Love them or hate them, recent reports of an "insect apocalypse" sounded alarm bells around the world as conservationists warned of dire repercussions for people and ecosystems.

But a wide-scoping meta-analysis of 166 long-term surveys across 1676 global sites shows the decline has considerable nuances – with some arthropods (insects and spiders) in free fall but others increasing – and is inconsistent across regions.

Previous analyses estimated that arthropod populations were dropping by 25% per decade, but the new European study, published in the journal Science, estimates that land-dwelling species have declined by 9% per decade on average.

They also found that freshwater insects like midges and mayflies have increased by around 11% per decade (although these only account for 10% of all species and don’t pollinate crops) which the authors suggest may partially neutralise the negative trend.

Although arthropods are diminutive in size, their biomass outweighs that of humans by a factor of 17.

These incredibly diverse critters are also highly industrious, providing vital services such as pollinating plants, aerating soil, decomposing waste and getting rid of pests.

Yet they are "critically understudied", note Maria Dornelas, from the University of St Andrews, and Gergana Daskalova, from the University of Edinburgh, Scotland, in a related commentary.

As an example, they cite the BioTIME biodiversity database that has records for 22% of bird species compared to just 3% of arthropods, which are tricky to study because of their short life spans and rapid population dynamics.

Variations also make it precarious to generalise from small scale studies.

"The drivers of biodiversity changes are almost never simple, and their discovery requires context," they note. "Thus, simple extrapolation from a handful of locations is unlikely to reveal the layers of complexity that underpin real-world biodiversity change."

For example, the meta-analysis found that overall declines were highest in North America, although these have plateaued, influencing the overall trends. They were also high in some European regions, where they have increased since 2005.

Falloffs were higher in unprotected than protected areas, which the authors, led by Roel van Klink from the German Centre for Integrative Biodiversity Research, link to urbanisation, possibly due to habitat loss and light and/or chemical pollution.

The true insect falloff could therefore have been underestimated because protected areas dominated the data, representing 34% of the sites whereas only 15% of the Earth’s land surface is under protection.

"This means that locations where human land use is most intensive, and thus where the strongest effects on insect trends might be expected, were underrepresented," they write, calling for more research from those regions to get a more accurate picture of human influences on insect losses.

The increase in freshwater insects they say could be partly due to successful clean water initiatives. That, combined with smaller declines in protected areas, gives some hope that conservation efforts such as habitat protection can help save insect species.

The losses are still disturbing, and because the analysis reports averages, some places have higher declines than others.

Dornelas and Daskalova draw comfort from the findings, urging scientists to factor variability into future biodiversity research, including drivers of risk, resilience and change.

"Embracing nuance allows us to balance accurate reporting of worrying losses with hopeful examples of wins. Hope is a more powerful engine of change than fear."