INTECOL 2013

London, UK, August 18-23, 2013

The 11th INTECOL Congress, Ecology: Into the next 100 years will be held in London as part of the centenary celebrations of the British Ecological Society.The theme of the Congress is advancing ecology and making it count, and will present world class ecological science that will truly move the science forward.

We will be very active at INTECOL this summer. Please come and visit us at booth E3 where you will be able to talk with our scientists and engineers about your needs for biodiversity data research tools. We shall have demonstrations of workflows, documentation on our products, a video, etc.

We will also have a joint-workshop with LifeWatch, 1:00-2:30 on Monday, Aug. 19

BioVeL partners are also giving talks and showing posters. The full schedule is still awaited from INTECOL's organization, but here are the topics we shall cover:

TALKS

A workflow framework for ecological niche modelling of interactions between forest insects and their host trees under various climate change scenarios

Päivi Lyytikäinen-Saarenmaa, Renato De Giovanni, Alan Williams, Robert Kulawik, Vera Hernandez-Ernst, Matthias Obst, and Hannu Saarenmaa

Ecological niche modelling (ENM) is a popular technique for analysing changes in the distribution of organisms. Literally hundreds of papers have been published on topics such as spread of invasive species, analysing the impact of climate change and habitat loss, understanding rare and endangered species’ distributions, and designing biodiversity conservation plans. Little work, however, has been done on economically important species, such as pest insects. Such analyses require modelling of the impact of the pest on host plant distribution.  This is difficult with today’s ENM tools, as they mostly deal with Hutchingson’s abiotic niche concept and abiotic layers only.

We present an analysis of change of distribution of ten major European forest insect pests under various climate change scenarios. This involves modeling their historical and future distribution, and also using the predicted changes in host tree distribution as environmental layers. In order to achieve that we needed to build a dynamic repository of modelled host and pest distributions historically and also into the future.  As data sources we use both GBIF occurrences and actual forest damage reports from the EFI-Alterra Database of Forest Disturbances in Europe.  The predictions show that the damages by most major pests will spread about 500 km towards north east by year 2050, but much uncertainly lies in how far the host trees will actually be capable of moving or will be planted.

As tool we used the workflow solutions of the BioVeL project (Biodiversity Virtual e-Laboratory; an FP7 project; see www.biovel.eu).  Taverna is the workflow engine that orchestrates the various modelling and data management tasks at remote servers through web services.  The openModeller ENM engine was used to execute the various modelling algorithms.  BioSTIF is a geographic repository of layers into which the results of the analyses are stored.  Statistical post-processing of the data was done with an R-server.  All the above tasks were chained using the pre-cooked ENM workflow package of BioVeL.  We conclude that BioVeL’s flexible tool framework is well suited for complex modelling tasks such as exploring dynamic interactions between organisms.  Still, more work is needed to bridge the gap between ENM and dynamic population modelling.

e-Science, methods and statistical approaches to study historical changes in community structure: Quantifying the taxonomic and geographical change over seven decades of anthropogenic pressures in the North East Atlantic

Saverio Vicario (1), Anna Karlsson (2), Cherian Matthew (3), Robert Haines (4), Matthias Obst (5)
(1) Institute of Biomedical Technologies, National Research Council, Via Amendola 122/D 70125 Bari, Italy
(2) Swedish Agency for Marine and Water Management, Gothenburg, Box 11930, SE-404 39 Göteborg, Sweden
(3) Botanic Garden and Botanical Museum Berlin-Dahlem, Königin-Luise-Str.6-8, 14195 Berlin, Germany
(4)  School of Computer Science, University of Manchester, Oxford Road, Manchester, M13 9PL, UK
(5) Department of Biological and Environmental Sciences, University of Gothenburg, Box 463, SE-40530, Gothenburg, Sweden

Museum collections serve as archives for ecological research on the long-term effects of global change and human impact on community structure in ecosystems. e-Science provides an innovative approach facilitating the large scale mobilization of historical collections and their effective harmonization and integration with recent data. In this study we developed an workflow (available at www.biovel.eu) able to generate comparable data sets from historical and contemporary species inventories and monitoring programs. The workflow includes a range of functionalities for data cleaning and refinement, and advanced routines for taxonomic name resolution. The workflow was tested with a large-scale comparison of over fifty thousand species observation records from two marine benthic inventories on the Swedish West coast, which are separated by more than 66 years (1921-1941 and 2007-2009).  

A range of statistical approaches were applied to analyze subsets of data that have been refined taxonomically and ecologically, with special regard to the differences in sampling regimes and sampling standards. We quantified the long-term change in terms of 1) species richness; 2) species turnover, and 3) geographical distribution of species richness. Firstly, we applied classical rarefaction analysis, re-sampling the observation and the localities to assess significance and the species richness ACE estimator for each inventory. Secondly, we took an information theory approach and measured the amount of mutual information, as index of beta diversity, between the observations species name and sampling time, parsing the species contribution to overall changes of the community structure. Significance was estimated with a permutation approach. Finally the overall geographical structure of the data was evaluated with a general additive model (GAM) with bi-dimensional smothers. To test for the change in geographical structure we compared the marginal values of the Akaike Information Criterion weight (wAIC), across over 40 models, using a unique geographical bi-dimensional smoother or one for each sampling time. To control for other nuisance parameters, different metrics of sampling effort and habitat description were added as predictors to the models with different degrees of interaction. The full analysis was done with R scripts or workflows to ensure replicability of the results.

