Fisheries forecasting in the Niger inner delta

The hydrological regime of the inner delta of the River Niger, situated in Mali, is subject to strong annual and indeed intra-annual variability. This delta ecosystem has a characteristic feature, a three-phase cycle. The first, a period of flood, starts in July marking the beginning of the cycle; then, after several months of rising water-levels, the flood recedes, between November and January; finally, a period of low water prevails between March and June.

The river’s various fish species are adapted to this cycle of alternating conditions. Feeding, growth and mortality depend on that rhythm. The flood scatters the fish away from the river bed and brings abundant food. It provides refuge areas, environments where reproduction can take place undisturbed. Growth then proceeds until the waters are in recession, a period of high natural mortality. Fishing effort has to follow the rhythm set by the succession of flood and recession. Most campaigns are concentrated in the period of flood retreat which heralds the return of fish into the fluvial zone and their unavoidable movement through the channels fishermen know well – and when their capturability is highest. Activity diminishes and the season ends with the onset of the next flood, when again the fish are dispersed into flooded areas.

Fishing activity is therefore dependent on the hydrological seasons. Two measurable hydrological parameters can express these: rainfall and river discharge. IRD scientists have sought to determine the extent to which these two variables can provide the basis for a model for predicting annual capturable fish stocks. The team focused first on defining which of the indicators was most pertinent, secondly on finding the number of years’ worth of data necessary for obtaining a reliable forecast.

No significant relationship was found between fishing and rainfall.1 This established, the investigation turned to the other parameter, the Niger river discharge. It appeared to correlate well with the volume of catches, which increased in proportion to the flood intensity. Two hydrological stations in Mali have been recording discharge rates since the beginning of the XXth century: Mopti, in the middle zone of the delta, and Koulikoro, upstream of it. Both these stations supply the necessary data, but Koulikoro’s site further upstream is more convenient. The Niger’s bed there is narrower and discharges are greater, rendering small changes easier to detect and record with finer accuracy. The flood peak occurs in September at that point, one month earlier than at Mopti. The Koulikoro station can hence also provide data sooner and potential catch estimates can be calculated earlier.

An important finding is that only two years of data are required for a reliable prediction to be made. An underlying biological factor is that 70% of the fish caught in the inner delta are less than a year old indicating that they arrived with the last flood, or with the previous one at the earliest.

The model in the end is extremely simple and involves just two easily recorded variables: the average discharge between July and September of the year in course and the same parameter between July and December of the previous year. So constructed it can predict the catches from September, 2 months before the start of the fishing season which runs from November to May. It can also bring into relief the immediate impact of any unusually smaller-scale flood. An intensive catch rate in one fishing season reduces the fish to a population composed mainly of juveniles (less than a year old). If a flood is not strong enough to allow renewal, the following season’s stocks will be depleted and the catches poorer.

This new forecasting model has proved to be extremely useful for giving warnings of insufficiency or overabundance of fish resources without the need for complex modelling systems. Its predictions could be published in the newsletter of the Fisheries observatory and distributed regularly throughout the region. With further development, the model could in the future help define –and hence predict- the places over the delta where fish are most abundant. It could then become an important tool in overall fisheries management.

1 Probably owing to exhaustion of the groundwater sources; any rainfall would be taken up in replenishing the water table rather than in increasing water volume in the River Niger.