Seasonal-to-interannual Prediction of Bottom Temperature on the Northeast US Continental Shelf

Abstract:

The Northeast U.S. Continental Shelf (NES) Large Marine Ecosystem is arguably one of the most oceanographically dynamic marine ecosystems and supports some of the most commercially valuable fisheries in the world. A reliable prediction of the NES environmental variables, such as ocean temperature, could lead to a significant improvement in fisheries stock assessment. However, the current generation climate model-based seasonal-to-interannual predictions exhibit limited prediction skill in this coastal environment. In this study, we developed a series of seasonal-to-interannual statistical predictions for the NES bottom temperature using the GLORYS12v1 ocean reanalysis dataset. A simple local damped persistence prediction provides significant skills for lead times up to ~5 months in the northern Mid-Atlantic Bight (NMAB) and up to ~11 months in the deep Gulf of Maine (DGoM). Both regions exhibit strong seasonality in the prediction skill. Incorporating an upstream or nearby subregion as a predictor largely improves the prediction skill, particularly for the forecasts of winter months in the NMAB and summer months in the DGoM over longer lead times, regardless of initialization month. We attribute the improved prediction skill primarily to advective processes from the continental slope. Spiciness of the slope and shelf waters are investigated to clarify the pathways of the predictive signals. Finally, using the Gulf Stream path index and the North Atlantic Oscillation index as additional predictors in the statistical models further improves the prediction skill.