Fish and flooding
Fish trade in the Inner Niger Delta have been registered by l’Opération Pêche de Mopti (OPM), now Direction Nationale de Pêche (DNP), since 1977. OPM also estimates the total annual catch, thus including fish eaten by fishermen and their families and other people in the Inner Niger Delta. This estimate is based on several untested assumptions, such as a fixed daily consumption by the fishermen themselves independent of the annual catch, and an overestimation of the annual increase of the number of fishermen (Zwarts & Diallo 2005). This chapter evaluates the raw data concerning the daily registered fish trade in Mopti, being either dry fish (smoked or dried) or fresh fish. The weight of dried fish is converted into fresh fish equivalents, so they can be taken together as fresh weight.
Most fish are captured, and brought to the market, during receding water when fish are easy to catch being concentrated in the last remaining water, being either creeks, the river bed or isolated lakes and depressions. This chapter describes the relation between the total annual fish trade and the flooding conditions some months before, thus in the previous year.
The bars in the graph below shows that the annual fish trade in Mopti since 1977 has varied between 5000 and 30 000 ton. The blue line in the same graph gives the variation in the maximal flood level such as measured in Akka. Obviously, the fish trade was very low in a series of dry years between 1983 and 1994 and high in wet years (1977-1980, 1995, 1996, 2004, 2009).
Annual fish trade in Mopti (ton fresh weight) since 1977 (data from the annual reports of OPM) shown as bars compared to the maximal flood level at Akka in the previous year (blue line).
The data from the graph above are plotted against each other in the next graph. The annual fish trade is a clear curvilinear function of the peak flood level in the year before. The blue line refers to the first 20 years, 1977-1996. Brown dots and yellow triangles show the fish trade in 1997-2006 and 2007-2016, respectively.
Theoretically, because the number of fishermen increased from 70 000 in 1967 (Gallais 1967) to 225 000 in 1987 (Morand et al. 1991) and to 268 000 in 2003 (Zwarts & Diallo 2005), the trade should have increased at the same rate. However, correcting for flood level, there was no increase in trade at all. When 270 000 fishermen are unable to bring more fish to the market than 70 000 fishermen, this strongly suggests that fish capture is constrained by an absolute ceiling in biological production. With as result for the individual fishermen: where the hogs are many, the wash is poor.
Annual fish trade (data of OPM) as a function of the peak flood level in the previous year, given separately for three periods: 1977-2006 (blue dots), 1997-2006 (brown dots) and 2007-2016 (yellow triangles). The regression line is calculated for 1977-1996. The relationship is the same for later years, although (for unknown reasons) the fit is less pronounced (i.e. the scatter around the drawn regression line is larger than in 1977-1996). Despite the increase of the number of fishermen during those 40 years, there is no increase of the fish trade independent of flood level.
The relation between annual fish trade and peak flood level may be described as an exponential function. Its exponent is 4.13. The flood level measures the flooding in one dimension (vertical), but the flood extent refers to two dimensions (width and length). Hence we may expect that the exponent would be lower when fish trade is related to the maximal flood extent. Indeed, the equation declines to 2.3 (see below).
Relation between annual fish trade (ton) and the maximal flood extent in the previous year; the exponential regression line refers to 1977-1996 (blue dots).
The total inflow of Bani and Niger may be a better measure for the flooding process since it gives the total water volume, and thus combines flood level and flood extent in the Inner Niger Delta. When the fish trade is plotted against the annual inflow during incoming water, the correlation is about the same as in the two graphs above. The function has an exponent of 1.69, thus (as expected) lower than for flood level and flood extent, but still higher than 1.
When the exponent would be 1, the total annual fish trade would be, on average, the same for all years when the fish trade is expressed as fish per km3 inflow. For instance, the graph shows that at a low inflow of 15 km3, the fish trade would be 5000 ton, being equivalent to 333 ton fish per km3 water. Assuming that this would be the same independent of the total inflow, one may expect an annual trade of 15000 ton fish at a high inflow of 45 km3. In fact, the annual trade is about twice as high. How to explain that the fish trade is so high in years with a large inflow?
Relation between annual fish trade (ton) and the total inflow of Niger and Bani during incoming water in the previous year; the exponential regression line refers to 1977-1996 (blue dots).
In years with a high inflow, the flood reaches a higher level and a larger part of the Inner Niger Delta is flooded, but there is still another dimension independent of measures of the amount of water: the flood duration. The higher the flood, the longer the floodplains remain covered by water. In a dry year, most floodplains are dry again in mid-December, but in a wet year, the floodplains remained flooded until early May. The date at which the water level in Akka has declined to below 200 cm is closely related to the peak water level some months before; see graph below.
Date at which the water during receding water is below 200 cm at the gauge of Akka as a function of peak flood level during the same flooding cycle.
