MARES_14_06 | PhD Research: Predictive Modeling for Marine Spatial Planning

Mapping the deep: the application of predictive modelling to European spatial planning
PhD Code: MARES_14_06:

Mobility

Host institute 1: P7 - University of Plymouth
Host institute 2: P13 - University of Aveiro
Host institute 3: P1 - University of Ghent

Research fields:

T6 - Habitat loss, urban development, coastal infrastructures and Marine Spatial Planning

Promotor(s):

Kerry Howell
Marina Cunha
Dr David Van Rooij ([email protected])

Contact Person and email: Kerry Howell - [email protected]

Subject description

Anthropogenic pressure on deep-sea habitats is growing, through fisheries and specifically bottom trawling (Rooper et al. 2011; Norse et al. 2012; Puig et al. 2012), climate change, including deep ocean temperature rising (Balmaseda et al. 2013) and ocean acidification (Form & Riebsell 2011), and more recently deep sea mining (Collins et al. 2013). Globally there is increasing awareness of the need to develop and implement biodiversity conservation policy to ensure appropriate and sustainable management of these ecosystems (Ban et al. 2013).

From an international perspective United Nations General Assembly resolution 61/105 (United Nations General Assembly 2003) requires the protection of Vulnerable Marine Ecosystems (VMEs) from damaging fishing practices. From a regional perspective, Annex V of the OSPAR Convention (On the Protection and Conservation of the Ecosystems and Biological Diversity of the Maritime Area) provides a list of Threatened and/or Declining Species and Habitats again for which conservation measures are required (OSPAR 2008b). At a European level the Marine Strategy Framework Directive (MSFD) requires an initial assessment of the current environmental status of European Union member’s marine waters. Effective implementation of these policies requires a sound understanding of the extent and distribution of benthic biological assemblages as a starting point.

Maps have proved to be a useful method of summarising biological information concerning the seabed. Mapping at the level of species or assemblages is a major problem in the deep sea, where sampling is expensive and logistically challenging due to its remoteness, depth and relatively poor ecological knowledge of the residing fauna. Recently the use of species distribution modelling (SDM) has been applied to assemblages to produce maps of their distribution over large spatial scales (Howell et al. 2011; Rengstorf et al. 2013; Ross & Howell 2013). In addition, smaller scale site based mapping of multiple assemblages has also applied predictive modelling techniques (Gonzalez-Mirelis & Lindegarth 2012; Piechaud et al. submitted). Predictive modelling is a promising tool in this area potentially reducing the cost of comprehensive field surveys by allowing targeting of important areas, and filling data gaps for large areas of un-sampled seabed (Galparoso et al. 2009; Elith & Leathwick 2009; Dambach & Rodder 2011; Robinson et al. 2011).

Recent studies from the deep-sea environment have highlighted an overall trend toward better model performance with increasing environmental data resolution, with significant differences in performance found between models of different resolution (Marshall, 2011; Rengstorf et al., 2012; Ross et al., submitted). Until recently, the lack of high-resolution environmental datasets has been a major restriction to the reliability and applicability of SDMs in the deep sea, as precise spatial matching between presence data and environmental variables is necessary in order to avoid an artificial expansion of the species niche width, especially when modelling the distribution of sessile organisms (Guisan and Thuiller, 2005). The emergence of large-scale multibeam derived high-resolution bathymetry surveys (e.g. the Irish National Seabed Survey, the UKs Maremap project, and the Norwegian Mareano project) has provided practitioners with the means to greatly increase SDM model resolution.

However, while increasing availability of multibeam data allows more fine-scale predictive modelling to be undertaken, other important environmental drivers of deep-sea species distributions within the modelling process have been overlooked. Recently Mohn et al. (2014) reviewed hydrographic observations from numerous cold-water coral locations in the NE Atlantic and showed that a common feature of the hydrographic regime at individual coral mounds and mound clusters was the presence of energetic near-bottom flow dynamics. They developed a high-resolution hydrodynamic model at three cold-water coral provinces and demonstrated intensified near-bottom currents in areas where living corals were observed by contrast with coral absence and random background locations. Rengstorf et al. (submitted) have developed high-resolution (250 m grid cell size) hydrodynamic variables based on the Mohn et al. (2014) model and incorporated them into a SDM to explore their model explanatory power with interesting results.

The incorporation of high resolution bathymetric data and hydrodynamic model output data into SDM modelling will move the field of predictive modelling into new territory potentially providing more accurate maps upon which to base marine environmental management decisions as well as improving our understanding of the role of hydrodynamics in observed deep-sea species distribution patters. Therefore we propose a PhD study to test the following hypotheses:

H1: Predictively modelled habitat maps provide an accurate reflection of the distribution of the benthic assemblages considered.

H2: Models constructed using high resolution multibeam bathymetry data perform significantly better than those constructed using coarser resolution GEBCO bathymetry data.

H3: Models which incorporate hydrodynamic variables perform significantly better than those that do not.

