Interactive comment on “ Calculating the water and heat balances of the Eastern Mediterranean basin using ocean modelling and available meteorological , hydrological , and ocean data ”

Dear Referee 1 Thanks for your comments on our manuscript: “Calculating the water and heat balances of the Eastern Mediterranean basin using ocean modelling and available meteorological, hydrological, and ocean data”. In general we find several of the comments good and useful and will respond to them one by one below. We fully agree that our modelling approach is a very simplified. However, we believe it is a useful start when analyzing the Mediterranean basin and we will later treat the dynamics in the different sub-basins in more details. As you stated a lot of data needs


Introduction
The Global Energy and Water Experiment (GEWEX) were developed within the World Climate Research Programme (WCRP).Within GEWEX, continental-scale experiments were initiated that included the Baltic Sea drainage basin.In the Baltic Sea experiment (BALTEX), several improvements have been developed with regard to scientific understanding, observations, and modelling considerations (Raschke et al., 2001;Omstedt et al., 2004).
New initiatives are now being taken in the Mediterranean region, where climate change may pose a severe threat.Within the HyMex programme (http://www.hymex.Introduction

Conclusions References
Tables Figures

Back Close
Full org/), the water cycle is of major concern.To support this initiative, we aim to transform knowledge gained from the BALTEX programme so it is applicable to the Mediterranean region.The present paper is the first such initiative and addresses the water and heat balances in the Eastern Mediterranean basin.The approach follows that of Omstedt and Nohr (2004), who used a process-based ocean model together with available meteorological, hydrological, and in situ ocean data to analyse the water and heat cycles.
The Eastern Mediterranean Basin (EMEB), which extends from 11 • E to 36 • E and from 30 • N to 46 • N, is a semi-enclosed basin with negative water balance (i.e.evaporation greater than precipitation plus river runoff).The Sicily Channel (149 km wide) and Strait of Messina (4 km wide) form the connection between the Eastern and Western Mediterranean Sea Basins (Fig. 1).Bosporus-Marmara-Dardanelles system connect Black Sea with EMEB.The exchange through Messina Strait is much smaller than the flow through the Sicily Channel and is therefore neglected.Black Sea will be modelled as a river runoff with a 18 psu smaller than the Mediterranean.The Eastern Mediterranean will be regarded here as a single natural basin with in-and outflows and with focus on processes such as: air sea interaction, land-sea interaction (i.e.river runoff), diapycnal mixing, overturning circulation (Atlantic water inflow modification, Intermediate and deep water formation), exchange through the Sicily Channel and the effect of Black Sea.The Nile River and the Black Sea play an important role in change the fresh water content in the study basin.Intermediate and deep water formations affect the EMED stratification.The model will be driven by available meteorological and hydrological data and validated with available oceanographic data.Based on the calculations, conclusions will be drawn regarding the water (salinity) and heat (temperature) balances.
The thermohaline water structure in the Eastern Mediterranean is an important climate issue, and any changes in it may greatly affect the marine system (e.g.change the deep water formation areas, current system, and sea level, and influences on the ecosystem).Fresh water input to the Eastern basin is mixed with sea surface water.Introduction

Conclusions References
Tables Figures

Back Close
Full Surface water flows from the Western Mediterranean basin of modified Atlantic water (MAW) through the Sicily Channel are mixed with Eastern basin surface water.Outflow flow water over sill depth (500 m) is responsible for water loss from the Eastern Mediterranean.The net precipitation rate (precipitation, P , minus evaporation, E ) is responsible for salinity increase inside the basin.In winter, and due to evaporation and cooling fluxes, density of Levantine surface water increase further until it become dense enough to forming Levantine intermediate water (200-500 m) or Levantine deep water.Roether and Schlitzer (1991) addressed that the average deep water formation rate in the Eastern Mediterranean is about 0.3 Sv.Malanotte-Rizzoli et al. (1999) indicated that deep water formation basically takes place in Adriatic, Aegean and Levantine subbasins.Zervakis et al. (2000) showed that enhanced negative water balance of Eastern Mediterranean leads to new source of deep water formation especially in Aegean Sea.
B éranger et al. ( 2002) investigated the mean inflow to the Eastern Mediterranean basin through the Sicily Channel using numerical modelling.They estimate that the mean inflow through the Channel is approximately 1.05 ± 0.35 Sv for 13 yr period.Stansfield et al. (2002) estimated the surface flow to the Eastern basin using observations from CTD data.They found that there is surface flow of MAW through the Sicily Channel to a depth of 150 m.The outflow of Levantine intermediate water (LIW) through the Sicily Channel takes place in deeper layers.Sorgente et al. (2003)  of 0.009 Sv).Recently estimates indicates a reduced inflow by about 0.003 Sv, which affect North Aegean Sea circulation (Stanev and Peneva, 2002).Nixon (2003) and Ludwig et al. (2009) mentioned that the average discharge of the Nile River to the Mediterranean basin before building Aswan high dam was about 2 times higher than after the building.This paper aims to: (1) examine the water exchange through the Sicily Channel, (2) calculate the long-term change of temperature and salinity distribution in the Eastern Mediterranean basin, and (3) examine the heat and water balances of the Eastern Mediterranean basin.

