DECLARATION BY THE SCIENTIST IN CHARGE
I, Prof. Carsten Dormann, as scientist in charge of the project (298380,
BAYESIANMETAFLATS), hereby confirm that:
- The attached periodic report represents an accurate description of the work carried out in this project for this reporting period;
- The project:
( ) has fully achieved its objectives and technical goals for the period;
- To my best knowledge, the financial statements which are being submitted as part of this report are in line with the actual work carried out and are consistent with the report on the resources used for the project and if applicable with the certificate on financial statement.
- The beneficiary, in particular non-profit public bodies, secondary and higher education establishments, research organisations and SMEs, have declared to have verified their legal status. Any changes have been reported under section 6 (Project
Management) in accordance with Article II.2.f of the Grant Agreement.
1

1. PUBLISHABLE SUMMARY
Understanding how the plants and animals that live in the seafloor vary in their spatial patterns of diversity and abundance is fundamental to improving knowledge of ecological processes that underpin communities, as well as advance the modelling of species distributions under realistic assumptions. Such critical knowledge, which was gathered during this project, informs both our fundamental understanding of ecosystem processes and community ecology, ecological monitoring, as well as the role of biodiversity in maintaining ecosystem functioning in coastal ecosystems.
1200 macrobenthic samples were collected in Manukau, Tauranga, and
Kaipara Harbour following a pre-determined sampling scheme (Fig. 1). We identified over 145 species and counted 73813 individuals . In addition, 960 Chlorophyll a samples as a proxy for food availability were analysed, 960 sediment samples to caption habitat variability were processed, and 960 photo-quadrats of the sampling locations were used to describe cover of seasgrass ( Zostera ), shellhash, and bare sand.
Using multivariate spatial models, we found that benthic communities were distinctly spatially structured at four spatial scales, i.e., >100 m, 50-100 m, 50-15 m, and <15 m, but there was no clear separation of the importance of environmental vs. biotic factors at these scales . A broad selection of environmental variables contributed to the large-scale variation, whereas a much-limited set explained part of the fine-scale community structure. Our results emphasize there is no prevailing scale to which environmental factors drive biodiversity and stress the importance of dissecting variation across scale to advance an understanding of processes structuring species communities and inform conservation measures .
Also, we used a recently developed Bayesian hierarchical multispecies model to accommodate both spatial correlations within and between locations and give full inference about model parameters and the characteristics of unobserved spatially structured factors. We demonstrated the utility of these models in defining the role of species interactions across large spatial scales. Our results show the presence of species interactions beyond scales generally amenable to manipulative studies and provide a new approach allowing conservation research and measures to acknowledge and assess important ecological phenomena.
2

Fig. 1.
Sampling design (top panel), encompassing 400 sampling stations, which are positioned to match a number of spatial lags ranging from 0.3 m to 1000 m. Illustrated are the abundances of
Macomona liliana (scaled dots; 0-40 ind./core) at Kaipara Harbour. The background displays a seascape of median grain size (darker grey indicates coarser sands; values range between 170-250 µm).
Bottom panels show a sampling area at low tide.
Furthermore, it is well recognized that the functional performance of individual species is affected by abundance and distribution, but the implication for multiple species that share traits influencing functional redundancy and resilience to stressors has not been resolved. Therefore, we assessed the degree of sharing of specific functional traits in these communities . We defined functional groups, based on biological traits that relate to how macrofauna modify sediment biogeochemistry and stability, and their resilience to disturbance. Results show clear spatial gradients and boundaries of abundance and distributions separating function.
Our findings provide a framework to investigate the resilience of benthic communities in the context of shifts in ecosystem functioning associated with different types and levels of disturbance and provide new insights into the resilience and health of seafloor communities .
Currently a number of publications have been mapped out underpinned by data obtained and we expect to finalize most of these in the coming year.
3

