Like several other members of the GCC Team at the University of Helsinki, I took part of the 11th International Congress of Ecology (INTECOL) in London, 18-23 August in 2013. Amid an overwhelming set of talks, workshops and symposia, I spent a noticeable fraction of my time learning about a few particular themes.
Data management
There are ever increasing amounts of ecological data collected, but much of it ends up in poorly understandable excel sheets at the researchers’ computers. The same destiny awaits most analysis code written. I’ve been affected by the open science movement enough to have started to feel this is poor use of public money and against the ideals of the reproducibility of science. Thus I’ve started to dream about publishing my field data and analysis code together with my research results whenever possible.
Much beyond this, however, the workshop “Managing ecological data for effective use and re-use” made me think about all kinds of things that could be done if datasets of the past were generally available, could be searched for and could be reused in imaginative ways as part of future studies. (Can you imagine?!) The workshop focused on how to store and document ecological data so it would become part of institutional memory, and thus reusable even after the retirement or death of its collector(s).
Much of the workshop focused on documenting the content and context of ecological datasets as metadata. Metadata, a structured description of what a dataset is about, can be submitted to a metadata library, and researchers can search these libraries for suitable datasets for their study. Much of this was new to me, so I share my learnings here:
- Several metadata search engines exist for browsing data descriptions, including Metacat, DataONE, and that of Global Biodiversity Information Facility (GBIF).
- To describe your dataset as metadata, your metadata entry might provide information such as title, keywords, data creator, time span of data collection, spatial coverage of data collection sites, location information, GPS points, experimental design used, current location of dataset, contact person, file type of data storage, variable names, measurement units used, level of completion of data, data omissions, usage rights and licenses for reuse. Metadata may also contain information on what the dataset was collected for, and ideas on possible future uses and extensions of the dataset.
- Metadata contents can follow and meet several alternative standards, such as relevant ISO metadata standards (like ISO 19115), Dublin Core Element Set, or Darwin Core, and they can be structured using various alternative metadata editors, like DataONE’s Morpho EML editor, and DataUp. Templates with checklists make it easier to make sure all relevant information has been entered.
- For more insight, the usage of metadata in ecological sciences has been discussed in detail in Michener et al. (1997): Nongeospatial Metadata for the Ecological Sciences.
The workshop also provided a few tips that can be utilized already at data collection phase in the field to prevent errors in data entry, such as using Google Forms, Excel pre-defined fields & data validation tools, or the DataUp extension to Excel. (This is of course only if you are not able to utilize more sophisticated biodiversity software for field data collection.)
Stephen Crouch’s talk “Software carpentry: or how to persuade ecologists that they actually want to learn to do this stuff properly”, a part of the symposium “Not just for Geeks: Broadening scope and participation in predictive ecology”, focused on much of the same thing from research software point of view. Crouch pointed out for starters that you have to start thinking about the reusability of your software already at the planning phase. Interesting content can be found in the web page of the Software Sustainability Institute and they also have an open call for co-operation projects to develop more sustainable & maintainable research software.
Citizen science & urban ecology
I am interested in citizen science projects both because their potential in data collection and because of their educational and inclusive nature that might help to mediate perceived conflicts between nature conservation and human land use.
In the symposium “The Future of Ecological Possibilities in Citizen Science”, Mordechai (Muki) Haklay made the topic ever more interesting by discussing their project Extreme citizen science; a project that develops citizen science tools and practices specifically to empower the less educated and illiterate people of the World to gain knowledge on their environments and to participate science-making. Haklay painted a future of co-created and self-created studies that people could utilize to monitor changes in their living environment (he gave an example of people following noise levels in their neighbourhood before and after an airport construction). He also mentioned that there are cheap contemporary tools for building self-assembled sensor devices to aid data collection. (I bet my colleague Wolfgang will be fascinated about this!) I was left wondering, however, if new kinds of issues with science ethics could emerge if more and more scientific studies would be made by public interest groups.
Nicely along the same lines, Finn Danielsen’s comparative field study, carried out by trained village locals and field ecologists, showed that uneducated people can collect data that is no less accurate or more biased than that collected by professionals. Danielsen talked about locally-based monitoring of carbon stocks, and I came to think that such locally-based monitoring would be a fascinating, democratic, inclusive way to gather data on global change. I didn’t quite grasp, though, how local communities would be compensated for their participation.
In a rapidly urbanizing World, urban ecology is an increasingly important meeting point of biodiversity and human land use. Nancy Grimm’s plenary showed how ecological modeling can help to build resilient urban water systems that can, for instance, mediate the effects of extreme weather events in the face of climate change. As an alternative to monolithic engineering solutions, she promoted risk-tolerant, resilient solutions, such as multifunctional wetlands that could serve as refreshment areas (and perhaps biodiversity havens?) but also as water canals in the case of storms and flooding. James Roquette’s talk on “Managing urban river corridors to enhance species richness across multiple taxonomic groups” looked at different parameters of urban water canals and how they were linked to the presence of species of different taxa. What if we would really start planning urban structures more and more so that they would support biodiversity in the cities?
