2:30 PM - 2:45 PM
[AOS18-10] Analysis of flow in Osaka Bay by using FVCOM ~Accuracy verification of tidal current reproducibility~
Keywords:FVCOM, tidal current, tide levels, HF radar
Material transport over periods longer than the tidal cycle is driven by residual currents; therefore, understanding the distribution of residual currents is useful for grasping the marine environment.
Fujiwara et al. (1989) presented the residual current distribution in Osaka Bay using several decades of tidal current observation data collected by national and local government agencies. In recent years, coastal development in Osaka Bay has progressed, leading to changes in coastal topography. However, few studies have examined the corresponding changes in distribution of residual current. This study aims to clarify the three-dimensional distribution and mechanisms of residual currents, including vertical seawater circulation, using the Finite-Volume, primitive equation Community Ocean Model (FVCOM). To achieve this, a simulation was conducted for Osaka Bay in 2015.
The period from 9:00 AM on January 1, 2015, to 9:00 AM on June 1, 2015, was set as the spin-up period, and calculations were performed until 9:00 AM on July 1. The analysis was conducted using the results from 9:00 AM on June 1 to 9:00 AM on June 30.
As boundary conditions for the calculations, flow fluxes and values of water temperature, salinity, and tide level obtained from the ocean model DREAMS were applied at the lateral boundaries, while CFSv2 heat flux data and MSM wind speed values were given at the sea surface boundaries. River water inflow was given values obtained from Water Information System, which is data from observations at stations under the jurisdiction of the Ministry of Land, Infrastructure, Transport and Tourism. To evaluate the reproducibility of the tidal currents, we used fixed-point observations of tide levels, current velocity values from current estimation, and surface current direction and velocity values from high frequency radar (HF radar). Both observed and inferred data were used for the same period of time as the analysis period.
The simulation results showed a large difference between the flow direction and velocity observed by HF radar and calculated by FVCOM. The correlation coefficients for the flow velocity were 0.22 for the east-west component and, 0.24 for the north-south component, which are very low values. Of the three components of the currents in Osaka Bay, namely, the blowdown current, density current, and tidal current, we first verified the reproducibility of the tidal current, which is the periodic component. Simulations were performed with FVCOM with only the DREAMS tidal values given to the boundaries, and the results were compared with the tidal current estimation and DREAMS. The period was generally consistent, but the phase lagged behind the tidal current estimation by about 2 hours. In addition, the FVCOM and DREAMS simulations showed that the flow velocities were underestimated compared to the tidal current estimation.
Comparison of the FVCOM tide level values with the actual observations showed that both the amplitude and phase of the FVCOM were close to the actual observations. Lag correlations were performed for the tide levels at Osaka and Kobe, which are the furthest from the lateral boundary, with a time lag of 1 hour and a correlation coefficient of 0.93. In Sumoto and Tannowa, which are the next furthest from the boundary after Osaka and Kobe, the time lag was 30-50 minutes and the correlation was 0.95. In Wakayama, which is the closest to the boundary, the time lag was 30 minutes and the correlation was 0.95. The closer to the boundary, the smaller the time lag and the higher the correlation.
These results indicate that while FVCOM reproduces tidal levels well, it does not reproduce tidal currents well and it correlates poorly with observations. This may be due to a problem in the process of determining tidal currents from the differences in tidal levels between locations.
Fujiwara et al. (1989) presented the residual current distribution in Osaka Bay using several decades of tidal current observation data collected by national and local government agencies. In recent years, coastal development in Osaka Bay has progressed, leading to changes in coastal topography. However, few studies have examined the corresponding changes in distribution of residual current. This study aims to clarify the three-dimensional distribution and mechanisms of residual currents, including vertical seawater circulation, using the Finite-Volume, primitive equation Community Ocean Model (FVCOM). To achieve this, a simulation was conducted for Osaka Bay in 2015.
The period from 9:00 AM on January 1, 2015, to 9:00 AM on June 1, 2015, was set as the spin-up period, and calculations were performed until 9:00 AM on July 1. The analysis was conducted using the results from 9:00 AM on June 1 to 9:00 AM on June 30.
As boundary conditions for the calculations, flow fluxes and values of water temperature, salinity, and tide level obtained from the ocean model DREAMS were applied at the lateral boundaries, while CFSv2 heat flux data and MSM wind speed values were given at the sea surface boundaries. River water inflow was given values obtained from Water Information System, which is data from observations at stations under the jurisdiction of the Ministry of Land, Infrastructure, Transport and Tourism. To evaluate the reproducibility of the tidal currents, we used fixed-point observations of tide levels, current velocity values from current estimation, and surface current direction and velocity values from high frequency radar (HF radar). Both observed and inferred data were used for the same period of time as the analysis period.
The simulation results showed a large difference between the flow direction and velocity observed by HF radar and calculated by FVCOM. The correlation coefficients for the flow velocity were 0.22 for the east-west component and, 0.24 for the north-south component, which are very low values. Of the three components of the currents in Osaka Bay, namely, the blowdown current, density current, and tidal current, we first verified the reproducibility of the tidal current, which is the periodic component. Simulations were performed with FVCOM with only the DREAMS tidal values given to the boundaries, and the results were compared with the tidal current estimation and DREAMS. The period was generally consistent, but the phase lagged behind the tidal current estimation by about 2 hours. In addition, the FVCOM and DREAMS simulations showed that the flow velocities were underestimated compared to the tidal current estimation.
Comparison of the FVCOM tide level values with the actual observations showed that both the amplitude and phase of the FVCOM were close to the actual observations. Lag correlations were performed for the tide levels at Osaka and Kobe, which are the furthest from the lateral boundary, with a time lag of 1 hour and a correlation coefficient of 0.93. In Sumoto and Tannowa, which are the next furthest from the boundary after Osaka and Kobe, the time lag was 30-50 minutes and the correlation was 0.95. In Wakayama, which is the closest to the boundary, the time lag was 30 minutes and the correlation was 0.95. The closer to the boundary, the smaller the time lag and the higher the correlation.
These results indicate that while FVCOM reproduces tidal levels well, it does not reproduce tidal currents well and it correlates poorly with observations. This may be due to a problem in the process of determining tidal currents from the differences in tidal levels between locations.