Hide

To give you the best possible experience, this site uses cookies. Continuing to use this site means you agree
to our use of cookies. If you'd like to learn more about the cookies we use please find out more

Session 3.5 abstracts

Symposium home | Session 2 | Session 3 | Session 4 | Abstracts by Author |

 

3.5 Marine Ecosystem Prediction: The Next Frontier of Ocean Forecasting – New Users and Societal Benefits

Session conveners: Rosa Barciela and Pierre Brasseur


The table below lists all abstracts for Session 3.5 by author. To read the full abstract click on the title-link.

The unique reference number (ref. no.) relates to the abstract submission process and must be used in any communications with the organisers.

All abstracts from session 3.5 are available for download - pdf.

 Ref.NoPrimary AuthorAffiliationCountryAbstract titlePosters
S3.5-01Chai, FeiUniversity of MaineUnited StatesIncorporating Optical Processes into Physical-Biogeochemical Models in the Pacific Ocean Oral
S3.5-02Gehlen, MarionIPSL/LSCEFranceBuilding the capacity for forecasting/predicting marine biogeochemistry: Recent advances and future developments of the Green Mercator initiative Oral
S3.5-03He, RuoyingNorth Carolina State UniversityUnited StatesPredicting Harmful Algal Blooms in the Gulf of Maine: From Event Hindcasting to Seasonal Forecasting 
S3.5-04Igarashi, HiromichiMRI-JMAJapanModeling optimal habitat suitability of neon flying squid in the central North Pacific by using 3-D ocean data assimilation product - An application of GOV product to fisheries community - 
S3.5-05Ji, XuanliangNMEFCChinaTemporal and Spatial Variability of Carbon Cycle in the China Sea: A Three-Dimensional Physical-Biogeochemical Modeling Study 
S3.5-06Kiyofuji, HidetakaMRI-JMAJapanInvestigating relationship between Skipjack Tuna (Katsuwonus pelamis) and environment from modeling approach - An application of spatial ecosystem and population dynamics model (SEAPODYM) and GOV product (MOVE) - 
S3.5-07Kwon, Jae-IlKIOSTKorea, Republic ofInterannual and Cross Shelf Variability in Sea Surface Temperature and Phytoplankton Scales in the East Sea 
S3.5-08Lehodey, PatrickCLSFranceOperational management of tuna fisheries in IndonesiaPoster-pdf
S3.5-09Nair, VijayalakshmiNational Institute of OceanographyUnited StatesPredictability of Physical Oceanography in structuring Zooplankton biodiversity - A vision for the Indian Ocean Cancelled
S3.5-10Nishikawa, HarukaMRI-JMAJapanImpact of paralarvae and juveniles feeding environment on the neon flying squid (Ommastrephes bartramii) winter-spring cohort stock - An application of GOV product to fisheries community - 
S3.5-11Perruche, CoralieIPSL/LSCEFranceAssessment of a global eddy-permitting biogeochemical hindcast of the ocean colour era 
S3.5-12Seferian, RolandIPSL/LSCEFranceMulti-year prediction of Marine Productivity in the Tropical Pacific 

 


ID 3.5-01

Incorporating Optical Processes into Physical-Biogeochemical Models in the Pacific Ocean