Citizen science and urban ecology, of course, have ideal common grounds to meet. Gesine Pufal gave a fascinating talk “Schoolyards as field sites – the challenges of investigating seed fate with 3rd grade students” on a project where ~10-year-old school kids carried out experiments on seed predation along a rural-urban gradient. According to Pufal, schoolyards often provide excellent sites for urban ecology projects, and they come with enthusiastic researchers! Such a study setting can meet two goals; you get data at low cost, but even more, children learn about their environment and about making science. As per Pufal advice, it is important to design this kind of projects so that they use language that children can understand. In Pufal’s study, for instance, percentages were not used as a measurement unit, as the children involved had not learned percentages yet.
Habitat dynamics & succession
Habitat dynamics and succession are key themes of my PhD project. Thus I chose to listen to as many speakers as many poster presenters as possible who work in this area.
The most interesting set of talks related to this theme took place in Friday morning’s symposium “Natural forest succession in the tropics – lessons and implications for forest restoration”. Some of the talks of this symposium presented findings from long-term follow-up studies of successional trajectories in regenerating tropical forests (15 years in Robin Chazdon’s and 20 years in Francis Brearley’s case). Helen Murphy had done an extensive literature review in the area. A key theme across the talks was that succession is not just as ecosystems attaining structure, biomass and species diversity in the course of time. Ecological and structural details of recovering forest sites―such as species composition, basal area of trees, estimated biomass, or the spatial patterns of ages of trees (as in Mart Vlam’s study on past canopy disturbance and forest recovery from it)―may also reveal interesting things about the past disturbance history and/or ongoing human land use of these sites; and these things can be modeled together. As such, studies of succession teach us about the relationship of human land use and patterns of species diversity in a dynamically changing World.
The symposium left me pondering two things. First, Chazdon and Murphy concluded that succession studies provide insight into how secondary succession can be accelerated by supporting those characters of late successional forests that are key to supporting high species diversity but take time to recover after disturbance. Chazdon’s results supported the idea of planting of long-lived trees with large crowns to regenerating sites. Working with the conservation of early-successional habitats myself, I couldn’t help thinking how well-grounded was the idea to maximize the coverage of late successional forests as opposed to other successional stages– it was here discussed as a self-evident goal. The presenters did have data on increasing species diversities in some taxa towards later successional stages, and apparently tropical forests are falling short of the late successional phases compared to the past status quo. Nevertheless, to protect species diversity, when should we accelerate succession, when should be provide disturbances, and how should we find balance between these two?
The second issue in my mind is related to the way of collecting data on succession and recovery from disturbances. Follow-up studies are undoubtedly the most reliable way of studying these things. Outside of the symposium, Konstantin Gongalsky presented a six-year experimental design follow-up study on the regeneration of soil macrofauna after wildfires in the boreal forests. He was nicely able to show how soil macrofauna re-established itself within six years from disturbance, though species dispersal ability and connectivity also had an effect on recovery rates. However, we don’t often have the time or resources to do this in a real conservation context for a particular habitat, as most recoveries take a much longer time. But, when can we gain enough understanding by comparing known past land uses against the present ecological conditions of study sites, and how to do so? In my own study, I study drivers of habitat change by taking advantage of past delineations of the habitats of endangered species, their present-day ecological condition and landowner interviews on land use histories of these sites. Several other speakers demonstrated other interesting ways to do this: Deepa Senapathi had used land use maps from 1936 and 2007 to study land use changes in Great Britain, to compare land use changes with changes of pollinator densities in the same time span. Keita Fukasawa had used archeological findings (!) as proxies for the temporal extent of human land use in his study areas and compared that with to the current presence of certain taxa. In Robin Chazdon’s project, a Master’s student had utilized aerial photos from 1976 to identify remnant trees to show that the present-day species richness was higher in the vicinity of these remnant trees compared to study plots further away from them.
More
The fourth theme I would have liked to focus on at INTECOL would have been the science-policy interface, on which there was a great number of interesting talks. Sadly, many of these talks coincided with other interesting talks, or I missed them because I needed to refuel myself with oxygen in the middle of the day.
Talks on metapopulation modeling and conservation planning in general were also on my priority list, and I did listen to many. One of my favourites was Lucie Bland’s talk on cost-effective global assessment of extinction risk under limited information. She pointed out that one in six (!!) species in the global Red List are classified as Data Deficient and looked at whether it would be sometimes more cost-effective to predict their status with machine learning models than to carry out an IUCN-style assessment on them. As machine learning tools become cheaper, it might be cost-efficient to use modeling in a wider scale to identify which are the species that should merit more study.
The great pleasure of working with ecology is that while learning about topics of practical interest, you regularly also encounter thoughts and ideas that help you understand life in general. In this sense, one of my favourite talks at INTECOL was Tim Clutton-Brock’s plenary on the ecology of cooperative breeding. He discussed the costs and benefits of cooperative breeding to the non-breeding members of the pack, how the inclusive fitness benefits are linked to the level of relatedness of pack members, and how the evolution of cooperative breeding must have been preceded by the evolution of monogamy. (Though, he also pointed out that after the evolution of cooperative breeding, monogamy may be lost.) He also presented humans as communal (not cooperative) breeders, and suggested that it is perhaps our language ability that has enabled us to develop such outstanding depths of non-kin cooperation that differentiates us from the “non-human animals”. (Only a biologist could use such a magnificent term!)