Tillin, H. & Tyler-Walters, H., 2014. Assessing the sensitivity of subtidal sedimentary habitats to pressures associated with marine activities. Phase 2 Report – Literature review and sensitivity assessments for ecological groups for circalittoral and offshore Level 5 biotopes. JNCC Report No. 512B, 260 pp.

Human activities within the marine environment give rise to a number of pressures on seabed habitats. Improved understanding of the sensitivity of subtidal sedimentary habitats is required to underpin the management advice provided for Marine Protected Areas, as well as supporting other UK marine monitoring and assessment work. The sensitivity of marine sedimentary habitats to a range of pressures induced by human activities has previously been systematically assessed using approaches based on expert judgement for Defra Project MB0102 (Tillin et al. 2010). This previous work assessed sensitivity at the level of the broadscale habitat and therefore the scores were typically expressed as a range due to underlying variation in the sensitivity of the constituent biotopes. The objective of this project was to reduce the uncertainty around identifying the sensitivity of selected subtidal sedimentary habitats by assessing sensitivity, at a finer scale and incorporating information on the biological assemblage, for 33 Level 5 circalittoral and offshore biotopes taken from the Marine Habitat Classification of Britain and Ireland (Connor et al. 2004). Two Level 6 sub-biotopes were also included in this project as these contain distinctive characterising species that differentiate them from the Level 5 parent biotope. Littoral, infralittoral, reduced and variable salinity sedimentary habitats were excluded from this project as the scope was set for assessment of circalittoral and offshore sedimentary communities. This project consisted of three Phases. • Phase 1 - define ecological groups based on similarities in the sensitivity of characterising species from the Level 5 and two Level 6 biotopes described above. • Phase 2 - produce a literature review of information on the resilience and resistance of characterising species of the ecological groups to pressures associated with activities in the marine environment. • Phase 3 - to produce sensitivity assessment ‘proformas’ based on the findings of Phase 2 for each ecological group. This report outlines results of Phase 2. The Tillin et al., (2010) sensitivity assessment methodology was modified to use the best available scientific evidence that could be collated within the project timescale. An extensive literature review was compiled, for peer reviewed and grey literature, to examine current understanding about the effects of pressures from human activities on circalittoral and offshore sedimentary communities in UK continental shelf waters, together with information on factors that contribute to resilience (recovery) of marine species. This review formed the basis of an assessment of the sensitivity of the 16 ecological groups identified in Phase 1 of the project (Tillin & Tyler-Walters 2014). As a result: • the state of knowledge on the effects of each pressure on circalittoral and offshore benthos was reviewed; • the resistance, resilience and, hence, sensitivity of sixteen ecological groups, representing 96 characteristic species, were assessed for eight separate pressures; • each assessment was accompanied by a detailed review of the relevant evidence; Assessing the sensitivity of subtidal sedimentary habitats to pressures associated with human activities • knowledge gaps and sources of uncertainty were identified for each group; • each assessment was accompanied by an assessment of the quality of the evidence, its applicability to the assessment and the degree of concordance (agreement) between the evidence, to highlight sources of uncertainty as an assessment of the overall confidence in the sensitivity assessment, and finally • limitations in the methodology and the application of sensitivity assessments were outlined. This process demonstrated that the ecological groups identified in Phase 1 (Tillin & Tyler-Walters 2014) were viable groups for sensitivity assessment, and could be used to represent the 33 circalittoral and offshore sediments biotopes identified at the beginning of the project. The results of the sensitivity assessments show: • the majority of species and hence ecological groups in sedimentary habitats are sensitive to physical change, especially loss of habitat and sediment extraction, and change in sediment type; • most sedimentary species are sensitive to physical damage, e.g. abrasion and penetration, although deep burrowing species (e.g. the Dublin Bay prawn - Nephrops norvegicus and the sea cucumber - Neopentadactyla mixta) are able to avoid damaging effects to varying degrees, depending on the depth of penetration and time of year; • changes in hydrography (wave climate, tidal streams and currents) can significantly affect sedimentary communities, depending on whether they are dominated by deposit, infaunal feeders or suspension feeders, and dependant on the nature of the sediment, which is itself modified by hydrography and depth; • sedentary species and ecological groups that dominate the top-layer of the sediment (either shallow burrowing or epifaunal) remain the most sensitive to physical damage; • mobile species (e.g. interstitial and burrowing amphipods, and perhaps cumaceans) are the least sensitive to physical change or damage, and hydrological change as they are already adapted to unstable, mobile substrata; • sensitivity to changes in organic enrichment and hence oxygen levels, is variable between species and ecological groups, depending on the exact habitat preferences of the species in question, although most species have at least a medium sensitivity to acute deoxygenation; • there is considerable evidence on the effects of bottom-contact fishing practices and aggregate dredging on sedimentary communities, although not all evidence is directly applicable to every ecological group; • there is lack of detailed information on the physiological tolerances (e.g. to oxygenation, salinity, and temperature), habitat preferences, life history and population dynamics of many species, so that inferences has been made from related species, families, or even the same phylum; • there was inadequate evidence to assess the effects of non-indigenous species on most ecological groups, and Assessing the sensitivity of subtidal sedimentary habitats to pressures associated with human activities • there was inadequate evidence to assess the effects of electromagnetic fields and litter on any ecological group. The resultant report provides an up-to-date review of current knowledge about the effects of pressures resulting from human activities of circalittoral and offshore sedimentary communities. It provides an evidence base to facilitate and support the provision of management advice for Marine Protected Areas, development of UK marine monitoring and assessment, and conservation advice to offshore marine industries. However, such a review will require at least annual updates to take advantage of new evidence and new research as it becomes available. Also further work is required to test how ecological group assessments are best combined in practice to advise on the sensitivity of a range of sedimentary biotopes, including the 33 that were originally examined.