Modeling forcing data and observations
Several data sources have been used in this study, they are: 1. Mediterranean Sea absolute dynamic topography data from May 2006 to October 2009 -these data were acquired from the AVISO database available on the internet at http://www.aviso.oceanobs.com/en/home/index.html.Absolute dynamic topography is the sea surface height above the geoid and is calculated as the sum of sea level anomaly and mean dynamic topography.The data were calculated using a one-day temporal scale and 1/3 • spatial scale and used to study exchange through the Sicily Channel.
2. Digitized bathymetric data obtained from the British Oceanographic Data Centre with a 0.5-min grid (http://www.bodc.ac.uk/data/online delivery/gebco/) -these data were used when calculating the cross-sectional area of the Sicily Channel and the area-depth distribution of the Eastern basin.and salinity  reanalyzed data for the Mediterranean Basin taken at a 1-month interval with a spatial distribution of 1 Both data sets were used to validate the model results.

Model description
The study relies on numerical modelling of the heat and the water balances of the Eastern Mediterranean basin and the water exchange through the Sicily Channel.
The model used calculates the horizontally integrated but vertically resolved resolution properties and is thus a transient one-dimensional model with in and out flows controlling the vertical circulation.The exchange through the Sicily Channel was modelled using current speed across the Channel calculated from satellite, evaporation, precipitation and river inflow data.The studied period was from 1958 to 2009.Introduction

Conclusions References
Tables Figures

Back Close
Full The model simulated the properties of the Eastern Mediterranean basin based on horizontally averaged advective-diffusive conservation equations for volume, heat, momentum, and salinity.A two-equation turbulent model was applied.The model is fully described in Omstedt and Axell (2003); here, however, only one basin with in-and outflows is considered.In the following, the basic equations are outlined: The transport equations for horizontal momentum read: where, U and V are the current components in east and north directions, respectively, W the vertical velocity calculated from the differences between in-and outflows, f the Coriolis parameter, ρ is the water density, and µ eff is the effective dynamic viscosity.
The conservation equation for heat reads: Where, T and C p are temperature of sea water and heat capacity respectively.σ effT is the turbulent Prandtl number. (5)

Conclusions References
Tables Figures

Back Close
Full where, Γ sum and Γ h are the source terms associated with sun radiation and in-and outflows, respectively, Γ w s the short-wave radiation through the water surface, η the infrared fraction of short-wave radiation trapped in the surface layer, and β the bulk absorption coefficient of the water.Q in (Q out ) is volume fluxes of in (out) flowing water.D is the total depth and z the vertical coordinate positive upwards.
The conservation equation for salinity reads: where, Γ s is the source term associated with in-and outflows, and σ effS is the turbulent Schmidt number.
The vertical turbulent transports in the surface boundary layer are calculated using the so-called k-ε model, which is a two-equation model of turbulence in which transport equations for turbulent kinetic energy, k, and its dissipation rate, ε, are calculated.The turbulence model reads:

Conclusions References
Tables Figures

Back Close
Full where σ k is the Schmidt number for k and σ ε is the Schmidt number for ε; C 1ε , C 2ε , and C 3ε are turbulent coefficients and P b is production/destruction due to stratification.α, α 1 , α 2 are constants, N −1 is the Buoyancy frequency and ν d t is the turbulent deep water eddy viscosity.The model use the general equation solver PROBE which is a well-documented and freely available program for studies of lakes and coastal seas (Omstedt, 2011).