The study will be focused on cold water coral mound habitats of the UK and Portuguese margin at the heads of the Dangeard & Explorer Canyons 48°45’N – 48°00’N – 10°15’W – 09°15’W and the Upper Ferrol Canyon 44°15’N – 43°45’N – 09°00’W – 08°15’W. This area has been selected for study as the University of Plymouth and the University of Ghent already have a funded multidisciplinary programme in this area that will provide data to the PhD. In addition the University of Plymouth already have a high resolution hydrodynamic model for the head of the Dangeard & Explorer Canyons that may be adapted for use in this study (Vlasenko et al., submitted). Hydrodynamic data for the Upper Ferrol Canyon will be derived from published regional hydrodynamic models.

In support of H1 the student will analyse video sample data collected from both the Dangeard & Explorer Canyons site and the Upper Ferrol Canyon site collected during a June 2014 sampling cruise. The will undertake multivariate analysis of the data to identify benthic assemblages for modelling purposes. The student will use widely available bathymetric datasets (General bathymetric chart of the oceans GEBCO) and coarse scale oceanographic datasets to produce a predictive model for the distribution of identified assemblages across both study areas. The model will be validated through internal validation. In support of H2 high resolution bathymetry data previously collected for this area will be used to create new predictive models of assemblages distribution for both areas. The performance of the high resolution bathymetry models will be compared with the GEBOC derived models to investigate changes in both performance and spatial output as a result of increasing data resolution. In support of H3 the student will rerun models incorporating both coarse and fine scale hydrodynamic model data and reassess model performance and spatial output considering both the inclusion of hydrodynamic data and scale. There will also be opportunity within the project for the student to consider other lines of investigation as the project progresses giving them academic freedom to develop. We believe the study is feasible within a 3 year time period since much of the data will already have been collected prior to the start of the PhD.

Preliminary work plan

Month 1-6: Literature review, Video analysis / develop taxonomic ID skills
Month 7-12: Video analysis / develop taxonomic ID skills
Month 13-18: Multivariate analysis, define benthic communities, draft of first paper - benthic assemblages of canyon mini-mounds, transfer to PhD report.
Month 19-24: Predictive modelling on coarse data and validation, predictive modelling on multibeam data and validation, production of 2nd paper - the effect of data resolution on model performance.
Month 25-30: Adaptation of existing hydrodynamic model, predictive modelling with hydrodynamic model data and validation, production of 3rd paper - comparison of models with and without hydrodynamic data.
Month 31-36: Possibility for student to expand research in direction of interest / contingency time, write up of thesis.

Synergy between collaborators

MC (Aveiro) will provide expertise on the deep-sea communities and fauna of the Portuguese margin aiding the student in analysis of the Upper Ferrol Canyon dataset. DVR (Ghent) will provide expertise in the acquisition and interpretation of multibeam data, he will also provide ship time for gathering data for use within the PhD in June 2015. KH (Plymouth) will provide expertise on the deep-sea communities and fauna of the UK margin and predictive modelling approaches. Vasily Vlasenko (Plymouth) will provide expertise on hydrodynamic modelling and model outputs.

References

Balmaseda, M. A., et al (2013) Geophysical Research Letters (2013); Ban, N. C. et al. (2013) Conservation Letters (2013); Dambach, J. & Rodder, D. (2011) Aquatic Conservation-Marine and Freshwater Ecosystems 21, 92-100; Elith, J. & Leathwick, J. R. (2009) in Annual Review of Ecology Evolution and Systematics Vol. 40, 677-697; Form, A. & Riebesell, U. (2011) Pangaea, doi:10.1594/pangaea.778439; Galparsoro, I., et al (2009) Ecological Modelling 220, 556-567; Gonzalez-Mirelis, G. & Lindegarth, M. (2012) Ecological Applications 22, 2248-2264; Howell, K. L., et al (2011) Biological Conservation 144, 2656-2665; Marshall, C.E., (2011). PhD Thesis, Plymouth University. 260pp; Mohn, C. et al., (2014). Progress in Oceanography, 122:92-104; Norse, E. A. et al. (2012) Marine Policy 36, 307-320; Puig, P. et al. (2012) Nature 489, 286-289; Rengstorf, A. M. et al., (2012); Marine Geodesy, 35(4), 343-361; Rengstorf, A. M., et al (2013) J. Biogeogr.; Rengstorf, A.M., et al., (submitted) Deep Sea Research II; Robinson, L. M., et al., (2011). Global Ecology and Biogeography, 20(6), 789-802; Rooper, C. N., et al (2011) Continental Shelf Research 31, 1827-1834; Ross, R. E. & Howell, K. L. (2013) Diversity and Distributions 19, 433-445.

Expected outcomes

1) At least 4 peer reviewed papers (both from the PhD research and interaction with the wider project)
2) Impact through uptake of maps produced in European Spatial Planning, and through application of methods in future spatial planning efforts.
3) Public outreach work, particularly interacting with school children, will be achieved through Plymouth University's annual Science and Technology Showcase http://www.plymouth.ac.uk/pages/view.asp?page=32025. In addition we will work with the media department at Plymouth University to create a short publically accessible film about the project similar to this http://youtu.be/Eu6snOCb3-Q produced by Dr Howell's research group and avaiable on youtube.
4) A new collaborative relationship between Plymouth, Ghent and Aveiro Universities.