Disentangling the impacts of heat wave magnitude, duration and timing on the structure and diversity of sessile marine assemblages

Extreme climatic events, including heat waves (HWs) and severe storms, influence the structure of marine and terrestrial ecosystems. Despite growing consensus that anthropogenic climate change will increase the frequency, duration and magnitude of extreme events, current understanding of their impact on communities and ecosystems is limited. Here, we used sessile invertebrates on settlement panels as model assemblages to examine the influence of HW magnitude, duration and timing on marine biodiversity patterns. Settlement panels were deployed in a marina in southwest UK for ≥5 weeks, to allow sufficient time for colonisation and development of sessile fauna, before being subjected to simulated HWs in a mesocosm facility. Replicate panel assemblages were held at ambient sea temperature (∼17 °C), or +3 °C or +5 °C for a period of 1 or 2 weeks, before being returned to the marina for a recovery phase of 2–3 weeks. The 10-week experiment was repeated 3 times, staggered throughout summer, to examine the influence of HW timing on community impacts. Contrary to our expectations, the warming events had no clear, consistent impacts on the abundance of species or the structure of sessile assemblages. With the exception of 1 high-magnitude long-duration HW event, warming did not alter not assemblage structure, favour non-native species, nor lead to changes in richness, abundance or biomass of sessile faunal assemblages. The observed lack of effect may have been caused by a combination of (1) the use of relatively low magnitude, realistic heat wave treatments compared to previous studies (2), the greater resilience of mature adult sessile fauna compared to recruits and juveniles, and (3) the high thermal tolerance of the model organisms (i.e., temperate fouling species, principally bryozoans and ascidians). Our study demonstrates the importance of using realistic treatments when manipulating climate change variables, and also suggests that biogeographical context may influence community-level responses to short-term warming events, which are predicted to increase in severity in the future.

