Integrating biogeochemistry and ecology into ocean data assimilation systems

ABSt rAct. Monitoring and predicting the biogeochemical state of the ocean and marine ecosystems is an important application of operational oceanography that needs to be expanded. The accurate depiction of the ocean’s physical environment enabled by Global Ocean Data Assimilation Experiment (GODAE) systems, in both real-time and reanalysis modes, is already valuable for various applications, such as the fishing industry and fisheries management. However, most of these applications require accurate estimates of both physical and biogeochemical ocean conditions over a wide range of spatial and temporal scales. In this paper, we discuss recent developments that enable coupling new biogeochemical models and assimilation components with the existing GODAE systems, and we examine the potential of such systems in several areas of interest: phytoplankton biomass monitoring in the open ocean, ocean carbon cycle monitoring and assessment, marine ecosystem management at seasonal and longer time scales, and downscaling in coastal areas. A number of key requirements and research priorities are then identified for the future. GODAE systems will need to improve their representation of physical variables that are not yet considered essential, such as upper-ocean vertical fluxes that are critically important to biological activity. Further, the observing systems will need to be expanded in terms of in situ platforms (with intensified deployments of sensors for O 2 and chlorophyll, and inclusion of new sensors for nutrients, zooplankton, micronekton biomass, and others), satellite missions (e.g., hyperspectral instruments for ocean color, lidar systems for mixed-layer depths, and wide-swath altimeters for coastal sea levels), and improved methods to assimilate these new measurements.

integrating Biogeochemistry and Ecology into Ocean Data Assimilation Systems ABStr Act. Monitoring and predicting the biogeochemical state of the ocean and marine ecosystems is an important application of operational oceanography that needs to be expanded. The accurate depiction of the ocean's physical environment enabled by Global Ocean Data Assimilation Experiment (GODAE) systems, in both real-time and reanalysis modes, is already valuable for various applications, such as the fishing industry and fisheries management. However, most of these applications require accurate estimates of both physical and biogeochemical ocean conditions over a wide range of spatial and temporal scales. In this paper, we discuss recent developments that enable coupling new biogeochemical models and assimilation components with the existing GODAE systems, and we examine the potential of such systems in several areas of interest: phytoplankton biomass monitoring in the open ocean, ocean carbon cycle monitoring and assessment, marine ecosystem management at seasonal and longer time scales, and downscaling in coastal areas. A number of key requirements and research priorities are then identified for the future. GODAE systems will need to improve their representation of physical variables that are not yet considered essential, such as upper-ocean vertical fluxes that are critically important to biological activity.
Further, the observing systems will need to be expanded in terms of in situ platforms (with intensified deployments of sensors for O 2 and chlorophyll, and inclusion of new sensors for nutrients, zooplankton, micronekton biomass, and others), satellite missions (e.g., hyperspectral instruments for ocean color, lidar systems for mixed-layer depths, and wide-swath altimeters for coastal sea levels), and improved methods to assimilate these new measurements. In this context, data assimilation is a relevant approach to achieve: (1) better control of the physical circulation that enhances the quality of biogeochemical dynamics, (2) initialization of the biological variables for prediction, (3) estimation of physical and biogeochemical model parameters, and (4) data-based assessments of modeling hypotheses.

