The interest of the international research and operational communities for seasonal and interannual simulations and/or predictions has been constantly increasing in the last few years. At the International Forum on ENSO, held in Washington DC in November 1995, representatives of 40 nations have acknowledged the importance of accurate, reliable and timely operational predictions at the seasonal time scale.
Such advances have been made possible by the discovery of the spectacular El Nino/Southern Oscillation events in the equatorial Pacific ocean and their impact on the general circulation. The demonstration that coupled dynamics is a fact of life and not a subject of theoretical investigation has contributed to focus the international research on climate simulations on the usage of coupled atmosphere-ocean global models. The amount of scientific evidence pointing to the importance of the coupled ocean-atmospheric dynamics for climate simulations at seasonal, interannual and longer time scales is impressive. Coupled ocean-atmosphere general circulation models have been demonstrated to be a necessary tool for a wide range of application ranging from short-term climate predictions (ENSO events, seasonal atmospheric anomalies, monsoon and sub-Saharan circulation) to simulations of long-term climate changes.
Subproject A will focus on the simulation of the present climate via global coupled numerical models of the atmosphere and the oceans. It is now widely accepted that both the intraseasonal and the interannual atmospheric variability is dominated by several interacting phenomena, such as ENSO and the Madden and Julian oscillation in the tropical regions and the apparently more unpredictable vagaries of the midlatitudes, blocking, storm tracks, teleconnections and weather regimes, all irregularly, but repeatedly, affecting our lives. Some of these processes have only recently been more precisely defined and observed. More careful work is needed in the direction of accurate statistical analysis of midlatitude and tropical weather regimes, including shedding light on the relations between global and regional regimes. Chemical processes have also been proved to be very effective in regulating climate fluctuations, but it has been clearly demonstrated that the interaction between dynamics and chemistry is so tight that it is necessary to develop interactive chemical processes parameterizations into general circulation models. Subproject A includes a Task (A.6) that has the responsibility to develop and implement into the GCM developed by Task A.2 such an interactive chemical package.
Numerical modeling is the cornerstone of climate studies, providing the framework to perform experiments in controlled conditions, sensitivity studies, and diagnostic analyses of process and phenomena. Advanced numerical models of the global atmosphere, oceans and chemical processes, are currently being employed in every leading research and operational institution in the world, but Italy has been lagging behind, mainly due to a lack of resources and coordination. It is important to make available to the Italian Scientific Community such an important tool. The data from the numerical integrations performed in this subproject will be available to the Italian scientific community for further studies and analyses. A deployment of modeling and simulation capabilities at the level of the other European nations should be placed at the highest national priority, but it requires a commitment of substantial financial resources outside the scope of a Progetto Strategico. However, this project can contribute a focal point for scientific discussions and a starting point for a more complete national program.
Task A.1 will concentrate on the study of the interannual midlatitude anomalies, whereas it will be the responsibility of Task A. 2 the basic construction of the coupled model and the simulation of the present climate variability and teleconnections patterns. The issues connected with the assessment of the climate predictability will be dealt with in Task A.3 and more theoretical details of the mechanism that could generate such variability will be analyzed in Task A.4. Task A.5 will attempt a study of the role of clouds and radiative feedback in the context of the climate problem. The components of the chemical model will be developed in Task A.6. All Tasks will contribute greatly to the advancement of our knowledge of the interannual variability and to our understanding of the various physical , dynamical mechanism that are responsible for it.
Subproject A will nicely fit within the directives and the recommendations of the international program CLIVAR and might constitute a substantial part of the Italian contribution to the CLIVAR program itself. At the end of subproject A, we will have greatly enhanced the level of research in the Italian community and the role of Italian science in the international competition.This results can be achieved with a comparatively small level of funding because of the positive feedback with the several international programs currently funded that this project nicelycomplement.