The results indicate an overall decline in species richness in the entire area of investigation. The overall mutual information across species and sampling was highly significant, and we identified those species that had changed their prevalence significantly across the two inventories. The overall results showed a great species turnover with few species that appear or increase greatly in their prevalence, while the large majority become less frequent or go locally extinct. Finally, the GAM approach indicates that it is very likely that a simplification of the spatial structure of species richness took place in the past seven decades.

This study shows that the analysis of changes in community structure of marine ecosystems are feasible over large temporal scales, and can go beyond the scope of conventional ecosystem assessments based on monitoring programs.

Large scale modelling of the distribution and impact of invasive species in the Baltic Sea

Matthias Obst (1), Sarah Bourlat (1), Renato De Giovanni (2), Robert Kulawik (3), Alan Williams (4)
(1) Department of Biological and Environmental Sciences, University of Gothenburg, Box 463, SE-40530, Gothenburg, Sweden
(2) Reference Center on Environmental Information, Campinas SP, Brazil
(3) Fraunhofer Institute for Intelligent Analysis and Informations Systems IAIS, Schloss Birlinghoven, 53754 Sankt Augustin, Germany
(4) School of Computer Science, University of Manchester, Oxford Road, Manchester, M13 9PL, UK

Biological invasions have dramatically amplified with the development of the trade network around the planet, increasing homogenization of communities, and representing one of the main causes of biodiversity decline. Invasions cause severe economic damage as they degrade valuable ecosystem services, generate human health problems, or enforce cost-intensive preventive and eradicative measures. In this study we developed an workflow (available at www.biovel.eu) for modeling invasive risk zones associated with climatic change and movement of biological material (e.g. soil, ballast water). The tool works across large geographical and taxonomic scales. We tested the workflow with a ‘black species list’ of 47 potentially invasive species for the Baltic Sea. Using more than 12.000 occurrence records, we were able to accurately model species distributions under present and 2050 climatic conditions, and analyzed geographical trends using the R-service deployed by the BioVeL project. Our preliminary analysis (at this point done for six species) identifies areas with accumulated risk for invasions, and gives a description of the invasion potential with respect to (i) the extent of the suitable habitat for invaders in the area of interest, as well as (ii) the intensity of the habitat suitability for invaders in the area of interest. More specifically, our results show that invasive species that are tolerant to a wide range of salinities, such as the chinese mitten crab Eriocheir sinensis or the Atlantic oyster drill Urosalpinx cinerea, show larger potential to establish wide-ranging viable populations in the Baltic Sea under 2050 climatic conditions, in comparison with salinity-intolerant species. This study shows the utility of e-science approaches to provide scalable tools for biodiversity modeling and rapid integration, visualisation and analysis of biodiversity data from multiple species and data sources, as well as over large spatio-temporal scales.

POSTERS
Deriving quantitative ecosystem service indicators and developing a new IT tool based on terrestrial ecosystem simulations

by Horváth, F., R. Aszalós, Z. Barcza, M. Kertész, P. Ittzés, D. Ittzés and B. Czúcz

Biogeochemical models are widely used to simulate the functioning of terrestrial ecosystems, including evergreen and deciduous forests, grasslands, shrubs and crops. These process-based models are driven by daily meteorological data, and handle the key ecophysiological, biogeochemical processes (i.e. partitioning of solar radiation; photosynthesis; carbon, water and assimilate allocation; initiation and cessation of growing season; autotrophic respiration; plant mortality; soil organic matter decomposition; interception; evapotranspiration; water movement and balance; runoff; environmental stress limitation; management and disturbances, etc.), and many different pools (i.e. foliage; aboveground and belowground biomass compartments; live and dead organic matter; different soil organic matter; soil water content, etc.), and a number of energy, water and carbon fluxes in details. Measured ecophysiological parameters specific to plant functional types are used to tune appropriately these models, and ecosystem scale flux measurements can provide independent dataset for sensitivity analysis, data-model integration (calibration) and/or validation. Long term runs can provide daily, monthly and annually aggregated ecosystem specific variables, such as net primary production (NPP), annual evapotranspiration (ET), annual maximum leaf area index (LAI), etc.

As the supply of ecosystem services is based on biophysical processes, well-calibrated ’in silico’ ecosystem models can be an appropriate tool to quantify meaningful but hard-to-measure ecosystem service indicators (ESI).