Fei Chai1, Peng Xiu1, Huijie Xue1, Curtis Mobley2, Yi Chao3,4

1 School of Marine Sciences, University of Maine, Orono, Maine, USA

2 Sequoia Scientific, Inc., Bellevue, WA, USA

3 Remote Sensing Solutions, Inc., Pasadena, California, USA

4 University of California, Los Angeles, California, USA

Abstract

Modeling interactions between physics and biology is important for understanding ocean processes. Existing models often use very sophisticated treatmes of the hydrodynamics and increasingly sophisticated biology, but still grossly oversimplify the optics. In this talk, we will report on how to bring optical computations up to the standards of the physical and biological components without unacceptable computational costs. We developed an extremely fast radiative transfer numerical model, EcoLight-S(ubroutine), which can compute spectral irradiance from 300-1000 nm throughout the euphotic zone in a few tenths of a second of computer time with errors of less than 10% in PAR or broadband Ed(300-1000). EcoLight-S was imbedded in a coupled ROMS-CoSiNE hydrodynamic-biological ocean ecosystem model to replace the simple analytic light models previously used to compute photosynthesis and water heating by short-wave radiation. Comparison runs for an idealized upwelling-downwelling channel geometry showed that both water heating and biological component concentrations (chlorophyll, nutrients, etc.) are significantly different after only two weeks of simulated time. Total chlorophyll concentrations differ by up to 50%, and sea surface temperatures differ by up to 0.3 deg C. The increase in total run time when using EcoLight-S to computed irradiances from 400-1000 nm was less than 20%. One of advantages integrating physical-biological-optical models together is to allow for model validation by direct comparison with satellite-derived or measured remote sensing reflectance Rrs(λ), without the intermediate step of converting Rrs(λ) to chlorophyll for comparison. Also, incorporating optical processes into physical-biological models can constrain ecosystem predictions using remote sensing or in-water optical data from moorings, gliders, and lead to better biological data assimilation.


ID 3.5-02

Building the capacity for forecasting/predicting marine biogeochemistry and ecosystems

M. Gehlen1, A. El Moussaoui2, C. Perruche2, E. Dombrowsky2,

P. Brasseur3, J. Le Sommer3, P. Lehodey4, L. Mémery5, J. Orr1

1 IPSL/LSCE, UMR CEA-CNRS-UVSQ, Gif-sur-Yvette, France

2 Mercator Océan, Ramonville St Agne, France

3 LGGE, CNRS, Grenoble, France

4 CLS, Ramonville-Saint-Agne, France

5 LEMAR, UMR UBO-CNRS-IRD, Plouzané, France

Abstract

Building the capacity for monitoring and forecasting marine biogeochemistry and ecosystem dyamics is a scientific challenge of strategic importance in the context of rapid environmental change and growing public awareness of its potential impacts on marine ecosystems and resources. National Operational Oceanography centers start to take up this challenge by integrating biogeochemistry to operational systems. Progress relies on the close collaboration with the academic community. The French Green Mercator project illustrates this approach. Its main objective is to provide the community with the capacity for hind-casting, real-time monitoring and forecasting of the ocean biogeochemical state. It builds on an approach combining (1) the development of an assimilation system for biogeochemical observations; (2) the integration of a biogeochemical component to the operational system; (3) the collaborative development of end-user targeted applications. These efforts are paralleled by the development of metrics for the systematic comparison of model output to observations, as well as objective benchmarking of technical choices and new developments. Expanding operational systems to biogeochemistry opens the possibility for wide ranging applications in environmental monitoring and research. Examples include the assessment of the variability of ocean biogeochemistry at seasonal to inter-annual time scales; the development of a capacity for the monitoring of the ocean carbon cycle, the development of a framework for the science-based monitoring/management of marine ecosystems and resources. The scope of this contribution is to present the current status of green operational oceanography, major challenges ahead, as well as potential end-user applications.


ID 3.5-03

Predicting Harmful Algal Blooms in the Gulf of Maine: From Event Hindcasting to Seasonal Forecasting

Ruoying He1, Dennis McGillicuddy2, Don Anderson2, Yizhen Li1

North Carolina State University1

Woods Hole Oceanographic Institution 2

 

Abstract

The Gulf of Maine (GOM) is a continental shelf sea that supports productive shellfisheries that are frequently impacted by Alexandrium fundyense blooms and outbreaks of paralytic shellfish poisoning (PSP). To predict HAB abundance and distribution in the GOM, we have coupled a population dynamics model for Alexandrium into a regional ocean circulation modeling framework. In the wake of the historic bloom of 2005 in the western GOM, this coupled model was used to diagnose the underlying causes, including 1) high abundance of resting cysts in fall 2004 that provided a large inoculum; 2) storms with strong northeast winds that carried toxic cells towards, and along the coast; and 3) abundant fresh water runoff, providing macro- and micro-nutrients, a stratified water column, and enhanced alongshore transport. Sensitivity experiments show that simulations initiated from A. fundyense cyst distributions capture large-scale seasonal patterns in the distribution and abundance of A. fundyense cells. We proposed that cyst abundance is a first-order predictor of regional bloom magnitude in the following year, and that cyst abundance may hold the key to interannual forecasts of PSP severity.This hypothesis was put to the test in our model hindcast experiments from 2004 through 2012. We will present these hindcast results, and also discuss our approaches to perform ensemble forecasts before the bloom season and near real-time weekly nowcast/forecast during the bloom season.

