Davant l’augment de les temperatures, conservar microclimes favorables pot marcar la diferència entre l’extinció local i la persistència d’una espècie. Hi ha grans solucions als mons petits que també ens beneficien els humans.
✍️ @jdiazcalafat.bsky.social, @uib.cat @culturacientifica.uib.cat
🌍 Take-home: fly pollination in boreal forests is shaped by climate and forest structure. As climates warm, local conditions (canopy cover, tree composition, microclimate) will strongly mediate how pollinators respond.
Read more about this piece of research here: doi.org/10.1002/oik.... 🪰🌡️🌼🌳🌲
☔ Rainfall mattered: More recent rains = fewer visits. But flies increased their activity after rain events.
🌼 Natural, real flowers found in each surveying plot had mixed effects:
- More species = longer fly foraging time, but fewer total (plastic!) flower visits.
- Abundant flowers = more visits, but shorter foraging per visit.
🌲 Forest composition matters: dense, spruce-dominated stands had fewer fly visits than more open, broadleaf-rich or mixed forests. Closed canopies reduce light and warmth at the forest floor, limiting activity of ectothermic pollinators.
On the other hand, microclimate (fine-scale forest temperature variations) better predicted how long flies stayed on the flowers once they landed.
🌡️ Macroclimate (broad temperature patterns) was the main driver of pollinator visitation rates. Plus, flies were most active within a narrow temperature window.
Roughly 6.5% of all the pictures had insects in them, of which about 97% were flies.
This resulted in many many pictures (2 185 092, representing 6 069.7 h of continuous sampling) that needed to be screened for insects. For this, we used a vision AI model (YOLOv5).
Then, we used time-lapse cameras to monitor insect activity every 10 seconds across a double gradient of forest density (dense vs. open plots) and tree species composition (pure spruce vs. pure birch). We also measured microclimatic temperatures at each camera.
So we borrowed the plastic flower design from Thomson et al. (2012), with some modifications. "Flowers" consisted of a jar filled with a sugary solution, with a small receptacle acting as the flower, in which the "nectar" was available through capillarity. The lid was painted in bright UV colors.
The problem with forest understories, though, is that flowers are not available in the shadiest areas, and we really wanted to measure pollinator foraging under different light conditions.
🐝 Unlike bees, many pollinators in boreal forests are flies (Diptera, especially Muscidae, Phoridae and Syrphidae). These ectotherms depend heavily on external heat for activity.
📢 New paper out in @oikosjournal.bsky.social !
We investigated how climate🌡️, forest structure🌲🌳 and flower resources🌼🌷 shape fly pollinator foraging in boreal forest understories 🪰🪰🪰
doi.org/10.1002/oik....
📖 Dive into the full paper for details on methods, stats, and ecological implications! onlinelibrary.wiley.com/doi/10.1002/...
🧩 Conclusions: Flower eDNA is a promising non-invasive tool for capturing arthropod–plant interactions in forest ecosystems. However, detection rates are patchy—likely influenced by flower traits, visitor behavior, and methodological nuances.
🔍 Method matters: pooling flowers before DNA extraction increased species detection. This is a key methodological insight for future eDNA sampling.
❓ Unexpectedly, microclimate (temperature) showed no significant effect on arthropod detection for either plant species. Forest structure mattered more.
🌳 For F. vesca, arthropod richness was higher in open, broadleaf-dominated plots—likely due to more light boosting pollinator activity. Forest density had a negative effect.
🔬 Using COI metabarcoding on washed flower samples, we detected 92 arthropod taxa—34 on F. vesca, 64 on T. pratense. Notably, T. pratense hosted nearly double the richness. Still, many taxa appeared only once per plot.
🌸 We translocated Fragaria vesca (wild strawberry) and Trifolium pratense (red clover) into 40 forest plots varying in density and tree composition. Each plant came paired: one accessible to pollinators, one covered as a control.
🌿 New paper alert! 📢
We used flower-based eDNA metabarcoding to uncover how forest structure and microclimate shape the diversity of flower-visiting arthropods in a Swedish boreal forest. #pollinators #eDNA 🧬🌼
🙏 Huge thanks to all the volunteers contributing to biodiversity platforms. Your observations matter, immensely.
We hope this work helps make your data even more valuable.
📘 Full paper: doi.org/10.1007/s135...
📂 Code & data: zenodo.org/record/15252...
Citizen science is powerful, but not perfect.
Ignoring temporal bias can lead to flawed conclusions about how pollinators respond to climate change, risking misguided conservation efforts.
We can do better. And we must. 🐝🌍
So, what can we do?
✅ Include weekday/weekend in statistical models
✅ Be transparent about biases
✅ Combine citizen science with other datasets
✅ Promote standardized protocols
Some species, like Bombus hypnorum, are especially prone to weekend bias, likely because they nest in vertical structures and often enter buildings. Their urban, visible habits make them easier to spot.
Bias isn’t just temporal: it’s tied to species traits too. 🐝🏙️
When we modeled the emergence date of Bombus hortorum without adjusting for weekend bias, we found a significant advance over time.
After adjusting for the day of the week?
🚫 The trend disappeared.
📉 Bias can send us down the wrong path.
We compared 2M+ citizen science records and 340k museum specimens.
🎯 Citizen science peaked on weekends.
🏛️ Museum data? The opposite—more records on weekdays.
Two biases, two directions. Both matter for how we interpret trends.
One-third (33%) of bumblebee records from citizen science platforms come from weekends. This is well above what random sampling would predict (!!!)
Why does this matter? Because weekend bias can distort our understanding of how pollinators respond to climate change.
As rising temperatures shift the timing of plant flowering and pollinator emergence, accurate phenological data is crucial, but hard to collect. Citizen science helps fill the gap, especially for tracking hibernating pollinators. 🐝🌸📅
But there's a catch...