Development of Water Quality Time Series Inputs for Hydrodynamic Modeling Based on Sparse Data

Hydrodynamic modeling is a widely accepted tool used in estuarine research. Hydrodynamic-water quality models have been used to assess ecosystem processes and interactions, simulate scenarios, and evaluate actions for management. However, a key limitation in hydrodynamic-water quality modeling development is sparse (limited) data. A basic step in the development of a model is to set up the water quality time series data at the boundary conditions. There exists a disconnect in the temporal resolution of observed and input model data. Observed water quality parameters are generally collected via grab samples and are often collected weekly or monthly. Water quality time series inputs for a hydrodynamic-water quality model are required hourly or sub-hourly. This study addresses this disconnection in data availability and develops the water quality time series data required as an input for a hydrodynamic-water quality model. Water sampling methods include (i) continuous and (ii) discrete. Hydrodynamic parameters that are often monitored continuously include surface water elevation, water temperature, and specific conductance. On the other hand, water quality parameters such as nutrients and algae are often monitored discretely via grab samples because they generally have to be processed in the laboratory. The time step of discretely collected data is not ideally suited for hydrodynamic-water quality model inputs as these models require time series data. This study uses a two-dimensional hydrodynamic-water quality model, Environmental Fluid Dynamics Code Plus (EFDC+), to simulate estuarine flow, salinity, and water quality within the model domain. The model domain for the study includes three estuarine systems: the Mississippi Sound, Mobile Bay, and Lake Pontchartrain. The main objective of this study is to construct a time series for water quality parameters from sparsely observed data using imputation methods. This assessment will help understand the availability of water quality data and compare different imputation methods. The study results will be used to develop time series inputs for the model. Future work will involve developing a model system to simulate scenarios for understanding the impacts of variable nutrient discharge from freshwater boundary locations on these estuaries.