Task A.1 Seasonal to interannual variability of atmospheric anomalies
Principal Scientist: Prof. S. Tibaldi (ARPA-SMR, Bologna)
Background
The ability to quantify the predictability of climate on seasonal to interannual time scales, and to develop an operational seasonal forecast capability, is feasible as a result of a number of major developments over the last decade or so. However, before we aim at predicting quantitatively seasonal to short-term climate fluctuations it is necessary to quantify how the atmospheric variability is partitioned among the various time scales and among the various atmospheric processes. It is , for example, well known that the idealized interpretative model which ascribes short-term atmospheric variance (less than 5 days) to cyclones and long-term variance (more than 5 days) to blocking is a gross oversimplification. In addition to this, it has been shown in recent years that both cyclone tracks and blocking are strongly influenced by teleconnection patterns and their precise relative phases. It has also recently been shown that observed diagnostic efforts in this area can produce solid benchmark tests for GCM quality control, providing hints and suggestions as to the possible causes of model systematic errors and climate drifts. The Italian scientific community has contributed much to this field of work in recent years and has all the necessary competence to continue along similar lines.
Scientific objectives
The main scientific objectives of task A1 are:
1) Quantify intraseasonal and interannual variability of cyclone activity and of blocking activity, with particular attention to the Atlantic area and to the interactions between the two phenomena.
2) Develop improved objective diagnostic tools to achieve goal 1) and to be used both on observed and model data.
3) Study the relationships between global (hemispheric) and regional weather regimes and how both relate to teleconnection patterns and subtropical rainfall anomalies, and therefore to more predictable tropical anomalies in general.
Work plan
1) Development of objective code to diagnose cyclone frequency and cyclone tracks (Units 1 and 2).
2) Development of subjective synoptic climatology of the Euro-Atlantic blocking in the period 1946-1995 and comparison with objective index proposed in task 1 (Unit 4).
3) Development of improved blocking objective indices (Units 1 and 2).
4) Study the relationship between storm tracks variability, blocking frequency, NAO and PNA
(Units 1 and 2).
4) Study the global/regional nature of blocking as a weather regime and its relationship with known indicators of atmospheric bimodality (Unit 3).
5) Study the relationship between Sahel rainfall and Euro-Atlantic blocking (Unit 4).
6) Study the relationship between Euro-Atlantic blocking and the development of some extreme weather patterns in the Mediterranean (unit 4)
Task A.2 Numerical Simulations of Climate Variability
Principal scientist: Dr. Antonio Navarra (IMGA-CNR, Bologna)
Background.
Several numerical models are currently in use at IMGA at the global scale. The atmospheric model is a spectral general circulation model resulting from a close collaboration with the Max-Planck Institute in Hamburg (ECHAM-4). The model is currently being used to perform simulation experiments at the decadal scale with prescribed SST distribution provided by the Hadley Center, UK. Main research subjects include (i) the representation of the orography, (ii) the study of tropical teleconnections, (iii) the understanding of the relation between tropical large-scale features ( monsoon precipitation, Sahel rains, and oceanic SST) with the general circulation. The resolution is currently being fixed at T30 ( Triangular 30) with 19 vertical levels.
The ocean model is the non-polar ocean model developed at GFDL (Princeton, USA). The model has 15 vertical levels and a resolution of 1 degree in the zonal direction, whereas the zonal resolution varies smoothly from 1.5 degrees at the poles to 0.3 degrees close to the equator. The enhanced equatorial resolution is necessary to represent the equatorial waves that are needed to simulate the ENSO events.
Both models represent the state of the art in the respective fields of application. The models can be speedily run on the CRAY computer of CINECA, where adequate archiving facilities and in-house modeling expertise are available.
Current coupled models present several deficiency, one of the most remarkable one is the inaccurate description of the seasonal cycle in the Eastern Pacific, that shows up as an insufficient upwelling close to the American coast. It is possible that the orographic representation of the Andes may be partially responsible for the error.
IMGA, UNIBO and CINECA are currently engaged in several EU projects that regards seasonal forecasting and the study of the interannual variability. The activities in this Task will complement and strengthen the participation in such projects.
Scientific Objectives.
The main scientific objectives of task A.2 are:
1. To improve the simulation of the seasonal cycle in the coupled model, especially in the Eastern Pacific.
2. To assess the effect of the mountains representation on the El Nino onset and evolution.
3. To develop a consistent, portable and accurate interaction package between the ocean and the atmospheric model.
4. To analyze the statistics of the anomalies and regimes in long runs of the coupled model
The main strategy will be the analysis of the results of simulations with the coupled model. Several integration at shorter time-scales (less than 2 years) will be performed starting from real initial conditions, but a long simulations ( O(20yr)) will be considered. The short integration will comprise also sensitivity studies to the mountains formulation, with a special emphasis to the Andes.