The data of the graph above are replotted in another graph showing the number of days during which the floodplains has been covered by water since 1 July. In dry and wet years, the flood duration varies between 170 and 300 days. The relationship between flood duration and maximal flood level would be the same, but at a higher level, if the selection criterium would have been 250 or 300 cm instead of 200 cm. For the same reason, the flood duration would be shorter when 100 or 150 cm would have been taken. However, 200 cm has been selected because at this level most of the floodplains in the central part of the Inner Niger Delta emerge about at this water level.
Same as the previous graph, but now given as the number of days since 1 July.
Flood duration is an extremely important factor because it determines the period during which the fish can grow. Fish more than one year old have become increasingly scarce in the Inner Delta (Laë 1995). The only way for a fish species to survive here is to reproduce as early in life as possible. Indeed, Bénech & Dansoko (1994) found that the fish species in the Inner Niger Delta have adapted to the extreme predation pressure by reducing their age of reproduction. Reproduction for most species is limited to the high water period (Bénech & Dansoko 1994). Therefore, the annual fish stock entirely depends on the spawn and fry produced by the few fish still alive at the end of their first year and by the very few fish over one year old. Hence the fish being captured in recent years are only some months old in a dry year, but the growing season is about twice as long in a wet year.
There is still another reason why so many fish are captured during receding water after a high flood. During flood recession, the water level in the river system of the Inner Niger Delta drops at 2-3 cm per day, but in lakes isolated from the river system only 1 cm evaporates daily. When the water level in Akka has receded from 300 to 100 cm, half of the floodplain get isolated from the river system after a high flood (such as in 1999), but the low flood level in 1984 was insufficient to reach many lakes and depressions that consequently were deprived of their seasonal water boost. Many of these lakes are liable to dry out completely if unreplenished over a number of years. In wet years, there are many depressions and lakes still filled with water (and thus with fish) when the water level in the river system has already declined to a low level. In contrast, the potential fishing grounds in a dry year are seriously limited during receding water because the flood level has not been high enough to temporary fill those depressions and lakes.
The seasonal catch data substantiate the conclusion given above that fish capture is constrained by an absolute ceiling in biological production. On average, the daily catch per fisherman decreases from 35 kg/day in early February to 7 kg/day at the end of June (Kodio et al. 2002). This decline is consistent with depletion of the available fish stocks; at the end of the fishing season, nearly all fish have been removed from the floodplains. When there would be plenty of fish during the entire flooding cycle, not a decline but an increase of the daily catch is to be expected due to the increase of the fish density in the last remaining water bodies.
One may doubt to what degree the relation between fish trade and several measures of flooding, given above for the years since 1977, is due to conditions which have nothing to do with flooding. It might be, for instance, that the fraction of fish sold on the market of Mopti has changed for some reason during the years, for instance, because fish trade beyond Mopti has increased or declined. Another question is whether the fish trade as fraction of the total capture has declined in the last 40 years because of the increase of the number of fishermen and/or had varied with the flood level (with less traded fish in a dry year). Hence to interpret the results given above, one needs to know how the total annual fish capture is related to the flooding of the Inner Niger Delta. This is unknown for the Inner Niger Delta as a whole, but it would be sufficient to know the capture for individual fishermen, not only during the course of the season, but also during different years with a different flood level. These data were collected by Béné et al. (2009) who found that in a dry year (1991/1992) fishermen in the Inner Niger Delta caught a meagre 9.2 kg fish per fishing trip, against 18.9 kg in two years with a moderate flooding (1996/7 and 1997/8) and 37.9 kg in three years with a higher flood (1994/5, 1995/6 and 1998/9). The graphs below plots these three figures against flood level, flood extent and flood volume (inflow).There is a perfect linear fit (R2 > 0.994), but the graphs show exponential relationships, to allow comparison with the graphs above. As the graphs show, the flooding conditions have a very large impact on the amount of fish captured by individual fishermen, even much larger than could be derived from the total annual fish trade.
The main conclusion of these analyses is that even small difference in the flooding of the Inner Niger Delta has a large impact of the amount of fish being captured. Consequently, infrastructures upstream of the Inner Niger Delta taking only a few percent of the river flow during incoming water, has a disproportionate, large negative impact on fish and fishermen in the Inner Niger Delta.
The amount of fish captured, on average, by individual fishermen during their fishing trip (ten per week, on average) as a function of the peak flood level (left), maximal flood extent (middle) and total flood volume as determined by the inflow of Niger and Bani during incoming water (right). Fish capture data from Béné et al. (2009).
Changes in the fisheries in the Inner Niger Delta
The fish population and also the exploitation techniques have changed in the Inner Niger Delta during the last dozen of years. The fish population depends now entirely on the young produced by the fish that managed to survive the previous year’s intensive fishing campaign during receding waster. As a consequence, fish species restricted in their distribution to flooded areas have decreased, whereas species able to reproduce at one year old became more abundant (Laë 1995).
