Tropical pollinator populations could be reduced by half this century

Three decades ago ecologists wading through Venezuela’s Moriche palm-punctuated swamps came across an ecosystem where multiple palm and non‑palm plant species shared groups of beetles, flies and bees, spinning a dense web of interactions across the flowering community.

Today, network datasets make those invisible connections visible and expose how fragile they can be, as global warming changes ecosystems worldwide. New research², using these open datasets, shows tropical plant–pollinator systems that are near their thermal limits could see pollinator populations, the lynchpin for biodiversity, crop yields, and food security, plunge by roughly 50% this century, under high-emissions scenarios.

“Tropical networks require urgent, intensive management,” says the study’s principal investigator Udit Bhatia at the Indian Institute of Technology, Gandhinagar. Temperate systems, such as those found in much of Europe and North America, by contrast, may face declines of around 5%, even under severe warming, though Mediterranean networks showed steep declines driven by intense summer heat and sharp seasonal fluctuations.

To understand how plant-pollinator networks respond to warming, researchers combined real ecological networks, temperature-dependent species dynamics, and future climate projections into a single modelling framework. The study drew on ecological data from 11 plant–pollinator networks in the Web of Life database, spanning tropical (including Venezuelan Morichal swamps), temperate and Mediterranean ecosystems, driven by monthly temperature projections from ten CMIP6 climate models.

Species growth, decline and interactions were simulated using nonlinear population dynamics informed by experiments on insect thermal responses, allowing the model to capture how warming affects species physiologically and how those effects cascade through networks of interactions.

Bhatia and co-authors at Northeastern University worry that pollinator declines are far more likely in India and most of Asia even under moderate warming. Tropical pollinator decline may accelerate through the 2040s, with the most severe impacts occurring after 2060. However, the study notes that relative temperate stability does not imply long-term safety.

“Temperate regions may have a larger window to mount solutions before reaching similar crisis points,” says quantitative ecosystem network modeller, Gemma Gerber, at the International Institute for Applied Systems Analysis (IIASA), Vienna. Lauren Talluto, IIASA’s quantitative macroecologist, adds: "There's no room for complacency because temperate systems may still need to tackle other stressors, given the multifaceted nature of climate change." Neither of the IIASA scientists was involved in the study.

One size does not fit all in conservation

Altogether, the findings reinforce the conclusion that conservation should be region-specific. Because plants are the heart of tropical ecosystems, boosting their diversity can set off powerful positive ripple effects, and these most at-risk systems also have the most to gain from a rich mix of plants, the simulation shows. “Gains from similar actions are much smaller in temperate and Mediterranean regions,” says co-author, Auroop Ganguly, at Northeastern University, in Boston, United States.

This mosaic led to a more balanced species distribution, supporting a wider range of pollinators and the system as a whole was less vulnerable to the loss of any single species. By contrast, efforts that focused directly on pollinators, such as bolstering the numbers of particular bees or butterflies, often produced only short-term gains. These tended to favour already dominant species, squeezing out specialists and, in some cases, making the network less stable overall.

Gerber acknowledges that a pivot from direct pollinator management to a plant-focused approach would require different infrastructure, skills, and timelines; the latter strategy, the study notes, would have to embrace diversified floral plantings, hedgerows, sequential bloom schemes, and long-term monitoring.

Practitioners may be locked into a pathway that prevents them from quickly shifting to a plant-focused approach. Also, plant communities take years to establish, whereas supplementing pollinator populations has an immediate effect, adds Gerber. “A hybrid strategy of keeping up critical direct pollinator support (managed colonies) while building plant diversity infrastructure and allowing gradual transitions as plant communities mature, may be more practical,” says Talluto, who notes that abandoning single-species management risks losing high-value individual species.

Vulnerability beyond temperature

In the networks studied, tropical ecosystems have historically experienced relatively stable temperatures, with average seasonal swings of about 1.28°C, compared with 4.59°C in temperate regions. As a result, tropical insects evolved within narrow thermal limits, leaving them less flexible in navigating fluctuations than their temperate counterparts.

“Their genetics are set around higher thermal optima, evolved in stable conditions. Temperate insects developed traits like diapause (dormancy periods) that buffer them against seasonal extremes,” says ecological and evolutionary physiologist, Subhash Rajpurohit at Ahmedabad University in Gujarat, India.

There could also be a fundamental difference in how the two groups cope with heat itself. Temperate insects experience regular, predictable heat stress seasonally, so they evolved acute heat shock responses, Rajpurohit, who was not associated with the study, notes. Tropical insects are increasingly exposed to chronic and unfamiliar heat stress, and scientists still do not fully understand how they will respond to sustained warming, particularly in high-altitude regions. According to Rajpurohit, plant–insect interactions in the tropics are shaped by multiple, interacting pressures, with pesticides in some cases posing an even greater threat than rising temperatures alone.

These uncertainties underscore the need for better tools to anticipate ecosystem tipping points before they are crossed, says Talluto. Current models often fail to account for local climate refuges, behavioural and timing shifts, vulnerabilities at different life stages, indirect effects such as reduced nectar production, and the combined impacts of habitat loss, and disease.

Using the present study as an example, she notes that very little is known about how flexible plant–pollinator relationships really are — including whether pollinators can readily switch to alternative plants if preferred species decline, and what proportion of pollinations for a plant species can be attributed to each potential pollinator species, and are these values sensitive to changes in the composition of pollinators? “This source of uncertainty in our forecasts is rarely incorporated into modelling efforts and thus not quantified, and it also represents a data gap.”

To build a more comprehensive risk model, next, the study authors plan to integrate additional pressures, including land-use change, chemical exposure, and fragmentation.

Ecological modeller, Partha Sharathi Dutta, who ran an earlier simulation³ assuming average temperature increases, says that the current modelling using CMIP6 projects a softer outlook for the temperate and Mediterranean regions — the likelihood of abrupt collapse decreases. “But it doesn’t remove the danger for the tropics. Tropical ecosystems remain at high risk of abrupt transitions similar to those initially predicted.” However, static interaction matrices often ignore the dynamic rewiring that may prevent collapse during heatwaves. "Experimental studies with realistic setups are necessary to assess these modelling results," says Dutta. "With machine learning and advanced statistical methods, we will be able to forewarn upcoming tipping points."