the Creative Commons Attribution 4.0 License.
the Creative Commons Attribution 4.0 License.
Seasonal forecasts of Northeast Pacific physics and biogeochemistry using MOM6-COBALT-NEP10k
Abstract. Seasonal ocean forecasts offer critical insights for marine resource management, particularly in regions like the Northeast Pacific where ecosystem variability can have profound impacts on marine life, fisheries, and coastal communities. Under NOAA’s Changing Ecosystems and Fisheries Initiative, we developed a seasonal ocean forecasting system for the Northeast Pacific based on the Modular Ocean Model version 6 (MOM6) ocean model coupled to the Sea Ice Simulator version 2 (SIS2) sea ice model and the Carbon, Ocean Biogeochemistry, and Lower Trophics (COBALT) biogeochemical model. The model domain extends from south of Baja California to the Chukchi Sea with ~10 km horizontal resolution. To evaluate forecast skill, we produced a 32-year set of retrospective forecasts with 10 ensemble members, four initializations per year, and lead times up to 12 months. Forcing for the forecasts was derived from the Geophysical Fluid Dynamics Laboratory's Seamless System for Prediction and Earth System Research (GFDL-SPEAR) global climate forecast system. Forecasts were assessed using ocean reanalyses and in situ observations, focusing on three ecologically- and societally-relevant variables: sea surface temperature, bottom temperature, and bottom oxygen concentration. Forecast skill varies by region, variable, and season, but generally remains high (anomaly correlation coefficient; ACC>0.5) for lead times of at least several months across all variables and regions. The downscaled forecasts reduce errors relative to the SPEAR global forecasts and in some cases increase ACC, while also enabling predictions of biogeochemical variables. These new forecasts compare favorably to previous downscaled ocean forecasts off the U.S. west coast, with increased ACC and small forecast errors that may be reduced further by future methodological changes. Routine real-time forecasts, available starting in October 2026, will support proactive decision making by managers, industry, and coastal communities.
Competing interests: At least one of the (co-)authors serves as editor for the special issue to which this paper belongs.
Publisher's note: Copernicus Publications remains neutral with regard to jurisdictional claims made in the text, published maps, institutional affiliations, or any other geographical representation in this paper. While Copernicus Publications makes every effort to include appropriate place names, the final responsibility lies with the authors. Views expressed in the text are those of the authors and do not necessarily reflect the views of the publisher.- Preprint
(11012 KB) - Metadata XML
- BibTeX
- EndNote
Status: open (until 03 Dec 2026)
- RC1: 'Comment on egusphere-2026-5844', Anonymous Referee #1, 04 Nov 2026 reply
-
RC2: 'Comment on egusphere-2026-5844', Pearse Buchanan, 11 Nov 2026
reply
Jacox et al. – Seasonal forecasts of Northeast Pacific physics and biogeochemistry using MOM6-COBALT-NEP10k.
Summary.
Jacox and colleagues present the development and evaluation of a regional ocean forecasting system for the Northeast Pacific Ocean.
The authors complete 32 years of retrospective forecasts, with four initializations per year, ten ensemble members, and 12-month lead times. This provides a substantial basis for assessing seasonal predictability. They compare model projections against persistence and the parent SPEAR model, and from these comparisons establish whether the regional forecast system adds value. The authors also account for temporal autocorrelation in statistical significance testing (although how they do this remains somewhat ambiguous).
The authors also combined an impressive set of datasets to ground truth their model projections including reanalysis products, other model output, moorings, shipboard observations and coastal monitoring datasets.
Extending seasonal prediction across the Northeast Pacific with a regional MOM6–COBALT configuration is valuable. The system enables forecasts of variables that are generally unavailable from the parent global prediction system.
The presentation is also generally good. The manuscript is concise, Figures 5–7 effectively communicate seasonal variations in skill, and the comparison with earlier regional forecasting systems provides useful context.
Main comments
The main weakness of the work is its demonstration of the added value of biogeochemical downscaling, for which the authors only focus on oxygen. Skillful biogeochemical prediction is not really established, for which I would consider at the very least NPP and particle fluxes to benthos. I would not ask the authors to include these variables, but I would like them to acknowledge in their introduction and discussion that they focus on oxygen and that this is a limitation, since oxygen is the most “physically” influenced of all biogeochemical properties/tracers (Buchanan et al., 2018 Global Biogeochemical Cycles).
The lack of biogeochemical focus is interesting to me, since they use a sophisticated BGC model with 40 tracers, representing multiple nutrients, multiple phytoplankton groups, zooplankton and bacteria. I admit that proper evaluation of the model’s forecast capability relies on having good quality time series of properties or fluxes, but surely some of the moorings have nutrients or DIC or pH? NPP algorithms from remotely-sensed data also exist, although they are complicated by clouds, but this could be handled by such a large and experienced team? At this point the authors cannot say that the system can predict changes in nutrient availability, primary production and chlorophyll variability, seasonal bloom events, changes in carbonate chemistry or POC fluxes to sediment ecosystems. An initial focus on oxygen is defensible for an initial evaluation, but it limits what can be concluded about broader biogeochemical predictability and the authors should tackle this as a key limitation in their discussion.My intuition at this point is that the authors may have excluded NPP and chlorophyll and other more “BGC” properties because of poor behaviour at the boundaries? The authors prescribe climatological open-boundary conditions for most biogeochemical variables, including oxygen and nutrients, except DIC and Alk are prescribed using ESPER relationships with T and S. This introduces a potentially important inconsistency. The physical ocean boundary conditions vary with the SPEAR forecasts, whereas most biogeochemical boundary conditions do not contain corresponding interannual variability, which may dislocate nutriclines with thermoclines/MLDs. Hence, changes in the physical properties of incoming water masses may not be accompanied by consistent changes in nutrient concentrations, which may cause funny behaviour at the boundaries such as sudden nutrient replete conditions or unrealistic nutrient limitation.
Minor comments
Line 147 – sentence needs rephrasing for clarity.
At lines 561–563, the authors write that the regional model consistently increased ACC relative to persistence. This is stronger than the results seem to support. Earlier, at lines 319–322, they acknowledge cases where persistence provides skill comparable to or greater than the dynamical model. A more precise conclusion would emphasise that the dynamical forecasts generally outperform persistence, with important regional and seasonal exceptions.
Thank you for considering my input to your research.
Citation: https://doi.org/10.5194/egusphere-2026-5844-RC2
Viewed
| HTML | XML | Total | BibTeX | EndNote | |
|---|---|---|---|---|---|
| 464 | 202 | 41 | 707 | 46 | 47 |
- HTML: 464
- PDF: 202
- XML: 41
- Total: 707
- BibTeX: 46
- EndNote: 47
Viewed (geographical distribution)
| Country | # | Views | % |
|---|
| Total: | 0 |
| HTML: | 0 |
| PDF: | 0 |
| XML: | 0 |
- 1
This paper will be a useful milestone / reference for applications of seasonal forecasts in the Northeast Pacific. Papers like these come from an enormous effort of many researchers which is to be commended. Section 4.4 and 5 document the significant improvement of this modelling system over previous efforts and provides wisdom for other modelling teams. Well done. Below are some comments that need addressing before publication.
Major comments.
Minor comments.