Modelling the Zambezi River plume

Abstract

A model for the Zambezi River plume, the largest on the Indian Ocean coast of Africa, is presented and the results of experiments with different discharges and wind forcings are analysed. Although the river plays an important role in the southern African economy through power generation on large dams, artisanal fisheries, and frequent flooding events that impact greatly on local populations, the plume has not been well studied. Observations during the period 2004–2007, when the winds were mainly easterly or south-easterly, indicated that the plume waters can extend both downstream (equatorwards) and upstream (polewards) of the Zambezi Delta with a recirculating bulge near the river mouth. The model is constructed using the Regional Ocean Modeling System (ROMS), with a 40-km long, 3-km wide river discharging into a rectangular coastal ocean with a linearly sloping bottom. When the model is forced only by a constant river discharge of 1 000 m³ s^–1 (typical of observed discharge amounts in summer), the Kelvin and Froude numbers for the resulting plume imply a ‘large-scale’ buoyant discharge with a coastal current that is close to being in geostrophic balance with the across-shore pressure gradient and a recirculating ageostrophic bulge near the mouth. The distributions of the bulge and plume waters are found to be relatively insensitive to the discharge amount. Under constant wind forcing, the plume distribution changes dramatically. Northerly and easterly winds produce the largest changes with the latter able to deflect the plume up to 180° due to Ekman drift. When sea breeze-like winds are imposed, accumulation of water in the bulge occurs with substantial spreading upstream. Stronger sea breezes lead to less downstream spreading of the plume than gentle winds. When the winds are mainly across-shore, Ekman drift dominates, but the dynamics become almost geostrophic when the winds are roughly aligned to the coast. These experiments suggest that the Zambezi River plume is sensitive to the winds on diurnal to synoptic time-scales.

Keywords: dynamics, ROMS modelling, wind forcing