 


ID 3.5-04

Modeling optimal habitat suitability of neon flying squid in the central North Pacific by using 3-D ocean data assimilation product

- An application of GOV product to fisheries community-

H. Igarashi1, T. Awaji1, M. Kamachi1,2, Y. Ishikawa1, N. Usui2, M. Iiyama3, M. Sakai 4, Y. Kato4, I. Alabia5, S-I. Saitoh5 and M. Seitoh6

1 Japan Agency for Marine-Earth Science and Technology (JAMSTEC) , Yokohama , Japan

2 Meteorological Research Institute (MRI), Japan Meteorological Agency (JMA), Tsukuba, Japan

3 Faculty of Economics, Kyoto University , Kyoto, Japan

4 Tohoku National Fisheries Research Institute (TNFRI), Fisheries Research Agency (FRA), Hachinohe, Japan

5 Graduate School of Fisheries Sciences, Hokkaido University, Hakodate, Japan

6 Fisheries Research Institute (FRI), Aomori Prefectural Industrial Technology Research Center, Aomori, Japan

Abstract

The neon flying squid (Ommastrephes bartramii) has a wide-spread distribution in subtropical and temperate waters in the world ocean. In the North Pacific, it plays an important role in the pelagic ecosystem and is one of the major targets in Japanese squid fisheries. There are two main fishing grounds for Japanese commercial vessels, east of northern Japan in winter and the wide north Pacific area (35-45N) around the dateline in summer. In this study, the suitable habitat area for the neon flying squid in central North Pacific in summer was investigated using MOVE-WNP (Meteorological Research Institute multivariate ocean variational estimation, GOV Japan National System) ocean reanalysis product, which can provide realistic fields of 3-dimensional ocean circulation and environmental structures including meso-scale eddies (Usui et al.,2006). The Japanese commercial fisheries data of neon flying squid from June to July during 2000-2012 and MOVE reanalysis have been applied to the identification and characterization of possible habitat suitable areas for the summer fishing ground of neon flying squid by using several kinds of habitat suitability models (HSI, GLM, GAM, etc.). The results indicate the 3-dimensional structure of temperature and salinity is highly correlated with the squid catch in the summer fishing area, which suggests that the position of the subarctic frontal zone (SAFZ) and the meso-scale eddy activity around SAFZ could control the formation of the suitable habitat for neon flying squid.


ID 3.5-05

Temporal and Spatial Variability of Carbon Cycle in the China Sea: A Three-Dimensional Physical-Biogeochemical Modeling Study

XL JI1, GM LIU1, S GAO1

1 National Marine Environmental Forecasting Center , Beijing, China

Abstract

The China Sea (CS) with a wide range of the continental shelf exhibits strong variations on seasonal, interannual and decadal time scale, and the monsoon has direct impacts on the nutrients cycle, as well as the carbon cycle. Based on the Regional Ocean Modeling System (ROMS), combining with the characteristics of marine ecosystem in the CS, a three-dimensional physical-biogeochemical model including carbon cycle with the resolution 1/12º×1/12º is established to investigate the physical variations, ecosystem responses, and carbon cycle consequences in the CS. The ROMS-NPZD model which is driven with daily air-sea fluxes derived from the National Centers for Environmental Prediction (NCEP) reanalysis between 1982 and 2005, has a good capable of reproducing the observed seasonal variation characteristics over the same period along 32ºN section in the East China Sea (ECS). The integrated air to sea CO2 flux over the entire CS reveals a strong seasonal cycle, playing as a source of CO2 to the atmosphere from June to October, but serving as a sink of CO2 to the atmosphere in the other months. The 24-year mean air to sea CO2 flux averaged over the entire CS is 1.06mol m-2 yr-1, which is equivalent to a regional total of 3.22Tg-C yr-1, indicating that the CS is a sink of CO2 to atmosphere. The partial pressure of carbon dioxide (pCO2sea) in the CS has an increasing rate of 1.15μatm yr -1, but pH in seawater has an opposite tendency, decreasing from 1982 to 2005 with a rate of -0.0013 yr-1. Biological activity is the dominant factor that controls the pCO2sea in the CS, while temperature is the second factor. The inverse relationship between the interannual variability of air to sea CO2 flux and NINO3 SST (Sea Surface Temperature) index indicates that the carbon cycle in the CS has a good connection with ENSO phenomenon (El Nino-Southern Oscillation).

