In the Mediterranean, the main demersal fisheries are localized on narrow continental shelves along the coasts. Experiences during the last decades in this area have shown that it was difficult to obtain a global estimate of the demersal resources from fishing activity, especially due to the very large dispersion of the landing places, the important diversity of the species caught and the scarceness of reliable statistics. This situation has induced different European Union States to conduct national programmes for the assessment of these resources from repetitive trawl surveys. A lot of the demersal resources in the Mediterranean are considered as fully or over-exploited. To support the regulation of these fisheries, particularly for the application of the common fishery policy in the Mediterranean, there was a need for standardised information on the status of these resources. In this context, the European Commission has incited to the implementation of a common programme for their assessment by trawl surveys. The need for this work has been confirmed during the last Diplomatic Conference on the fishery management in the Mediterranean (Venice, 1996). To complete this job, different institutes from the four Mediterranean European countries gathered together in 1993 to build the Mediterranean international trawl survey (MEDITS) programme. The characteristics of this programme are presented below.

The general objectives of the programme defined during an ad hoc working group in 1993 are the next : "The motivation for establishing this survey lies in the fact that comprehensive biological studies of the biological status of most of the demersal fish stocks in the Mediterranean are entirely lacking. The Commission wishes to promote such studies and one way of doing so is establish an international survey of the demersal stocks. It is hoped and expected that the collection and analysis of appropriate survey data will allow the Commission to formulate scientifically based proposals for improved conservation of the stocks"(anon., 1993). From this general goal, the present programme has been designed with the following basic aims : (i) to contribute to the characterization of bottom fisheries resources in the Mediterranean in term of population distribution (relative abundance indices) as well as demographic structures (length distributions), (ii) to provide data for modelling the dynamic of the studied species. In this scope, estimation of total mortality of the exploited species constitutes an important aim. The programme had also to take into consideration different observations. A simple analysis of the geography and the bathymetry of the zone shows the very great diversity of the different sub-areas. For example, one can underline differences of hydrological conditions between the waters in the Alboran sea marked by the Atlantic influence and those in the Aegean sea, in direct contact with the Black sea, or the relative monotony of bottoms of the High Adriatic opposed to their very wild aspect in the Aegean sea. Finally, the diversity of the exploited species contributes to the fisheries' wealth in the Mediterranean. If a limited number of species produces an important part of the landings value, the existence of this great species diversity needs however a special attention for the fishery management in the area.

The general orientations of this action have been defined in 1993 by an ad hoc working group managed by the European Commission (Directorate of Fisheries) and opened to scientists from all the Community countries (Anon., 1993). Then the project has been formalized in the Medits programme (Mediterranean International Trawl Survey) which began at the end of the year 1993. At the beginning of the programme, it has been decided to call up as well as possible the competencies available in the different countries concerned by the project. So, from its beginning, the programme is managed by four main partners respectively in each of the four Mediterranean European countries, the Spanish Institute of Oceanography (IEO, Spain), the French Research Institute for the Exploitation of the Sea (IFREMER, France), the Italian Society of Marine Biology (SIBM, Italy) and the National Center of Marine research (NCMR, Greece). Those partners have been chosen for their own competency and their ability to mobilize at the national level the technical and financial means required for the programme. In each of the countries, regional co-ordinations are defined when necessary. A general co-ordination is assumed by one of the four main partners (IFREMER since the beginning of the programme till now). Since 1996, the activity of the programme has been enlarged in the Adriatic through the participation of three newcomer countries: Albania, Croatia and Slovenia. For the time being, about twenty institutes and laboratories from the Northern Mediterranean contribute in the programme. The activity of the group is managed through a Co-ordination Committee and a Steering Committee. The partners organize ad hoc working groups when necessary and meet once a year in a general meeting. For the time being, the programme has no formal link with other international bodies, but some of the partners are mainly involved in different other groups or organizations (GFCM, ICSEM, etc.) and other international research programmes. Those networks are used to favour exchanges of information and collaboration in the area. One of the main challenges of the programme was to define and apply standardized protocols for the whole area, despite the great diversity of situations encountered. For this reason, the co-ordination group defined in detail a common protocol to conduct the surveys in the different areas. Then, one survey has been carried out each year with the same rules since 1994.

Even when they had yet organized their own survey series in some areas and when important improvements were introduced compared with theirs, all the partners accepted to fully adopt the new standardised protocols defined for the MEDITS surveys. These standardized protocols include the sampling gear (feature and handling), the design of the survey, the information collected, the management of the data as far as the common standard analysis of the data. Before the first survey, all the common protocols have been brought together in a "Manual of protocols" (Bertrand et al., 1994) agreed by the Steering Committee and distributed to the participants. This manual has been established from different experiences, and particularly the one of the IBTS Group (anon., 1992). The protocols have been amended when necessary for the following surveys.

The working zone is defined as the totality of the trawlable areas off the coasts of the partners' countries (fig. 1), by depths from 10 to 800 meters. These limits have been adopted to cover at best the distribution areas of the main exploited - or potentially exploitable - species, considering the administrative and technical constraints of the project.

