WORKSTREAM 4 Ecological Consequences of wave and tidal energy extraction (lead HWU)
The mean and variability in tidal and residual current flows (e.g. Osalusi et al., 2009) in sea levels and wave heights, fresh water inputs, stratification and the mixing generated by wave and current shear turbulence, interaction of waves and tidal streams and these combined with arrays of energy extraction devices will all influence the oceanographic processes and hence the ecology of a region (Shields et al., 2011). This includes indirect effects on biota, mediated by changes in sediment dynamics and suspended particle loads.
Task WS4.1 Statistical modelling for benthic biotope characterisation: The task of assessing ecological consequences for the benthos (littoral and sub-littoral) will use the outputs from the 3D hydrodynamic models, as inputs to statistical models. In their simplest form statistical models will synthesise given physical parameters (depth, slope, aspect etc.) combined with 3D model outputs which will act as proxies for hydrokinetic energy at given locations where detailed field data on benthic biotopes are available. A range of outputs from the hydrodynamic models will be used, allowing means of wave and tidal combinations in addition to extremes. Deliverables from WS2 and WS3 will be used to identify associations of benthic biotope types with sediment dynamics and suspended particle loads, Statistical methods include: supervised classification; canonical variates analysis (the combinations of physical variables which differentiate biotope occurrence); modelling the incidence of key individual species using maximum entropy methods; and the application of generalised additive models (GAMs) to species abundance data (where available). Similar work has proven very successful in providing a general biotope characterisation of EMEC’s Billia Croo site.
Task WS4.2 Model validation: Data available for the development of statistical models for biotope and species distribution will be partitioned randomly into two sets. The first set (two-thirds of the data) will be used directly in the model fitting procedures; the second set (the remaining third) will be used to validate predictions from the fitted models, allowing a direct assessment of confidence in model outputs. Outputs from other studies, e.g. undertaken as part of Remit 2 of this Supergen Marine Challenge call, and from the Hebridean Futures Project may also provide new information on benthic biotopes and their environmental associations that could be of use in a ground-truthing exercise.
Task WS4.3 Model re-runs with extraction of hydrokinetic energy: Following validation for present conditions, the statistical models will then be applied to the various scenarios developed in WS1 where hydrokinetic energy is extracted in order to establish patterns of change in the benthic biotope distribution. These will be assessed, in conjunction with Marine Scotland, to establish the extent to which Good Environmental Status under the European MSFD may be threatened or impaired. A limited range of scenarios would include deliverables from WS2 and WS3 providing predictions on changes in bottom sediment types and suspended particle distributions. Using a Monte Carlo modelling approach, statistical uncertainties from all three sources (sediment model, suspended particle model and ecological model) would be propagated through to overall uncertainties in predictions of changes in biotope distributions. In addition to being an assessment of confidence in model predictions, this would also allow the probabilities of particular outcomes (e.g. outcomes that have consequences under MSFD) to be considered under a risk assessment framework.
Task WS4.4 Extended studies of ecological change: Other studies of potential ecological change will focus on the near and far field, as localised device scale ecological effects will be considered as part of the EIA process. Highlighted by Bell and Side (2011) these include consideration of (i) effects
of any changes in suspended sediment on spatial variations in the penetration depth of photosynthetically active radiation (PAR) using data from WS3.3; (ii) vertical upwelling effects on nutrient availability and whether there are regional changes that may affect for example seabird foraging patterns; (iii) larval transport for keystone species, particularly fishes, and effects on larval dispersal and competence; (iv) exploration of the use generalised ecosystem models and the approaches and data requirements for these. Finally potential climate change drivers will be investigated within the modelling approaches adopted, for comparison, and selected models re-run for combined scenarios. WS4 Deliverables
DW4.1. Validated statistical models for benthic biotope characterisation;
DW4.2. Probabilistic outputs of change in benthic biotopes from various energy extraction scenarios; DW4.3. Methods description for such assessments and means of incorporation/development in other regional and shelf-wide models
The mean and variability in tidal and residual current flows (e.g. Osalusi et al., 2009) in sea levels and wave heights, fresh water inputs, stratification and the mixing generated by wave and current shear turbulence, interaction of waves and tidal streams and these combined with arrays of energy extraction devices will all influence the oceanographic processes and hence the ecology of a region (Shields et al., 2011). This includes indirect effects on biota, mediated by changes in sediment dynamics and suspended particle loads.