S3.5-04-Ji

Fig.1. The modeled surface climatological distributions of pCO2sea averaged from 1982 to 2005 (a: February; b: May; c: August; d: November). The unit isμatm.

 


ID 3.5-06

Investigating relationship between Skipjack Tuna

(Katsuwonus pelamis) and environment from a modeling approach

– An application of spatial ecosystem and population dynamics model (SEAPODYM) and GOV product (MOVE) –

H. Kiyofuji1, H. Okamoto1, I. Fusejima2, T. Oshima2, T. Kimura2, I., Senina3, P. Lehodey 3,

N. Usui4, and M. Kamachi4

1 National Research Institute of Far Seas Fisheries (NRIFSF),

2 Marine Fisheries Research and Development Center (JARMARC)

Fisheries Research Agency (FRA), Japan

3 CLS, Toulouse, France

4 Meteorological Research Institute (MRI), Japan Meteorological Agency (JMA), Tsukuba, Japan

Abstract

The relationship between skipjack tuna (Katsuwonus pelamis) and ocean environment is investigated with numerical model experiments. The numerical experiments were carried out using the Spatial ecosystem and population dynamics model (SEAPODYM) developed by CLS, the vertically generalized production model (VGPM), and the three-dimensional variational data assimilation system (MOVE-WNP: GOV Japan National System) for ocean environment. The SEPODYM is characterized as the spatial dynamics model for both micronecton functional groups and detail age structured population of skipjack tuna with a system of advection-diffusion-reaction equations and environmental forcing (temperature, currents, primary production and dissolved oxygen concentration). Young and adult skipjack biomass around Japan were estimated and compared to the fisheries catch date. Three month forecasting of skipjack biomass were also conducted in order to investigate model predictability for near real time fishing operation.


ID 3.5-07

Interannual and Cross Shelf Variability in Sea Surface Temperature and Phytoplankton Scales in the East Sea

J. Kwon1,*, H. Kim2, K. Park1, J. Jeong1, and S. Son3

1 Coastal Disaster Research Center, Korea Institute of Ocean Science & Technology, Ansan, Republic of Korea

2 I.M. Systems Group, Rockville, MD, USA

3 CIRA, Colorado State University, Fort Collins, CO, USA

*jikwon@kiost.ac

Abstract

The East Sea has dynamic environmental conditions in respect to its physical phenomena and biological characteristics, showing many characteristics of large ocean. Climate change and associated changes in physical oceanography impact the structure and functioning of marine ecosystems. They can also affect biological production, carbon budget, and fisheries resources in the East Sea. In addition, the mesoscale variability in the system may be as important as the chl-a biomass in determining the potential productivity of higher trophic levels.

Time series of 10 years high-resolution sea surface temperature and chlorophyll-a data measured from the NASA Moderate Resolution Imaging Spectroradiometer (MODIS) on Aqua are used to assess dominant time scales of variability using a wavelet analysis in the East Sea. Dominant periods of sea surface temperature and chlorophyll-a variances, and how these changes in time, are quantified.

The interannual and decadal patterns of the sea surface temperature and phytoplankton chlorophyll-a in the East Sea are also compared with the large scale climate patterns such as the El-Nino Southern Oscillation (ENSO) and the Pacific Decadal Oscillation (PDO) data to investigate how the large scale climate changes link to the interannual and decadal phytoplankton variability in the regional oceans. Possible reasons of the decadal changes in the temporal and spatial distributions will be further discussed.


ID 3.5-08

Operational management of tuna fisheries in Indonesia

Patrick Lehodey 1, Inna Senina 1, Olivier Titaud 1, Beatriz Calmettes1, Teja Arief 2, Marion Gehlen 3, Gilles Larnicol 1 and Philippe Gaspar 1

1 CLS Space oceanography division, 8-10 rue Hermes, 31520 Ramonville Saint-Agne, France

2 SEACORM (Southeast Asia Center for Ocean Research and Monitoring), Perancak, Bali, Indonesia

3 LSCE/IPSL, Gif-sur-Yvette, France

Abstract

Tuna fisheries are a major economical sector in Indonesia. Despite the increasing huge amount of tuna catch in the Indonesian EEZ (annual catch > 600,000 t), very few is known about the dynamics and real abundance of tuna in the region. Fisheries statistics are still incomplete and lack accuracy. With the INDESO project, the Indonesian government is implementing a chain of operational regional models from physics to fish at a resolution of 1/12°x day.

