A generic framework for individual-based modelling and physical-biological interaction

Coupling of potential habitat models with particle tracking experiments to examine larval fish dispersal and connectivity in deep water regions

Computing Lagrangian trajectories with ocean circulation models is a powerful way to infer larval dispersal pathways and connectivity. Defining release areas and timing of particles to represent larval habitat realistically is critical to obtaining representative dispersal pathways. However, it is challenging due to spatial and temporal variability in larval density. Forward-tracking particle experiments were conducted to study larval connectivity of four species (neritic or mesopelagic) in the Gulf of Mexico's (GoM) deep-water region. A seasonal climatology coupled with predicted potential larval habitat models based on generalized additive models was used to delimit the particle dispersal origin. Two contrasting mesoscale circulation patterns were examined: (1) high Loop Current (LC) intrusion, absence of recently detached LC anticyclonic eddies (LC-ACE), and no interaction between LC-ACEs and the semi-permanent cyclonic eddy (CE) in the Bay of Campeche (BoC), and (2) limited LC intrusion, a recently detached LC-ACE, and interaction between LC-ACEs and the BoC's CE. To simulate larval transport, virtual larvae were randomly released in the potential habitats and advected for 30 days with the velocity fields of the HYbrid Coordinate Ocean Model with hourly-resolution assimilation. Potential habitat location and size played a major role in dispersal and connectivity. A greater percentage of particles were retained in potential habitats restricted to the southern BoC, suggesting lower connectivity with other GoM regions than those encompassing most of the BoC or the central Gulf. Mesoscale feature interactions in the western GoM and BoC led to greater dispersion along the western basin. By contrast, the absence of ACE-CE interaction in the BoC led to greater retention and less connectivity between the southern and northern GoM. Under high LC intrusion, particles seeded north of the Yucatan Shelf were advected through the Florida Straits and dispersed within the GoM. Coupling potential habitat models with particle experiments can help characterize the dispersal and connectivity of fish larvae in oceanic systems.

John M. Simulating transport and distribution of marine macro-plastic in the Baltic Sea

We simulated the spatial distribution and dynamics of macro plastic in the Baltic Sea, using a new Lagrangian approach called the dynamical renormalization resampling scheme (DRRS). This approach extends the super-individual simulation technique, so the weight-per-individual is dynamic rather than fixed. The simulations were based on a mapping of the macro plastic sources along the Baltic coast line, and a five year time series of realistic wind, wave and current data to resolve time-variability in the transport and spatial distribution of macro plastics in the Baltic Sea. The model setup has been validated against beach litter observations and was able to reproduce some major spatial trends in macroplastic distributions. We also simulated plastic dispersal using Green's functions (pollution plumes) for individual sources. e.g. rivers, and found a significant variation in the spatial range of Green's functions corresponding to different pollution sources. We determined a significant temporal variability (up to 7 times the average) in the plastic concentration locally, which needs to be taken into account when assessing the ecological impact of marine litter. Accumulation patterns and litter wave formation were observed to be driven by an interplay between positive buoyancy, coastal boundaries and varying directions of physical forcing. Finally we determined the range of wind drag coefficients for floating plastic, where the dynamics is mostly directly wind driven, as opposed to indirectly by surface currents and waves. This study suggests that patterns of litter sorting by transport processes should be observable in many coastal and off-shore environments.

A versatile marine modelling tool applied to arctic, temperate and tropical waters

The improved understanding of complex interactions of marine ecosystem components makes the use of fully coupled hydrodynamic, biogeochemical and individual based models more and more relevant. At the same time, the increasing complexity of the models and diverse user backgrounds calls for improved user friendliness and flexibility of the model systems. We present FlexSem, a versatile and user-friendly framework for 3D hydrodynamic, biogeochemical, individual based and sediment transport modelling. The purpose of the framework is to enable natural scientists to conduct advanced 3D simulations in the marine environment, including any relevant processes. This is made possible by providing a precompiled portable framework, which still enables the user to pick any combination of models and provide user defined equation systems to be solved during the simulation. We here present the ideas behind the framework design, the implementation and documentation of the numerical solution to the Navier-Stokes equations in the hydrodynamic module, the surface heat budget model, the pelagic and benthic equation solvers and the Lagrangian movement of the agents in the agent based model. Five examples of different applications of the system are shown: 1) Hydrodynamics in the Disko Bay in west Greenland, 2) A biogeochemical pelagic and benthic model in the inner Danish waters, 3) A generic mussel farm model featuring offline physics, food levels and mussel eco-physiology, 4) Sediment transport in Clarion-Clipperton zone at the bottom of the Pacific and 5) Hydrodynamics coupled with an agent based model around Zanzibar in Tanzania. Hence we demonstrate that the model can be set up for any area with enough forcing data and used to solve a wide range of applications.

Weak genetic structure despite strong genomic signal in lesser sandeel in the North Sea

Sandeels are an ecologically important group of fishes; they are a key part of the food chain serving as food for marine mammals, seabirds and fish. Sandeels are further targeted by a large industrial fishery, which has led to concern about ecosystem effects. In the North Sea, the lesser sandeel Ammodytes marinus is by far the most prevalent species of sandeel in the fishery. Management of sandeel in the North Sea plus the Kattegat is currently divided into seven geographical areas, based on subtle differences in demography, population dynamics and results from simulations of larval dispersal. However, little is known about the underlying genetic population structure. In this study, we used 2,522 SNPs derived from restriction site-associated DNA sequencing (RADseq) typed in 429 fish representing four main sandeel management areas. Our main results showed (a) a lack of a clear spatially defined genetic structure across the majority of genetic markers and (b) the existence of a group of at least 13 SNPs under strong linkage disequilibrium which together separate North Sea sandeel into three haplotype clusters, suggestive of one or more structural variants in the genome. Analyses of the spatial distribution of these putative structural variants suggest at least partial reproductive isolation of sandeel in the western management area along the Scottish coast, supporting a separate management. Our results highlight the importance of the application of a large number of markers to be able to detect weak patterns of differentiation. This study contributes to increasing the genetic knowledge of this important exploited species, and results can be used to improve our understanding of population dynamics and stock structure.