1. Introduction
Air–sea heat fluxes are critical for both ocean and atmosphere dynamics and play a key role in defining the global climate change processes. They feed the ocean circulation and weather prediction numerical models. Surface heat exchanges are useful in providing initial conditions for coupled ocean–atmosphere models for forecasts from nowcasting to climate time scales (Yu et al. 2004). It is also known that the surface fluxes are a strong driving force for the intermediate-water and deep-water formation, an important mechanism for the deep water oxygenation. The surface fluxes are mostly regulated by several meteorological factors; for example, cloud fraction affects the radiative components (shortwave and longwave radiation) and wind speed affects the turbulent components (latent and sensible heat). However, little attention has been drawn on the fluxes dependence on the sea level pressure (SLP).
On synoptic time scales, SLP anomalies (SLPA) show direct relationship to episodic extreme events of air–sea heat fluxes over the Gulf Stream region (Zolina and Gulev 2003; Shaman et al. 2010), along the Kuroshio recirculation gyre in the northwest Pacific (Bond and Cronin 2008), and in the South Atlantic (Sterl and Hazeleger 2003). In the North Atlantic and North Pacific, SLPA are associated with wind speed anomalies, which result in a latent and sensible heat flux anomalous regime (Konda et al. 2010). Cayan (1992) and Alexander and Scott (1997) found significant relationships between SLP field and surface fluxes over the North Pacific and Atlantic Oceans on various time scales. Notwithstanding the strong episodic events of surface heat exchanges trigger the convective processes on synoptic time scale, it is the low-frequency variability of these exchanges that provokes distinct climatological influence and affects decisively the intermediate-water and deep-water convection. The persistence of such episodes over prolonged time periods makes their influence critical for the marine ecosystems. Therefore, the climatology of the atmospheric factors affecting the air–sea heat exchanges is more conclusive when studied on a monthly or seasonal instead of the synoptic time scale.
The influence of the large-scale atmospheric variability over the North Atlantic and Europe on air–sea heat exchanges in the Mediterranean Sea is rather complex and area dependent (Josey et al. 2011). Ruiz et al. (2008) describe the impacts of large-scale atmospheric circulation on the surface heat exchanges in the Mediterranean Sea using climatic indices such as the North Atlantic Oscillation (NAO) and the Mediterranean oscillation index. Paz et al. (2003) link the SLP oscillations between northwest Africa and West Asia (NAWA index) with the temperature and wind regime of the eastern Mediterranean. In many cases, the atmospheric variables and the surface heat exchanges exhibit a strong relationship. For example, the correlation between total precipitable water over Europe and North Africa and surface heat exchanges over the Aegean Sea is particularly strong, especially during the cold season of the year (Papadopoulos et al. 2011). Conversely, the turbulent fluxes in the western Mediterranean affect the precipitation regime over the eastern Mediterranean (Lolis et al. 2004). From a different point of view, Gilman and Garrett (1994) examine the impacts of large-scale atmospheric patterns on the heat budget of Mediterranean in terms of transport and dispersion of the atmospheric anthropogenic and mineral aerosols. Although the topographic relief and the frequent land/ocean transition of the Mediterranean Basin favor the modulation of local air masses, it is the large-scale circulation that enhances such local effects.
The Aegean is a typical semienclosed sea and represents an area of the Mediterranean basin where intermediate and deep waters are formed (Roether et al. 1996; Theocharis et al. 1999; Wu et al. 2000; Nittis et al. 2003, Beuvier et al. 2010). Intermediate-water and deep-water formation are related to conditions dominated by anomalously high rates of heat losses during the cold season of the year. Events of deep convection are important for the eastern Mediterranean climatology because they provide the necessary aeration to the deep water environment. At times, strong and persistent heat losses in the Aegean Sea trigger intermediate-water/deep-water formation events. Josey (2003) attributes these events to distinct large-scale atmospheric patterns. Gündüz and Ozsoy (2005) demonstrate the effects of the North Sea–Caspian pattern (NCP) on several surface variables over the Mediterranean. In general, the patterns associated with anomalously high heat loss over the Aegean Sea are characterized by strong pressure gradient across the Aegean Sea, which favors persistent, dry, and cold northerly winds (Chronis et al. 2011a) that, in turn, cause intense heat loss via latent and sensible heat. In addition, Romanou et al. (2010) identify localized winds extrema across the Cretan Straits (south Aegean Sea) and large values of evaporation rates over the neighboring Levantine Sea. Although the same authors exclude the Aegean Sea, their findings are potentially evident over the Aegean, resulting in anomalous heat loss. The Aegean Sea can further affect the climatic regime on a regional scale. It is considered as one of the cyclogenesis areas in the Mediterranean Sea, especially during the cold season (Trigo et al. 2002; Flocas et al. 2010); hence, the turbulent fluxes are expected to play a dominant role.
In the study at hand, we examine the potential influence of widely used climatic indices and SLP variability on surface heat exchanges over the Aegean Sea. We further investigate the atmospheric patterns dictating the minima and maxima of the air–sea heat fluxes and the associated with them physical mechanism. The analysis of each component’s contribution to the extreme events of the heat exchange is also needed in order to gain a physical insight of the involved driving mechanism. Finally, specific attention is given to extreme winter heat loss events that are directly related to the intermediate- and deep-water formation over the Aegean Sea.
The study is divided into the following sections: Section 2 presents the employed datasets. Section 3 introduces two testing climatic indices, whereas section 4 discusses the applied methodology. Section 5a examines the possible influence of the climatic indices, and section 5b describes the correlation between SLP and each component of the air–sea heat fluxes. Section 5c analyzes the extreme events of heat exchanges and addresses the associated composites of SLP and SLPA fields. In the same section, the contribution of the radiative and turbulent components to the net heat fluxes anomaly during the extreme events is also investigated. Finally, the concluding remarks are highlighted in section 6.
