Version classiqueVersion mobile
OpenEdition Books

Avenir du fleuve Niger

Jérome Marie
Pierre Morand
Hamady N’Djim

Synopsis and Recommendations

Proposals, conclusions and recommendations

Texte intégral


Simulation models

1Scientific knowledge needs to be more integrated and more available for decision-makers. One of the options would involve using models to generate simulations to support strategic decisions. We believe that three such models will prove particularly useful. They follow:

  • a hydraulic model simulating the river’s flow and, in particular, allowing researchers to assess the impact of schemes currently on the plans. There is one already, but it does not cover the whole river and has never been used since it was developed in 1988. NBA is sponsoring plans to upgrade this model and stretch it to encompass the whole river (Guinea and Nigeria included). We hope this plan reaches implementation – and that it reaches the relevant agencies thereafter;
  • a water-use model, to list and integrate every river-water usage, should also prove useful. This model should generate scenarios to optimise development strategies. It should interface with the hydraulic model and use the information in it. It should also include hypotheses covering major climatic and demographic forcing factors to generate long-term scenarios. There are two of these models on the plans or underway: an NBA-hosted project for the entire basin, and an IRD-backed plan in Mali involving local and international partner scientific organisations (Kuper et al., 2003). Connecting these two projects seems highly desirable;
  • local models tailored around specific issues, and interfacing with the models above to support decisions, should be included in this scheme. Examples include projects underway at the Office du Niger and desirable projects such as an environmental information system for the Niger River’s inland delta or an information system covering flood hazards in urban areas (Bamako).

An information system to bolster monitoring capabilities and support decisions

2A comprehensive and sustainable approach to administrating the environment and renewable resources necessarily involves acquiring, harnessing, storing and circulating generous amounts of high-quality data, information and knowledge. So it involves setting up and maintaining more than one environmental information system. An information system is a structured set of data, processing procedures and communication protocols. It involves a series of technical requirements (the IT system, strictly speaking). But it also involves a series of documents and reprographic and communication equipment to generate, store, process, transfer and use the information to meet defined goals.

Tallying goals and needs

3There are a number of goals that will plausibly meet across-the-broad consent – and are necessary. They follow:

  • strengthen/develop environmental-information collection and storage capabilities;
  • increase the availability of processed information that can be directly used in the fields in which it is lacking (i.e., in particular, using more efficient processing systems);
  • facilitate public-sector access to environmental information (public-sector, that is, more or less broadly speaking, encompassing national decision-makers, research organisations, and other State-run agencies producing and administrating information). Eventually, allow other stakeholders and audiences to access this information.

4Some decisions on this issue are still pending. In particular, what river-related information should be accessible at which levels in the territorial-community hierarchy is an open-ended question. This also applies outside governmental circles (meaning NGOs, associations and unions), in business circles (entrepreneurs and merchants) and in operations (fishers, farmers, etc.). Efforts to define the information products generated by the information systems will have to take into account the fact that the list of target users might grow.

5The final decision on the information system’s architecture and technology will have to wait until these questions are answered. At this point in time, the architecture we suggest below is based on the assumption that the information system or systems will chiefly target high-level public-sector decision-makers. The suggestions also involve drawing a line between information that is relevant in national and international spheres (decisions on river improvement and management, for example) and information that is relevant in regional and local spheres (office management, for instance) which involve decisions over a much more contained area.

6Requirements also hinge on the time factor (i.e. on whether decisions enlightened by the information system require longer or shorter forecasts). From this angle, we can talk about short-term or tactical timeframes (organising annual campaigns) and more long-term or strategic timeframes (river improvement and investment). Long-term forecasts, in turn, can cover large national or international areas or smaller regional or local ones. It seems clear that the information system that the Niger River needs should address the first as much as the second, even if doing so involves packaging information in different ways.