2. PROJECT OBJECTIVES FOR THE PERIOD
- Objective 1: Multi-scale benthic sampling
Fieldwork to collect the necessary data to support Objectives 2 and 3.
-Objective 2: Scale-dependent dynamics of species distributions
Apply multivariate spatial models scale to assess which environmental factors drive biodiversity and at which spatial scale, to advance an understanding of processes structuring species communities and inform conservation measures.
-Objective 3: Bayesian modelling and forecasts
Use recently developed Bayesian hierarchical multispecies model to accommodate both spatial correlations within and between locations and give full inference about model parameters and the characteristics of unobserved spatially structured factors.
This Objective initially was not part of the reporting period, but since first steps have been taken to employ such models it is included here.
3. WORK PROGRESS AND ACHIEVEMENTS DURING THE PERIOD
At present all tasks assigned to Objectives 1 and 2 have been completed. All samples collected for multi-scale analyses have been processed. Data have been made retrievable in databases (Access) or geographical information systems (ArcView or Quantum GIS).
Analyses are being done at present, which will result in further publications.
Originally we proposed to use Generalised Estimating Equations to assess scale-dependent dynamics of species. However, we opted to use recent multivariate spatial models, i.e. Moran’s Eigenvector Mapping, which enables to study hierarchical structures of the complete community, rather than species-by-species assessments. Our multi-scale sampling effort offers a unique data set to test such
4

methods.
Generally data are extracted from existing data-portals, typically satellite images, and downscaled, rather than collected at different spatial scales as we have.
Furthermore, we assessed spatial variation in functional group distributions.
The latter resulted in an MSc-thesis by Barry Greenfield and has been offered for publication. A second MSc-student, Laura Hines, has submitted her thesis recently and is expected to receive her degree soon.
Bayesian model development (Objective 3) has been started, but these prove to be quite challenging to fit. 2 manuscripts have been prepared and Andrew Finley, an expert on Bayesian modelling, is currently running the final models. Despite access to a cluster mainframe, running these models takes several weeks. One paper is a review dealing with studying species interactions in species distribution modelling and the many mistakes that are made and how improvements can be made . The other paper provides a detailed case study how to employ spatially explicit Bayesian models to assess species interactions. The latter has been submitted to the high-impact journal Methods in Ecology and Evolution.
Upon request by the editor we are now adding much more modelling details (which we initially left out to keep the focus on the ecology), before it will be submitted again.
4. ADDITIONAL INFORMATION
Comments: No comments
5. DISSEMINATION ACTIVITIES
Grants:
An ERC Starting Grant proposal has been submitted (February 2015) to allow continuation and expansion of the research carried out during this Marie-Curie IOF.
A Sofja-Kovalevskaja grant proposal has been submitted, but unfortunately not won
(received that news Early April 2015).
Papers:
5

Kraan C , Finley A, Dormann C, Thrush S. Capitalizing on information from ecological data: revealing processes across different scales. In preparation.
Many species shift their distribution as a result of changes in environmental conditions and biological factors. Understanding current and forecasting future distributions of species is complicated by environment-biology interactions, but critical for better-integrated ecosystem management. Most analyses of large-scale species distributions ignore the many scales of spatial autocorrelation in spatial patterns, scale-dependent variability, and the importance of biotic interactions. Such omissions affect statistical analyses and the ecological inferences drawn from them.
Here we review recent methodological advances that encompass these shortcomings and address their benefits.
Kraan C , Finley A, Thrush S, Piersma T, Dormann C. Putting biotic interactions into distribution analysis: co-occurrence modelling beyond experimental scales. In preparation.
Management and conservation strategies use species distribution models to help define threats to habitats and species ecology. Most of these models, however, ignore biology encompassed by spatial autocorrelation and species co-occurrences. Such omissions affect both the statistical analyses and the ecological inferences drawn from them. To capture this complexity we use a recently developed Bayesian hierarchical multispecies model. These models accommodate both spatial correlations within and between locations and give full inference about model parameters and the characteristics of unobserved spatially structured factors.
Douglas E, Pilditch C, Thrush S, Kraan C , Lohrer A, Schipper L. Denitrifier response to nutrient enrichment controlled by macrofauna community composition. In preparation.
The aim is to determine the importance of functional group diversity and abundance for Denitrification Enzyme Activity and benthic ecosystem resilience to nutrient
6

enrichment. Resilience to nutrient enrichment for benthic ecosystems means getting rid of excess nutrients and restoring stability of local nutrient cycling in order to resist consequences such as anoxia and loss of species. This experiment is an assessment of the ecological significance of functional abundance versus functional diversity within a physically homogeneous site, allowing a test of these community characteristics without the background environmental variability that commonly determines their distribution.
Andresen H, Thrush S, Lohrer A, Kraan C . Adult-juvenile interactions in bivalves: large-scale spatial associations disentangled from environmental influences. In preparation.
Biotic interactions shaping species distributions are currently a hot topic. Here we aim to assess such for adult and juvenile bivalves in New Zealand sandflat habitats.
Key will be to do this in a spatially explicit fashion.
Kraan C , Greenfield B, Dormann C, Thrush S. Cross-scale variation in biodiversityenvironment links illustrated by coastal sandflat communities. To be submitted to Ecosphere (currently waiting on final comments from coauthors).
Community composition encompasses many biotic and environmental elements. A neglected factor in its analysis is the multi-scale nature of data, which affects speciesenvironment relationships and consequently the development of conservation policies and assignment of protected areas. Drawing on recent multivariate spatial analysis, we investigate whether including spatial structure and environmental conditions enable to better differentiate between the ecological processes that underlie biogeographical patterns of species communities across scales. Our results emphasize there is no prevailing scale to which environmental factors drive biodiversity and stress the importance of dissecting variation across scale to advance an understanding of processes structuring species communities and inform conservation measures.
7

