ABSTRACT:

Dataset (splitted into 14 files due to file size limitation) for training, validating, and evaluating AI model which was developed for daily hypoxia prediction in the Louisiana-Texas shelf. The dateset is derived from a coupled hydrodynamic-biogeochemical model embedded in the Regional Ocean Modeling System. External dataset from three independent hydrodynamic forecast products: HYCOM, NEMO, and FVCOM.

ABSTRACT:

A 3-dimensional, free surface, topography following numerical model – the Regional Ocean Model System (ROMS, version 3.7) was adapted to the Barataria Estuary. The model used a curvilinear coordinate with a horizontal resolution of ~200 m and eight vertical sigma layers. The model hindcast covered the period of August 1, 2013–January 2, 2018. The south boundary was the only open boundary and was forced by hourly water level, 3-D current velocity components, depth-integrated horizontal current velocity components, 3-D salinity and temperature. The boundary water levels were derived from the tidal station at Grand Isle, LA (073802516 Barataria Pass) maintained by U.S. Geological Survey (USGS) National Water Information System (NWIS). Other boundary conditions were derived from a parent nGoM with a horizontal resolution of 1 km (see Zang’s et al., 2019 for details of model setup and validation). Instantaneous measurements of DPFD discharges were added to the model as point sources on the northern boundary near Lake Cataoutche with a 15-min interval. The DPFD transported horizontal momentums and freshwater to the study domain. The discharge measurements were obtained from site 295501090190400 near Boutte, LA provided by the USGS NWIS. More details for model set-ups can be found in Ou et. al., (2020). Monthly averaged fields are archived in the NetCDF format.