Seasonal shelf-sea front mapping using satellite ocean colour and temperature to support development of a marine protected area network

Front detection and aggregation techniques were applied to 300m resolution MERIS satellite ocean colour data for the first time, to describe frequently occurring shelf-sea fronts near to the Scottish coast. Medium resolution (1km) thermal and colour data have previously been used to analyse the distribution of surface fronts, though these cannot capture smaller frontal zones or those in close proximity to the coast, particularly where the coastline is convoluted. Seasonal frequent front maps, derived from both chlorophyll and SST data, revealed a number of key frontal zones, a subset of which were based on new insights into the sediment and plankton dynamics provided exclusively by the higher-resolution chlorophyll fronts. The methodology is described for applying colour and thermal front data to the task of identifying zones of ecological importance that could assist the process of defining marine protected areas. Each key frontal zone is analysed to describe its spatial and temporal extent and variability, and possible mechanisms. It is hoped that these tools can provide guidance on the dynamic habitats of marine fauna towards aspects of marine spatial planning and conservation.

Investigation of benthic community change over a century-wide scale in the western English Channel

Since the early part of the 20th Century the impact of a range of anthropogenic activities in our coastal seas has steadily increased. The effect of such activities is a major cause for concern but in the benthic environment few studies exist that date back more than a few decades. Hence understanding long term changes is a challenge. Within this study we utilized a historic benthic dataset and resurveyed an area west of Eddystone reef in the English Channel previously investigated 112 years ago. The aim of the present work was to describe the current benthic community structure and investigate potential differences between 1895 and 2007. For each of the four major phyla investigated (Polychaeta, Crustacea, Mollusca and Echinodermata), multivariate community analysis showed significant differences between the historic and contemporary surveys. Echinoderm diversity showed a clear reduction between 1895 and 2007. The sea urchins Echinus esculentus, Spatangus purpureus, and Psammechinus miliaris and large star-fish Marthasterias glacialis showed reductions in abundance, in some cases being entirely absent from the survey area in 2007. Polychaetes showed a shift from tubiculous species to small errant and predatory species such as Glycera, Nephtys, and Lumbrineris spp. Within the group Mollusca large species such as Pecten maximus and Laevicardium crassum decreased in abundance while small species increased. Crustaceans in 1895 were dominated by crab species which were present in similar abundances in 2007, but, the order Amphipoda appeared to show a significant increase. While some of the differences observed could stem from differences in methodologies between the surveys, in particular increases of small cryptic species, the loss of large conspicuous species was judged to be genuine. The study area is an important beam trawling and scallop dredging ground; the differences observed are concomitant with changes generally associated with disturbance from demersal fishing activities such as these.

Probabilistic Models to Describe the Dynamics of Migrating Microbial Communities

In all but the most sterile environments bacteria will reside in fluid being transported through conduits and some of these will attach and grow as biofilms on the conduit walls. The concentration and diversity of bacteria in the fluid at the point of delivery will be a mix of those when it entered the conduit and those that have become entrained into the flow due to seeding from biofilms. Examples include fluids through conduits such as drinking water pipe networks, endotracheal tubes, catheters and ventilation systems. Here we present two probabilistic models to describe changes in the composition of bulk fluid microbial communities as they are transported through a conduit whilst exposed to biofilm communities. The first (discrete) model simulates absolute numbers of individual cells, whereas the other (continuous) model simulates the relative abundance of taxa in the bulk fluid. The discrete model is founded on a birth-death process whereby the community changes one individual at a time and the numbers of cells in the system can vary. The continuous model is a stochastic differential equation derived from the discrete model and can also accommodate changes in the carrying capacity of the bulk fluid. These models provide a novel Lagrangian framework to investigate and predict the dynamics of migrating microbial communities. In this paper we compare the two models, discuss their merits, possible applications and present simulation results in the context of drinking water distribution systems. Our results provide novel insight into the effects of stochastic dynamics on the composition of non-stationary microbial communities that are exposed to biofilms and provides a new avenue for modelling microbial dynamics in systems where fluids are being transported.