Outputs of the regional ocean model and satellite-derived primary production and euphotic depth are used to drive the Spatial Ecosystem And Population Dynamics Model (SEAPODYM) for three tuna species (skipjack, yellowfin and bigeye) and their fisheries. Primary production predicted from the coupled regional biogeochemical model will be also tested. The SEAPODYM model includes a representation of several functional groups of micronekton from surface to ~1000m and a detailed age structured spatial population dynamics. Open boundary conditions of the regional model are provided from a global operational tuna model at resolution ¼° x week driven by the ocean model PSY3 of Mercator-Ocean. The operational model is updated every week with 10 days of forecast. The ultimate goal of this effort is to have an integrated management system to assist in the day-to-day monitoring of Indonesian tuna fisheries, to improve reporting and to deter illegal fishing in order to maintain the Indonesian tuna resources to a sustainable level.


ID 3.5-09

Predictability of Physical Oceanography in structuring Zooplankton Biodiversity – A Vision for the Indian Ocean

Vijayalakshmi R. Nair1 (retired) and K. K. Kusum2

1 National Institute of Oceanography), and

2 National Institute of Oceanography, Salim Ali Road, Kochi 682018, India

Abstract

Even after the lapse of 50 years of International Indian Ocean Expedition (IIOE,1960 -65), the intricacies of physical factors on zooplankton ecology are yet to be fully ascertained. The IIOE data on zooplankton had not been or could not be analysed effectively to evolve a clear scenario on the disposition on these passively drifting animal community in the oceanic realm. The over powering monsoon regime of the northern Indian Ocean is a unique phenomenon and basic information on the back and forth movements of zooplankton associated with current reversal and vertical stratification in relation to the thermocline & upwelling are a few of the important findings. The remote sensing data also show a correlation between zooplankton and physical characteristics of the northern Indian Ocean. The Hydro-chemical front at 100 S Lat. of the Indian Ocean limits the spreading of many species of zooplankton towards further south.

Later studies along the EEZ of India and restricted zones in the Indian Ocean such as JGOFS, CMLRI , etc. added further information to the ecology of the zooplankton. Comparison of IIOE and CMLRI data on zooplankton in the Bay of Bengal shows that in the southern part towards the Equator the zooplankton (particularly Chaetognatha, a predominant group of zooplankton) has decreased while in the northern part it has increased showing poleward movement. Also, higher abundance of mesopelagic species is observed in the 200m water column of the Bay of Bengal and the Arabian Sea. The increase of SST and spreading of OMZ thickness in the mid depth layer were found to be the major responsible factors. These studies emphasise the short term and long term influences of different mesoscale and basin scale physical events on the ecology of zooplankton. Despite these outcomes, we still have not been able to develop a complete understanding of this unique basin. Probably, we are better equipped currently compared to IIOE era to unravel the outstanding scientific questions using a combination of new technologies, model studies and physical oceanography. May be a global effort for a repeat of IIOE can lead to a much better understanding and forecasting of the Indian Ocean ecosystem.

 


ID 3.5-10

Impact of paralarvae and juveniles feeding environment on the neon flying squid (Ommastrephes bartramii) winter-spring cohort stock

