Using ROMS Lagrangian-Float Simulations to Estimate Exchange
Rates Between Spatial Compartments in a Tidal Estuary
Mark Hadfield
National Institute for Water and Atmospheric Research
New Zealand
My colleagues and I are attempting to build a spatially
aggregated ecosystem model to estimate aquaculture
sustainability in Beatrix Bay, which is a branch of Pelorus
Sound, New Zealand. Pelorus Sound is a drowned river valley
system subject to tidal forcing and occasional large freshwater
inputs. The water is normally stratified and the tidal flow along
the contorted main channel generates internal tides and seiches.
Attempts have been made to estimate exchange rates between the
spatial compartments in the ecosystem model using high-resolution
hydrodynamic simulations with Eulerian tracers. At a given time
each compartment is filled with a tracer, then the tracer field
is examined one or two tidal cycles later to see how much of each
tracer has escaped into the other boxes. From this information it
is straightforward to estimate a matrix of exchange rates. This
approach produces reasonable and self-consistent rates for
horizontal exchange, provided the compartments are chosen
appropriately. For vertical exchange the estimated rates are much
too high. The main problem appears to be that undulations in the
material surfaces due to internal waves are mis-interpreted as
large vertical exchanges.
The ROMS model is being set up for the Pelorus Sound system in
order to carry Lagrangian-float simulations of horizontal and
vertical exchange. Given the extra information available from
Lagrangian trajectories (relative to Eulerian concentration
fields for the corresponding dispersion experiment) it should be
possible to correct for some of the effects that generate
excessively large vertical exchange.