Artificial Intelligence Accidentally Learned Ecology through Video Games.

An advanced artificial intelligence (AI) system defeated the best human players in StarCraft II, a popular real-time strategy game. In a virtual ecosystem, players compete for habitats and resources, unintentionally reproducing many ecological phenomena. We propose to repurpose this AI to test ecological hypotheses that have been intractable using traditional approaches.

Transdisciplinary Bioblitz: Rapid biotic and abiotic inventory allows studying environmental changes over 60 years at the Biological Field Station of Paimpont (Brittany, France) and opens new interdisciplinary research opportunities.

Background: The Biological Field Station of Paimpont (Station Biologique de Paimpont, SBP), owned by the University of Rennes and located in the Brocéliande Forest of Brittany (France), has been hosting student scientific research and field trips during the last 60 years. The study area of the SBP is a landscape mosaic of 17 ha composed of gorse moors, forests, prairies, ponds and creeks. Land use has evolved over time.

Passive rewilding may (also) restore phylogenetically rich and functionally resilient forest plant communities.

Passive rewilding is increasingly seen as a promising tool to counterbalance biodiversity losses and recover native forest ecosystems. One key question, crucial to understanding assembly processes and conservation issues underlying land-use change, is the extent to which functional and phylogenetic diversity may recover in spontaneous recent woodlands. Here, we compared understorey plant communities of recent woodlands (which result from afforestation on agricultural lands during the 20th century) with those of ancient forests (uninterrupted for several centuries) in a hotspot of farmland abandonment in western Europe.

Unexpected spatial stability of water chemistry in headwater stream networks.

Understanding how water and solutes enter and propagate through freshwater landscapes in the Anthropocene is critical to protecting and restoring aquatic ecosystems and ensuring human water security. However, high hydrochemical variability in headwater streams, where most carbon and nutrients enter river networks, has hindered effective modelling and management. We developed an analytical framework informed by landscape ecology and catchment hydrology to quantify spatiotemporal variability across scales, which we tested in 56 headwater catchments, sampled periodically over 12 years in western France.

Functionally dissimilar neighbors accelerate litter decomposition in two grass species.

Plant litter decomposition is a key regulator of nutrient recycling. In a given environment, decomposition of litter from a focal species depends on its litter quality and on the efficiency of local decomposers. Both may be strongly modified by functional traits of neighboring species, but the consequences for decomposition of litter from the focal species remain unknown.