Bayesian statistical data assimilation for ecosystem models using Markov Chain Monte Carlo

This study considers advanced statistical approaches for sequential data assimilation. These are explored in the context of nowcasting and forecasting using nonlinear differential equation based marine ecosystem models assimilating sparse and noisy non-Gaussian multivariate observations. The statistical framework uses a state space model with the goal of estimating the time evolving probability distribution of the ecosystem state. Assimilation of observations relies on stochastic dynamic prediction and Bayesian principles. In this study, a new sequential data assimilation approach is introduced based on Markov Chain Monte Carlo (MCMC). The ecosystem state is represented by an ensemble, or sample, from which distributional properties, or summary statistical measures, can be derived. The Metropolis-Hastings based MCMC approach is compared and contrasted with two other sequential data assimilation approaches: sequential importance resampling, and the (approximate) ensemble Kalman filter (including computational comparisons). A simple illustrative application is provided based on a 0-D nonlinear plankton ecosystem model with multivariate non-Gaussian observations of the ecosystem state from a coastal ocean observatory. The MCMC approach is shown to be straightforward to implement and to effectively characterize the non-Gaussian ecosystem state in both nowcast and forecast experiments. Results are reported which illustrate how non-Gaussian information originates, and how it can be used to characterize ecosystem properties.     

Modelling the hydrodynamics and ecosystem of the North-West European continental shelf for operational oceanography

This paper outlines an approach to complex spatio-temporal marine ecosystem modelling as applied to the North Western European Continental Shelf. The model presented here combines an eddy-permitting (approximately 6 km horizontal resolution) baroclinic model, the Proudman Oceanographic Laboratory Coastal Ocean Modelling System (POLCOMS), with the European Regional Seas Ecosystem Model (ERSEM). This has been run within an operational framework using operationally available high resolution atmospheric and lateral boundary forcing, allowing hindcast and near-real time nowcast simulations to be performed. The modelled surface temperature and chlorophyll distributions are presented, and interannual variations discussed. Validation of both the physical and ecosystem submodels show the system to be effective, whilst highlighting areas where improvements in the system can be made. Distinct regional differences in predictive skill are shown. The system presented is ready for operational implementation to provide products and services for use both scientifically and in coastal zone and shelf seas management activities. A programme of work to update the system is already in place.     

Modelling the spatial and temporal variability of the Cretan Sea ecosystem

The ecosystem function of the oligotrophic Cretan Sea is explored through the development and application of a 3D ecological model. The simulation system comprises of two on-line coupled submodels: the 3D Princeton Ocean Model (POM) and the 1D European Regional Seas Ecosystem Model (ERSEM) adapted to the Cretan Sea. For the tuning and initialisation of the ecosystem parameters, the 1D version of the biogeochemical model is used.After a model spin up period of 10 years to reach a quasi-steady state, the results from an annual simulation are presented. A cost function is used as validation method for the comparison of model results with field data. The estimated annual primary and bacteria production are found to be in the range of the reported values. Simulation results are in good agreement with in situ data illustrating the role of the physical processes in determining the evolution and variability of the ecosystem.     

Modelling Pseudocalanus elongatus stage-structured population dynamics embedded in a water column ecosystem model for the northern North Sea

This paper outlines an approach to couple a structured zooplankton population model with state variables for eggs, nauplii, two copepodites stages and adults adapted to Pseudocalanus elongatus into the complex marine ecosystem model ECOHAM2 with 13 state variables resolving the carbon and nitrogen cycle. Different temperature and food scenarios derived from laboratory culture studies were examined to improve the process parameterisation for copepod stage dependent development processes. To study annual cycles under realistic weather and hydrographic conditions, the coupled ecosystem–zooplankton model is applied to a water column in the northern North Sea. The main ecosystem state variables were validated against observed monthly mean values. Then vertical profiles of selected state variables were compared to the physical forcing to study differences between zooplankton as one biomass state variable or partitioned into five population state variables. Simulated generation times are more affected by temperature than food conditions except during the spring phytoplankton bloom. Up to six generations within the annual cycle can be discerned in the simulation.     

