Only a handful of studies demonstrate the existence of hysteresis in bistable systems. In a follow-up from our earlier work, we study the recovery trajectory of a light-stressed plankton population in a chemostat experiment (early view in Oikos). We find that reverse trajectories can be explained by hysteresis, time-delays and adaptive process, all of which pose interesting questions for the behavior of bistable systems under changing conditions.
Recently, we summarized a set of measures that can be used as spatial indicators for detecting loss of resilience. We now add another measure that can be used as early warning of critical transitions in spatially explicit systems: spatial heteroskedasticity. In short, this indicator is the analog of conditional heteroskedasticity in timeseries (the non constant variance along a timeseries). We now expand its use from indicator of critical transitions in timeseries to spatial data (early view in Ecology & Evolution).
Recently our paper on spatial indicators for critical transitions was published in PloS One. In this paper we summarize methods and create a flowchart for looking for indicators of upcoming transitions in spatial data. It is a natural follow-up paper from our previous work on methods for timeseries. The methods of the paper are now summarized in the spatial indicators section of the EWS toolbox website together with the actual R code.
Together with Leo Lahti, we fixed bugs and moved the earlywarnings toolbox in R. It is now a library ready to be installed from your preferable CRAN repository. In the process, we also migrated the earlywarnings toolbox to github for shifitng towards open-source, community-based project development. We hope this will facilitate the use of the toolbox both for research and education. More details on the Early Warning Signals Toolbox webpage.
The methodological paper that we worked on during the Santa Fe Institute workshop last October became available few weeks ago. It presents a suite of most developed methods for identifying early-warnings and provides clear step-by-step examples on how to apply them, in an attempt to offer a protocol to the uninitiated into the field. Most of the content of the paper is presented in simple form- together with more supporting material- on the newly launched Early Warning Signals Toolbox. This website accompanies the paper and aims to be a source of existing methods, newly developed methods, and a repository of case studies. Have a look and if you are interested to contribute, don’t hesitate to contact me!