Lateral diffusivity coefficients from the dynamics of a SF6 patch in a coastal environment

The dispersion of a patch of the tracer sulfur hexafluoride (SF6) is used to assess the lateral diffusivity in the coastal waters of the western part of the Gulf of Lion (GoL), northwestern Mediterranean Sea, during the Latex10 experiment (September 2010). Immediately after the release, the spreading of the patch is associated with a strong decrease of the SF6 concentrations due to the gas exchange from the ocean to the atmosphere. This has been accurately quantified, evidencing the impact of the strong wind conditions during the first days of this campaign. Few days after the release, as the atmospheric loss of SF6 decreased, lateral diffusivity coefficient at spatial scales of 10 km has been computed using two approaches. First, the evolution of the patch with time was combined with a diffusion-strain model to obtain estimates of the strain rate (γ = 2.5 10- 6 s- 1) and of the lateral diffusivity coefficient (Kh = 23.2 m2s− 1). Second, a steady state model was applied, showing Kh values similar to the previous method after a period of adjustment between 2 and 4.5 days. This implies that after such period, our computation of Kh becomes insensitive to the inclusion of further straining of the patch. Analysis of sea surface temperature satellite imagery shows the presence of a strong front in the study area. The front clearly affected the dynamics within the region and thus the temporal evolution of the patch. Our results are consistent with previous studies in open ocean and demonstrate the success and feasibility of those methods also under small-scale, rapidly-evolving dynamics typical of coastal environments.

Ocean acidification has different effects on the production of DMS and DMSP measured in cultures of Emiliania huxleyi and a mesocosm study: a comparison of laboratory monocultures and community interactions

The human-induced rise in atmospheric carbon dioxide since the industrial revolution has led to increasing oceanic carbon uptake and changes in seawater carbonate chemistry, resulting in lowering of surface water pH. In this study we investigated the effect of increasing CO2 partial pressure (pCO2) on concentrations of volatile biogenic dimethylsulfide (DMS) and its precursor dimethylsulfoniopropionate (DMSP), through monoculture studies and community pCO2 perturbation. DMS is a climatically important gas produced by many marine algae: it transfers sulfur into the atmosphere and is a major influence on biogeochemical climate regulation through breakdown to sulfate and formation of subsequent cloud condensation nuclei (CCN). Overall, production of DMS and DMSP by the coccolithophore Emiliania huxleyi strain RCC1229 was unaffected by growth at 900 matm pCO2, but DMSP production normalised to cell volume was 12% lower at the higher pCO2 treatment. These cultures were compared with community DMS and DMSP production during an elevated pCO2 mesocosm experiment with the aim of studying E. huxleyi in the natural environment. Results contrasted with the culture experiments and showed reductions in community DMS and DMSP concentrations of up to 60 and 32% respectively at pCO2 up to 3000 matm, with changes attributed to poorer growth of DMSP-producing nanophytoplankton species, including E. huxleyi, and potentially increased microbial consumption of DMSand dissolvedDMSPat higher pCO2.DMSandDMSPproduction differences between culture and community likely arise from pH affecting the inter-species responses between microbial producers and consumers.

The combined effects of seasonal community succession and adaptive algal physiology on lipid profiles of coastal phytoplankton in the western English channel