Water exchange through the Sicily Channel
To calculate the heat and water balances of the Eastern Mediterranean basin, the water exchange through the Sicily Channel is needed.This exchange is considered a twolayer exchange, comprising surface flow (Q in ) from the Western Mediterranean basin and a deeper flow (Q out ) from the Eastern to Western basins over sill depth (150 to 500 m).Surface flow is calculated using surface satellite sea level data (η) across the Channel.Climatological oceanographic data are used to calculate the deeper flow.
The surface flow from the Western to the Eastern Mediterranean basin (Q in ) was calculated as follows: 1.The horizontal distance from northern to southern Sicily Channel was divided into 17 cross-sectional sectors based on sea level spatial distribution (1/3 • ).
2. The surface geostrophic current speed for each sector was calculated based on geostrophic approximation, or

Conclusions References
Tables Figures

Back Close
Full Where f is the Coriolis parameter, g the gravity force, U g and V g the velocity components in the x and y directions, respectively, and W g the surface geostrophic speed.
-For simplification, we assumed that the depth of surface layer was 150 m (see, e.g.Stansfield et al., 2002).In addition, the upper 150 m was divided into five vertical layers of water, each 30 m deep.We then assumed that the surface current was decreasing linearly toward the bottom of the surface layer.
-A Cartesian grid of the Sicily Channel with an 8.76-km horizontal resolution and 0.03-km vertical resolution was created, and the surface flow from the Western Mediterranean basin was calculated as a summation over the Channel: Where, W n is now the normal cross-Channel component of the geostrophic flow and A is the corresponding cross-sectional area (Fig. 1).
From the water volume conservation throughout the Eastern Mediterranean basin, the rate of surface elevations change can be calculated as Where A s is the Eastern Mediterranean surface area (= 1.67 × 10 12 m 2 ) and ( ∂η ∂t ) is the change in sea level with time.Q f is the river discharge to the basin and calculated as a sum of total amount of river runoff to the Eastern Mediterranean basin and the Black Sea brackish water.In the present model, we assume that the volume fluxes related to surface elevation changes are small relative to the other contributions, which means that the left hand side of Eq. ( 17) is zero; this is valid for a long-term scale.This equation can therefore be used to calculate Q out .Introduction

Conclusions References
Tables Figures

Back Close
Full The heat balance of the water body can be calculated according to Omstedt and Nohr (2004) as: where H = ρC p T d zd A s is the total heat content of the Eastern Mediterranean, F in and F out the heat fluxes associated with in-and outflows through trhe Sicily Channel, respectively, and calculated according to respectively, T in the surface inflowing temperature from the Western Mediterranean basin, and T out temperature of the outflow.C p is heat capacity (= 4200 J kg F loss is the total heat loss through the sea surface (W m −2 ) and calculated as Fluxes are negative when going from atmosphere to the ocean.

Forcing data and later boundary conditions
The Eastern Mediterranean basin was modelled using the PROBE equation solver, the present application being called PROBE-EMED.The model used the area-depth distribution of the Eastern Mediterranean Basin and was forced with meteorological and river runoff data.Satellite sea level observations across the Sicily Channel and surface temperature and salinity at the western side of the Channel were used as lateral boundary conditions.Meteorological data including surface air temperatures ( • C), surface winds (m s −1 ), percentage of total cloud cover, and the relative humidity and precipitation rates (m s −1 ) were used.The air temperature data were corrected for land influence by comparing them with sea surface temperatures.River runoff data is used as fixed values for each month during the all running time and calculated from the available data.Annual average value of river runoff is 5000 and 3800 m 3 s −1 before and