Description of a flexible and extendable physical–biogeochemical model system for the water column

A modelling system for coupled physical–biogeochemical simulations in the water column is presented here. The physical model component allows for a number of different statistical turbulence closure schemes, ranging from simple algebraic closures to two-equation turbulence models with algebraic second-moment closures. The biogeochemical module consists of models which are based on a number of state variables represented by their ensemble averaged concentrations. Specific biogeochemical models may range from simple NPZ (nutrient–phytoplankton–zooplankton) to complex ecosystem models. Recently developed modified Patankar solvers for ordinary differential equations allow for stable discretisations of the production and destruction terms guaranteeing conservative and non-negative solutions. The increased stability of these new solvers over explicit solvers is demonstrated for a plankton spring bloom simulation. The model system is applied to marine ecosystem dynamics the Northern North Sea and the Central Gotland Sea. Two different biogeochemical models are applied, a conservative nitrogen-based model to the North Sea, and a more complex model including an oxygen equation to the Baltic Sea, allowing for the reproduction of chemical processes under anoxic conditions. For both applications, earlier model results obtained with slightly different model setups could be basically reproduced. It became however clear that the choice for ecosystem model parameters such as maximum phytoplankton growth rates does strongly depend on the physical model parameters (such as turbulence closure models or external forcing).     

A generalized model of pelagic biogeochemistry for the global ocean ecosystem. Part II: Numerical simulations

This paper presents a global ocean implementation of a multi-component model of marine pelagic biogeochemistry coupled on-line with an ocean general circulation model forced with climatological surface fields (PELAgic biogeochemistry for Global Ocean Simulations, PELAGOS). The final objective is the inclusion of this model as a component in an Earth System model for climate studies. The pelagic model is based on a functional stoichiometric representation of marine biogeochemical cycles and allows simulating the dynamics of C, N, P, Si, O and Fe taking into account the variation of their elemental ratios in the functional groups. The model also includes a parameterization of variable chlorophyll/carbon ratio in phytoplankton, carrying chl as a prognostic variable. The first part of the paper analyzes the contribution of non-local advective–diffusive terms and local vertical processes to the simulated chl distributions. The comparison of the three experiments shows that the mean chl distribution at higher latitudes is largely determined by mixing processes, while vertical advection controls the distribution in the equatorial upwelling regions. Horizontal advective and diffusive processes are necessary mechanisms for the shape of chl distribution in the sub-tropical Pacific. In the second part, the results have been compared with existing datasets of satellite-derived chlorophyll, surface nutrients, estimates of phytoplankton community composition and primary production data. The agreement is reasonable both in terms of the spatial distribution of annual means and of the seasonal variability in different dynamical oceanographic regions. Results indicate that some of the model biases in chl and surface nutrients distributions can be related to deficiencies in the simulation of physical processes such as advection and mixing. Other discrepancies are attributed to inadequate parameterizations of phytoplankton functional groups. The model has skill in reproducing the overall distribution of large and small phytoplankton but tends to underestimate diatoms in the northern higher latitudes and overestimate nanophytoplankton with respect to picoautotrophs in oligotrophic regions. The performance of the model is discussed in the context of its use in climate studies and an approach for improving the parameterization of functional groups in deterministic models is outlined.     

Weaving marine food webs from end to end under global change

Marine food web dynamics are determined by interactions within and between species and between species and their environment. Global change directly affects abiotic conditions and living organisms, impinging on all trophic levels in food webs. Different groups of marine researchers traditionally study different aspects of these changes. However, over medium to long time scales perturbations affecting food webs need to be considered across the full range from nutrients to top predators. Studies of end-to-end marine food webs not only span organism sizes and trophic levels, but should also help align multidisciplinary research to common goals and perspectives. Topics are described that bridge disciplinary gaps and are needed to develop new understanding of the reciprocal impacts of global change on marine food webs and ocean biogeochemistry. These include (1) the effects of nutrients on biomass and production, (2) the effects of varying element ratios on food web structure and food quality, (3) bulk flows of energy and material in food webs and their efficiencies of transfer, (4) the ecological effects of species richness and the roles of microbial organisms, (5) the role of feeding behaviour in food web dynamics and trophic controls, (6) the spatial dynamics of communities and links between different food webs, (7) the combined effects of body size and behaviour in determining dynamics of food webs, and (8) the extent to which the ability of marine organisms (and communities) to adapt will influence food web dynamics. An overriding issue that influences all topics concerns the time and space scales of ecosystem variability. Threads link different nodes of information among various topics, emphasizing the importance of tackling food web studies with a variety of modelling approaches and through a combination of field and experimental studies with a strong comparative approach.     