In the analysis, the European sector and the possibility of teleconnections leading to the area will be carefully examined and the prospects for seasonal forecasts will be assessed. There is wide evidence now that a biennial cycle exist in the monsoon precipitation and recently several studies have pointed out that the monsoon rain impacts the circulation over the Siberian plains, Africa and the Mediterranean region. The physical mechanism for this set of teleconnections is still obscure and further study is keenly required.
Another area that will be investigated, subject to the realization of the long experiment, will be the possibility of a mechanism for decadal variability in the North Pacific. Currently competing theories exist (Gu and Philander; Latif and Barnett) and a coupled model is the only way to try to unravel the puzzle. Both theories involve propagation of anomalies in the ocean, either via a north-south mechanism or a east-west propagation mediated presumably by Rossby waves, that can be traced into the coupled model.
Workplan
Subtasks.
1. Development of the interaction module and development of the coupled general circulation model.
2. Sensitivity experiments with the coupled model to the orographic formulation.
3. Identification of tropical teleconnections and their interaction with the Mediterranean region.
4. Assessment of the interannual variability simulation with a special emphasis to the Mediterranean region and for surface parameters ( precipitation, soil moisture, surface temperature, Tmax Tmin).
Description of team
Institution Personnel Position Man Months
IMGA-CNR, Unit 1 A. Navarra Scientist 2
IMGA-CNR, Unit 1 K. Miyakoda Visiting Professor 3
IMGA-CNR, Unit 1 S. Gualdi Scientist 3
IMGA-CNR, Unit 1 S. Masina Scientist 2
IMGA-CNR, Unit 1 Post-doc Scientist 6
CINECA, Unit 2 F. Molteni Scientist 2
Univ. of Bologna, Unit 3 S. Tibaldi Professor 1
Financial budget *
Total
1997
45
1998
45
Total
90
*All costs are in Millions of Lire
Budget allocated to each team and detailed explanation of costs
Principal scientist: Dr. Franco Molteni (CINECA, Bologna)
Background.
In the international community, research on climate predictability has dramatically increased in the last decade, focusing on two main subjects: a) the predictability of interannual to interdecadal variability arising from coupled oscillations (or regimes) of the ocean-atmosphere system; b) the prediction of climatic changes (possibly) arising from variations in the atmospheric composition or in land-surface properties as a result of human activities.
Subject (a) includes, for example, the predictability of the ENSO (El Nino - Southern Oscillation) phenomenon on the interannual time scale, while on the interdecadal time scale considerable progress has been made on the connections between modes of variability of the global ocean and the decadal variations of rainfall in sub-Saharan Africa. Subject (b) has been recently dominated by research on the effect of the increased concentration of carbon dioxide and other greenhouse gases; the impact of tropical deforestation and other man-induced changes in vegetation is still a fairly active field of investigation.
The important political and economical implications of subject (b) have been much debated; however, research on point (a) also has enormous potential benefits, with the considerable advantage that its results can be objectively tested against observations. In fact, the promising results of predictability studies on ENSO have led the National Center for Environmental Prediction of the USA to develop an experimental climate prediction system based on the coupling of an atmospheric general circulation model (GCM) to a numerical model of the tropical Pacific Ocean. The European Center for Medium-Range Weather Forecasts (ECMWF) has also started a research project aimed at quantifying predictability on the seasonal to interannual time scale through the use of coupled atmosphere-ocean GCMs, while the Hadley Centre of the U.K. Met. Office is currently involved in experimental predictions of rainfall in various tropical regions using a mixture of statistical and dynamical methods.
Scientific Objectives.
The main scientific objectives of task A.3 are:
1. to improve our understanding of the physical and dynamical mechanisms which are responsible for potentially predictable oscillations of the climate system;
2. to quantify the predictability of the atmosphere-ocean systems using stochastic-dynamic methods and dynamical system theory.