2. Datasets
We examine the magnitudes of the air–sea heat fluxes over the Aegean Sea with respect to the spatial and temporal variability of the atmospheric pressure field. First, our analysis employs a dataset based on five buoy observations from the Hellenic Center for Marine Research (HCMR) Poseidon network (Nittis et al. 2001; Soukissian et al. 2002), spanning the period 2000–08. Specifically, 3-h values of atmospheric pressure, air temperature, sea surface temperature, and wind speed are considered. To calculate the specific humidity, we use satellite observations from the Special Sensor Microwave Imager platform and implement the Bentamy model (Bentamy et al. 2003). In addition, we use 3-h cloud cover observations from coastal and island stations in the Aegean Sea. The combination of the above datasets, after making use of appropriate bulk formulas, results in the monthly-mean values of the four heat flux components, shortwave and longwave radiation and latent and sensible heat, at each buoy location.
Because the observational time series are limited to nine years, we employ a well-established and much more extended dataset. Therefore, our monthly mean heat fluxes values are compared against three well-documented gridded datasets: namely, the National Oceanography Centre Southampton (NOCS) (Berry and Kent 2009), objectively analyzed air–sea fluxes (OAFlux) (Yu and Weller 2007), and Hindcast of Dynamic Processes of the Ocean and Coastal Areas of Europe (HIPOCAS) (Sotillo et al. 2005; Ratsimandresy et al. 2008). These comparisons consider the closest to the buoy grid points and involve statistical analysis such as correlation coefficients, biases, and rms errors. The results indicate that, over the Aegean Sea, the Mediterranean HIPOCAS (0.5° × 0.5° resolution) is the most appropriate dataset for the study of shortwave and longwave radiation and the OAFlux (1° × 1° resolution) for the latent and sensible heat (Papadopoulos et al. 2010). Since the Aegean Sea approximately covers 6° in latitude and 4° in longitude, the spatial resolution of the selected datasets is considered adequate for depicting the regional heat fluxes characteristics. The monthly mean shortwave and longwave radiation values cover the period 1958–2001, while the latent and sensible heat extend over the period 1958–2008. In addition, the monthly mean air temperature and sea surface temperature data are also retrieved from the OAFlux archives.
Furthermore, the well-documented indices, the North Atlantic Oscillation (NAO) and east Atlantic–western Russian pattern (EAWR), are retrieved from the NOAA Climate Prediction Centre. In terms of the atmospheric pressure dataset, we use monthly mean SLP values retrieved from the 40-yr European Centre for Medium-Range Weather Forecasts Re-Analysis (ERA-40) for 1958–2001, 2.5° × 2.5 (Uppala et al. 2005). The monthly mean SLP values for Marseille, Moscow, and Heraklion needed for the introduced indices calculation are retrieved from NOAA/National Climatic Data Center (NCDC) monthly bulletin of climatic data for the world (http://www7.ncdc.noaa.gov/IPS/mcdw/mcdw.html). Finally, wind data (zonal and meridional wind components) are also retrieved from ERA-40 dataset.
3. The new indices
In this study, we introduce two testing indices and examine their correlation with air–sea heat fluxes over the Aegean Sea. The first index is considered as an alternative to the Mediterranean oscillation index (Conte et al. 1989). This index, hereafter called the “Mediterranean index” (MI), corresponds to the monthly-mean anomaly of SLP difference between Marseille in southern France and Heraklion in the island of Crete (Fig. 1a). MI does not significantly depart from two similar indices: the Mediterranean pressure index (Raicich et al. 2003) referred to as the SLP difference between Marseille and Mersa Matruh (southeastern Mediterranean) and the Mediterranean circulation index (Brunetti et al. 2002) referred to as the SLP difference between Marseille and Jerusalem. The second experimental index, hereafter called the “Eastern Europe index” (EEI), represents the anomaly of SLP difference between Moscow and Heraklion (Fig. 1a). We select Heraklion because it is located at the south Aegean, and therefore we anticipate a better response of the proposed indices to the regional characteristics of the SLP variability. In addition, Heraklion is a high-quality meteorological station with a long and continuous operation. The choices of Marseille and Moscow are reasoned by the directions of the lines connecting them with Heraklion, which are almost perpendicular to each other and reflect SLPA gradients of nearly zonal and meridional directions. Using these testing indices, we aim to underline the different impact of an along-Mediterranean pressure gradient and a transversal one on the surface heat exchanges in the Aegean Sea. Both testing indices are normalized divided by their standard deviation to obtain zero mean and unit variance.
(a) The new index baselines and the Aegean Sea region and (b) the representative sites in the north Aegean (Athos) and south Aegean (Avgo); squares correspond to buoys, and triangles to the closest grid point of the databases used.
Citation: Journal of Climate 25, 2; 10.1175/2011JCLI4197.1
4. Methodology
The influence of the atmospheric pressure on the Aegean Sea surface heat fluxes is investigated separately for the cold and the warm season and for the northern and the southern subbasin. Monthly values are used in the analysis, from November to March for the cold season and from May to September for the warm season, whereas April and October are excluded as transitional months. As representative sites for the two subbasins, we select the locations of two Poseidon buoys, Athos and Avgo (Fig. 1b). At the respective closest grid points (Fig. 1b), we calculate the linear correlation coefficient between the seasonal time series of the heat fluxes and the climatic indices of the NAO, EAWR, and the North Sea–Caspian pattern (NCP) as well as the MI and EEI. Each seasonal time series comprises 220 monthly values for the radiative and 255 for the turbulent terms. We calculate the NCP according to Kutiel and Benaroch (2002) using the ERA-40 database. We select the NAO, EAWR, and NCP because they are generally considered as the most influential indices for the eastern Mediterranean climate (Eshel and Farrell 2000; Ben-Gai et al. 2001; Kutiel and Benaroch 2002; Dünkeloh and Jacobeit 2003; Gündüz and Ozsoy 2005; Lionello and Sanna 2005; Hatzaki et al. 2007; Feliks et al. 2010; Josey et al. 2011). Thus, the influence of NAO, EAWR, NCP, and the two testing indices on the air–sea heat fluxes regime over the Aegean is examined. Next, in order to define potential centers of action of the atmospheric pressure field, which affect the surface fluxes over the Aegean, we correlate the anomalies of each heat flux component with the atmospheric pressure anomalies over the northeast Atlantic, Europe, and North Africa. Monthly SLPAs are calculated at each grid point by subtracting the monthly-mean values from the SLP time series. At both sites, a point-to-point correlation is performed for the area extending from 25° to 70°N, 15°W to 35°E with a 2.5° × 2.5° resolution. The period of correlation is 1958–2001 for the radiative components and 1958–2008 for the turbulent ones. Furthermore, in order to provide possible explanations for the featured correlation patterns, we perform seasonal EOF analysis of SLP and present the EOF modes as homogeneous correlation maps (Venegas 2001).
