with respect to the cruise measurements of -1.3 W m-2.

The new formula still exhibits significant biases under certain situations, in particular overcast, low cloud base conditions at high latitudes. However, by modify-

Fig. 6.3 Comparison of the atmospheric component of the net longwave flux estimated using Eq. (6.6) with measurements made on a research cruise in the North Atlantic. (Modified version of figure from Josey et al. (2003), copyright American Geophysical Union)

Fig. 6.3 Comparison of the atmospheric component of the net longwave flux estimated using Eq. (6.6) with measurements made on a research cruise in the North Atlantic. (Modified version of figure from Josey et al. (2003), copyright American Geophysical Union)

250 300 350

Measured Longwave, W m-2

250 300 350

Measured Longwave, W m-2

ing this formula to include a dependence on the dew point depression, good agreement between the measured and estimated mean longwave over the full range of observations can be obtained and the mean bias error reduced to 0.2 W m-2 (see Fig. 6.3). The resulting formula for the net longwave flux is as follows:

Ql = sbT - (1 - «L)ffSB{Ta + an2 + bn + c + 0.84(D + 4.01)}4 (6.6)

where s is the emissivity of the sea surface, taken to be 0.98, aL = 0.045 and n is the fractional cloud cover. The terms a, b and c are empirical constants and D is the dew point depression, D=TDew - Ta, where TDew is the dewpoint temperature (i.e. the temperature at which it becomes saturated) of the air in the surface layer. The new formula was tested using independent measurements made on two more recent cruises and found to perform well, agreeing to within 2 W m-2 in the mean, at mid-high latitudes.

In contrast, to the formulae for the sensible, latent and longwave fluxes which may be used with individual ship meteorological reports, widely-used formulae for the net shortwave flux typically provide monthly mean values. In particular, the following formula of Reed (1977) provides the monthly mean net shortwave flux,

where a is the albedo, Qc is the clear-sky solar radiation, n is the monthly mean fractional cloud cover and 0 N is the monthly mean local noon solar elevation. Gil-man and Garrett (1994) note that under conditions of low cloud cover, the Reed formula estimate of the mean incoming shortwave can become greater than the clear-sky value if 0N is sufficiently large. and suggest that the incoming shortwave be constrained to be less than or equal to Qc.

Finally, the net heat flux, QNet, is given by the sum of the four individual components,

where QE is the latent heat flux; QH, the sensible heat flux; QL, the longwave flux and QSW, the shortwave flux.

6.2.4 Wind Stress

Estimates of the zonal, tx, and meridional, xy, components of the sea surface wind stress are typically obtained using the following equations,

0 0

Post a comment