Greenfield B, Kraan C , Pilditch C, Thrush S. Incorporating spatial variation into functional diversity to gauge community resilience. Revision resubmitted to
Biology Letters .
Species that share traits may perform similar functions and this redundancy can confer resilience in ecosystem function. It is well recognised that the functional performance of individual species is affected by abundance and distribution, but the implication for multiple species that share traits influencing functional redundancy has not been resolved. Our findings provide a framework to investigate the resilience of benthic communities in the context of shifts in ecosystem functioning associated with different types and levels of disturbance and provide new insights into the resilience and health of seafloor communities.
Lohrer A, Thrush S, Hewitt J, Kraan C . 2015. Ecological self-organisation and the up-scaling of ecosystem functions in a heterogeneous world. Accepted by
Scientific Reports .
Here, by integrating copious observational and experimental field data from marine soft-sediment habitats in New Zealand, we convincingly demonstrate the scalability of short-term single-site primary productivity patterns. Our study documents consistent positive relationships between ecological drivers (e.g., the biological mixing of sediments by mobile burrowing fauna) and seafloor primary productivity across space and time scales in a heterogeneous coastal seascape, and identifies mutually reinforcing feedbacks and cross-scale interactions contributing to this critical element of coastal marine ecosystem functioning.
Pratt D, Pilditch C, Lohrer A, Thrush S, Kraan C . 2015. Spatial distributions of grazing activity and microphytobenthos reveal scale-dependent relationships across a sedimentary gradient. Estuaries and Coasts , in press.
The density, spatial structure and functional roles of macrofaunal and
8

microphytobenthic (MPB) communities change across sedimentary gradients.
Grazing by macrofauna can impose considerable top-down control on MPB biomass at the scale of the animal’s feeding ambit (cm scale), yet how relationships between deposit feeders and MPB scale up across such transitional environments (10’s m scale) is poorly understood. Our results demonstrate that significant effects on MPB biomass can emerge across large, spatially heterogeneous areas of tidal flat, despite appearing stochastic at small scales.
Thrush S, Hewitt J, Parkes S, Lohrer A, Pilditch C, Woodin S, Wethey D, Chiantore
M, Asnaghi V, de Juan S, Kraan C , Rodil I, Savage C, van Colen C. 2014.
Experimenting with ecosystem interaction networks in search of threshold potentials in real world marine ecosystems. Ecology 95, 1451-1457.
Thresholds profoundly affect our understanding and management of ecosystem dynamics, but we have yet to develop practical techniques to assess the risk that thresholds will be crossed. By demonstrating loss of positive feedbacks and changes in the architecture of the network, we provide mechanistic evidence that stressors lead to break points in dynamics, which theory predicts predispose a system to a critical transition.
Calabrese J, Certain G, Kraan C , Dormann C. 2014. Stacking species distribution models and adjusting bias by linking them to macroecological models. Global
Ecology and Biogeography 23, 99-112.
Species distribution models (SDMs) are common tools in biogeography and conservation ecology. It has been repeatedly claimed that aggregated (stacked) SDMs
(S-SDMs) will overestimate species richness. Here, we examine current practice in the development of S-SDMs to identify methodological problems, provide tools to overcome these issues, and quantify the performance of correctly stacked S-SDMs alongside macroecological models.
9

