For modelling uses, we gathered all the rainfall records available at ORSTOM (today IRD), the CRU ("Climate Research Unit "University" East Anglia") and other sources and put them in a form which enables us to establish, for West and Central Africa countries, monthly rainfall grids of 0.5° * 0.5°, that is more or less 2750 km˛.

Window and period of validation of the rainfall grids

According to various elements, it seemed to us judicious to build grids in a "spaces" window limited to the terrestrial part between the latitudes 2°S and 18°N and longitudes 18°W and 17°E and in a "time" window which begins at the decade 1940 and ends at the decade 1990 included. However, in order to avoid effects edge during the space interpolation, we used rainfall stations which exceed these limits and which cover 21 countries.

Data and rainfall stations selected for the grid construction

The data come mainly from 2 sources: the IRD and the CRU.

- IRD: the existing data, with a daily or monthly time step, constitute a little more than 89% of the data used.
- CRU: the data of the CRU constitute nearly 8.5% of the data used.
- For the remainder, i.e. a little more than 2% of the data base SIEREM, they come from data edited in various bulletins or specialized works.

The stations selected are:

- primary stations (synoptic): these are the basic stations whose observation is permanent and on which the statistical study makes it possible to define the climatic characteristics of a given area.
- secondary stations (climatological and agro-meteorological): these stations are removable. They are observed a few number of years in order to be significantly correlated the primary stations.
- These stations are supplemented by tertiary stations: they have an exclusively practical vocation and are specific to a study. They do not have any reason to be perennial if their goal does not require it.

In 1977 then in 1984, the "Comité Interafricain d' Etudes Hydraulique", funding by the French Overseas Affairs Ministry, entrusted to ORSTOM and ASECNA, for all State Members of the Committee, the publication of the daily precipitations from the origin of the stations to 1980 for the synoptic stations of the national meteorological network, for the climatological stations and the rainfall stations, approved by the national meteorological services. ORSTOM carried out work of keyboarding comparison - original, "initial" criticism of the data and of data - processing preparation of the publication. "initial" criticism was based on a visual and automatic criticism. Visual criticism was a subjective method because based on comparisons inter-stations and the knowledge of the person in charged of this work. An automatic criticism was based on tests highlighted during the visual criticism and was thus objective. It was based on the number of days of rain per year (> 0.1 mm), on the ratio of the number of days of small rains (<10 mm) on year number of days of rain. It was also based on a research of systematic reading errors.

Beyond 1980, such a work was not continued; however, punctually, for research programs, checks and corrections could be made. But they are certainly not same level of rigour that those done before.