In turn, an extreme values analysis is carried out by taking into account the total heat exchange anomaly, namely, the sum of the four air–sea heat flux components anomalies. As “extreme” values, we consider the outermost 10% (upper and lower decile) of the monthly mean net heat flux anomalies for each season. Each case corresponds to 22 individual months for the period 1958–2001. The 22 monthly mean SLP and SLPA patterns that correspond to each decile are averaged, identifying the general characteristics of the atmospheric circulation favoring the extreme heat fluxes in the Aegean Sea. We further question our findings by using the composite wind vector map and the Tair – Tsea anomalies map corresponding to each decile. Next, we examine the contribution (in percent) of each heat flux component to the aggregated energy anomaly of these extremes. In fact, the contribution of each component reflects the physical process leading to anomalous heating or cooling of the sea surface. Moreover, by highlighting the lowermost winter values, this study aims at identifying the specific regime under which the atmospheric circulation, along with the hydrology, can trigger the mechanism of intermediate-water and deep-water formation in the Aegean Sea. It should be noted that the study at hand refers to the oceanographic convention, in that positive air–sea heat fluxes represent oceanic heat gain.
5. Results and discussion
a. Climatic indices and surface heat flux
Table 1 summarizes the correlation coefficients between the atmospheric indices and heat flux anomalies for each component at the two sites in the Aegean Sea during the cold season from November to March (NDJFM). Namely, the NAO index, whose general influence on the eastern Mediterranean is inconclusive and controversial (Eshel and Farrell 2000; Dünkeloh and Jacobeit 2003; Hatzaki et al. 2007; Feliks et al. 2010; Chronis et al. 2011b), exhibits a general weak influence on surface fluxes. NAO shows a weaker relation than the EAWR, the NCP, and the two testing indices on all components except the shortwave radiation. The MI exhibits the strongest influence on the turbulent components with negative correlation coefficients ranging from −0.51 to −0.64. The EEI has the highest impact on the radiative components at both sites. The EEI is correlated negatively with shortwave radiation with coefficient values −0.49 for Avgo and −0.62 for Athos and positively with longwave radiation with coefficient values 0.53 and 0.59, respectively. In every case, during the warm season (not shown) the correlation is consistently weaker. In summary, the EEI and NAO, both of which represent a meridional pressure gradient, show stronger influence on the radiative fluxes. This reflects an impact of the two indices on the regional cloud cover. For example, high values of EEI correspond to a strong anticyclone over western Russia, implying fair weather there, and a deep low pressure system over Crete, causing cloudiness all over the Aegean. Hereby, the higher (lower) the EEI, the lower (higher) the incoming shortwave radiation in the Aegean. The opposite behavior is expected for the longwave radiation since higher cloud cover is associated with higher longwave radiation values (lower heat loss). On the other hand, MI, EAWR, and NCP, which represent a zonal pressure gradient, affect the turbulent components of the air–sea heat fluxes. For example, high values of MI correspond to higher pressure over Western Europe and lower pressure over the eastern Mediterranean. This synoptic condition creates an almost zonal pressure gradient, which favors the heat loss by the prevailing northerly winds.
Correlation coefficients between atmospheric indices and heat fluxes anomalies during the cold season (NDJFM) at the Athos and Avgo site in the Aegean Sea (correlation coefficients not statistically significant at 95% confidence level are not presented); Heat flux components are shortwave (SW) and longwave (LW) radiation and sensible (SH) and latent (LH) heat flux: QNET = SW + LW + LH + SH.
The two testing indices MI and EEI show different impacts on the heat flux components. The MI is associated with the meridional wind component and affects the turbulent components regime over the Aegean Sea. The EEI, representing a nearly north–south SLP gradient, affects the radiative components and in fact reflects the regime of the cloud cover and specific humidity. The cloudiness regime is mostly regulated by the track of North Atlantic and Western Europe depressions, which can shift between a Mediterranean route and a north/central Europe route. The corroboration of such a hypothesis is provided by the first EOF mode of SLP variability during the cold season (Fig. 2a). The first EOF mode displays an out-of-phase relationship between the Mediterranean region and North Europe. On the contrary, windy conditions favoring heat loss over the Aegean Sea, mainly by means of the latent heat, are attributed to an SLP gradient developing across the Aegean Sea. The latter can be identified by the second EOF (Fig. 2b). Similar findings are shown for the warm season (Figs. 2c,d). The overall behavior underlines a remarkable spatial variability of the correlation between SLPA and heat fluxes. This finding further motivates the employment of a point correlation between the surface fluxes and a large-scale SLPA field in order to identify the areas of the strongest influence on both radiative and turbulent heat flux components over the Aegean Sea.
The first two EOFs of SLP during (a),(b) the cold season of the year (NDJFM) and (c),(d) the warm season of the year, May–September (MJJAS), presented as homogeneous correlation maps between the EOF and the SLP field. The fraction of variance (%) explained by the respective mode is indicated in the upper right corner of each map. Dashed lines represent negative correlations.
Citation: Journal of Climate 25, 2; 10.1175/2011JCLI4197.1
b. Heat fluxes and SLPA
This correlation is performed separately for the northern and for the southern part of the Aegean Sea. Figures 3–6 present the correlation pattern for the north and south Aegean Sea, for both seasons (cold and warm). Especially for the sensible heat, only the winter correlation is presented owing to its frequent sign variability between positive and negative values during the warm season of the year. This inconsistency is justified by the temperature difference between the sea surface and the atmosphere. The sensible heat can be negative or positive, and the wind magnitude regulates the amount of heat loss/gain by the sea. As a result, during the warm season of the year, two opposite signs of correlation can occur, leading to a misrepresented correlation.
Spatial distribution of statistically significant (at 95% confidence level) correlation coefficients between shortwave radiation anomalies in the north Aegean and SLP anomalies during (a) winter and (b) summer: (c),(d) for the south Aegean. The dashed lines illustrate negative correlations.
Citation: Journal of Climate 25, 2; 10.1175/2011JCLI4197.1
As in Fig. 2, but for longwave radiation anomalies.