OSD Introduction Conclusions References
Tables Figures

Back Close
Full after 1964 respectively.This is due to building of Aswan high dam and its effect on Nile river discharge.Black Sea is modelled as river runoff.The most significant rivers flow to the Black Sea are Danube (6766 m 3 s −1 ), Dnieper (1506 m 3 s −1 ), Rioni (408 m 3 s −1 ), Dniester (375 m 3 s −1 ), Kizilirmak (202 m 3 s −1 )), Sakarya (193 m 3 s −1 ) and southern Bug (110 m 3 s −1 ).Fresh water discharge into the Eastern Mediterranean basin (rivers of Eastern basin plus Black Sea) addressed a negative trend of −26.14 m 3 s −1 ) yr −1 due to building of Aswan high dam and increasing in evaporation rates.This may leads to increasing Eastern Mediterranean basin salinity.
Lateral boundary parameters of surface salinity and water temperature were applied representing the southern Tyrrhenian basin.Surface salinity was calculated as monthly means using data obtained from the National Oceanographic Data Center.Surface air temperature was calculated using ECWMF database with 6 h temporal resolution.
Three years (May 2006 to June 2009) of daily sea level difference across the Sicily Channel were used to calculate daily Q in values for the same period.The approach assumes that the daily Q in values is used for the May 2006-June 2009 period and monthly average values of Q in can be used to cover the rest of the study period (January 1958-April 2006).This assumption is reasonable as: (1) tide (which highly influence sea level fluctuation) is a periodic parameter; (2) the difference between the monthly average values of surface temperature and salinity for the eastern and western side of the Sicily Channel are small; and (3) the vertical difference in the Eastern Mediterranean basin between surface and 500-m-depth salinity is only approximately 0.3 m.In future work, the Mediterranean climate system will be modelled using several coupled subbasin models, which will calculate the Sicily Channel flow as a baroclinic exchange flow.Moreover, we will create another four experiments with the PROBE-EMED model for 1958-2009 to analyse the sensitivity of this assumption.This step will be used to examine the error in modelling the Eastern Mediterranean due to the assumption of Q in as a fixed monthly variable during the years 1958-2005.

Conclusions References
Tables Figures

Back Close
Full

Eastern Mediterranean basin geometry
Bathymetric information and the area-depth distribution of the studied basin are illustrated in Figs. 1 and 2. The surface area is 1.67 × 10 12 m 2 , the water volume is 2.4 × 10 15 m 3 , the average depth is 1430 m, and the maximum depth is 5097 m.The annual average fresh water runoff is 12 943 m 3 s −1 , and average precipitation and evaporation are 1.58 and 3.76 mm day −1 , respectively.Moreover, average monthly surface salinity and water temperature over the entire basin ranged from 38.3 to 38.8 and 14.8 to 27 • C, respectively.The cross-sectional area of the Sicily Channel is calculated from bathymetric data (Fig. 2b).This figure illustrates: (1) the Channel width from the southern to the northern parts is approximately 149 km; (2) the southern part of the Channel is deeper than the northern part; (3) the maximum depth across the Channel is 830 m; and (4) the outflow volume flux is concentrated in the southern part.

Water exchange through the Sicily Channel
Satellite data on the sea level across the Sicily Channel were used to examine surface current flow from the Western to Eastern basins using Eq.(15) (Fig. 3). Figure 3 depicts how the surface current of MAW from the Western to Eastern basins can take various routes, such as: 1. two routes, one in the northern and the other in the southern part of the Channel (Fig. 3a), 2. water transport across the entire Channel (Fig. 3b), 3. water transport concentrated on the northern part (Fig. 3c), and 4. water transport concentrated on the southern part (Fig. 3d).

Conclusions References
Tables Figures

Back Close
Full These four routes must be kept in mind when measuring the exchanges.So we shall divide the cross section area across the Channel into small grid size (17 grid cell are used) to get accurate Q in values.Also, we cannot use only one current meter across the Channel but we have to use series of current meter across the Channel (more than 10) which is more difficult.This is why, calculating the exchange across Sicily Channel using satellite data is the best way.Using 17 grids is in agreement with Buongiorno Nardelli et al. ( 2006) who addressed the mesoscale process importance across the channel.
The temporal variations in the surface-and deep-layer flows are depicted in Fig. 4, showing several features.The surface flow over the period (early June 2006-late June 2009) ranged from 0.25 to 2.56 Sv with average values of 1.16 ± 0.34 Sv, while the deep flow ranged from 0.25 to 2.56 Sv with average values of 1.13 ± 0.36 Sv.The deep flow (150: 500 m) was slightly smaller than surface flow, indicating a loss of water inside the Eastern Mediterranean basin probably due to evaporation.There is a significant difference of Q in monthly average values.This difference is ranged between 1.4 ± 0.46 Sv during April and 0.98 ± 0.20 Sv during July.Q in average value is in good agreement with B éranger et al. ( 2002) value (1.05 ± 0.35 Sv) and larger than Ferjani and Gana (2010) values (by about 0.6 Sv).This 0.6 Sv different with Ferjani and Gana (2010) in calculating Q in values come from their method that calculating the exchange through the channel from western side to 60 km wide (not to 149 km as the present study).
The agreement with B éranger et al. ( 2002) and Buongiorno Nardelli et al. ( 2006) and disagreement with Ferjani and Gana (2010) indicates that the present study is useful.

