© Photo : J.P. Venot, IRD

This area is characterized by a network of man-made earth drainage and irrigation channels, natural streams and low lying wetlands and by a diversified and intensive agricultural landscape and a diversity of capture fishery practices. Infrastructure development projects are currently implemented though the impact of rehabilitating and building new water control infrastructures on this socioecosystem are unknown given current lack of knowledge on hydrology, agricultural systems and modalities of natural resources management.

This is an interdisciplinary research project drawing from disciplines as diverse as hydrology, hydrogeology, agricultural economics, geography, anthropology of development, and Science and Technology Studies. The project combines remote sensing analysis, modeling, qualitative interviews and participatory approaches. It is implemented in parallel to, and close connection with, a development project financed by AFD that involves rehabilitating selected preks for agriculture intensification. The project, hence, has a dual aim: (1) generating new knowledge on the socio-hydrological dynamics of a relatively little researched and complex area; (2) support the implementation of a development project by generating knowledge that can be used “along the way” so that infrastructure rehabilitation happens in a socially and environmentally just way. 

This project contributes to a “structural action” of the COSTEA program entitled “planning and managing floodplains in a global change context” and in parallel to two other similar projects conducted in Morocco and Ecuador. The COSTEA structural action’s objective is to contribute to renewing current approaches to floodplain development that center on the construction of grey infrastructures and the idea of controlling water. This will be achieved by focusing and better accounting for the multiples uses and services that floodplain support and provide, and that can form part of more endogenous answers to the multiple pressures floodplains face.