Citation: Journal of Climate 25, 2; 10.1175/2011JCLI4197.1
As in Fig. 2, but for latent heat anomalies.
Citation: Journal of Climate 25, 2; 10.1175/2011JCLI4197.1
As in Fig. 2, but for sensible heat anomalies in winter only.
Citation: Journal of Climate 25, 2; 10.1175/2011JCLI4197.1
The correlation maps for the radiative and turbulent components exhibit substantial differences. These are consistent with the previously documented discrepancies on how the zonal and meridional pressure gradient indices affect the components. A meridional out-of-phase dipole dominates the radiative forcing (Figs. 3 and 4), thus explaining the major influence of the EEI. This dipole, much stronger during the cold period of the year, particularly for the shortwave radiation, exhibits its centers over Scandinavia and North Africa with small shifts for the longwave radiation. The latter is the most complex among the heat flux components as many parameters affect its variability. The cloud fraction is prominent during winter, whereas the specific humidity is believed to play the key role during summer when the cloud fraction is systematically very low. Lower than normal pressure values over the Aegean Sea (and the eastern Mediterranean) imply higher values of cloud cover, which decrease the shortwave radiation (reduced heat gain by the sea) and increase the longwave radiation (reduced heat loss by the sea). Thus, SLPA and shortwave radiation covariate over the Aegean (positive correlation coefficients), whereas the opposite is true for SLPA and longwave radiation (negative correlation coefficients). Thence, it is evident that the radiative fluxes are directly related to the two distinct routes of the atmospheric depressions: namely, over the Mediterranean and over Northern Europe. Metaxas et al. (1993) link these two routes to dry and wet winters in Greece. These two route domains are anticorrelated and decisive for the cloud fraction distribution. Figure 7 shows the correlation between the cloud fraction and SLPA. The two maps substantiate the direct association of the two parameters mainly during winter. Moreover, these maps are similar to the correlation maps for the radiative components (Figs. 3 and 4). As it was initially demonstrated in section 5a, the correlation coefficients in all cases decrease during the warm period of the year. This is a strong evidence of less variability of the SLP field during summer in comparison with winter. The standard deviation of the SLP field provides additional information of this seasonal variability displaying lower values during the warm season (Fig. 8). It is also consistent with the potential transition from the European to the Indian monsoon influence regime between the cold and warm periods of the year over the eastern part of the Mediterranean Sea (Trigo et al. 2002; Ziv et al. 2004).
Correlation maps between cloud cover in the north Aegean Sea and SLP anomalies during (a) winter and (b) summer.
Citation: Journal of Climate 25, 2; 10.1175/2011JCLI4197.1
Spatial distribution of standard deviation of SLP (hPa) during the (a) cold and (b) warm season of the year.
Citation: Journal of Climate 25, 2; 10.1175/2011JCLI4197.1
Regarding the turbulent components, their direct dependence on the wind regime is evident during both seasons of the year over the Aegean Sea (Figs. 5 and 6). A domain of significant negative correlation with values lower than −0.6 is observed during the cold season, indicating an increase in surface heat loss over the Aegean with simultaneous SLP increase (taking into account the negative sign for energy loss from the sea). Increasing heat loss occurs as the SLP increases mainly over an area encompassing France and Germany. SLPAs over the above regions govern the turbulent heat flux regime in terms of producing or not a strong pressure gradient over the Aegean Sea. In the cases of high atmospheric pressure over the aforementioned area, strong, dry, and cold winds advect from higher latitudes over the Aegean Sea and induce surface heat loss. In contrast, lower than normal SLP over the same area suppresses the northerlies and modulates the turbulent fluxes to exhibit their maxima (reducing the heat losses). This correlation pattern is spatially modified and becomes weaker during the warm season, still controlling the presence of northerlies over the Aegean Sea. The previously described patterns for the turbulent fluxes are approximately the same for both the north and south Aegean Sea, with negligible differences between the two subbasins. They are consistent with the stronger influence of the MI, EAWR, and NCP on the latent and sensible heat over the Aegean Sea.
Table 2 presents the correlation coefficients between the two first EOFs of the SLP and the correlation maps during winter. Clearly, the correlation patterns of the radiative components are associated with the first mode of the SLP variability. This is also in accordance with the stronger EEI/NAO influence, as stressed in section 5a. Conversely, the second EOF characterized by a coherent domain over the entire Europe is associated with the turbulent components in accordance with the stronger influence of MI, EAWR, and NCP.
Correlation coefficients between EOFs and correlation maps of each air–sea heat flux component for the cold season (NDJFM). Bold font indicates the highest coefficient.
c. Extreme events analysis
The analysis of the uppermost and lowermost deciles (10%) of the net heat flux anomaly reveals the characteristics of the SLP patterns and the corresponding physical mechanisms associated with the occurrence of the surface heat exchange extremes. Figures 9 and 10 present the composite maps for the SLP and SLPA fields resulting from the average of 22 SLP and SLPA maps for each outermost mode. In general, the SLP fields over Europe and North Africa are similar for the north and south Aegean extreme events. For this reason, we show only the composite maps for the south Aegean for the warm and the cold season of the year.
Composite SLP (hPa) distribution for the uppermost 10% extreme values of the net air–sea heat flux occurring within the period 1958–2001 in the south Aegean during the (a) cold and (b) warm seasons of the year: (c),(d) for the SLP anomalies.
Citation: Journal of Climate 25, 2; 10.1175/2011JCLI4197.1
As in Fig. 9 but for the lowermost 10%.
Citation: Journal of Climate 25, 2; 10.1175/2011JCLI4197.1
Figure 9 portrays the SLP and SLPA composite fields for the uppermost extremes during the cold and warm seasons of the year. For the cold season, the atmospheric low over the Gulf of Genoa (Fig. 9a) is the prominent feature over the Mediterranean Sea. This pattern, along with a recession of the Atlantic subtropical anticyclone (Azores high) and the absence of the Cyprus low, generates weak to moderate southerlies, accompanied by high air temperature and specific humidity values, over the eastern Mediterranean region. Under these conditions, the turbulent components are expected to be weak. For the warm season (Fig. 9b), the Aegean Sea is affected by the thermal lows of northwest Asia and North Africa, whereas the extension of the Atlantic subtropical anticyclone is hardly detected over Central Europe. The composite field for the warm season results in weakened northerlies over the Aegean, hence reducing the intensity of the turbulent components. In terms of the SLPA during winter (Fig. 9c), the uppermost heat fluxes in the Aegean Sea are associated with a strong negative anomaly of atmospheric pressure (lower than −6 hPa) centered over Ireland and Britain. On the contrary, the summer uppermost extremes (Fig. 9d) occur under slight negative SLPA (−0.5 to −1 hPa) over Europe. They are characterized by a weaker positive net heat flux anomaly in comparison to their winter counterpart.