To reach both these objectives, two approaches will be followed:
1. to perform and/or analyze ensembles of long-range integrations of atmospheric GCMs either forced by observed sea-surface temperatures (SSTs) or coupled to oceanic GCMs; this will be mainly performed in collaboration with ECMWF and other major European institutions which cooperate in a
European program on seasonal predictions;
2. to develop a hierarchy of simplified models of the atmospheric and oceanic circulation, which can be more easily diagnosed using dynamical system theory; these models can be used to perform very long integrations from which dynamical signal can be detected with good statistical significance.
Workplan
Subtasks.
1. Diagnosis of ensembles of integrations of atmospheric GCMs forced by observed SSTs to quantify the potential predictability arising from oceanic boundary conditions.
2. Multiple integrations of atmospheric and/or coupled GCMs to address specific dynamical problems identified in Subtask 1.
3. Development of simplified quasi-geostrophic and primitive equation models of the atmospheric and oceanic circulation, and of low-order models which reproduce specific dynamical processes.
4. Analysis of the predictability properties of the models developed in Subtask 3 using dynamical system tools.
Description of team
Institution Personnel Position
CINECA (Unit 1) F. Molteni Scientist
IMGA-CNR (Unit 2) A. Navarra Scientist
FISBAT-CNR (Unit 3) A. Trevisan Scientist
Univ. of Bologna (Unit 4) S. Tibaldi Professor
Financial budget *
Total
1997
75
1998
75
Total
150
*All costs are in Millions of Lire
Budget allocated to each team and detailed explanation of costs
Task A.4. Physical processes and idealized studies.
Principal scientist: Dr. Piero Malguzzi, (FISBAT-CNR, Bologna)
Background
The low-frequency variability of the atmospheric general circulation shows multimodal probability density distribution (regimes) in the amplitude indicators of the ultralong, planetary waves; the understanding of the physical processes determining the existence, predictability and transitions from weather regimes is of paramount importance for successful modeling and prediction of climate variability.
One hand, it has recently been shown (Malguzzi et al, 1995 a,b) that models of intermediate complexity of the general circulation (barotropic and 2-level baroclinic models in spherical geometry) exhibit a multiplicity of stationary solutions in near resonant conditions when realistic profiles of zonal wind and bottom orography are imposed. The role played by the (unstable) stationary solutions in determining the probability density function has, however, yet to be determined.
On the other hand, there is evidence that simulation models show multimodal distributions in the amplitude parameters of the planetary scale wave activity (Hansen and Sutera, 1990). In particular, the three-level quasigeostrophic model of Molteni, used in many idealized studies of the general circulation, shows a stable bimodality in the above parameters.
The reduction of low frequency variability to a few important recurrent features has been addressed in terms of EOF and cluster analysis (Kimoto and Ghil, 1993; Trevisan, 1995). Spells of low frequency oscillations have been recently identified by means of singular spectrum analysis (Plaut and Vautard, 1994)
Scientific objectives
Given the previous considerations, we plan to study the structure of the stationary solutions of Molteni's model in order to link its statistical properties to the above theories of multiple equilibria. If necessary, periodic orbits originated from the unstable stationary states will have to be determined
for this task. This will be done by using a pseudo-arclength continuation algorithm already developed and used in dynamical systems having several hundreds of equations.
More generally, we will investigate the possibility of constructing the skeleton of the dynamically relevant components of low frequency variability in terms of stationary and periodic solutions.
Workplan
Subtasks year 1 year 2
Model acquisition X
Statistical properties X
search for stationary sol. X
search for periodic orbits X
Description of team
Institution Personnel Position Man/month
CNR-FISBAT P. Malguzzi Scientist 4
CNR-FISBAT A. Trevisan Scientist 4
CNR-FISBAT Maurizio Fantini Scientist 2
Financial budget *
Total
1997
50
1998
60
Total
110
*All costs are in Millions of Lire
Budget allocated to each team and detailed explanation of costs
Principal scientist: Dr. Rolando Rizzi (Dept. of Physics, Univ. of Bologna)
Background
There is a considerable growth of interest in the development of parametrization of realistic cloud fields inside numerical weather prediction and climate model. The 'validation' of these parametrization is generally obtained by comparison of model products to data measured during field campaigns, and to comparison of mean quantities (like the radiative flux at the top of the atmosphere) to the corresponding measured quantities. The first method allows the testing of the time evolution of several model parameters corresponding to the measurements, but it is very localized both in space and also in time. The second method allows to draw conclusion on the mean, say over a month, behavior of the model, in other words it is a test of some aspects of the model climate.