Lipids are key constituents of marine phytoplankton, and some fatty acids (key constituents of lipids) are essential dietary components for secondary producers. However, in natural marine ecosystems the interactions of factors affecting seasonal phytoplankton lipid composition are still poorly understood. The aim of this study was to assess the roles of seasonal succession in phytoplankton community composition and nutrient concentrations, on the lipid composition of the phytoplankton community. Fatty acid and polar lipid composition in seston was measured in surface waters at the time series station L4, an inshore station in the Western English Channel, from January to December 2013. Redundancy analyses (RDA) were used to identify factors (abiotic and biotic) that explained the seasonal variability in phytoplankton lipids. RDA demonstrated that nutrients (namely nitrogen) explained the majority of variation in phytoplankton lipid composition, as well as a smaller explanatory contribution from changes in phytoplankton community composition. The physiological adaptations of the phytoplankton community to nutrient deplete conditions during the summer season when the water column was stratified, was further supported by changes in the polar lipid to phytoplankton biomass ratios (also modelled with RDA) and increases in the lipid to chlorophyll a ratios, which are both indicative of nutrient stress. However, the association of key fatty acid markers with phytoplankton groups e.g. 22:6 n-3 and dinoflagellate biomass (predominant in summer), meant there were no clear seasonal differences in the overall degree of fatty acid saturation, as might have been expected from typical nutrient stress on phytoplankton. Based on the use of polyunsaturated fatty acids (PUFA) as markers of ‘food quality’ for grazers, our results suggest that in this environment high food quality is maintained throughout summer, due to seasonal succession towards flagellated phytoplankton species able to maintain PUFA synthesis under surface layer nutrient depletion.

The assessment of a global marine ecosystem model on the basis of emergent properties and ecosystem function: a case study with ERSEM

Ecosystem models are often assessed using quantitative metrics of absolute ecosystem state, but these model-data comparisons are disproportionately vulnerable to discrepancies in the location of important circulation features. An alternative method is to demonstrate the models capacity to represent ecosystem function; the emergence of a coherent natural relationship in a simulation indicates that the model may have an appropriate representation of the ecosystem functions that lead to the emergent relationship. Furthermore, as emergent properties are large-scale properties of the system, model validation with emergent properties is possible even when there is very little or no appropriate data for the region under study, or when the hydrodynamic component of the model differs significantly from that observed in nature at the same location and time. A selection of published meta-analyses are used to establish the validity of a complex marine ecosystem model and to demonstrate the power of validation with emergent properties. These relationships include the phytoplankton community structure, the ratio of carbon to chlorophyll in phytoplankton and particulate organic matter, the ratio of particulate organic carbon to particulate organic nitrogen and the stoichiometric balance of the ecosystem. These metrics can also inform aspects of the marine ecosystem model not available from traditional quantitative and qualitative methods. For instance, these emergent properties can be used to validate the design decisions of the model, such as the range of phytoplankton functional types and their behaviour, the stoichiometric flexibility with regards to each nutrient, and the choice of fixed or variable carbon to nitrogen ratios.

A molecular method for the identification of resting eggs of acartiid copepods in the Thau lagoon, France

Acartia and Paracartia species, often known to co-occur, can exhibit complex life cycles, including the production of resting eggs. Studying and understanding their population dynamics is hindered by the inability to identify eggs and early developmental stages using morphological techniques. We have developed a simple molecular technique to distinguish between the three species of the Acartiidae family (Acartia clausi, A. discaudata and Paracartia grani) that co-occur in the Thau lagoon (43�250N; 03�400E) in southern France. Direct amplification of a partial region of the mitochondrial cytochrome oxidase I gene by polymerase chain reaction and subsequent restriction fragment length polymorphism results in a unique restriction profile for each species. The technique is capable of determining the identity of individual eggs, including resting eggs retrieved from sediment samples, illustrating its application in facilitating population dynamic studies of this ubiquitous and important member of the zooplankton community.

Abundance of a chlorophyll a precursor and the oxidation product hydroxychlorophyll a during seasonal phytoplankton community progression in the Western English Channel