Building on the considerable resources that already exist

7Broadly speaking, the Niger River already counts a considerable number of environmental-information systems. A number of organisations already gather a wide variety of information, and package it and circulate in a variety of formats, usually on a regular basis (daily, monthly or annually). The “information actors”, however, vary considerably. Their institutional status and level of investment in information varies. They use the information in different ways and their grasp of information technology varies.


8Agrhymet, a Niamey-based CILSS offshoot specialising in agricultural weather monitoring and forecasts, is worth mentioning. It has robust satellite-information processing capabilities and can generate situation maps within short timeframes.

9NBA (which is also based in Niamey) has an information system in the form of a resource and information centre. It stores records and survey documents, and houses an inter-State hydrological-forecast centre (the Hydroniger programme) which gathers, stores and processes hydrological data from the different countries along the basin.

National and local

10The agencies working under national departments (DNH for hydraulics and DNM for the weather, for example) are the core country-level “information actors”. But there are also local development offices (Office de développement rural de Sélingué, Office du riz Mopti, Office du Niger and Opération pêche Mopti), laboratories or research teams such as IER (Institut d’économie rurale/Institute of rural economics) and permanent programmes such as SAP (Système d’alerte précoce/advance warning system). All these organisations gather, store and publish information on a regular basis. Some of them (Office du Niger, for example) are building proper modern information systems.

The different types and levels of information that have to be used and processed

11It is important to pinpoint the different information categories and information levels that an information system needs if it is to produce information that will be relevant to administrating the Niger River’s resources. These categories will shape some of the functional architecture. The information categories follow.

  • Static facts gathered “once and for all” (topography, for example). Here, gathering information is not an issue in itself. But very accurate information is vital.
  • Information that can be gathered once and used for five to ten years (population, land use and cadastres). This information is best gathered in a decentralised scheme and does not even have to be gathered simultaneously (countries or regions in different parts of the basin can gather and pool this information at different points in time).
  • Yearly information, typically covering farming, fishing and other campaign results. This information can also be gathered through decentralised channels. At least minimally standardising data presentation and format, however, is vital.
  • Very dynamic facts that have to be collected on an ongoing basis (rainfall, water levels, flow rates, etc.). This information can only be gathered efficiently using sophisticated technology (sensors, transmitters, satellite-image receptors, and databases fed in real time). These systems have to be set up and operated centrally, and require considerable resources and highly-qualified staff.

A two-tier architecture

12Choosing an information system’s architecture is tantamount to choosing the general geometry that will shape the organisation of the various components. The goal is to interface components in such a way as to achieve overall efficiency. Taking into account, on the one hand, the existing systems, and, on the other, target data-collecting and information-circulating efficiency levels, a realistic architecture could involve two integration levels. These levels follow.

  • The first, single and relatively central level would be an information system covering the basin as a whole, focusing on hydraulic variables (flow rates, swell levels and flood spread). This level would use the most efficient resources to process information from numerous sensors and transmitters, and from satellites. It would interface with a hydraulic model generating short-term forecasts as well as simulations.
  • The second level of integration would involve regional or local information systems. This level would monitor environmental variables, support environmental-management decisions, support production systems, and address risks in clearly-identified, high-stake areas (the Office du Niger, Bamako-Koulikoro stretch and Niger River inland delta, for example). This is where the hydraulic variables (flow rates, water levels and flooded areas) recorded or forecast mainly by the first-level information system would interface with the information about usage systems. As these second-level information systems would serve different goals (they may be designed to provide advance warning, forecast floods, administrate land and monitor environmental impacts, for example), their architecture could vary.

13How these two tiers interface warrants special attention, especially as regards using shared repositories (rating systems and geographic systems).

14Technically speaking, the first tier’s level- and flow-monitoring systems and its mathematical modelling capabilities tally fairly well with the system that NBA is trying to build through its various projects.