The Biome-BGC model provides a good tool to quantify a broad range of quantitative ESIs: annual wood production; yearly maximum of total above ground biomass of grasslands; yearly maximum of total above ground biomass of selected crops – as biomass provisioning ESI, and annual net primary production; total average carbon stock – as global climate regulation ESI; energy absorption by evapotranspiration – as micro and regional scale climate regulation ESI; damping of ecosystem daily water outflow – as hydrological cycle and water flow maintenance ESI; sum of living and dead biomass protecting the soil against erosion – as mass stabilization and control of erosion rates ESI; litter and coarse woody debris decomposition rate – as weathering processes ESI; Nitrogen leaching / deposition rate – as mediation (purification) by ecosystems ESI. Using these indicators in real ecosystems or hypothetical reference ecosystem simulations, different climate or land use scenarios can be applied and compared. The calculation of ESI’s are based on indicator specific algorithms and aggregation functions of internal model variables.

The ecosystem service indicator IT tool is under development within the frame of the Biodiversity Virtual e-Laboratory (BioVeL) project, which consists of internet-based web services, Taverna workflow system and myExperiment virtual laboratory. The BioVeL and OpenNESS (Operationalisation of Natural Capital and EcoSystem Services) FP7 projects are planning to collaborate in developing and testing of Biome-BGC model based ecosystem service indicators, which can help scientists to inform stakeholders and policy-makers about unsustainable and sustainable options of development in our human dominated changing world.

An e-science approach for studying microbial communities through metagenomics and a functional trait-based ecology

Antonio Fernàndez-Guerra (1), Pelin Yilmaz (1), Norman Morrison (2), Frank Oliver Glöckner (1, 3) and Renzo Kottmann (1)
(1) Microbial Genomics and Bioinformatics Research Group. Max Planck Institute for Marine Microbiology, Celisusstr. 1, 28359 Bremen, Germany
(2) School of Computer Science, The University of Manchester, Oxford Road, Manchester, M13 9PL, UK.
(3) Jacobs University Bremen gGmbH, Campusring 1, 28750 Bremen, Germany

Metagenomics is a sequence based approach to study microbial communities in the environment. The rate of information collection generated by metagenomics is increasing with the development of more cost effective and faster sequencing technologies. This allows microbial ecologists to formulate new questions and develop new approaches to study microbial communities. However, despite the enormous quantity of data acquired through metagenomics, the analytical part is still uncoupled from meaningful ecological interpretations.

A first attempt to interpret metagenomics data to gain a full understanding of microbial community patterns in a rigorous ecological framework using a functional trait-based ecology approach was proposed by Barberán et. al 2012. This work extends the classical trait approach to the community level in vast and complex environmental genetic data sets, proposing inter-trait relationships that could be further used as habitat descriptors or indicators of artefacts during sample processing.

In parallel to the metagenomics revolution, e-science is witnessing similar advances providing the computer infrastructures and analytical tools to get the most of the new sequenced data. A new generation of ecological analyses through metagenomics can greatly benefit from e-science approaches. The Biodiversity Virtual e-Laboratory FP7 project developed data-driven “metagenomics workflows” expanding on the work of Barberan et al. 2012.  These workflows allow the standardized, repeatable, and comparable utilization of metagenomic data while hiding the complexity of analytical procedures behind well defined web services. These web services allows the calculation of the traits described in the work of Barberan et al. 2012 like GC contents, ORF prediction, functional and taxonomic diversity, furthermore, the workflow is flexible enough to include new traits of interest that will be included in the later analyses performed by the workflow. Therefore, these workflows offer the microbial ecology community easy access to a new repertoire of tools for better understanding of microbial community patterns. Moreover, workflows enhance the comparability and repeatability of results from different sampling campaigns. An example is the Ocean Sampling Day (OSD), a simultaneous sampling campaign of the world’s oceans which will take place on the summer solstice (June 21st) in the year 2014 organized by the Micro B3 FP7 project. OSD will make use of BioVeL’s standardized and repeatable metagenomic workflows for the analysis of the  cumulative OSD samples, related in time, space and environmental parameters. OSD will provide insights into fundamental rules describing microbial diversity and function and has the potential to contribute to a blue economy through the identification of novel, ocean-derived biotechnologies. The use of e-science workflows for rigorous ecological analysis of metagenomic analysis has significance, as it is likely that environmental sequencing in natural and disturbed habitats will become a routine part of future monitoring programs.

European Biodiversity Informatics Conference 2013

BioVeL will co-sponsor, along with LifeWatch, ViBRANT and other projects, a European Biodiversity Informatics Conference to be held in Italy 3-6th September 2013.

The purpose of the conference is to review Biodiversity Informatics in the context of the LifeWatch vision, the decadal priorities for biodiversity informatics expressed in the BMC Ecology Whitepaper (in press) and the EC roadmap workshop on biodiversity infrastructures (took place 19-20 March 2013). The meeting may also act as a springboard to forming and coordinating consortia to build proposals for Horizon 2020 funding.

Watch this space for further announcements.