Modelling the Eastern Mediterranean basin
Simulating the Eastern Mediterranean basin water and heat cycles was done using the PROBE-EMED model from 1958 to 2009 with a 3-h temporal resolution.Water

Conclusions References
Tables Figures

Back Close
Full masses were modelled using a vertically resolved grid with 190 grid cells, expanding from surface to bottom.
The PROBE-EMED model results were examined by comparing the reanalysed data and modelled long-term monthly surface and 500-m-depth temperatures, surface and 500-m-depth salinities, and evaporation (Figs.5-8).Note that the model simulated data every 3 h, while reanalysed data were based on a 6 h interval for evaporation and a monthly interval for temperature and salinity.Reanalysed data used for validation were taken as average values for the entire Eastern Mediterranean basin.
The modelled and reanalysed temperature and salinity (MEDAR and NODC datasets) at the surface and at a depth of 500 m are presented in Figs. 5 and 6, respectively.The modelled surface temperature follows the reanalysed temperature closely with a correlation (R) of 0.98; n (observations) = 612, SE (Standard Error) = 0.7 • C and the average modelled and reanalysed surface temperatures over the study period were 20.65 ± 3.7 • C and 20.3 ± 3.7 • C, respectively.Average modelled and reanalysed surface salinities were 38.32 ± 0.14 and 38.39 ± 0.13, respectively (R = 0.6; n = 540; SE = 0.11).This slight difference between modelled and reanalysed salinities may be because the reanalysed data are used 1-yr running window.At a water depth of 500 m, the average modelled and reanalysed temperatures (salinity) over the study period were 14.9 ± 0.2 • C (38.65 ± 0.16) and 14.1 ± 0.1 • C (38.8 ± 0.05), respectively.
The monthly, seasonal, and yearly cycles of surface temperature and salinity were realistically simulated by the model.To investigate the heat balance in greater detail, PROBE-EMED modelled evaporation rates were compared with meteorological modelled evaporation data.For the meteorological data, we now used the NCEP reanalysed data, so an independent dataset was used.The PROBE-EMED model results for surface temperature, salinity, and evaporation rates were also modelled as monthly means (Fig. 9).This figure indicates that the monthly average surface temperature ranged from 15.8 ± 0.32 A final test of the water and heat simulations was to investigate the water mass structure in the whole Eastern Mediterranean basin.By comparing modelled and reanalysed data, an independent test of the approach could be performed.The results are presented in Fig. 10a.In this figure, three water masses (i.e.MAW at the surface, LIW at the intermediate level, and deep water) are clearly seen in the T -S diagram.Deep water masses are more clearly visible in the reanalysed data than the modelled data.This may be because the Eastern basin has at present only been modelled as a single basin, but will in the future be dived into a number of sub-basins.This figure implies that the deep water formation is partly captured despite the simple approach.
To indicates more sensitivity to meteorological data than to exchange through the Sicily Channel.The Eastern basin may therefore be modelled using a lateral boundary of fixed monthly Q in .

Modelled water balance components
The water balance of the Eastern Mediterranean basin is controlled by the Sicily Channel exchange (Q in and Q out ), river runoff (Q f ), and net precipitation (difference between precipitation and evaporation rates, Eq. 18).The various water balance components, except precipitation and river runoff, are modelled using the PROBE-EMED model.Table 1 and Fig. 11 show the estimated monthly and annual mean water balances of the Eastern Mediterranean basin averaged over 52 yr.Moreover, annual means of the difference between inflow and outflow and the net precipitation flow, i.e.A s (P − E ), are shown in Fig. 12.
The results indicate that the in-and outflows are of the order of 1 Sverdrup, while the difference between them is about two orders less.This difference between the in-and outflows was mainly balanced by net precipitation and river runoff.The water balance is thus mainly controlled by the in-and outflows through the Sicily Channel and by net precipitation.The results also indicate that the maximum monthly mean values of Q in occur in April (1.43 Sv), December (1.28 Sv), and November (1.22 Sv), while Q out maximum monthly mean values were 1.42, 1.31, and 1.20 Sv, occurring also in April, December, and November, respectively.The monthly net precipitation rates were in