All the hauls are carried out using the same sampling gear. The adopted gear (fig. 2) constitutes a compromise between the different constraints above mentioned. To increase the catch of demersal species, it has a vertical opening slightly superior to the most common professional gears used in the area. The design of the gear has been drawn up by fishery technologists (P.Y. Dremiere, IFREMER-Sete) from specifications defined by the biologists. The gear has been tested from a model in an artificial flume then in full-size at sea, before its production for the first survey. Then, specific studies have been conducted to complete the knowledge about the efficiency of the gear (Fiorentini, 1996; Fiorentini and Dremiere, 1996 and in progress).

The stations are distributed applying a stratified sampling scheme with simple random drawing inside each stratum. The stratification parameter adopted is the depth, with the following bathymetric limits: 10, 50, 100, 200, 500 and 800 meters (fig. 3). Each position has been selected randomly in small sub-areas defined so as to get a compromise between the constraints of statistics based on random sampling and those of geostatistics (Green, 1979; Hilborn and Walters, 1992). The forseen average sampling rate is one station per 60 square nautical miles in all the areas except in the Adriatic where it is laid down to one station per 200 square nautical miles because of the relative monotony of the substratum. The same positions are visited each year. A total of about one thousand hauls are carried out during each annual survey. A example of final sampling rate is given in table 3. The duration of the hauls is fixed to half an hour on depths less than 200 meters and one hour on more important depths.

A list of about thirty common target species (including fish, molluscs and crustaceans) has been adopted. This list of species has been established with reference to their commercial production, their accessibility for a bottom trawl and their potential interest as biological indicator. Observations on these species are the count of individuals, length frequency distribution, sex (including sexual maturity stage) and total weight. The characteristics of each kind of observation are specified in the manual of protocols (Bertrand et al., 1994 & 1996). For all the other species of commercial interest (fish, crustacean and mollusc), the total number and total weight are collected for each haul. During each annual survey, a total of approximately 150 species are identified abord each of the vessels.

The data are put in computer files by the teams in charge of the survey. Three standard exchange formats (in ASCII) including normalized coding are defined. A specific software has been written (Souplet, 1996a) for an automatic checking of the data. This checking is done by each of the partners for its own data before their regrouping. After a second validation in the regrouping place (IFREMER-Sete), duplicates of the total set of data files are deposited on diskettes at the Co-ordinators and EC-DG XIV offices. A specific chart defines the rules for the distribution of the data. Since 1997, an outstanding project is in growing to develop a data base for a full management of the MEDITS data (EC-DG XIV project 96-016). The objective is to constitute a unique data base and common statistical algorithms to serve the demand of the MEDITS programme. The main functions will be the control-validation of the data before storing, the storage of the data, their exploitation using standard statistics and the performance of specific requests. The new common tool will be available to the Medits group by the end of 1998.

At the end of each survey, standard analyses are produced on the data. These analyses include the production of biomass and abundance indices (in kg/km2 and in N°/km2) as well length frequency distribution for each of the reference species and each of the strata. These analyses are made using a classical statistical method approved by the Steering Committee. A specific software has been written (Souplet, 1996b) for the computerization of the calculations. Routines are computerized on Excelâ spreadsheets (Bertrand) for the graphic presentation of the results obtained so.

The objectif is to conduct one survey per year. This yearly survey occurs during the spring and the beginning of summer. Till now, four common Medits surveys have been carried out, in 1994, 1995, 1996 and 1997. The partners are preparing a new project to continue the programme after 1997. A working group is organized after each survey for a common analysis of the results and the estimators produced from the survey series.

Each survey has been carried out aboard eight or nine vessels, according to the year. Each of those vessels works at sea during about one month per year. Research vessels and chartered fishing vessels are used depending on the local possibilities. The organisation of the work at sea mainly depends on the facilities given aboard the vessels. In some cases, the samples are only taken and preserved on board and all the biological analyses are carried out in the laboratories. On the contrary in other situations, particularly aboard the research vessels, the whole analysis of the samples, including the data input in computer files, are conducted aboard. Generaly speaking, it is considered that the Medits survey mobilizes five equivalent scientist-days (at sea or in the laboratories) for one vessel-day at sea.

The results of the surveys are used mainly at two levels by the scientists involved in the programme. At a general level, a global description on the distribution of the reference species is produced. This information is given using abundance and biomass indices as well length distribution of those species by stratum (fig. 4). Another objective is to use this information to estimate demographic parameters like recruitment and mortality. Nevertheless, for the moment, the series (three years available till the preparation of this paper) has been considered too short to introduce such analyses. In the different sub-areas, the MEDITS results are introduced in various analyses. For instance, they have been added in a composite series of trawl surveys, from 1957 to 1995, to study the biodiversity trend of the demersal species in the Gulf of Lions (Aldebert, in press). In some places, they are used to study the structure of demersal fish communities and more generally to analyse the spatial distribution of the species by different spatial methods (Lembo, 1997; Corsi et al., 1997). It is also proposed to use the MEDITS data for preliminary stock assessments by composite production model. Apart from the biological observation, trawl surveys constitute an opportunity to assess the human refuses on the shelves and slopes (Galgani et al., 1996).