Measurements from current operational infrared sounding sensors, like HIRS/2, contain a wealth of information on the geographical distribution of clouds and their top temperature and pressure. Future sensors will allow the determination of spectral cloud properties at several levels in the atmosphere. Although the details of cloud geographical distribution can best be achieved using image data, because of its higher horizontal resolution, what is required is the determination of cloud properties at a resolution comparable to the model's, and HIRS/2 resolution, 40 km on average, is perfectly adequate for today's and tomorrow's needs. The use of this type of data for testing is extremely demanding, it is a comparison of values in a specific point and time, it can be extended over any length in time, and is global in nature. It is however limited to a comparison of spectral radiance’s, or to a quantity derived from a set of measured spectral radiances, like, for example, the outgoing longwave flux at the top of the atmosphere.
Validation of clouds parametrization, and of the parametrization of their radiative properties, is only a first step toward assimilating radiances into a NWP assimilation scheme. The direct assimilation of satellite radiance measurements requires in any case the ability to simulate the model-equivalent of the observation with the necessary accuracy.
Although current progress in assimilation in the major NWP centers, notably the ECMWF, are impressive, it is apparent that the assimilation of cloud observation is still premature at the present time. Operational centers are awaiting further advancements in the development of adjoins of the physical processes within their analysis cycle prior to this effort. It is foreseen that NWP centers will be anyhow ready, within the next three years, to experiment with variational assimilation of cloudy radiances.
Scientific objectives
The main scientific objectives are the improvement of the parametrization of cloud radiative properties and the determination of problem areas within the parametrization of dynamical cloud processes. This may be achieved via a comparison of measured TOVS lB radiances and simulated values based on fields, including cloud fields, produced by a model short range forecasts.
A spectral cloud transmittance model (CTM) will be defined, based, as a first step, on Rizzi (1994a and 1994b). The latter has some similarity to the way cloud emittance is treated inside the radiation scheme of the ECMWF. However several improvements are envisaged:
- the introduction of spectral variability;
- the use of diversified cloud types, and therefore radiative parameters, but only as a function of height since the IFS system post processing does not allow to retrieve cloud type as a function of model level;
- initially cloud radiative properties will be treated in terms of equivalent blackbody clouds; but also a model in which each cloud layer is assigned an emissivity, which can handle semi-transparent clouds will be developed. It will be then possible to test the different results obtained with the two models.
The allowance for semi-transparent clouds would make the cloud model suitable for incorporation into fast radiative transfer models developed for microwave sensors.
Workplan
W1. Generation of 'model' fields.
A short-range forecast (two days), possibly using the ECMWF IFS forecasting system, with spectral resolution T213 and 31 levels in the vertical, would be used to produce global 3-dimensional fields of temperature, humidity, water content, ice content and cloud percentage at model levels, and 2- dimensional fields of skin temperature and surface type.
W2. A model for radiative transfer in a cloudy atmosphere.
It is assumed that a Fast Gaseous Transmittance Model (FGTM) will be available at this stage as part of the software being produced by the IASI (Infrared Atmospheric Sounding Interferometer) Sounder Science Working Group (ISSWG). If such model is not available at the time, HARTCODE (Miskolczi et al. 1988, Miskolczi 1994) will be used to produce line-by-line gas transmittances at each model grid point. Of course the computing time (and costs) would be far higher.
The CTM will have no hard-coded level definition so it will easily adapt to FGTM. The model vertical layering is used to compute layer cloud transmittance from model fields and the final cloud transmittance profile is then interpolated to FGTM layering for the subsequent radiance computation.
The treatment of overlapping model clouds will be initially analogous to the one used in Rizzi (1994a). With the development of the semi-transparent cloud model a critical overview will be made also using radiances computed from the two methods against raw HIRS/2 radiances.
W3. Development of an orbit (fov) simulator and of a time/space interpolator of model fields: the resolution issue.
