Two kinds of forecasts are important for flood management in Mozambique: the long-term and the short-term forecasts.The long-term forecasts cover the whole wet season, from October to March. Usually INAM prepares these forecasts in co-ordination with the corresponding institutions from the neighboring SADC countries (SARCOF вЂ“ Southern Africa Climate Forum).
A first forecast is issued in September (the next will take place in Malawi at 17-19 September) and is later updated in November and in January. This seasonal forecast is quite important. On one hand, it indicates the possibility or not of a very wet season and therefore is the first warning in terms of flood alert, which is particularly important for the INGC to mobilize resources and prepare for a possible flood. On the other hand, it also allows the managers of large dams to refine their operation for that particular season, eventually increasing the discharges to create additional flood storage capacity. Besides, other institutions can look more closely into the state of their infrastructures that can be affected by a flood (dykes, irrigation areas, roads and bridges).
The short-terms forecasts of extreme rainfall events (cyclones, tropical depressions, and cold fronts) are even more important as these events are the origin of any large flood. However, it is important that these forecasts consider not only the territory of Mozambique but also what will happen in the upstream countries as most of the floodwaters come through the Mozambican borders. Therefore, close links with the meteorological institutions of those countries and well-established channels of communications and procedures during floods are essential.
Besides the forecasts, it is also fundamental to collect as quickly as possible data on the rainfall that is occurring through traditional rain gauge, meteorological radars (fig. 3 вЂ“ South Africa System; 4 вЂ“ Proposal for Mozambique and 5 вЂ“ Integrated System), as these data form the basis for any reliable analysis of the incoming flood and its propagation along the river.
All these data must be assisted to follow hydrological events integrating telemetric