The introduction of nylon nets in the 1960s enabled the near-depletion of fish stocks in the Inner Niger Delta, thus changing the exploitation system significantly (Laë et al. 1994). Concomitant with the steady decline of the size of captured fish, the mesh size of nylon nets decreased: before 1975, most nets had a mesh width of 50 mm, but this declined to 41-50 mm between 1976 and 1983, and to 33-41 mm between 1984 and 1989 (Laë et al. 1994). This downtrend trend has continued: in 2007 many nets had a mesh size of only 10 mm.
There is also a shift where the fish are captured. In the past, nearly all fish were captured in the creeks and in the river bed after the fish have left the drying-up floodplains, or they were taken from depressions when the water has declined enough to make it worthwhile to capture them. As a consequence, fish were captured relatively late during the flooding cycle. This gives a maximal result, because the fish have grown during a longer period, while at the same time they are more easy to catch because they are more concentrated.
However, in recent years, fish are, on average, captured earlier in the season, because more and more fishermen attempt to take them before other fishermen can capture them. For instance, people dig small grooves in the levees along the creeks which allow them to catch fish easily with a small net and do that before fish will arrive in the creeks in a natural situation. However, what they do in fact is draining the floodplain and thus have a large negative impact on the growing conditions of all fish still present on the floodplains.
Fish can be captured still earlier, but then it is necessary to trap them on the floodplains themselves. That is what people do, using km’s of plastic nets (mesh size of 10 mm) to enclose all fish in the still flooded bourgou fields. This allows fishermen to capture fish before other fishermen can do, but the strategy is very harmful, since only a part of the enclosed fish can be captured, whereas all fish captured within these nets cannot escape and thus die.
The overall trend is that during the last dozens of years, fish are captured in an earlier stage during their local migration within the Inner Niger Delta. The history of fish exploitation in the Inner Niger Delta is a classic example of overexploitation, but it is also a typical “tragedy of the Commons”, where an individual takes a (small) benefit at the expense of the community as a whole. Annual catches have become more variable in volume as most fish are now less than one year old and the total fish population is determined by flood level. Meanwhile, the future prospects for fishermen have become precarious – they are with more, use more nets, but catch fewer and smaller fish, whereas at the same time flooding will be less in the future due to more irrigation and dams upstream of the Inner Niger Delta.
Sources and more information
- Béné C., Kodio A., Lemoalle J., Mills D., Morand P., Ovie S., Sinaba F. & Tafida A. 2009. Participatory diagnosis and adaptive management of small-scale fisheries in the Niger River Basin. CPWF project report 92. CGIAR Challenge Program on Water & Food. 83 pp.
- Bénech, V. & D.F. Dansoko. 1994. Reproduction des espèces d'intérêt halieutique. In: J. Quensière (ed.). La Pêche dans le Delta Central du Niger, pp. 213-227. Karthala, Paris.
- Gallais, J. 1967. Le Delta Intérieur du Niger. Etudes de géographie régionale. Paris: Larose.
- Kodio A., Morand P., Dienepo K. & Laë R. 2002. La dynamique de la pêcherie du delta intérieur du Niger revisitée à la lumière des données récentes. Implications en termes de gestion. In Orange D., Arfi R., Kuper M., Morand P. and Poncet Y. (eds.) Gestion Intégrée des Ressources naturelles en Zone Inondable Tropicale. pp. 431-453. Editions IRD Colloques et Séminaires.
- Laë R. 1992. Influence de l’hydrologie sur l’évolution des pêcheries du Delta central du Niger de 1966 à 1989. Aquatic Living Resources 5(2): 115-126.
- Laë R. 1993. Impact des barrages sur les pêcheries artisanales du Delta Central du Niger. Agricultures 1: 256-263
- Laë, R. 1995. Climatic and anthropogenic effects on fish diversity and fish yields in the Central Delta of the Niger River. Aquat. Living Resour. 8: 43-58.
- Laë, R., M. Maïga, J. Raffray & J. Troubat. 1994. Évolution de la pêche. In: J. Quensière (ed.). La Pêche dans le Delta Central du Niger: 143-163. Paris: Karthala.
- Morand, P., J. Quensière & C. Herry. 1991. Enquête pluridisciplinaire auprès des pêcheurs du delta Central du Niger: plan de sondage et estimateurs associés. Le Transfert d’Echelle, Séminfort 4: 195-211.
- Welcomme R.L. 1986. The effects of the Sahelian drought on the fishery of the Central Delta of the Niger river. Aquatic Fisheries Management 17: 147-154.
- Zwarts L. & Dialo M. 2005. Fisheries in the Inner Niger Delta. In: Zwarts, L., van Beukering P., Kone B. and Wymenga E. (eds.). The Niger, a lifeline. Effective water management in the Upper Niger Basin. RIZA, Lelystad / Wetlands International, Sévaré / Institute for Environmental studies (IVM), Amsterdam / A&W ecological consultants, Veenwouden. Mali / the Netherlands: 89-107.