Task WS4.1 Statistical modelling for benthic biotope characterisation: The task of assessing ecological consequences for the benthos (littoral and sub-littoral) will use the outputs from the 3D hydrodynamic models, as inputs to statistical models. In their simplest form statistical models will synthesise given physical parameters (depth, slope, aspect etc.) combined with 3D model outputs which will act as proxies for hydrokinetic energy at given locations where detailed field data on benthic biotopes are available. A range of outputs from the hydrodynamic models will be used, allowing means of wave and tidal combinations in addition to extremes. Deliverables from WS2 and WS3 will be used to identify associations of benthic biotope types with sediment dynamics and suspended particle loads, Statistical methods include: supervised classification; canonical variates analysis (the combinations of physical variables which differentiate biotope occurrence); modelling the incidence of key individual species using maximum entropy methods; and the application of generalised additive models (GAMs) to species abundance data (where available). Similar work has proven very successful in providing a general biotope characterisation of EMEC’s Billia Croo site.
Task WS4.2 Model validation: Data available for the development of statistical models for biotope and species distribution will be partitioned randomly into two sets. The first set (two-thirds of the data) will be used directly in the model fitting procedures; the second set (the remaining third) will be used to validate predictions from the fitted models, allowing a direct assessment of confidence in model outputs. Outputs from other studies, e.g. undertaken as part of Remit 2 of this Supergen Marine Challenge call, and from the Hebridean Futures Project may also provide new information on benthic biotopes and their environmental associations that could be of use in a ground-truthing exercise.
Task WS4.3 Model re-runs with extraction of hydrokinetic energy: Following validation for present conditions, the statistical models will then be applied to the various scenarios developed in WS1 where hydrokinetic energy is extracted in order to establish patterns of change in the benthic biotope distribution. These will be assessed, in conjunction with Marine Scotland, to establish the extent to which Good Environmental Status under the European MSFD may be threatened or impaired. A limited range of scenarios would include deliverables from WS2 and WS3 providing predictions on changes in bottom sediment types and suspended particle distributions. Using a Monte Carlo modelling approach, statistical uncertainties from all three sources (sediment model, suspended particle model and ecological model) would be propagated through to overall uncertainties in predictions of changes in biotope distributions. In addition to being an assessment of confidence in model predictions, this would also allow the probabilities of particular outcomes (e.g. outcomes that have consequences under MSFD) to be considered under a risk assessment framework.
Task WS4.4 Extended studies of ecological change: Other studies of potential ecological change will focus on the near and far field, as localised device scale ecological effects will be considered as part of the EIA process. Highlighted by Bell and Side (2011) these include consideration of (i) effects
of any changes in suspended sediment on spatial variations in the penetration depth of photosynthetically active radiation (PAR) using data from WS3.3; (ii) vertical upwelling effects on nutrient availability and whether there are regional changes that may affect for example seabird foraging patterns; (iii) larval transport for keystone species, particularly fishes, and effects on larval dispersal and competence; (iv) exploration of the use generalised ecosystem models and the approaches and data requirements for these. Finally potential climate change drivers will be investigated within the modelling approaches adopted, for comparison, and selected models re-run for combined scenarios. WS4 Deliverables
DW4.1. Validated statistical models for benthic biotope characterisation;
DW4.2. Probabilistic outputs of change in benthic biotopes from various energy extraction scenarios; DW4.3. Methods description for such assessments and means of incorporation/development in other regional and shelf-wide models