Impact of sensor location on assimilated hydrodynamic model performance

A data assimilation (DA) framework assessed with an observing system simulation experiment (OSSE) can ensure reliable predictions for a water body, yet its application is very limited in shallow estuaries. In this study, we implemented an ensemble-based DA system to improve the accuracy of a hydrodynamic model of a micro-tidal estuary. Synthetic water level and velocity data were assimilated into the model in both single and dual variable DA forms. To evaluate the sensitivity of DA performance to the location of the hypothetical water level and velocity sensors, an OSSE assessment was used. Results revealed that DA performance is significantly sensitive to location of velocity observations, while relatively insensitive to location of water level observations. Our analysis suggests that the assimilation of velocity data at a location close to the downstream boundary (i.e., water level boundary) can result in a significant improvement in model estimates.