1st meeting: 07.-11.11.2016
Ecology has a tradition of using predictions to test hypotheses and theories, however, anticipating how global environmental change will affect natural ecosystems requires a new set of tools and thinking to help safeguard biodiversity for future generations. Currently, quantitative statements of how well community ecology does in prediction are hampered by a lack of quantitative data. However, even with the few syntheses that are available, it is hard to judge whether predictive skill (i.e. realized predictability) in ecology is good or bad, because we do not know how much predictive structure (i.e. intrinsic predictability) ecological data contains and hence how difficult it is to make ecological predictions. We aim to develop a theory of intrinsic predictability and thus provide ecologists with a tool to address when and why their forecasting strategies succeed or fail. We will synthesize methods developed in fields such as computer science and complex systems to infer intrinsic predictability on simulated ecological datasets and then apply these metrics to empirical datasets from Lake Constance and experimental mesocosms. By issuing a data request to practitioners, we will collect datasets with existing realized predictions, with which we can apply the metrics of intrinsic predictability broadly across ecosystem types. With this analysis, we aim to establish the baseline of intrinsic and realized predictability in community ecology, and start a dialogue about the data needed for improving predictive models. Our workshop is timely, because we not only critically need to quantify, but anticipate the effects of global change on global biodiversity to implement successful conservation and management.
Georgina Brennan (Bangor University); Ulrich Brose (iDiv); Ursula Gaedke (Potsdam University); Joshua Garland (Santa Fe Institute); Alison Iles (iDiv); Ute Jacob (University of Hamburg); Pavel Kratina (Queen Mary University of London); Blake Matthews (Eawag); Mark Novak (Oregon State University); Frank Pennekamp (University of Zurich); Owen Petchey (University of Zurich); Benjamin Rosenbaum (iDiv); Andrea Tabi (University of Zurich); Colette Ward (University of Zürich); Richard Williams (Slice Intelligence); Gian Marco Palamara (University of Zurich); Stephan Munch (UC Santa Cruz); Björn Rall (iDiv)
2nd meeting: 23.-27.10.2017
Alison Iles (Oregon State University); Frank Pennekamp (University of Zurich); Ulrich Brose (iDiv); Joshua Garland (Santa Fe Institute); Owen Petchey (University of Zurich); Benjamin Rosenbaum (iDiv); Stephan Munch (UC Santa Cruz); Andrea Tabi (University of Zurich); Richard Williams (Slice Intelligence); Hao Ye (University of California San Diego)
Pennekamp F., Iles A. C. et al. (2018) The intrinsic predictability of ecological time series and its potential to guide forecastin. Ecological Monographs. See here