Multi-threaded Congo River channel hydraulics: Field-based characterisation and representation in hydrodynamic models

Hydrodynamic processes that occur along the Congo Middle Reach are a key
determinant of risks pertaining to biogeochemical cycling, ecology, public health,
transportation, and flood risk. Knowledge of channel hydraulics is paramount to
understanding and modelling these hydrodynamic processes, yet such knowledge is
severely lacking here.

The aims of the research presented in this thesis were twofold. The first aim was
to assess the water surface and in-channel hydraulic conditions along the Congo Middle
Reach, and the capacity of satellite observations to determine these conditions. The
second aim was to evaluate methods of channel geometric representation in
hydrodynamic models of the multichannel Congo mainstem. Fieldwork was central to
achieving these aims; the field data having been used to characterise hydraulics, assess
satellite altimetry datasets, model bathymetry, and model fluvial hydraulics and
hydrodynamics.

A key finding of the hydraulic characterisation was a complete absence of river
flow constrictions that cause backwater effects, which partly explains the relatively subtle
nature of inundation here. Assessment of existing satellite profiling altimetry datasets
showed their spatial coverage adequately captures the water surface profile along more
than 1,200 kilometres of the middle reach. However, coverage was insufficient through
the Chenal entrance, where a downstream increase in bed-slope generates a significant
drawdown effect. Satellite altimetry deviated from field measurements by two metres
here, which is half the annual flood wave amplitude. The findings show that these satellite
profiling altimeters cannot be relied on to capture significant water surface slope
variability resulting from gradually varied flow conditions, even on the world’s largest
rivers.

Modelling work showed that the Congo’s multi-threaded channel geometry can
be simplified to an effective single channel in a hydrodynamic model, without introducing
significant error. The resultant root mean square error in water surface elevation was
estimated to be less than 0.25 metres, providing channel friction and shape parameters
are calibrated to observations obtained across the entire flow range. This finding may
apply to other large multi-threaded channel reaches, which are commonly found on the
world’s largest rivers.