15Second-tier regional and local information systems are still embryonic at this point (with the exception of the Office du Niger’s system, which is already at an advanced stage). To nurture information systems at this level, decentralised State agencies working in the areas under review should be encouraged to work closer together collecting, processing and interpreting information. Doing so could involve streamlining tools, methods and repositories and go as far as pooling information and resources. Shared information products could then be developed. To back these integration efforts, information-system units (local observatories) could be created and put under local development or management organisations when these exist.

16These regional or local information systems should also belong to peer federations or “clubs” across the country and, if possible, across the basin. They could meet regularly and be backed in their efforts to streamline information collection and restitution, and to bolster their ability to interface their information with first-tier information.

Figure 12. A example showing cooperation between information systems at different levels
In this case, local (or national) information systems and/or information systems dealing with more restricted areas can operate practically autonomously, i.e. run their own acquisition and restitution loop (6) through their own databases (7) and generate their own presentation formats (1) tailored to specific target needs. But these systems also feed some information (5) into the first-level information system (2) which, in turn, is also fed (3) by specific globally-relevant sensors. The first-level information system, here, has its own restitution formats (again, tailored to specific targets) and feeds some of the information it produces (4) back to specific local and regional information systems, completing the information that the latter generates (1).

The two conditions

17Two issues come up every time an organisation decides to develop an information system. First, the different parties have to agree on the rules for circulating and using information. Second, the organisation needs to make sure it will be able to count on sufficient human and non-human resources over the long term.

Information-sharing rules

18Sharing data, information and models is not the most natural thing in the world. That is why, early on in the process, parties need to decide how they will provide access to, circulate and use raw data and packaged information in the system. The first question is about the sort of information that parties will pool, and when they will do so. They can sign a data policy agreement governing information circulation and usage (and outside support could hinge on them doing so).

Long-term resources

19It is not so much the total amount of funding that a country’s authorities and outside organisations are willing to provide that is important: it is how long they are willing to back the project that will make or break efforts to improve production of and access to information about the Niger River and its environment. The usual project-based approach is an issue: in this case, the usual timeframes are not long enough (projects only last 3 or 4 years in general).

20Self-generated funding (for example charging outside users for access to information) is another option. International institutions also have to understand that monitoring the world’s large tropical ecosystems is a global concern, and that the burden of doing so can not be on southern-hemisphere countries alone. So they could sign a contract agreeing to buy some of the information that this system generates, on condition that they can monitor the quality of that information on a regular basis.

21To conclude, it is not the amount of resources as much as the way in which these resources are obtained and used that will no doubt prove decisive and ensure that the environmental information needed to monitor and administrate the Niger River in a suitable manner becomes increasingly available.


22The Niger River basin’s water is used for a variety of purposes. It is used to generate electricity, for irrigation, for navigation, to feed towns and industrial operations, for fishing, for cattle farming, for flood-fed farming, for receding-water farming, and for ecosystems.

23These uses have been clearly identified. But some are better known than others. Some, in particular, are not very well known and poorly quantified (or not quantified at all). Requirements and demands – which we have tried to review in detail in this paper – vary and indeed sometimes conflict. We will not review every one of these areas of friction, but the main ones follow. Generating electricity involves using water all year round (to provide consumers on a regular basis). These requirements, however, peak during the hot and dry season, when the river’s flow withers to its lowest. So generating electricity involves storing water in the high-water stage and using it in the low-water stage. Irrigation schemes need water all year round as well, if the goal is to harvest twice. But these schemes are especially exposed in the dry season when the Office du Niger, for example, diverts almost 80% of the available flow. Cities need a very steady supply of high-quality water. Navigation involves keeping water levels at guaranteed levels for as long as possible. Conversely, ecosystems (inland-delta ecosystems in particular) as well as fishing, cattle-farming and high-water and receding-flood rice farming operations need healthy swells; their requirements do not hinge as much on flow rates as much as they hinge on water levels (which determine the size of the flooded areas) and on how long flooding lasts. Also, if the inland delta’s ecosystems and economic endeavours (fishing, cattle farming, etc.) need healthy, regular annual swells, they also need sharp recession and low-water stages. The ecosystems derive their health and wealth from the river’s ample seasonal variations (Gepis, 2000).