海洋生态系统及海岸工程生态影响预测模型研究进展

对海洋生态系统的研究已形成了以实验观测为基础、建立生态模型的方法。文章对已有的海洋生态模型研究成果进行了归纳总结,认为初级生产力模型是纯生态模型的主要发展方向,综合考虑物理生化作用的生态系统动力学模型将是未来海洋生态研究的必然趋势。同时,由于我国大规模海岸工程建设对近岸海域生态系统造成了较大影响,而相关的生态影响预测模型研究较少,为有效预测评价海岸工程建设对近岸海域生态系统的影响,进一步深入研究海岸工程建设对近岸海洋生态影响预测模型将显得尤为重要。     

Gaussian anamorphosis extension of the DEnKF for combined state parameter estimation: Application to a 1D ocean ecosystem model

We consider the problem of combined state-parameter estimations in biased nonlinear models with non-Gaussian extensions of the Deterministic Ensemble Kalman Filter (DEnKF). We focus on the particular framework of ocean ecosystem models. Such models present important obstacles to the use of data assimilation methods based on Kalman filtering due to the non-linearity of the models, the constraints of positiveness that apply to the variables and parameters, and the non-Gaussian distribution of the variables in which they result.We present extensions of the DEnKF dealing with these difficulties by introducing a nonlinear change of variables (anamorphosis function) in order to execute the analysis step with Gaussian transformed variables and parameters. Several strategies to build the anamorphosis functions are investigated and compared within the framework of twin experiments realized in a simple 1D ocean ecosystem model. A solution to the problem of the specification of the observation error for transformed observations is suggested. The study highlights the inability of the plain DEnKF with a simple post-processing of the negative values to properly estimate parameters when constraints of positiveness apply to the variables. It goes on to show that the introduction of the Gaussian anamorphosis can remedy these assimilation biases.     

Biological control of harmful algal blooms: A modelling study

A multispecies dynamic simulation model (ERSEM) was used to examine the influence of allelopathic and trophic interactions causing feeding avoidance by predators, on the formation of harmful algal blooms, under environmental scenarios typical of a Mediterranean harbour (Barcelona). The biological state variables of the model included four functional groups of phytoplankton (diatoms, toxic and non-toxic flagellates and picophytoplankton), heterotrophic flagellates, micro- and mesozooplankton and bacteria. The physical–chemical forcing (irradiance, temperature and major nutrient concentrations) was based on an actual series of measurements taken along a year cycle in the Barcelona harbour. In order to evaluate potential effects of advection, some runs were repeated after introducing a biomass loss term. Numerical simulations showed that allelopathic effects of a toxic alga on a non-toxic but otherwise similar competitor did not have appreciable influence on the dynamics of the system. However, induction of avoidance of the toxic alga by predators, which resulted on increased predation pressure on other algal groups had a significant effect on the development of algal and predator populations. The presence of advection overrided the effect of these interactions and only allowed organisms with sufficiently high potential growth rates to thrive.     

Future aspects in marine ecosystem modelling

Existing ecosystem models are briefly presented and summarised. The problem of coupling physical and biological models as well as aspects of prediction and predictability are discussed. The general perception that marine ecosystems are inherently unpredictable due to non-linearity becomes questionable if the response of climate variability in marine ecosystems is analysed. Many authors have shown correlations between climate variability and the variability of abundance or biomass of marine organisms such as phytoplankton, zooplankton, benthos or fish recruitment in different parts of the world ocean. In the northern hemisphere, certain species show a linear response to climate variability mainly during winter and spring. However, the underlying mechanisms are not well understood. Often, a phase lag can be observed between climate variability and the reaction of organisms. The identification of a plausible mediator between climate and biology is difficult, since all possible physicochemical mechanisms having a direct or indirect influence on the variability of abundance or biomass of marine organisms have to be considered as mediator.The understanding of the reason of the phase lag, which possibly implies a “biological memory”, and the identification of all possible mediators are necessary to predict the response of marine organisms to climate variability. The identification of mediators will result in an improvement of coupled models, a deeper understanding of physical–biological interaction and the improvement of predictive capability of marine ecosystem models.     

