Influence of river routing resolution and parameterization on ocean circulation and sea surface temperature in seasonal prediction system

Abstract River discharge represents an important freshwater forcing in coupled climate systems, yet the river routing components used in operational seasonal prediction systems often remain at substantially coarser spatial resolutions than their atmospheric, land, and ocean model components. This study investigates the influence of river routing resolution and associated parameterization on ocean simulation in the Korea Meteorological Administration (KMA) Global Seasonal Forecasting System version 6 (GloSea6). The TRIP river routing model, originally hard-coded at a fixed 1 ° resolution, was modified to accommodate flexible spatial resolutions, and a higher-resolution configuration of 0.5 ° was implemented using refined river direction, flow sequence, and river-mouth information. Increasing the river routing resolution provides a more detailed representation of river networks and river mouths, leading to changes in freshwater discharge and coastal salinity. These changes modify upper-ocean stratification and circulation, reducing regional SST biases around the Korean Peninsula, Japan, and the Sea of Okhotsk. Although salinity changes are relatively limited over the open ocean, associated adjustments in mixed layer depth, circulation, and SST extend into parts of the tropical Pacific and Indian Oceans. These results demonstrate that river routing representation constitutes a non-negligible source of freshwater forcing uncertainty in coupled seasonal prediction systems. While increasing routing resolution alone is insufficient to eliminate systematic ocean biases, improved representation of river networks and appropriate parameter optimization can influence ocean salinity, stratification, circulation, and SST from coastal to broader oceanic scales. The results therefore emphasize the importance of a more physically realistic treatment of river routing and freshwater discharge for improving ocean simulations in future coupled seasonal prediction systems.