24In other words, some systems need regular variations and others need wide variations. So it will be difficult to accommodate every river user’s needs. Added to that, users need security. And upstream usage impacts downstream users in a number of different ways. This is true on a national scale, but even more so in the international arena enveloping Mali’s water-management policy. This river’s flow rates in Niger or entering Nigeria depend on the flow rates in Guinea and Mali and, therefore on existing and planned infrastructure.

25Finding a suitable compromise is all the harder in the present context. Outside forces are exercising constraining on the Niger River and on its environment. First of all, there are climatic forcing factors and the fact that the trends shaping them are greatly unfathomable. In particular, the river’s flow will hinge on rainfall (which is difficult to forecast). But it will also depend on the basin’s soil (deforestation, erosion and other developments might alter the link between rainfall and flow). Besides being difficult to characterise, soil conditions also hinge on rainfall and land-use patterns. Which, in turn, depend on the fast-growing population’s needs and on that population’s desire to enjoy a better lifestyle.

26How to generate more electricity or grow more rice without burdening fish production (fish is the Malian population’s main source of protein) and without jeopardising the ecosystems that directly keep one million inhabitants and which are indispensables to a fauna or worldwide importance is one open-ended question. How to build river infrastructure in Guinea or Mali without penalising countries downstream is another. These are two of the challenges facing authorities administrating water in the Niger River. Addressing those questions will involve bold and creative choices. And the choices will have to lean on specific and valid information.

27This paper shows that there is a considerable body of information about the river, but that this information is neither complete nor homogeneous. The river’s hydrology is well known. How its ecosystems work is not. The only comprehensive survey of inland-delta vegetation, for example, is more than 20 years old. On another front, the notion of “environmental flows” is a recent concept in IWRM and warrants deeper analysis, bearing in mind that, in the case of the inland delta, it would have to focus more on “environmental water levels” if it is to help the ecosystem enjoy a healthy, stable future.


28The conclusions above lead us to make the following recommendations on three key areas.

First recommendation: consolidate scientific knowledge

29Efforts to acquire scientific information about the river basin have to continue. In particular, this involves clearing and updating information about the following:

  • developments in the state of the basin surface, as changes could alter the link between rainfall and flow rates;
  • the Niger River valley’s ecosystems, their structure and the processes underlying them. These ecosystems are evolving constantly, in particular due to pressure from usage. Existing surveys, in some cases, are incomplete or obsolete;
  • production systems, which are evolving fast, in step with the ecosystems and Mali’s society as a whole. There is considerable information about some of these systems (fishing, irrigated rice farming, etc.), but that information has to continue to evolve. In other cases, currently-available information is inaccurate or obsolete (information about livestock farming in the inland delta or rain-fed rice farming, for example);
  • water uses across the basin will have to be listed, analysed and, especially, better quantified (including details on their seasonal variations). There is a fair amount of information about how much water major users such as the Office du Niger divert. But there is a lot less information about how much the smaller users (village irrigation areas) peppered across the valley use. Taken in isolation, these small diversions do not seem to have a noticeable effect on the river’s flow. But their compounded effect can no longer be neglected;
  • lastly, efforts to research and develop information systems and models, and circulate those systems’ and models’ findings to management organisations, need to move ahead. Information systems should simulate scenarios to support decision-making processes. There should be three distinct models: a hydraulic model covering the entire river (to run current plans though impact assessments) a water-usage model (representing and integrating all requirements and associated impacts), and local models interfacing with the above and tailored around specific issues.

