Ecosystems are widely interconnected by spatial flows of material, but the overall importance of these flows relative to local ecosystem functioning remains unclear. Here we provide a quantitative synthesis on spatial flows of carbon connecting ecosystems worldwide. Cross-ecosystem flows range over eight orders of magnitude, bringing between 10⁻³ and 10⁵ gC m⁻² year⁻¹ to recipient ecosystems. Magnitudes are similar to local fluxes in freshwater and benthic ecosystems, but two to three orders of magnitude lower in terrestrial systems, demonstrating different dependencies on spatial flows among ecosystem types. The strong spatial couplings also indicate that ecosystems are vulnerable to alterations of cross-ecosystem flows. Thus, a reconsideration of ecosystem functioning, including a spatial perspective, is urgently needed.

Water pollution is ubiquitous globally, yet how the effects of pollutants propagate through natural ecosystems remains poorly understood. This is because the interactive effects of multiple stressors are generally hard to predict. Agriculture and municipal wastewater treatment plants (WWTPs) are often major sources of contaminants for streams, but their relative importance and the role of different pollutants (e.g. nutrients or pesticides) are largely unknown. Using a 'real world experiment' with sampling locations up- and downstream of WWTPs, we studied how effluent discharges affected water quality and macroinvertebrate communities in 23 Swiss streams across a broad land-use gradient.
Variation partitioning of community composition revealed that overall water quality explained approximately 30% of community variability, whereby nutrients and pesticides each independently explained 10% and 2%, respectively. Excluding oligochaetes (which were highly abundant downstream of the WWTPs) from the analyses, resulted in a relatively stronger influence (3%) of pesticides on the macroinvertebrate community composition, whereas nutrients had no influence. Generally, the macroinvertebrate community composition downstream of the WWTPs strongly reflected the upstream conditions, likely due to a combination of efficient treatment processes, environmental filtering and organismal dispersal. Wastewater impacts were most prominently by the Saprobic index, whereas the SPEAR index (a trait-based macroinvertebrate metrics reflecting sensitivity to pesticides) revealed a strong impact of arable cropping but only a weak impact of wastewater.
Overall, our results indicate that agriculture can have a stronger impact on headwater stream macroinvertebrate communities than discharges from WWTP. Yet, effects of wastewater-born micropollutants were clearly quantifiable among all other influence factors. Improving our ability to further quantify the impacts of micropollutants requires highly-resolved water quality and taxonomic data with adequate spatial and temporal sampling. These improvements would help to better account for the underlying causal pathways that drive observed biological responses, such as episodic contaminant peaks and dispersal-related processes.

Individual-level variation arising from responses to environmental gradients influences population and community dynamics. How such responses empirically relate to the mechanisms that govern species coexistence is, however, poorly understood. Previous results from lake phytoplankton communities suggested that the evenness of organismal traits in multiple dimensions increases with resource limitation, possibly due to resource partitioning at the individual level. Here we experimentally tested the emergence of this pattern by growing two phytoplankton species (Pseudokirchneriella subcapitata and Microcystis aeruginosa) under a gradient of light intensity, in monoculture and jointly. Under low light (resource) conditions, the populations diversified into a wide range of phenotypes, which were evenly distributed in multidimensional trait space (defined by four pigment-related trait dimensions), consistent with the observed field pattern. Our interpretation is that under conditions of light limitation, individual phytoplankton cells alter photosynthetic traits to reduce overlap in light acquisition, acquiring unexploited resources and thereby likely maximising individual success. Our results provide prime experimental evidence that resource limitation increases the evenness of conspecific and heterospecific microbial phenotypes along trait axes, advancing our understanding of trait-based coexistence.