Modelling the impact of dams on flood-recession agriculture in the Omo valley: a satellite-data based analysis

Conflicting stakeholder interests in water systems such as power generation, agriculture and local livelihoods have required the development of an integrated approach to water resources management. An important livelihood for many African rural communities is flood-recession agriculture. Especially in monsoonal climates, river adjacent sites that are inundated by seasonal flood pulses provide humid and fertile soils of high value for small-scale agriculture. Alterations to natural flood regimes due to the construction of water infrastructures (e.g. dams) threaten this practice by reducing flooding of riparian areas. Artificial flood releases from reservoirs have the potential to counter such alteration, but in order to maximize their effectiveness many aspects are yet to be studied. In particular, in a context where resources are shared among multiple stakeholders, little research has been done on how to ensure sufficient flood magnitude to protect communities from the risk of crop failure.

As part of the national hydropower development strategy, Gibe III dam is in operation on the Omo river in southern Ethiopia since 2015, and local populations practicing flood-recession agriculture in the downstream Omo valley have been exposed to reduced or absent seasonal floods. The development of a large, state-owned irrigation district along the river course further reduced water availability in the region of its delta, where flood-recession agriculture was practiced the most. For artificial floods from Gibe III dam to be effective, we developed an indicator to assess water needs for flood-recession agriculture and to include them in reservoir policy optimizations. Lack of ground data and remoteness of the area were the main challenges of this work, preventing direct data acquisition and extensive stakeholder participation. We used high-resolution satellite imagery taken annually to quantify the yearly extent of flood-recession agriculture in the region and linked it to estimated past streamflow magnitudes simulated by means of a distributed hydrological model.

We observed a strong correlation between historical extents of flood-recession agriculture fields in the study area and river streamflow, allowing to build an indicator for livelihood flood requirements that was included in the evaluation of alternative development pathways. We used the designed indicator to assess the impact of alternative management strategies with varying sectoral trade-offs, combined with multiple system configurations representing present and planned infrastructural development of the region. Preliminary results show that appropriately designed development pathways can substantially limit damages to flood-recession agriculture practices. This indicator will contribute to planning effective artificial flood releases and to capturing rural communities’ agriculture needs.