The composite SLP pattern for the lowermost extremes during the cold season is characterized by high pressures covering almost the entire continental Europe and low pressures (the Cyprus low) covering the eastern Mediterranean (Fig. 10a). Under these synoptic conditions, the cold, dry northerlies over the Aegean Sea are strengthened and induce greater heat loss. During the warm season, the Azores high extends to reach the Balkan Peninsula (Fig. 10b). The northwest Asia thermal low is always present during this period of the year. The combination of the two systems affects the Aegean Sea with persistent, dry, and relatively cold northerlies, known as the Etesian winds (Metaxas and Bartzokas 1994). These winds enhance the heat loss over the Aegean by means of turbulent fluxes. The extended anticyclonic system over Europe during winter is associated with a strong positive SLPA (+7 to +8 hPa) occurring again over Ireland and Britain (Fig. 10c). During the warm season (Fig. 10d), the positive SLPA field linked to the lowermost extremes is suppressed (+1 to +2 hPa) and moves eastward. It should be noted that the winter SLP pattern favoring the lowest extremes is consistent with the general description of the atmospheric pattern associated with events of extreme surface cooling and dense water formation over the Aegean Sea (section 1). In particular, this SLP pattern has very similar characteristics with the patterns featured by Josey (2003) and Josey et al. (2011). As a result, the aforementioned pattern transfers cold and dry air masses from higher latitudes into the Aegean region, promoting additional heat loss.
The composites of the wind regime during each extreme case for the south Aegean Sea (Fig. 11) explain also the physical processes under which the extreme heat fluxes can occur. The composite wind fields comply well with the corresponding SLP patterns and underline the role of these patterns on the air–sea heat fluxes variation. The presence of high pressures over Europe and the low pressure centers over the North Atlantic and North Africa are the controlling factors for these extremes. During the cold season, the negative SLPAs located to the southwest of the British Isles and over western Europe favor the cyclogenesis over the Gulf of Genoa (Trigo et al. 1999, 2002). This affects the Aegean Sea by giving rise to weak and wet southwest wind (Fig. 11a), which in turn minimize the heat loss by the turbulent fluxes. In contrast, the positive SLPAs over the same area lead to the dominance of high pressures over continental Europe and generate cold and dry northerlies associated with anomalous high surface cooling over the Aegean Sea (Fig. 11c). The two opposite patterns are in accordance with the influence of the MI, EAWR, and NCP. These indices are linked through a zonal SLP gradient, the MI over the Mediterranean and EAWR and the NCP over higher latitudes. The two opposite patterns of SLPA correspond to the seesawlike fluctuation of the west flank of the mentioned indices. The Azores high extension is the prominent regulator during the warm season. A suppression of the Azores high extension west of the Balkan Peninsula favors the weaker wind that increases the turbulent fluxes values (reduces the heat loss) over the Aegean Sea (Fig. 11b). In contrast, the extension of the anticyclone as far as western Russia generates stronger cold and dry northerly winds decreasing the specific humidity and increasing the heat loss from the sea surface, mostly by means of latent heat (Fig. 11d). As these conditions are commonly accompanied by low humidity and clear skies, the longwave radiation also enhances the amount of heat loss. The Etesian winds are present in both uppermost and lowermost events and the difference in the composite wind magnitudes between maxima and minima of the heat exchanges is barely 1.5 m s−1. This occurs as the exchange coefficients are more sensitive to the wind variation under stable conditions (Kara et al. 2000), which frequently appear during summer (SST is lower than air temperature). In contrast to the winter regime, the conditions favoring the warm season extremes are associated with a weaker SLPA field. This is slightly negative over Europe in the case of the uppermost extremes and positive over Northern Europe in the case of the lowermost.
Composite wind vectors for the (a),(b) uppermost and (c),(d) lowermost 10% extreme values of the net air–sea heat flux occurring within the period 1958–2001 in the south Aegean during the (left) cold and (right) warm season of the year.
Citation: Journal of Climate 25, 2; 10.1175/2011JCLI4197.1
In addition to the wind behavior, the difference between the air and the sea temperature during each extreme case corroborates the findings herein. Figure 12 illustrates the composites of Tair – Tsea anomalies in the entire Mediterranean Sea for the lowermost and uppermost deciles of the net air–sea heat flux in the south Aegean Sea during the cold season of the year. Obviously, in the case of lowermost extremes, the Aegean Sea is the region of highest negative anomalies in the Mediterranean Sea, reaching values of −2°C (Fig. 12a). The opposite is observed during the uppermost extremes with anomalies exceeding +1.5°C over the Aegean Sea (Fig. 12b).
Composites of Tair – Tsea anomalies for the (a) lowermost and (b) uppermost 10% extreme values of the net air–sea heat flux occurring within the period 1958–2001 in the Mediterranean Sea during the cold season of the year.
Citation: Journal of Climate 25, 2; 10.1175/2011JCLI4197.1
The SLP and SLPA composite patterns show that the atmospheric pressure gradient across the Aegean Sea is the prominent regulating factor that enhances the heat exchange extremes over the area. This becomes more evident if we examine the total contribution of the radiative and turbulent components during the extremes. The contribution of the radiative and turbulent components reflects their percentage to the total amount of energy flux anomaly (aggregation of the 22 net heat flux anomalies for each case). The turbulent components are the determining factor in the production of the net heat flux anomalies. They consistently contribute 80%–90% to the total energy anomaly of the extremes (Fig. 13). The radiative components are considered as the main source of the heat gain by the sea, especially during the warm season (Matsoukas et al. 2005; Bond and Cronin 2008; Weller et al. 2008; Ghate et al. 2009). Nevertheless, their anomalies exhibit a very narrow fluctuation over time in the Aegean Sea. Therefore, the radiative components have only a minor contribution in determining the extreme events throughout the year. Finally, by examining the aggregated surface flux anomaly, we infer that the extremes are stronger during the cold season, whereas the lowermost events are marginally stronger than their uppermost counterpart.