Axiomatic design approach for marine design problems

The axiomatic design approach proposed by N.P. Suh consists of the Independence Axiom and the Information Axiom. The Independence Axiom assists a designer in generating good design alternatives by considering the relationships between the functions and the physical product using a hierarchical mapping procedure. The Information Axiom, which is related to the probability of achieving the given functional requirements, can be used as a criterion for the selection of the best solution among the proposed alternatives in the conceptual or preliminary design stage.In the early stages of marine design, especially ship design, there exists a lot of uncertainty because of the size and complexity of a marine vehicle. This uncertainty often leads to a probabilistic approach rather than a deterministic approach. Ship designs are often routine, requiring small modifications of an existing, successful design. The availability of axiomatic design in this marine design field has been investigated through four examples: a conceptual thruster design, a foil-strut optimization problem, a main engine selection problem, and a barge design problem. By these examples, the possibility of the application of Design Axioms in marine design has been illustrated.     

组态式船舶电力推进系统仿真软件设计与实现

针对船舶电力推进系统方案论证效率低、成本高等问题,开发了一款组态式船舶电力推进系统仿真软件。对船舶电力推进系统进行结构划分,并建立了系统的数学模型和仿真模型,并封装形成模型库。使用Matlab/GUI开发界面层,用M语言对Simulink模型进行调用与控制。使用SQL Server 2008创建数据库,实现数据的存取和管理功能。基于该软件对某船舶电力推进系统设计方案进行组态式的图形化建模,并设计了典型工况的仿真实验。结果表明,仿真模型较好预报了实船运行工况。由此可见,该软件可有效验证船舶电力推进系统设计方案。     

Monitoring marine plankton ecosystems: Identification of the most relevant indicators of the state of an ecosystem

The number of variables involved in the monitoring of an ecosystem can be high and often one of the first stages in the analysis is to reduce the number of variables. We describe a method developed for geological purposes, using the information theory, that enables selection of the most relevant variables. This technique also allows the examination of the asymmetrical relationships between variables. Applied to a set of physical and biological variables (plankton assemblages in four areas of the North Sea), the method shows that biological variables are more informative than physical variables although the controlling factors are mainly physical (sea surface temperature in winter and spring). Among biological variables, diversity measures and warm-water species assemblages are informative for the state of the North Sea pelagic ecosystems while among physical variables sea surface temperature in late winter and early spring are highly informative. Although often used in bioclimatology, the utilisation of the North Atlantic Oscillation (NAO) index does not seem to provide a lot of information. The method reveals that only the extreme states of this index has an influence on North Sea pelagic ecosystems. The substantial and persistent changes that were detected in the dynamic regime of the North Sea ecosystems and called regime shift are detected by the method and corresponds to the timing of other shifts described in the literature for some European Systems such as the Baltic and the Mediterranean Sea when both physical and biological variables are considered.     

Analysis of Ocean-Space and Sea-Level Rise Policy in Two Coastal Cities

Global sea-level rise (SLR) is among the most alarming aspects of anthropogenic climate change. The human impacts of SLR are experienced unequally between and within municipalities. Existing research has identified social variables that predict municipal adoption of adaptive SLR policy, but this work does not account for the locally specific social factors that shape particular policies to fit particular cases. This study describes social conceptions of the ocean and SLR policy for two coastal cities in western Washington: Aberdeen and Bainbridge Island. Examining conceptions of marine spaces provides insight into the complex process by which local physical and socio-demographic characteristics shape local policy. The study uses a grounded theoretical approach to content analyses, resulting in localized typologies of marine spaces as well as SLR policy profiles for each case. Results indicate that municipalities vary by both social conception of the ocean and SLR policy form. These findings elaborate upon the relationship between socioeconomic conditions and municipal climate change policy adoption, suggesting that in local media and policy discourse, the absence of relevant resources is related to adversarial conceptions of local marine spaces, while abundance of relevant resources is associated with a conception of marine spaces as natural resources.     

The Necessity of Establishing a Regional Marine Research Program for the U.S. West Coast

The welfare of the marine environment of the U.S. West Coast of California, Oregon, and Washington is hindered by insufficient understanding of ecosystem dynamics and human pressures. Greater coordination, integration, and support of mutually identified research priorities across research organizations on the West Coast are necessary. This level of collaboration is essential to responding to regional ecosystem threats, such as the impacts of human occupation of coastal areas, over-harvesting of marine resources, and climate change. To address this need for collaboration and build on the accomplishments of existing research organizations, this article advocates the establishment of a West Coast regional marine research program (RMRP).     