In addition plans have been made to write the following papers:
Describe abundance thresholds along environmental gradients (Kraan et al.) to focus on habitat preferences of functional groups across seascapes.
Describe network complexity (Kraan et al.) by employing recent structural equation models, which allow looking for indirect interactions between, e.g., diversity and ecosystem functioning.
Paper describing functional group accumulation curves (Kraan et al.). This is more a theoretical approach to study where along the curve functional groups occur.
Scientific Communications:
Kraan C , Greenfield B, Dormann C, Thrush S. 2014. Spatial structuring of biodiversity. Ecological Society of Australia, Alice Springs, Australia.
This represents the largest meeting for ecologists in the Oceania region.
Kraan C , Finley A, Thrush S, Piersma T, Dormann C. 2014. Putting biotic interactions into distribution analysis. Ecological Society of America,
Sacramento, USA.
This represents the largest meeting for ecologists worldwide.
Kraan C , Finley A, Thrush S, Piersma T, Dormann C. 2014. Conservation strategies should encompass species interactions and spatial autocorrelation. Australian
Marine Science Association, Canberra, Australia.
The largest meeting of marine researchers in Oceania.
Hines L, Pilditch C, Kraan C , Lohrer D, Thrush S. 2014. The effect of enhanced nutrient levels on coastal ecosystem functions. Golden Key International
Summit, Boston, USA.
A global meeting of young talents (MSc-level) across all disciplines.
10

Invited lecture at Massey University, Palmerston North, New Zealand (2014).
Invited to the workshop: Matrix of Marine Ecosystem Services, Auckland, New
Zealand (2014).
Thrush S, Hewitt J, Parks S, Lohrer D, Pilditch C, van Colen C, Woodin S, Wethey
D, Chiantore M, Asnaghi V, Kraan C . 2013. Integrating causal and exploratory techniques to indicate threshold potentials in real world ecosystem dynamics. Coastal and Estuarine Research Federation, San Diego, USA.
Largest meeting of researchers in coastal and estuarine systems.
Participated in the Eco-Stats workshop: New opportunities at the interface between ecology and statistics, Sydney, Australia (2013).
Invited lecture at the University of Waikato, Hamilton, New Zealand (2013).
Kraan C , Finley A, Dormann C, Thrush S. 2013. Spatial perspectives in ecology.
New Zealand Marine Science Society conference, Hamilton, New Zealand.
Annual meeting of marine sciences in New Zealand.
Greenfield B, Pilditch C, Kraan C , Hewitt JE, Thrush S. 2013. Resilience of sandflats indicated by variation in functional group diversity. New Zealand
Marine Science Society conference, Hamilton, New Zealand.
Annual meeting of marine sciences in New Zealand.
In addition we plan to present at:
Andresen H, Thrush SF & Kraan C.
2015. Adult-juvenile associations in bivalves disentangled from environment. European Marine Biology Symposium .
Helgoland, Germany.
Largest marine ecological meeting in Europe.
Pratt D, Pilditch C, Lohrer AM, Thrush SF & Kraan C . 2015. Deposit feeding activity and microphyte biomass relationships across a benthic sediment gradient. European Marine Biology Symposium . Helgoland, Germany.
11

Largest marine ecological meeting in Europe.
Douglas E, Pilditch C, Kraan C , Schipper L, Lohrer AM & Thrush SF. 2015.
Macrobenthic community composition drives denitrifier response to nutrient loading. European Marine Biology Symposium . Helgoland, Germany.
Largest marine ecological meeting in Europe.
The next Estuarine and Coastal Science Conference in England (Douglas et al.). Title unknown currently.
A large meeting of coastal researchers.
Book Chapters:
Thrush S, Kraan C , Hewitt J, Lohrer A. Habitat complexity and marine soft-substrate systems. In: Tokeshi M (ed). Habitat Complexity in Aquatic Systems:
Ecological Perspectives. Springer. Accepted for publication.
In this chapter we focus on the processes by which habitat heterogeneity and complexity structure biodiversity and ecological functioning in marine soft-substrate systems. We traverse across spatial and temporal scales to illustrate how small-scale patterns and processes impact large-scale, ecosystem-wide, processes. In particular, we provide some detailed examples of the influence of habitat structure on nutrient fluxes, of self-organisation and of habitat heterogeneity on disturbance recovery dynamics.
Other:
Participating expert during the Miranda BioBlitz. 2013. A day and night survey of an area for all forms of life, with the participation of as many people as possible, including local schools, Iwi (Maori), district and regional councils, community groups, searching for species which are then identified by invited on-site experts. Miranda, New Zealand.
12

6. PROJECT MANAGEMENT
Based on discussions, as well as joint manuscripts, the project planning and status is in agreement with the designed time-plan. Thus far the only problem that occurred has been delayed payment of salary to the researcher at the start of the project, due to a human error in the Project Management department at the University of Freiburg.
There are no changes to the legal status of any of the beneficiaries, no deviations from the planned milestones and deliverables. Management costs do not deviate from the budget put forward in the Grant Proposal.
13