Impact of Submesoscale Eddies on Synoptic and Mesoscale Oceanic Structures in a Continental Shelf Margin Analyzed with a Downscaling Ocean Model of Japan Sea

Wednesday, 17 December 2014
Dai Miyazaki1, Yusuke Uchiyama1, Ryosuke Kanki1 and Yasumasa Miyazawa2, (1)Kobe University, Kobe, Japan, (2)JAMSTEC, Yokohama, Japan
Japan Sea (JPS) is connected to other seas by five narrow and shallow straits with a minimal depth of the order of 100 meters or less, resulting in limited water exchange thereby isolating the water and aquatic ecosystem. The modeling and observational studies on quantifying the dynamics of JPS are still undergoing (e.g., Hirose et al., 2007), whereas effects of submesoscale dynamics on the mean structure, eddies, and material dispersal in JPS have not been extensively investigated yet. In the present study, we conduct a detailed oceanic downscaling numerical experiment using ROMS in a double nested configuration bounded on the assimilative JCOPE2 (Miyazawa et al., 2009) reanalysis at horizontal resolutions of 3 km (ROMS-L1) and 1 km (ROMS-L2).

The L1 and L2 models are compared to the observed data to show a good agreement with an appropriate parameter choice. Our models sufficiently reproduce the overall frontal structure and associated major currents in JPS consisting of the Liman Cold Current along the Russian coast and the Tsushima Warm Current along the Japanese coast. Surface normalized relative vorticity fields demonstrate that both the mesoscale and submesoscale eddies are apparently enhanced, as the model grid resolution is finer. In summer and fall, mesoscale eddies are evident in L1 and L2. In contrast in winter and spring, submesoscale eddies are significantly energized in the whole JPS particularly in L2 due to the surface cooling that preconditions symmetric instability (e.g., Thomas et al., 2012). The enhancement of EKE appears around Tsushima strait and along the Korean Peninsula in L1, while EKE in L2 is extensively increased in the most part of the southern JPS. On the other hand, a SSH variance, a proxy of mesoscale variability, is more realistically distributed in L2 than L1, suggesting a potential importance of submesoscale eddies on the mesoscale dynamics.