Salt intrusion in the Pungue estuary, Mozambique: effect of sand banks as a natural temporary salt intrusion barrier

Abstract. This paper presents a salt intrusion model for the Pungue estuary with the aim to determine the minimum discharge required to prevent the salt intrusion from reaching the water intake situated 82 km from the estuary mouth. The Pungue river is shared between Zimbabwe and Mozambique and has a large variation in precipitation and runoff. The mean monthly discharge can be as low as 8 m³/s and as high as 893 m³/s. The second largest city of Mozambique, Beira, relies on the Pungue for its water supply. In the dry season it frequently occurs that the water intake has to be ceased because the salinity of the Pungue is too high.

The salt intrusion model used in this paper is based on a fully analytical and predictive theory which is confronted with measurements of salt intrusion and estuary topography. The paper presents the collection of estuary characteristics and the salt water intrusion measurements that were obtained by field measurements in 1993 and 2002. Using these data the salt intrusion model has successfully been applied.

During salinity intrusion measurements in the dry season of 1993 it was observed that sand banks in the middle zone of the estuary prevented the salt water from intruding further upstream, resulting in lower salinity levels upstream than the theoretical salt water intrusion model predicts. This effect occurs during ebb of neap and average tides and can reduce the salt water intrusion by 10 km.

The model indicates that in a natural situation a minimum monthly discharge of 12 m³/s is required to maintain acceptable salinity levels during high water and spring tide near the water intake. The actual water discharge upstream of the water intake has to be higher, since this minimum discharge does not take into account the water abstracted for irrigation and/or urban water supply. Current water abstractions lead to salt water intrusion near the water intake at approximately 10% of the time. The model indicates that an additional water abstraction of 5 m³/s will lead to an increase in salt water reaching the intake at 10% of the time. During neap tide the sand banks act as a temporary natural salt intrusion barrier reducing the chance of salt water reaching the water intake.