Bar chart representation of the contribution of each component anomaly to the total energy anomaly of the uppermost and lowermost 10% extreme values for the north Aegean (N) and south Aegean (S) Sea seasonal uppermost (+) and lowermost (−) extreme cases.
Citation: Journal of Climate 25, 2; 10.1175/2011JCLI4197.1
6. Concluding remarks
The influence of climatic indices on the surface fluxes over the Aegean Sea highlight a different response between the radiative and turbulent components. This is expected since the radiative terms are regulated by the cloud fraction, whereas the turbulent terms are regulated mainly by the wind speed. The EEI exhibits the strongest influence on the radiative components and the MI and NCP on the turbulent ones. The strongest influence of the EEI and MI depicts the regional effect of the SLP variability. Notwithstanding the examined indices individually explaining a rather small fraction of the total variance of heat fluxes, they are almost uncorrelated (independent) with each other. The highest correlation coefficient is 0.5 between the EAWR and NCP indices during the cold season. For this reason, the climatic indices have the potential to affect the heat fluxes in a complementary mode. Periods of strong influence of one index can coincide with periods of weak influence of another index. This has been particularly proven for the uncorrelated NAO and EAWR (H. Kontoyiannis et al. 2011, unpublished manuscript), which at times complementarily affect the air temperature over the Aegean and Black Seas. As a result, a much greater amount of the total variance can be explained. From the authors’ perspective, the sign and absolute value of the correlation coefficients provide useful information for the influence exerted by the different indices.
The one-point correlation between SLP and each component of the surface fluxes corroborates the diversity of each component’s behavior. The radiative fluxes present a seesawlike correlation pattern with the SLPA, a fact that implies opposite SLP phases between the Mediterranean Sea and Scandinavia. As a physical explanation, we consider the cloud cover associated with the different tracks of the east Atlantic and western Mediterranean barometric lows. These depressions can potentially propagate from west to east over a Mediterranean track or follow a higher-latitude track over central and northern Europe. Although some of these atmospheric depressions originate from the North Atlantic, the NAO exhibits a poor correlation with the air–sea heat fluxes over the Aegean. In addition, the importance of the eastern Mediterranean and northwest Asian lows is prominent during the warm season. At the same time, the correlation between SLPA and the turbulent air–sea fluxes is stronger, and it is related to the SLPA over the central Europe during the cold season and over the central Mediterranean during the warm season. Especially for the warm season, the extension, or not, of the subtropical anticyclone of the Atlantic (Azores high) as far as over eastern Europe governs the occurrence of the heat flux extreme values. Such extension of the Azores high over central and northern Europe is frequently observed, especially during the boreal summer (Davis et al. 1997), and is associated with intensification of the SLP gradient over the eastern Mediterranean. The Azores high extension is combined with the eastern Mediterranean low and causes a cold, dry, northerly wind that enhances the turbulent fluxes over the Aegean Sea.
The contribution of each component highlights the dominance of the turbulent fluxes and particularly of the latent heat in the extremes modulation as it commonly occurs over the global ocean (Alexander et al. 2002). Although the shortwave radiation is the main source of heat gain by the sea over the global ocean, in the Aegean Sea it exhibits surprisingly poor variation and contributes very little to the surface heat flux anomaly throughout the year. The energy anomaly of the extreme events is larger during the cold season of the year when sensible heat occupies a significant portion of the total energy anomaly. In contrast, during the warm season of the year, the latent heat is by far the prominent regulator of the extremes. A weak wind regime implies small amounts of heat loss and positive heat flux anomalies. Conversely, a strong wind regime increases the amount of latent heat and leads to negative heat flux anomalies. This is considered among the determining factors for the intermediate-water and deep-water formation during winter, when the heat loss is significantly enhanced by both latent and sensible heat. Occasionally, persistent and higher than normal SLP over central Europe supports a cold and dry airflow of continental origin over the Aegean Sea, resulting in large heat loss and triggering the mechanism of deep water convection. This process requires the combination of enhanced surface heat loss by the turbulent fluxes and specific oceanographic conditions, for instance, lateral advection of warm and salty water (Schroeder et al. 2010). This combination leads to the initial formation of surface dense water and to buoyancy anomalies. Thus, the dense water sinks and produces intermediate or deep water, depending not only on the duration and intensity of severe surface cooling but also on the physical characteristics (temperature and salinity) of the water column (Wu et al. 2000).
In conclusion, this study addresses the atmospheric driving factors that modulate the heat fluxes in the Aegean Sea. The analysis of the featured indices and SLP effects reveals some of the basic mechanisms that govern such energy exchanges. However, a small fraction the total variance of heat fluxes still remains unaccounted for. Further investigation is needed to address possible additional factors modulating the surface heat exchanges over the Aegean Sea and the broader area of eastern Mediterranean. For example, contributions of the Siberian thermal anticyclone and the North African depressions as well as the impact of the West Asia monsoon system have not been extensively clarified. Finally, more study should be devoted to elucidate the potential mechanisms through which local effects modulate the air–sea heat fluxes in the Aegean Sea.
Acknowledgments
The authors thank Christopher Fairall (NOAA, United States) and Simon Josey (NOC, United Kingdom) for their valuable advice. Simon Ruiz (IMEDEA, Spain) is also acknowledged for kindly providing the HIPOCAS dataset for the Aegean Sea. The HIPOCAS dataset has been produced by Puertos del Estado (Spain). In particular, the three anonymous reviewers are acknowledged for their valuable contribution in improving the initial manuscript.
REFERENCES
Alexander, M. A., and J. D. Scott, 1997: Surface flux variability over the North Pacific and North Atlantic Oceans. J. Climate, 10, 2963–2978.
Alexander, M. A., I. Blade, M. Newman, J. R. Lanzante, N.-C. Lau, and J. D. Scott, 2002: The atmospheric bridge: The influence of ENSO teleconnections on air–sea interaction over the global oceans. J. Climate, 15, 2205–2231.