The model radiance computation is performed using the vertical layering of the FGTM and at every grid point of a T213 grid. It is proposed that the radiances at single field of view to be derived from an horizontal interpolation of grid-point radiances.
HIRS/2 data and model data at T213L31 possess different horizontal and vertical resolution. A study will be needed to generate ' HIRS clouds' at HIRS horizontal spatial resolution from the fields generated by IFS (about 90 km at mid latitudes). This requires an independent knowledge of typical statistical properties of real cloud fields below the model resolution. This task may require the use of AVHRR data and it is expected that will be one of the research activities within the ISSWG. If results from this work is not available in time for being used in the work here proposed than the technique already used in Rizzi (1994a, 1994b) will be applied. It consists of filtering measured HIRS/2 data by a bidimensional gaussian filter to produce measured data at the model resolution.
Description of team
Institution Personnel Position Man/month
Univ. Of Bologna R. Rizzi Doctor 6
Univ. Of Bologna to be funded Senior Post Doc 6
Univ. Of Bologna to be funded Computing System Ad. 1
Univ. Of Bologna to be funded Secretarial staff 1
Financial budget *
Total
1997
50
1998
50
Total
100
*All costs are in Millions of Lire
Task A.6 Stratospheric chemistry and dynamics
Principal scientist: Prof. Guido Visconti ( Dept. Of Physics, Univ. of L’Aquila)
Background.
Several models have been developed around the world for simulation and prediction of the stratosphere chemistry. The studies have been concentrated on the improvement of the data base on the circulation and composition of the atmospheric constituents and on modeling studies. The dynamics of the stratosphere has an important role in the modulation of the ozone distribution and on its interannual variability. Such interactions are so intense that they can be described only via a dynamical model ( general circulation model) coupled to a detailed chemistry package describing the interactions between the various chemical species present in the atmosphere. The objective of this task is to construct such a model introducing the chemistry package into the coupled model to be constructed in Task A.2 to simulate realistically the interannual variability of the atmospheric constituents in a realistic way.
Scientific Objectives.
The main scientific objectives of task A.6 are:
1. To improve the simulation of the interannual variability of atmospheric constituents.
2. To assess the effect of El Nino variability on the chemistry.
3. To develop a consistent, portable and accurate interaction package between the chemistry and the dynamics.
The work will be divided in two parts. In the first part the CTM (chemical transport model) will be used off-line on outputs from the long simulation of Task A.2 for calibration and testing and for an assessment of the overall sensitivity of the CTM to the atmospheric variability. The period will be focused on 1960-1994 that will include some of the most relevant event and should give ample opportunities to analyze the development of decadal-scale depletion phenomena.
In the second phase, the CTM will be inserted into the GCM developed in Task A.2 and simulation mode experiments will be performed for the same period.
Validation of the experiments will take place both via a comparison between the off-line results and the interactive results, but also with comparison with observed data. In situ data will provided by the experimental group at Aquila University, but other data will also be used. Particular emphasis will be placed on the Mediterranean region and to the Italian area.
Workplan
Subtasks.
1. Collection of the off-line atmospheric data from the long simulations
2. Calibration of the CTM with the off-line data.
3. Insertion of the CTM in the IMGA GCM and execution of the simulations.
4. Collection of observed data for the Mediterranean area and globally.
5. Final comparison between simulation, off-line analysis and observed data.
Description of team
Institution Personnel Position Man Months
Univ. L’Aquila, Unit 1 G. Visconti Professor 2
Univ. L’Aquila, Unit 1 G. Pitari Researcher 2
Univ. L’Aquila, Unit 1 V. Rizi Researcher 2
Ist. Naz. Geofisica, Unit 2 F. Masci Scientist 2
Ist. Naz. Geofisica, Unit 2 B.Zalesi Scientist 2
Financial budget *
Total
1997
80
1998
80
Total
160
*All costs are in Millions of Lire
Budget allocated to each team and detailed explanation of costs
There is a clear and persistent indication coming from the past decade of oceanographic studies: the Mediterranean behaves as a reduced scale ocean where most of the principal processes affecting the world ocean are present and can be studied with relatively small effort and investments. In addition, the Mediterranean offers the opportunity of directly testing the OGCM (Ocean General Circulation Models) on the seasonal and interannual time scales for an almost closed basin system. All that makes the Mediterranean research important as a world ocean counterpart of climate studies.
