
A €700 camera powered by a solar panel recorded visiting insects
Five cameras recorded insects in German meadows. A computer classified hoverflies well, but beetles and true bugs less well.
A field measurement, an experiment and a scenario model can answer different questions about the environment. These readings bring together research on ecosystems, energy and changes on our planet.

Separate what was observed from what was projected. A finding from one place or period does not represent every region, and projections depend on assumptions. The sources let you revisit the design, scale and limits behind each finding.

Five cameras recorded insects in German meadows. A computer classified hoverflies well, but beetles and true bugs less well.

A U.S. estimate found that producing discarded food required less land but more water and pesticides.

In Paris, beetles, large solitary bees, and bumblebees visited flowers less where more hives were nearby.

Speakers played healthy-reef sounds over dead coral. After 40 days, those patches had more fish and more species.

A global estimate found that 92% of microfibers estimated to reach land traveled in treatment-plant sludge.

The tool combines temperature, vulnerability, and space for trees. In Boston, one map was not enough.

Asia had the largest share of potential, followed by North America and Europe, with differences across months.

Over 78 days, curtailed turbines generated 5% to 15% less electricity. With wind-only curtailment, even a reduction in deaths remained uncertain.

In simulations with 4°C more warming, waiting for usable thickness took nearly 20 extra days with white ice and about three with black ice, relative to historical black ice.

Combining more than 100 studies, the decline appeared for earthworms, especially in forests and farmland, but not for cranefly larvae.

In simulations of dry forests in Oregon, moderate-severity fire came closest to historical tree abundance. Species composition did not.

Young oysters from a Chilean estuary were studied for two months. Estimated energy available for growth did not differ between treatments.

At recovering sites, species composition explained only 2% of differences in ecosystem functions. The relationship was stronger for biomass and structure.

By 2050, the model estimated the smallest land share in India and the largest in Japan and South Korea; results also varied with solar’s share of electricity.

In models of 11 western U.S. states, stronger protections kept solar and wind projects out of intact landscapes.

In Stockholm student housing, the seasonal increase was smaller with both messages. Among higher-use homes, the difference was clearer with practical tips.

Both kinds of diversity went along with ecosystem functioning across 101 grasslands. In containers, they jointly supported production and nutrients.

In plots warmed for years, an index of processes and nutrients was higher in fall. Warming did not detectably change it.

After five years in experimental ponds four degrees warmer, diversity was 67% higher and production increased. Larger types became more common.

In freshwater containers, its mass fell by 85% to 95% with predators or without herbivores. The decline was smaller with herbivores and no predators.

Across 28 tanks, drought affected algae, invertebrates, and leaves in shallow areas. Combined with nutrients, it increased one fish species’ survival.

Where plots were cut four times a year, vegetation declined after drought. No decline was detected with two cuts. Species number had little effect.

In chambers with grassland blocks, species number showed no marked differences. Carbon went along with leaf nitrogen and height.

In the lab, ferrocene and phenothiazine lowered charged-cell voltage to about 0.1 V. In another test, phenothiazine mobilized deposited lithium.

Less spring rain did not always coincide with less photosynthesis. The response depended on water availability: at many sites without strong water limitation, dry springs coincided with greater carbon uptake. The opposite pattern appeared where water was scarce.

In the lab, 94% to 97% of measured metals were recovered from cathode waste. Water and energy savings came from a model.

By combining samples, shipboard observations, and a model, microplastics made up 92% of pieces while large pieces accounted for 75% of the mass.