This study presents the first in-situ measurements of the chlorophyll a oxidation product, hydroxychlorophyll a as well as the chlorophyll a precursor, chlorophyll aP276 conducted over an annual cycle. Chlorophyll a oxidation products, such as hydroxychlorophyll a may be associated with the decline of algal populations and can act as an initial step in the degradation of chlorophyll a into products which can be found in the geochemical record, important for studying past climate change events. Here, hydroxychlorophyll a and chlorophyll aP276 were measured at the long-term monitoring station L4, Western Channel Observatory (UK, www.westernchannelobservatory.org) over an annual cycle (2012). Weekly measurements of phytoplankton species composition and abundance enabled detailed analysis of possible sources of hydroxychlorophyll a. Dinoflagellates, 2 diatom species, the prymnesiophyte Phaeocystis spp. and the coccolithophorid Emiliania huxleyi were all associated with hydroxychlorophyll a occurrence. However, during alternate peaks in abundance of the diatoms, no association with hydroxychlorophyll a occurred, indicating that the oxidation of chlorophyll a was dependant not only on species but also on additional factors such as the mode of mortality, growth limiting factor (i.e. nutrient concentration) or phenotypic plasticity. Surface sediment samples contained 10 times more hydroxychlorophyll a (relative to chlorophyll a) than pelagic particulate samples, indicating that more chlorophyll a oxidation occurred during sedimentation or at the sediment–water interface, than in the pelagic environment. In addition, chlorophyll aP276 correlated with chl-a concentration, thus supporting its assignment as a chl-a precursor.

Dynamic responses of the benthic bacterial community at the Western English Channel observatory site L4 are driven by deposition of fresh phytodetritus

The impact of the seasonal deposition of phytoplankton and phytodetritus on surface sediment bacterial abundance and community composition was investigated at the Western English Channel site L4. Sediment and water samples were collected from January to September in 2012, increasing in frequency during periods of high water column phytoplankton abundance. Compared to the past two decades, the spring bloom in 2012 was both unusually long in duration and contained higher than average biomass. Within spring months, the phytoplankton bloom was well mixed through the water column and showed accumulations near the sea bed, as evidenced by flow cytometry measurements of nanoeukaryotes, water column chlorophyll a and the appearance of pelagic phytoplankton at the sediment. Measurements of chlorophyll and chlorophyll degradation products indicated phytoplankton material was heavily degraded after it reached the sediment surface: the nature of the chlorophyll degradation products (predominantly pheophorbide, pyropheophorbide and hydroxychlorophyllone) was indicative of grazing activity. The abundance of bacterial 16S rRNA genes g−1 sediment (used as a proxy for bacterial biomass) increased markedly with the onset of the phytoplankton bloom, and correlated with measurements of chlorophyll at the surface sediment. Together, this suggests that bacteria may have responded to nutrients released via grazing activity. In depth sequencing of the 16S rRNA genes indicated that the composition of the bacterial community shifted rapidly through-out the prolonged spring bloom period. This was primarily due to an increase in the relative sequence abundance of Flavobacteria.

Effect of ocean acidification and elevated fCO2 on trace gas production by a Baltic Sea summer phytoplankton community