Ben-Gai, T., A. Bitan, A. Manes, P. Alpert, and Y. Kushnir, 2001: Temperature and surface pressure anomalies in Israel and the North Atlantic Oscillation. Theor. Appl. Climatol., 69, 171–177.
Bentamy, A., K. B. Katsaros, A. M. Mestas-Nuñez, W. M. Drennan, E. B. Forde, and H. Roquet, 2003: Satellite estimates of wind speed and latent heat flux over the global oceans. J. Climate, 16, 637–656.
Berry, D., and E. Kent, 2009: A new air–sea interaction gridded dataset from ICOADS with uncertainty estimates. Bull. Amer. Meteor. Soc., 90, 645–656.
Beuvier, J., and Coauthors, 2010: Modeling the Mediterranean Sea interannual variability during 1961–2000: Focus on the Eastern Mediterranean Transient. J. Geophys. Res., 115, C08017, doi:10.1029/2009JC005950.
Bond, N. A., and M. F. Cronin, 2008: Regional weather patterns during anomalous air–sea fluxes at the Kuroshio Extension Observatory (KEO). J. Climate, 21, 1680–1697.
Brunetti, M., M. Maugeri, and T. Nanni, 2002: Atmospheric circulation and precipitation in Italy for the last 50 years. Int. J. Climatol., 22, 1455–1471.
Cayan, D. R., 1992: Latent and sensible heat flux anomalies over the northern oceans: The connection to monthly atmospheric circulation. J. Climate, 5, 354–369.
Chronis, T., V. Papadopoulos, and E. I. Nikolopoulos, 2011a: QuickSCAT observations of extreme wind events over the Mediterranean and Black Seas during 2000–2008. Int. J. Climatol., 31, 2068–2077, doi:10.1002/joc.2213.
Chronis, T., D. Raitsos, D. Kassis, and A. Sarantopoulos, 2011b: The summer North Atlantic Oscillation influence on the eastern Mediterranean. J. Climate, 24, 5584–5596.
Conte, M., S. Giuffrida, and S. Tedesco, 1989: The Mediterranean oscillation: Impact on precipitation and hydrology in Italy. Proc. Conf. on Climate and Water, Vol. 1, Helsinki, Finland, Academy of Finland, 121–137.
Davis, R., B. Hayden, D. Gay, W. Phillips, and G. Jones, 1997: The North Atlantic subtropical anticyclone. J. Climate, 10, 728–744.
Dünkeloh, A., and J. Jacobeit, 2003: Circulation dynamics of Mediterranean precipitation variability 1948-98. Int. J. Climatol., 23, 1843–1866.
Eshel, G., and B. F. Farrell, 2000: Mechanisms of Eastern Mediterranean rainfall variability. J. Atmos. Sci., 57, 3219–3232.
Feliks, Y., M. Ghil, and A. W. Robertson, 2010: Oscillatory climate modes in the Eastern Mediterranean and their synchronization with the North Atlantic Oscillation. J. Climate, 23, 4060–4079.
Flocas, H. A., I. Simmonds, J. Kouroutzoglou, K. Keay, M. Hatzaki, V. Bricolas, and D. Asimakopoulos, 2010: On cyclonic tracks over the Eastern Mediterranean. J. Climate, 23, 5243–5257.
Ghate, V., B. Albrecht, C. Fairall, and R. Weller, 2009: Climatology of surface meteorology, surface fluxes, cloud fraction, and radiative forcing over the southeast Pacific from buoy observations. J. Climate, 22, 5527–5540.
Gilman, C., and C. Garrett, 1994: Heat flux parameterization for the Mediterranean Sea: The role of atmospheric aerosols and constraints from the water budget. J. Geophys. Res., 99 (C3), 5119–5134.
Gündüz, M., and E. Ozsoy, 2005: Effects of the North Sea Caspian pattern on surface fluxes of Euro-Asian-Mediterranean seas. Geophys. Res. Lett., 32, L21701, doi:10.1029/2005GL024315.
Hatzaki, M., H. A. Flocas, D. N. Asimakopoulos, and P. Maheras, 2007: The Eastern Mediterranean teleconnection pattern: Identification and definition. Int. J. Climatol., 27, 727–737.
Josey, S. A., 2003: Changes in the heat and freshwater forcing of the Eastern Mediterranean and their influence on deep water formation. J. Geophys. Res., 108 (C7), 3237–3259.
Josey, S. A., S. Somot, and M. Tsimplis, 2011: Impacts of atmospheric modes of variability on Mediterranean Sea surface heat exchange. J. Geophys. Res., 116, C02032, doi:10.1029/2010JC006685.
Kara, B. A., P. Rochford, and H. Hurlburt, 2000: Efficient and accurate bulk parameterizations of air–sea fluxes for use in general circulation models. J. Atmos. Oceanic Technol., 17, 1421–1438.
Konda, M., H. Ichikawa, H. Tomita, and M. F. Cronin, 2010: Surface heat flux variations across the Kuroshio Extension as observed by surface flux buoys. J. Climate, 23, 5206–5221.
Kutiel, H., and Y. Benaroch, 2002: North Sea–Caspian Pattern (NCP)—An upper level atmospheric teleconnection affecting the Eastern Mediterranean: Identification and definition. Theor. Appl. Climatol., 71, 17–28.
Lionello, P., and A. Sanna, 2005: Mediterranean wave climate variability and its links with NAO and Indian Monsoon. Climate Dyn., 25, 611–623.
Lolis, C. J., A. Bartzokas, and B. D. Katsoulis, 2004: Relation between sensible and latent heat fluxes in the Mediterranean and precipitation in the Greek area during winter. Int. J. Climatol., 24, 1803–1816.
Matsoukas, C., A. C. Banks, N. Hatzianastassiou, K. G. Pavlakis, D. Hatzidimitriou, E. Drakakis, P. W. Stackhouse, and I. Vardavas, 2005: Seasonal heat budget of the Mediterranean Sea. J. Geophys. Res., 110, C12008, doi:10.1029/2004JC002566.
Metaxas, D. A., and A. Bartzokas, 1994: Pressure covariability over the Atlantic, Europe and N. Africa. Application: Centers of action for temperature, winter precipitation and summer winds in Athens, Greece. Theor. Appl. Climatol., 49, 9–18.