Increases in the potential temperature and salinity of the Western Mediterranean Deep Waters on time of decades have been proposed by several authors. These slow trends might be coupled to the recently discovered interannual changes in the deep Eastern Mediterranean Basin, where the Adriatic Deep Waters were found to be replaced by the Aegean waters. The changes in the water column of the Eastern Mediterranean basin (Roether et al., 1995) are correlated with sediment temperature anomalies recorded in the past years. It is interesting to remark that the actual intensity of these changes might also be controlled through the observation of the bottom waters flowing through the Sicily Channel, at the western boundary of the Eastern Mediterranean.
The debate on the decadal and interannual changes of the hydrographic structure of the basin, e.g. a significant part of the thermohaline circulation of the Mediterranean Sea, concentrates now on the possible causes, from global greenhouse warming, to local atmospheric and oceanic conditions, to man induced river input changes. The long term current time series in the Western Mediterranean Sea indicate seasonal and interannual changes in the upper thermocline circulation, such as a general decrease of the kinetic energy of the circulation in the past decade. The proposed explanation in terms of general improvement of the atmospheric conditions in the past decade (due both to a decrease of the heat losses or of the wind stress amplitude), needs to be confirmed and understood in terms of dynamical processes. Specific experimental measurements indicate that the long term evolution of the thermohaline circulation in the Tyrrhenian Sea is such that it is increasing the amount and size of the double diffusion steps in the central part of the sea, and this also is consistent with a general improvement of the climatic conditions over the Mediterranean.
The numerical simulations have established that wind and thermohaline fluxes at the air-sea interface are responsible for the seasonal changes of the circulation. The interannual variability in the atmospheric parameters could be also responsible for the major interannual changes in strength and structure of the surface circulation. Many questions, such as the connection between the seasonal cycle and the interannual variability of the basin, the basic dynamical mechanisms which 'memorize' the atmospheric information in the ocean and the importance of the preconditioning in the circulation for interannual changes have arose and should be examined in the next years. There is a chance that, focusing the attention to the seasonal and interannual variability, we could come up with a complete picture of the general circulation and its evolution. At the regional scale, a high resolution climatological data set nearly available for the Tyrrhenian will allow us to set up a very detailed case study specifically devoted to this basin.
The necessity of making a realistic estimation of the fluxes at the air-sea interface, poses the question of a better definition of the atmospheric parameters at the sea surface. The knowledge of coefficients specifically tailored to the Mediterranean is still rather coarse, and it has prevented, so far, the estimation of the actual energies involved at the sea interface. In order to provide a first direct approach to this aspect, direct marine measurements and analysis will be carried out in critical Sea regions, such as the Ligurian-Provencal Basin and the Adriatic Sea. The other aspect to be considered is the building of numerical modeling activities which will use the knowledge of air-sea interaction physics in order to reproduce the observed thermohaline changes and thus give a quantitative explanation of the mechanisms responsible for the circulation variability.
Finally this subproject proposes to carry out two multidisciplinary observational experiments. The first one wants to follow the climatic change occurred in the past five years in the Eastern Mediterranean basin, involving the intrusion of Aegean deep waters in the abyssal plain of the Ionian basin. This climatic change involves modifications of the temperature and salinity of the deep water column which are stored in the sediment temperatures and in the fluctuations of organic matter fluxes under anoxic conditions. The second experiment is designed to observe the changes in the pelagic ecosystem structure and in particular the thermohaline cell of the Ionian basin and its connections with the Sicily Strait, occurring after the Eastern Mediterranean climaitc event discussed above.
In general the aim of subproject B is to collect and build the data and models capable of defining the mechanisms responsible for the long term changes of the Mediterranean ecosystem, especially its physical component. This task will be accomplished through the extension and analysis of existing long term time series of oceanographic parameters, the better definition of the energy exchange at the air-sea interface through observations and modeling, the analysis of general circulation numerical experiments and the execution of specific multidisciplinary oceanographic surveys in the Eastern Mediterranean basin, where climatic changes on the interannual time scales occurred in the past decade.