The Baltic Sea is a unique environment as the largest body of brackish water in the world. Acidification of the surface oceans due to absorption of anthropogenic CO2 emissions is an additional stressor facing the pelagic community of the already challenging Baltic Sea. To investigate its impact on trace gas biogeochemistry, a large-scale mesocosm experiment was performed off Tvärminne Research Station, Finland in summer 2012. During the second half of the experiment, dimethylsulphide (DMS) concentrations in the highest fCO2 mesocosms (1075–1333 μatm) were 34 % lower than at ambient CO2 (350 μatm). However the net production (as measured by concentration change) of seven halocarbons analysed was not significantly affected by even the highest CO2 levels after 5 weeks exposure. Methyl iodide (CH3I) and diiodomethane (CH2I2) showed 15 % and 57 % increases in mean mesocosm concentration (3.8 ± 0.6 pmol L−1 increasing to 4.3 ± 0.4 pmol L−1 and 87.4 ± 14.9 pmol L−1 increasing to 134.4 ± 24.1 pmol L−1 respectively) during Phase II of the experiment, which were unrelated to CO2 and corresponded to 30 % lower Chl-ɑ concentrations compared to Phase I. No other iodocarbons increased or showed a peak, with mean chloroiodomethane (CH2ClI) concentrations measured at 5.3 (± 0.9) pmol L−1 and iodoethane (C2H5I) at 0.5 (± 0.1) pmol L−1. Of the concentrations of bromoform (CHBr3; mean 88.1 ± 13.2 pmol L−1), dibromomethane (CH2Br2; mean 5.3 ± 0.8 pmol L−1) and dibromochloromethane (CHBr2Cl, mean 3.0 ± 0.5 pmol L−1), only CH2Br2 showed a decrease of 17 % between Phases I and II, with CHBr3 and CHBr2Cl showing similar mean concentrations in both Phases. Outside the mesocosms, an upwelling event was responsible for bringing colder, high CO2, low pH water to the surface starting on day t16 of the experiment; this variable CO2 system with frequent upwelling events implies the community of the Baltic Sea is acclimated to regular significant declines in pH caused by up to 800 μatm fCO2. After this upwelling, DMS concentrations declined, but halocarbon concentrations remained similar or increased compared to measurements prior to the change in conditions. Based on our findings, with future acidification of Baltic Sea waters, biogenic halocarbon emissions are likely to remain at similar values to today, however emissions of biogenic sulphur could significantly decrease from this region.

Defining scenarios of future vectors of change in marine life and associated economic sectors

Addressing the multitude of challenges in marine policy requires an integrated approach that considers the multitude of drivers, pressures, and interests, from several disciplinary angles. Scenarios are needed to harmonise the analyses of different components of the marine system, and to deal with the uncertainty and complexity of the societal and biogeophysical dynamics in the system. This study considers a set of socio-economic scenarios to (1) explore possible futures in relation to marine invasive species, outbreak forming species, and gradual changes in species distribution and productivity; and (2) harmonise the projection modelling performed within associated studies. The exercise demonstrates that developing interdisciplinary scenarios as developed in this study is particularly complicated due to (1) the wide variety in endogeneity or exogeneity of variables in the different analyses involved; (2) the dual role of policy decisions as variables in a scenario or decisions to be evaluated and compared to other decisions; and (3) the substantial difference in time scale between societal and physical drivers.

Spatio-temporal Bayesian network models with latent variables for revealing trophic dynamics and functional networks in fisheries ecology

Ecosystems consist of complex dynamic interactions among species and the environment, the understanding of which has implications for predicting the environmental response to changes in climate and biodiversity. However, with the recent adoption of more explorative tools, like Bayesian networks, in predictive ecology, few assumptions can be made about the data and complex, spatially varying interactions can be recovered from collected field data. In this study, we compare Bayesian network modelling approaches accounting for latent effects to reveal species dynamics for 7 geographically and temporally varied areas within the North Sea. We also apply structure learning techniques to identify functional relationships such as prey–predator between trophic groups of species that vary across space and time. We examine if the use of a general hidden variable can reflect overall changes in the trophic dynamics of each spatial system and whether the inclusion of a specific hidden variable can model unmeasured group of species. The general hidden variable appears to capture changes in the variance of different groups of species biomass. Models that include both general and specific hidden variables resulted in identifying similarity with the underlying food web dynamics and modelling spatial unmeasured effect. We predict the biomass of the trophic groups and find that predictive accuracy varies with the models' features and across the different spatial areas thus proposing a model that allows for spatial autocorrelation and two hidden variables. Our proposed model was able to produce novel insights on this ecosystem's dynamics and ecological interactions mainly because we account for the heterogeneous nature of the driving factors within each area and their changes over time. Our findings demonstrate that accounting for additional sources of variation, by combining structure learning from data and experts' knowledge in the model architecture, has the potential for gaining deeper insights into the structure and stability of ecosystems. Finally, we were able to discover meaningful functional networks that were spatially and temporally differentiated with the particular mechanisms varying from trophic associations through interactions with climate and commercial fisheries.