Modelled heat balance components
The monthly means of heat budget components are presented in Table 2 and Fig Channel.Finally, it is clear that using ocean modelling together with meteorological and river runoff data provides a powerful method for simulating heat and water cycles.
The water and heat balances, together with trend analysis of a long time series, will be used as climate change tools in future studies.Introduction

Conclusions References
Tables Figures

Back Close
Full  Full  Full Screen / Esc Printer-friendly Version Interactive Discussion Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | Screen / Esc Printer-friendly Version Interactive Discussion Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | used the Princeton Ocean Model (POM) to study the flow through the Sicily Channel.This modelling detected two main MAW veins, one in the south along the African coast and the other in the north along the Sicilian coast.Based on geostrophic calculation using CTD data from April, 2003 to October, 2003, Ferjani and Gana (2010) indicated that the mean inflow and outflow through western side of Sicily Channel (60 km wide only) were 0.5 and 0.4 Sv respectively.Stanev et al. (2000) characterized the water exchange at Bosporus-Marmara-Dardanelles system as a two layer flow, where Black Sea water occupies the surface layer (average flow of 0.019 Sv) and Mediterranean occupies deep layer (average flow Discussion Paper | Discussion Paper | Discussion Paper | 3. Data on river runoff into the Eastern Mediterranean basin and the Black Sea over the 1976-1987 periods taken from the SAGE Center for Sustainability and the Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | Screen / Esc Printer-friendly Version Interactive Discussion Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | Screen / Esc Printer-friendly Version Interactive Discussion Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | Screen / Esc Printer-friendly Version Interactive Discussion Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | where F h is the sensible heat flux, F e the latent heat flux, F prec heat flux associated with precipitation (neglected in the study), F l the net long-wave radiation, and F o s the solar radiation to the open water surface.Heat loss from the open sea is calculated as Screen / Esc Printer-friendly Version Interactive Discussion Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | Screen / Esc Printer-friendly Version Interactive Discussion Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | Screen / Esc Printer-friendly Version Interactive Discussion Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | Figure 7 depicts the monthly and yearly average values of modelled evaporation rates based on the PROBE-EMED simulations.Figure 8 depicts the scatter plot of modelled and NCEP reanalysed data for the Eastern Mediterranean basin.Over the study period, Modelled evaporation rates ranged from a minimum of 0.2 to a maximum of 11.3 mm day −1 , with an average value of 3.1 ± 1.5 mm day −1 .The monthly average evaporation rates over the study period Discussion Paper | Discussion Paper | Discussion Paper | ranged from 4.95 ± 1.8 mm day −1 in August 1985 to 1.31 ± 0.45 mm day −1 in May 1993, while the yearly average evaporation rates ranged from 3.26 in 1961 to 2.74 mm day −1 in 1972.The Reanalyzed and modelled monthly evaporation rates agreed fairly well (R = 0.76, n = 618, SE = 0.5 mm day −1 ).
examine the sensitivity of the PROBE-EMED model and confirm our assumption of Q in as a fixed monthly variable from 1958 to 2005 and running daily variable from 2006 to 2009, the results of four sensitivity runs of the model were analysed by adding ±15 % and ±50 % of the average value of Q in (average Q in = 1.16 Sv) to all Q in values, as shown in Figs.10-11.We can conclude that any change of Q in in the ±15 % range indicates insignificant change in the vertical distribution of salinity and temperature.Any change of Q in in the ±50 % range indicates significant change in salinity distribution and negligible change in temperature.Modelling the Eastern basin from 1958 to 2009 Discussion Paper | Discussion Paper | Discussion Paper | the −0.0068:−0.0679Sv (−0.35:−3.5 mm day −1 ), reaching a maximum in August and minimum in December.The Eastern Basin monthly mean river runoff varied between a maximum of 0.019 Sv in April and a minimum of 0.008 Sv in August, with an annual average of 0.013 Sv.Over the studied 52-yr period, Q in −Q out averaged 0.023 ± 0.84 Sv, while A s (P − E ) averaged −0.03 ± 0.04 Sv, the difference being balanced by the river discharge.Discussion Paper | Discussion Paper | Discussion Paper | . 