Sensitivity study on durability variables of marine concrete structures

In order to study the influence of parameters on durability of marine concrete structures, the parameter’s sensitivity analysis was studied in this paper. With the Fick’s 2nd law of diffusion and the deterministic sensitivity analysis method (DSA), the sensitivity factors of apparent surface chloride content, apparent chloride diffusion coefficient and its time dependent attenuation factor were analyzed. The results of the analysis show that the impact of design variables on concrete durability was different. The values of sensitivity factor of chloride diffusion coefficient and its time dependent attenuation factor were higher than others. Relative less error in chloride diffusion coefficient and its time dependent attenuation coefficient induces a bigger error in concrete durability design and life prediction. According to probability sensitivity analysis (PSA), the influence of mean value and variance of concrete durability design variables on the durability failure probability was studied. The results of the study provide quantitative measures of the importance of concrete durability design and life prediction variables. It was concluded that the chloride diffusion coefficient and its time dependent attenuation factor have more influence on the reliability of marine concrete structural durability. In durability design and life prediction of marine concrete structures, it was very important to reduce the measure and statistic error of durability design variables.     

The Importance of Brokers in Collaborative Marine Ecosystem-Based Management of Social Ecological Systems

Abstract

Practitioners play a critical yet largely unexamined role in translating collaborative, ecosystem-based management (EBM) for social-ecological systems from theory to practice. We paired mental models and social network analytical methods and applied them to two cases of marine EBM in Rhode Island and New York, focusing on practitioners themselves, to understand the relationship between practitioners’ mental models of marine ecosystems and the extent and nature of collaboration within each network. Mental models analysis was used to assess the comprehensiveness and balance of practitioners’ mental models, and social network analysis was used to assess the role and influence of practitioners within each network. A comparative statistical analysis was then performed to understand the relationship between mental models and network measures. Research revealed a statistically significant correlation between practitioners’ mental model comprehensiveness and practitioners’ influence within the network. In other words, practitioners with comprehensive mental models of the ecosystem for which they were planning were found to act as “brokers,” connecting those who were not otherwise connected, bridging jurisdictions, sectors, and disciplines. Results underscore the importance of brokers in achieving the collaborative and integrated goals of EBM and suggest the need for greater attention to practitioners’ role in EBM implementation.     

基于蜂窝单元的海上交通风险分析

传统的船舶交通风险分析是将研究水域作为一个整体来评估其风险状况的。基于蜂窝单元的船舶交通风险分析,把研究水域按一定标准划分为若干地理单元,将历史数据、数学模型以及专家学者的经验和判断有机地结合起来,用以评估每一地理单元船舶碰撞和搁浅事故的概率,其输出结果以地理分布的形式突出了高风险区域。相对于传统的风险分析,该理论更能够全面、准确地描述水域的风险状况,为采纳和实施正确的安全措施提供了强有力的支持。     

Planktonic community structure and carbon cycling in the Arabian Sea as a result of monsoonal forcing: the application of a generic model

The Arabian Sea exhibits a complex pattern of biogeochemical and ecological dynamics, which vary both seasonally and spatially. These dynamics have been studied using a one-dimensional vertical hydrodynamic model coupled to a complex ecosystem model, simulating the annual cycle at three contrasting stations. These stations are characterised by seasonally upwelling, mixed-layer-deepening and a-seasonal oligotrophic conditions, respectively, and coincide with extensively measured stations on the two JGOFS ARABESQUE cruises in 1994. The model reproduces many spatial and temporal trends in production, biomass, physical and chemical properties, both qualitatively and quantitatively and so gives insight into the main mechanisms responsible for the biogeochemical and ecological complexity. Monsoonal systems are typified by classical food web dynamics, whilst intermonsoonal and oligotrophic systems are dominated by the microbial loop. The ecosystem model (ERSEM), developed for temperate regions, is found to be applicable to the Arabian Sea system with little reparameterisation. Differences in in-situ physical forcing are sufficient to recreate contrasting eutrophic and oligotrophic systems, although the lack of lateral terms are probably the greatest source of error in the model. Physics, nutrients, light and grazing are all shown to play a role in controlling production and community structure. Small-celled phytoplanktons are predicted to be dominant and sub-surface chlorophyll maxima are robust centers of production during intermonsoon periods. Analysis of carbon fluxes indicate that physically driven outgassing of CO₂ predominates in monsoonal upwelling systems but ecological activity may significantly moderate CO₂ outgassing in the Arabian Sea interior.