Assimilating Multi-source Data in a Highly Variable Coastal Water

To improve the forecasting performance of numerical simulation of ocean circulation and dynamics in the wind-tide-freshwater forced coastal waters around Hong Kong, a multivariable data assimilation (DA) system using the ensemble optimal interpolation (EnOI) method has been developed and implemented to assimilate hydrographic data from in situ ship-board mapping, time series buoy mooring and remote sensing sea surface temperature (SST) data into a high-resolution estuary-shelf circulation model around Hong Kong. We found that the selection of the assimilation time window and the number of the observation samples are two key factors to improve assimilation in the unique dynamic estuary-shelf system. The DA with a varied assimilation time window following the intra-tidal variation reduces the errors caused by the flood-ebb tidal variability, resulting in a remarkable improvement of model performance. Statistically, the overall root-mean-square errors between the DA for temperature and salinity have been reduced by 31.81% and 33.42% in the experiment period, respectively. Moreover, when the wind forcing is relatively weak, a 3-day observation window provides a favourable outcome, leading to a better improvement in the simulation. By assimilating higher resolution remote sensing SST data, this EnOI method improves surface temperature simulation, while by assimilating near real-time buoy mooring data, it provides a continuous correction to the model bias both around the buoy location and beyond. Assimilation with combined data from ship-board mapping, time series mooring and remote sensing SST provides an overall improvement of the three-dimensional model solution. Unlike that in the open ocean, assimilations with the spatiotemporally well covered data are essential in the model forecasting in the dynamically active estuary-shelf circulation system under multi-forcing of winds, tides and river discharge.