14, while the annual means of F n , F o s , and F loss are shown in Fig. 15.The heat balance simulations indicate that the heat loss from the open sea was almost balanced by the solar radiation to the open water surface.Heat loss from the open sea ranged from 134.9 W m −2 to 229.8 W m −2 , while solar radiation to the open water surface ranged from −300.3 W m −2 in July to −73.3 W m −2 in December.The total heat loss from the Eastern Basin surface was negative (negative fluxes indicate fluxes into the water body) from March to August, while it was positive in the rest of the year.Latent heat flux and net long-wave radiation is important than sensible heat flux in controlling the variability of heat loss from the open sea.The annual average value of F loss was 8.7 W m −2 , which needs to be balanced by the difference in heat transported by the in-and outflowing water.During the study period, the annual average value of F n and F o s were 195.6 W m −2 and −186.9W m −2 respectively.Modelled data of F n shows increasing trend of 0.07 W m −1 .The trend thus indicates increased sun radiation into the water body and increased net heat loss, probably due to increased evaporation rates.In addition, the positive values of annual average F loss imply that the Eastern Mediterranean basin imports heat from the Western basin.Moreover, the figures indicate the close relationship between the meteorological ECMWF data and the present modelled heat balance components (i.e.F n , F o s , and F loss ), with a bias of 4, 2.7, and 3.2 W m −2 , respectively.4 Discussion This work analysed the Eastern Mediterranean basin heat (temperature) and water (salinity) balances for the early 1958-June 2009 period in relation to current climate conditions and exchange through the Sicily Channel.The modelling work was based on one-dimensional equations of momentum, heat, and salt conservation, turbulence Discussion Paper | Discussion Paper | Discussion Paper | being modelled using a two-equation model of turbulence.The exchange through the Sicily Channel connecting the Eastern and Western Mediterranean basins was modelled from satellite altimeter data, assuming a geostrophic flow balance.The model was forced with available meteorological and hydrological datasets and the results were validated with independent datasets for salinity, temperature, evaporation and heat fluxes components reanalysed data for the Eastern Mediterranean.Independent datasets of salinity and temperature are taken from MEDAR (annual) and NODC (monthly) reanalyzed datasets while evaporation and heat fluxes components taken from ECMWF dataset.The paper analyses how the individual terms of the water and heat balances change with time and how the climate change signals affect the heat and water cycles.Individual water and heat component values are presented in figures and tables.There are strong monthly and inter-annual variations in most of the water and heat balance components.The most interesting results are (1) climate conditions are dominating the heat and water balances in the eastern Mediterranean than exchange through the Channel, (2) satellite dynamic height observations across the whole channel is a useful and more comprehensive tool to study the exchange through Sicily Channel instead of using series of current meters across the Channel.Modelled surface long-term temperature and salinity follow the reanalysed data with a bias of −0.4 • C and −0.004, respectively.Monthly and yearly cycles of temperature and salinity are well simulated.Reanalysed and modelled water mass structure and heat balance components display good agreement, indicating that the air-sea interaction and turbulent mixing are realistically simulated.The water balance in the Eastern Mediterranean basin is controlled by the difference between the in-and outflows through the Sicily Channel and by the net precipitation rates.The heat balance is controlled by the heat loss from the water surface, sun radiation into the sea, and heat flow through the Channel, the first two displaying both climate trends.An annual net heat loss of approximately 9 W m −2 was balanced by net heat flow through the Discussion Paper | Discussion Paper | Discussion Paper | Screen / Esc Printer-friendly Version Interactive Discussion Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | Screen / Esc Printer-friendly Version Interactive Discussion Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper |

Fig. 15 .
Fig. 15.Eastern Mediterranean basin annual means for calculated and reanalysed heat balance components: Heat loss from the open sea (F n ) (a), solar radiation to the open water surface (F o s ) (b), and total heat loss through the sea surface (F loss ) (c).The modelled values are taken from the present model study, while the reanalysed data were extracted from the meteorological database (ECMWF).
• C in March to 25.98 ± 0.44 • C in August.The monthly average surface salinity ranged from 38.19 ± 0.09 in Mai to 38.5 ± 0.09 in September.The monthly average evaporation rates over the study period ranged from 1.78 ± 0.78 to 3.91 ± 1.08 mm day −1 in April and August, respectively.In the summer, surface temperature and evaporation reached their maximum values, as did surface salinity values.

Table 1 .
Modelled monthly and annual mean water balance of the Eastern Mediterranean basin averaged over 52 years.Storage is change in Eastern basin volume.

Table 2 .
Estimated monthly and annual means of heat balance components (unit: W m −2 ).