Impact of Ocean Resolution on Navy ESPC Forecast Skill

Abstract The impact of an eddy-resolving ocean as compared to an eddy-permitting ocean in the U. S. Navy’s Earth System Prediction Capability (ESPC) global coupled system on ocean and atmospheric forecast skill is examined. The use of an eddy-resolving (1/12°) ocean provides clear benefits to ocean forecasting skill compared to the use of an eddy-permitting (1/4°) ocean initialized using the eddy-resolving ocean analysis. The eddy-resolving ocean reduces RMSE for mixed layer depth, sea surface height anomaly, and upper-ocean temperature and salinity by as much as 20% throughout the 45-day forecast integration lead times over forecasts using the eddy-permitting ocean. The impact of eddy-resolving scales is particularly important in the energetic western boundary currents, with over 50% reduction in the SST biases in the Gulf Stream and Kuroshio Extension by the end of the 45-day forecasts, although the warm bias along the immediate east coasts of Asia and North America is increased. These ocean improvements result in reductions of air temperature bias magnitudes at long lead times, most notably over the tropics and southern extratropics. However, the impact on other atmospheric skill metrics is near neutral. Significance Statement Using an eddy-resolving ocean model as compared to a lower-resolution eddy-permitting ocean model in a global coupled system has clear benefits for the upper-ocean forecast skill on subseasonal forecasts. It also results in improvements to atmospheric forecasts, particularly for low-level atmospheric temperature biases, but the impact on other atmospheric metrics is close to neutral. The decision on how to spend limited computational resources will depend on the metrics that are a priority for the forecasting center.