B y p i E r r E B r A S S E u r , N i c O l A S G r u B E r , r O S A B A r c i E l A , K E i t h B r A N D E r , M A é VA D O r O N , A B D E l A l i E l M O u S S A O u i , A l i S tA i r J . h O B D Ay, M A r t i N h u r E t, A N N E -S O p h i E K r E M E u r , pAt r i c K l E h O D E y, r i c h A r D M At E A r , c y r i l M O u l i N , r A G h u M u r t u G u D D E , i N N A S E N i N A , A N D E i N A r S V E N D S E N
The importance of mesoscale activity in primary production is the subject of much discussion in the literature (e.g., Oschlies and Garçon, 1998), and the example below confirms that the resolution of GODAE products is key to realistically monitoring phytoplankton biomass in ocean basins.
During the EU-funded MERSEA (Marine Environment and Security for the European Area) project, a prototype of a coupled physical/biological assimilation system based on the PISCES model (Aumont et al., 2003) has been iNtrODuctiON In its original design, the Global Ocean Data Assimilation Experiment (GODAE) was conceived as a practical demonstration of real-time ocean data assimilation in order to provide a regular and complete depiction of global ocean circulation at eddy resolution and better consistency with observations of physical parameters and dynamical constraints (Le Traon et al., 1999). This type of oceanic information was identified in the early stages of GODAE as potentially very valuable to different applications related to the "living ocean, " such as the fishing industry and fisheries management (Griffin et al., 2002). Numerous examples of GODAE products used for environmental applications are found in the literature, ranging from real-time temperature products for monitoring seasonal fish migration (Hobday and Hartmann, 2006) and ecological regime shifts (Brander, in press), to ocean currents and velocities used to understand the transport and spread of fish larvae (Bonhommeau et al., 2009;Johnson et al., 2005), to sources and sinks of atmospheric CO 2 . Some of these examples will be discussed in the following sections.
Today, advances in modeling biogeochemical processes and increased computer power have made it possible to couple physical and biogeochemical models online to address wider environmental and societal issues by providing hindcasts, nowcasts, and forecasts from short lead times (Siddorn et al., 2007) to climate time scales (Johns et al., 2006).  (Gregg et al., 2005;Behrenfeld et al., 2006;Polovina et al., 2008). Accurate reanalyses of the physical ocean during the past 60 years, such as those expected from the Mercator reanalysis systems, will offer new insight into how phytoplankton biomass and production has varied over this period.

OcEAN cArBON cyclE MONitOriNG AND ASSESSMENt
The operational systems available as GODAE ended are not yet able to accurately provide real-time monitoring of global or basin-scale air-sea carbon fluxes as required (e.g., for attempts to obtain reliable regional carbon budgets).

MAriNE EcOSyStEM MANAGEMENt
Marine exploitation issues need to be addressed from an ecosystem perspective, that is, using an ecosystem-based management approach (e.g., Svendsen et al., 2007). The long-term vision is to develop the systems required for the study, management, and monitoring of exploited and protected marine species. populations  that are interacting with their modeled prey. sea-to-air CO 2 Flux (Pg C yr -1 ) s ea-to-air CO 2 Flux (Pg C yr -1 ) sea-to-air CO 2 Flux (Pg C yr -1 ) Figure 2. comparison of the ocean inversion estimate of the contemporary sea-to-air cO 2 flux    Gruber et al. (2009) other approaches, for example, a continuous size spectrum (Maury et al., 2007).
Ultimately, collected data should be used for parameter optimization and data assimilation .
From the user's perspective, oceanographic information is used in fisheries management in two basic ways. First, traditional fisheries oceanography has involved a search for explanation of historical catch-related patterns.
Stock assessment scientists then use this information to correct observed catch rates, and derive historical population estimates (e.g., Bigelow et al., 2002).
These estimates lead to an understanding of how fishing mortality changes population size, and the mortality is the control variable that management seeks to adjust via gear restrictions, quotas, and spatial management.
The relationships with the environment are often quite weak, because although the selected environmental variables may be correlated with fish distributions and abundances, they are not strong drivers (Myers, 1998).
Derived variables such as mixed layer depth, productivity, and prey biomass may be more relevant to the distribution and abundance of fish. Thus, improved "physical" and "biological" variables from coupled models may improve understanding of historical patterns, and offer improved management for ocean resources (e.g., Senina et al., 2008).  (Brander, in press). Decadal changes are observed in the North Pacific as well, with significant impacts on fisheries (Chavez et al., 2003;Brander, in press).
Because give rise to prolonged system changes.  (Huret et al., 2007;Gruber et al., 2006). Biological tracers or particles need to be transferred in a conservative way at the chosen boundary between the global and regional models in a two-way mode. For that to occur, the slope region, with its complex exchange processes of biological material, should not represent a boundary for regional applications, but rather it needs to be part of the regional high-resolution model.