Second recommendation: push ahead with developments in the institutional framework

30As most of its neighbours, Mali has embarked on far-reaching institutional reform. On the one hand, decentralisation has reshaped the country’s territorial, administrative and political organisation. On the other, IWRM has also changed its approach to water governance. Mali has adapted its institutions, but the changes are arguably still underway or not yet entirely effective. We believe two points warrant special attention:

  • clarifying the relationships between organisations in charge of administrating water in national and international spheres. On this front, we can hope that reviving NBA will revive cooperation between national organisations (which must not be divested of their roles) and international cooperation (which needs to assume a clear federating and arbitrating role as required);
  • we also believe that society as a whole (through elected or association representatives) should have more of a say in decisions about how to use water and in the strategic choices that will shape the country’s and it’s people’s future for years to come.

Third recommendation: a contribution to balanced strategic choices

31The Niger River’s upstream and middle reaches are still relatively undeveloped. Existing infrastructure brings undeniable benefits. But it has also cast light on real issues (even if the consequences are still limited). The river’s situation is not yet compromised, and surveys show ample room to manoeuvre and open up a wide variety of strategic options for the future.

32To protect the river’s major balance, it is no longer possible to consider sector-specific plans independently. Doing so could be tantamount to taking the risk of achieving limited benefits in local areas while compounding disadvantages across the entire basin.

33Developing water use and river infrastructure of course seem necessary. But this can only be done in a global and integrated manner, weighing the expected benefits and the possible interaction with other uses and the river’s environment as a whole.

  • 1 One possible option would be a downstream dam (Taoussa maybe) to compensate the negative impact of (...)

34This is obvious in the case of river infrastructure on a regional level. So, if we consider the large dams that are currently on the plans (and often have been for many years), we can only understand and gauge their impact properly by studying them together, taking into account all the interactions that might be generated by the different management decisions made when running them. In particular, the effects on the inland delta warrant serious analysis. As indeed does the possibility of guaranteeing an “environmental water level”, especially during the high-water stage1.

35Such an approach is obviously fairly difficult and complex to implement. It can therefore only be considered in an international context. It could contribute to reviving NBA and dispel the reservations that some funding agencies have voiced since World commission on dams report (WCD, 2000).

36On a national scale, the big options need to be fine-tuned. All observers agree on the need to push ahead with efforts to develop power generation and farming production. However, giving rice priority, for example, involves providing for the fact that producing 1 kg of rice on an irrigated paddy will require 5 m2 of water (in the best of cases). This options seems costly. Rice, however, is only one of the options: other options use water more efficiently when it is scarce. Moreover, within the rice sector, if the extension of the Office du Niger (a major water user) seems desirable from certain stand-points, it is not the only option. Further developing other rice-farming systems – from rain-fed rice in the south of the country to flood-fed rice in the inland delta – is also highly desirable.

37Strategic decisions will shape the Niger River’s long-term future. Often irreversibly. That is why every possible scenario has to be considered and studied beforehand. This makes developing and installing the simulation models discussed above all the more urgent: these models will be able to help decision-makers choose the development options that provide the best possible compromise, with clear and full knowledge of the facts.


1 One possible option would be a downstream dam (Taoussa maybe) to compensate the negative impact of upstream ones, provided it does not, in turn, penalise users further downstream (in Niger and Nigeria, especially).

Table des illustrations

Légende Figure 12. A example showing cooperation between information systems at different levelsIn this case, local (or national) information systems and/or information systems dealing with more restricted areas can operate practically autonomously, i.e. run their own acquisition and restitution loop (6) through their own databases (7) and generate their own presentation formats (1) tailored to specific target needs. But these systems also feed some information (5) into the first-level information system (2) which, in turn, is also fed (3) by specific globally-relevant sensors. The first-level information system, here, has its own restitution formats (again, tailored to specific targets) and feeds some of the information it produces (4) back to specific local and regional information systems, completing the information that the latter generates (1).
Fichier image/jpeg, 95k

© IRD Éditions, 2007

Conditions d’utilisation :