Author(s): Takanori Nagano; Masayasu Irie
Linked Author(s):
Keywords: Hypoxia; Osaka Bay; 4DVar; Variational method; Dual number
Abstract: This study aimed to assess the impact and performance of stratification and biogeochemical model parameters on improving the accuracy of hypoxia spatial distribution simulations in estuaries. Hypoxia poses a significant threat to estuarine aquatic environments. Although density stratification primarily influences the extent of hypoxia in estuaries, accurate modeling of these unsteady and spatially variable density structures using ocean models is challenging. Furthermore, the inherent uncertainty in biogeochemical model parameters demands diverse parameterization procedures. These issues may lead to complementary errors, where the errors in the physical field are compensated by adjustments to the model parameters. Such compensations can result in unrealistic predictions, especially during extrapolation and future forecasting. Therefore, the model parameters and density distributions must be evaluated independently to maintain realism. In this study, the Regional Ocean Modeling System were employed to simulate hydrodynamics and hypoxia. The performance of the model was validated by comparing it with hourly monitoring data from Osaka Bay, Japan, during August 2012. The monitored data on water temperature, salinity, dissolved oxygen, and chlorophyll content were assimilated using a dual-number-based four-dimensional variational data assimilation (DN-4DVar) method. This method was employed to correct the physical field and optimize the model parameters. Additionally, a new parameter estimation term was incorporated into the 4DVar evaluation function to enable simultaneous optimization of the physical field and model parameters. This study focused on the role of these uncertainty factors in accurately simulating the size of hypoxia. In total, four data assimilation experiments were conducted by correcting the initial values of density, biogeochemical state variables, and model parameters. The corrections to the initial density improved the horizontal extent of hypoxia, whereas the optimization of biogeochemical model parameters related to oxygen consumption by remineralization enhanced the thickness of the hypoxic layer. The findings also emphasized the uncertainty in parameters associated with primary production and highlighted the modeling challenges (especially, incompleteness) related to the surface mixing layer during runoff. This research deepens our understanding of hypoxic dynamics in the bay and poses broader implications for ecosystem management and climate change research.
Year: 2024