From the beginning of the programme, all the basic information related to the programme is presented in annual reports (Bertrand et al., 1994, 1995, 1996, 1997). Generally, these reports comprise two main parts. One part describes in detail the technical organisation of the surveys with a general information on the survey progress aboard each of the vessels, a description of the vessels used, the survey calendar, the lists of the scientists who took part in the survey aboard each vessel, a presentation of the final sampling scheme as well as the meeting reports. The common biological results including formal description, data tables and length distribution graphs are grouped in a second part of the reports. This information is completed with the bringing up to date of the manual of protocols, when necessary. Other results obtained by the different teams from the MEDITS data are presented and diffused through their usual channels. To reinforce the communication and the diffusion of the results obtained from the programme, the Medits partners decided to organize a symposium on "Assessment of demersal resources by direct methods in the Mediterranean and the adjacent seas". This symposium will occur in Pisa (Italy) by the end of March 1998, i.e. allowing to incorporate four annual MEDITS surveys. It is anticipated that this meeting will favour relationships between the scientists involved in different research surveys programmes and in different analytical approaches.

Regarding the reach of the objectives and the prospect of the programme, the following points may be highlighted. The MEDITS programme was the first one in the Mediterranean where a collaboration was developed at such a scale and in such an integrated way for the assessment of the demersal resources. Particularly, with the production of a common data bank, the partners have created the basis for a common working framework. The results obtained so far are very promising. So, at the time when the officials intend to harmonize the fishery regulation in the Mediterranean, the standardization adopted inside the MEDITS programme makes possible from now to introduce a general description of the demersal resources all along the coasts of the four Mediterranean European countries as well as in the Adriatic sea. The first surveys gave very interesting information for the description and the production of general indicators concerning the demersal resources, particularly on their distribution in the different areas. For the time being, the series is yet too short to permit a full analysis of the data obtained, especially for time trend analyses. Nevertheless it is anticipated that the symposium in Pisa will offer an opportunity to progress to this aim. This meeting will also give a chance for a debate on the future of the programme. Naturalists have mainly to take part in the trawl surveys, particularly to ensure the quality of the biological data introduced in the data banks. Statisticians, data analysts, fishery biologists are also strongly required to contribute in the valorization of surveys. One of the challenges for the research survey programmes is to favour a well-balanced collaboration between these different specialists. This is particularly important to give the best conditions for the continuation of the surveys. Inside the MEDITS programme, different attempts have been done to favour this kind of collaboration (working groups, etc.). It is expected that the symposium in Pisa (March 1998) will contribute to progress on that question. Till now, the common data collection inside the MEDITS programme has been focused on biological data. For the future, it is expected to include the collection of basic environmental data (particularly temperature) in the standard observations at each station. The Medits Steering committee is preoccupied with the intercalibration between the different vessels. This question is all the more important because each of the vessels has to work in an independent area. During the last few years, the Italian partners have carried out a national and the international Medits surveys concurrently in same areas. As they have started a peculiar study for the calibration between the different vessels and methods used, the MEDITS group is waiting for the result of this work before further reflection. The EC-DGXIV initiative has been a deciding factor for the establishment of the MEDITS programme. The first results obtained inside the programme can be considered very encouraging. Till now, the programme is supported through a structure of short duration research contracts. This organization does not offer the best conditions to manage lasting data collection. For the future, the need for a stabilized common management structure would be considered.

Co-ordination Committee:
Jacques BERTRAND, general co-ordinator of MEDITS program (IFREMER), Nantes
Arnauld SOUPLE,new director of program (IFREMER), Nantes
Luis GIL DE SOLA (IEO), Fuengirola
Giulio RELINI (SIBM), Genova
Costas PAPACONSTATINOU (NCMR), Atena

Steering Committee:
Gianni ARDIZZONE, Ligurian Sea, North and Central Tyrrhenian Sea (Rome)
Dino LEVI, South Tyrrhenian Sea and Sicilian Channel (Mazara del Vallo)
Angelo CAU, Sardinia (Cagliari)
Giovani MARANO, West Ionian Sea and part of South Adriatic Sea (Bari)
Corrado PICCINETTI, North, Central and part of South of Adriatic Sea (Fano)
Stanko CERVEK, Nort-East Adriatic Sea (Ljubljana)
Stjepan JUKIC-PELADIC, Central-East Adriatic Sea (IOR-Split)
Alexander FLLOKO, Sout-East Adriatic Sea (Drac)
Argyris KALLIONIOTIS, North Aegean Sea (Kavala)
Georges TSERPES, Sout Aegean Sea (Kreta)
Chrissi-Yianna POLITOU, West Aegean Sea and Est Ionian Sea (Atena)

Institute of Oceanography and Fisheries - Participants:
Stjepan JUKIC-PELADIC, program co-ordinator
Nedo Vrgoc
Miro Tonkovic
Vjekoslav Ticina
Svjetlana Krstulovic-Sifner
Vlado Dadic
Miroslav Kozuh