Metaxas, D. A., A. Bartzokas, C. C. Repapis, and N. R. Dalezios, 1993: Atmospheric circulation anomalies in dry and wet winters in Greece. Meteor. Z., 2, 127–131.
Nittis, K., V. Zervakis, L. Perivoliotis, A. Papadopoulos, and G. Chronis, 2001: Operational monitoring and forecasting in the Aegean Sea: System limitations and forecasting skill evaluation. Mar. Pollut. Bull., 43, 154–163.
Nittis, K., A. Lascaratos, and A. Theocharis, 2003: Dense water formation in the Aegean Sea: Numerical simulations during the Eastern Mediterranean Transient. J. Geophys. Res., 108, 8120, doi:10.1029/2002JC001352.
Papadopoulos, V. P., T. G. Chronis, S. Ruiz, E. Tragou, S. Josey, and G. Ferentinos, 2010: Air-sea heat fluxes estimation in the Aegean Sea (Eastern Mediterranean). Rapp. Comm. Int. Mer Medit., 39, 156.
Papadopoulos, V. P., A. Bartzokas, T. Chronis, S. Ruiz, and G. Ferentinos, 2011: Correlation between air-sea heat fluxes over the Aegean Sea and the total precipitable water over Europe and North Africa. Adv. Sci. Res., 6, 63–67.
Paz, S., Y. Tourre, and S. Planton, 2003: North Africa–West Asia (NAWA) sea-level pressure patterns and their linkages with the Eastern Mediterranean (EM) climate. Geophys. Res. Lett., 30, 1999, doi:10.1029/2003GL017862.
Raicich, F., N. Pinardi, and A. Navarra, 2003: Teleconnections between Indian Monsoon and Sahel rainfall and the Mediterranean. Int. J. Climatol., 23, 173–186.
Ratsimandresy, A. W., M. G. Sotilo, J. C. Carretero, E. Alvarez, and H. Hajji, 2008: A 44-year high resolution ocean and atmospheric hindcast for the Mediterranean Basin developed within the HIPOCAS project. Coastal Eng., 55, 827–842.
Roether, W., B. Manca, B. Klein, D. Bregant, D. Georgopoulos, V. Beitzel, V. Kovacevic, and A. Luchetta, 1996: Recent changes in Eastern Mediterranean deep waters. Science, 271, 333–335.
Romanou, A., G. Tselioudis, C. S. Zerefos, C.-A. Clayson, J. A. Curry, and A. Andersson, 2010: Evaporation–precipitation variability over the Mediterranean and the Black Seas from satellite and reanalysis estimates. J. Climate, 23, 5268–5287.
Ruiz, S., D. Gomis, M. Sotillo, and S. Josey, 2008: Characterization of surface heat fluxes in the Mediterranean Sea from a 44-year high-resolution atmospheric data set. Global Planet. Change, 63, 258–274.
Schroeder, K., S. A. Josey, M. Herrmann, L. Grignon, G. P. Gasparini, and H. L. Bryden, 2010: Abrupt warming and salting of the Western Mediterranean Deep Water after 2005: Atmospheric forcings and lateral advection. J. Geophys. Res., 115, C08029, doi:10.1029/2009JC005749.
Shaman, J., R. M. Samelson, and E. Skyllingstad, 2010: Air–sea fluxes over the Gulf Stream region: Atmospheric controls and trends. J. Climate, 23, 2651–2670.
Sotillo, M. G., A. W. Ratsimandresy, J. C. Carretero, A. Bentamy, F. Valero, and F. Gonzalez-Rouco, 2005: A high-resolution 44-year atmospheric hindcast for the Mediterranean Basin: Contribution to the regional improvement of global reanalysis. Climate Dyn., 25, 219–236.
Soukissian, T., G. Chronis, K. Nittis, and C. Diamanti, 2002: Advancement of operational oceanography in Greece: The case of the Poseidon system. Global Atmos. Ocean Syst., 8, 93–107.
Sterl, A., and W. Hazeleger, 2003: Coupled variability and air-sea interaction in the South Atlantic Ocean. Climate Dyn., 21, 550–571.
Theocharis, A., K. Nittis, H. Kontoyiannis, E. Papageorgiou, and E. Balopoulos, 1999: Climatic changes in the Aegean Sea influence the Eastern Mediterranean thermohaline circulation (1986-1997). Geophys. Res. Lett., 26, 1617–1620.
Trigo, I. F., T. Davies, and G. Bigg, 1999: Objective climatology of cyclones in the Mediterranean region. J. Climate, 12, 1685–1696.
Trigo, I. F., G. Bigg, and T. Davies, 2002: Climatology of cyclogenesis mechanisms in the Mediterranean. Mon. Wea. Rev., 130, 549–569.
Uppala, M., and Coauthors, 2005: The ERA-40 Re-Analysis. Quart. J. Roy. Meteor. Soc., 131, 2961–3012.
Venegas, S. A., 2001: Statistical methods for signal detection in climate. Danish Center for Earth System Science Rep. 2, 96 pp.
Weller, E., M. Nunez, G. Meyers, and I. Masiri, 2008: A climatology of ocean–atmosphere heat flux estimates over the Great Barrier Reef and Coral Sea: Implications for recent mass coral bleaching events. J. Climate, 21, 3853–3871.
Wu, P., K. Haines, and N. Pinardi, 2000: Toward an understanding of deep-water renewal in the Eastern Mediterranean. J. Phys. Oceanogr., 30, 443–458.
Yu, L., and R. A. Weller, 2007: Objectively analyzed air–sea heat fluxes for the global ice-free oceans (1981–2005). Bull. Amer. Meteor. Soc., 88, 527–539.
Yu, L., R. A. Weller, and B. Sun, 2004: Improving latent and sensible heat flux estimates for the Atlantic Ocean (1988–99) by a synthesis approach. J. Climate, 17, 373–393.
Ziv, B., H. Saaroni, and P. Alpert, 2004: The factors governing the summer regime of the Eastern Mediterranean. Int. J. Climatol., 24, 1859–1871.
Zolina, O., and S. K. Gulev, 2003: Synoptic variability of ocean–atmosphere turbulent fluxes associated with atmospheric cyclones. J. Climate, 16, 2717–2734.