- An application of GOV product to fisheries community-

H. Nishikawa1, H. Igarashi1, Y. Ishikawa1, M. Sakai2, Y. Kato2, M. Ebina3, N. Usui 4, M. Kamachi4,*, and T. Awaji1,5

1 Japan Agency for Marine-Earth Science and Technology (JAMSTEC), Yokohama, Japan.

2 Tohoku National Fisheries Research Institute (TNFRI), Fisheries Research Agency (FRA), Hachinohe, Japan

3 Fisheries Research Institute (FRI), Aomori prefectural Industrial Technology Research Center , Aomori, Japan

4 Meteorological Research Institute (MRI), Japan Meteorological Agency (JMA), Tsukuba, Japan

5 Kyoto University, Kyoto, Japan

Abstract

We found that significant positive correlations between the CPUE (squid abundance index) of the neon flying squid (Ommastrephes bartramii) winter-spring cohort and satellite derived chlorophyll-a density in their spawning grounds, in the area of 140–160ºE and 21ºC < SST < 25ºC, from February to May. Since the spawning grounds of the winter-spring cohort are located on a quiet stream region, particle tracking experiment, which is based on the velocity field obtained from an ocean data assimilation system MOVE-WNP (Meteorological Research Institute multivariate ocean variational estimation, GOV Japan National System), shows that younger than 90 day-old paralarvae and juveniles stay at their spawning grounds and the chlorophyll-a density in their habitat has a significant positive correlation with the CPUE. Also, backward particle tracking experiment shows that the chlorophyll-a density in the spawning grounds has a significant positive correlation with the autumn-winter mixed layer depth. From these results, we hypothesize that the CPUE interannual variability is a result of variations in the paralarvae and juveniles feeding environment which can be linked to the autumn-winter mixed layer depth variations.


ID 3.5-11

Assessment of a global eddy-permitting biogeochemical hindcast of the ocean colour era

C. Perruche1, M. Gehlen2, A. Daudin2, A. El Moussaoui1, E. Greiner3, C. Ethé4

1 Mercator-Ocean, Toulouse, France

2 IPSL/LSCE, UMR CEA-CNRS-UVSQ, Gif-Sur-Yvette, France

3 CLS, Toulouse, France

Abstract

 

The combination of climate change and various anthropogenic drivers, such as e.g. changes in external nutrient inputs and exploitation of marine resources drive important changes in marine ecosystems. These changes occur against the background of natural variability. The retrospective analysis of global ocean biogeochemical state holds promise for identifying the response of marine ecosystems and biogeochemical fluxes to natural climate variability and, potentially, allows to detect trends driven by global climate change. Ideally, such a biogeochemical hindcast simulation should resolve the mesoscale and span multiple decades. Here, we present a biogeochemical simulation at 1/4° resolution for the period between 1994 to 2010 with NEMO/PISCES. The biogeochemical model PISCES was forced off-line by weekly fields provided by a physical simulation at 1/4° resolution (orca025) over the same period. The model was initialized with global climatologies. The spin-up involved 20 years of biogeochemical off-line simulation forced by a climatology of ocean physics. The inter-annual simulation (1994-2010) followed on the spin-up. The analysis of our spin-up strategy is presented with focus on the adjustment of model fields. The inter-annual simulation is evaluated by systematically comparing model fields to observations at global and regional scales. We draw on EOF (Empirical Orthogonal Functions) analysis to evaluate spatial/temporal variability. We focus on links between biogeochemical and physical variables in order to identify underlying common modes of variability for multiple variables. Finally, to complete the assessment, we compare EOF modes for Globcolour chlorophyll estimates (a merged Seawifs-Meris-Modis product) and model output over the period of observations.


ID 3.5-12

Multi-year prediction of Marine Productivity in the Tropical Pacific

R. Séférian1, 3, L. Bopp1, M. Gehlen1, D. Swingedouw1, J. Mignot2, E. Guilyardi2, J. Servonnat1

1 IPSL/LSCE, UMR CEA-CNRS-UVSQ, Gif sur Yvette, France

2 IPSL/LOCEAN, Paris, France

3 CNRM-GAME/GMGEC/ASTER, Toulouse, France

Abstract

Phytoplankton is at the base of the marine food web. Their carbon fixation, the net primary productivity (NPP), sustains most living marine resources and global fisheries. In certain regions, e.g. the tropical Pacific, NPP exhibits natural fluctuations at interannual to decadal time scales that have large impacts on marine ecosystems and fisheries. The predictions of NPP fluctuations could be of major relevance to the science-based management of marine resources. Yet, at present, the predictive capacity is hampered by the ability of Earth system models to reproduce the phasing and the amplitude of NPP variations. Here, we use observed sea surface temperature as a simple approach to partly overcome this difficulty. We present the first retrospective prediction of NPP over the last decades (i.e., from 1997 to 2010) with an Earth system model. Our analyses focus on the tropical Pacific, a region hosting the world largest fisheries. Results suggest a predictive skill for NPP of 3 years, which is higher than that of physical ocean fields such as SST (1 year). As opposed to SST, biogeochemical fields are isolated from the stochastic noise of the atmosphere. The increased predictability arises from the poleward advection of surface nutrients and iron anomalies, which sustain fluctuations of ocean productivity over years. These results open novel perspectives to the development of science-based management approaches to marine resources relying on integrated physical-biogeochemical forecasting systems.