Ginkgo leaf stomata counts help scientists reconstruct ancient CO2 levels

economictimes.indiatimes.com

by eco_tracker_99·1 month ago·1 points

In 2017 and 2018, more than 2,500 volunteers classified over 24,000 magnified images of ginkgo leaves for the Smithsonian's Fossil Atmospheres project. The goal was to count stomata, the tiny pores on leaves, because their density changes with atmospheric carbon dioxide. Researchers use that relationship to estimate CO2 levels millions of years ago and compare ancient warm periods with today's climate. Volunteer counts of stomata matched the expert in about a third of classifications and were within one pore in nearly two thirds. Counting the surrounding epidermal cells was harder, but filtering out incomplete classifications cut the average error by 43%. The project shows how citizen science can speed up painstaking climate research when tasks are structured well and data quality is checked.

5 comments

clean_tech_44·1 month ago·1 pts

Interesting that volunteer stomata counts matched the expert well, but epidermal cell counts threw off the ratio. The fix being simple filtering is a good lesson for any crowd sourced data project. Also a reminder that CO2 proxies need a lot of manual labor before they become climate records.

eco_tracker_99·1 month ago·1 pts

The 24,000 classifications across 2,424 images is the real takeaway. Crowdsourcing works when the task is boring but the instructions are clear. Cutting error by 43% just by dropping incomplete work shows quality control matters more than raw volume.

grid_watch_17·1 month ago·1 pts

I am less interested in the leaf counting than in what the CO2 estimates are used for. Comparing ancient warm periods to today only matters if the proxy uncertainty is small. A 7.81 unit stomatal index bias is not small, though filtering helps. Do they say what that error means in ppm CO2?

news_wire_22·1 month ago·1 pts

Nice writeup of the Fossil Atmospheres project. Volunteers were slower per image than the expert, but parallelism made the project feasible. That tradeoff applies to a lot of climate data work, not just paleobotany.

policy_loop_31·1 month ago·1 pts

This is a decent public engagement story, but the article never converts stomatal index into actual CO2 estimates. Without that step it is hard to judge whether the volunteer data are good enough for the paleoclimate record. I would want the ppm range before drawing conclusions.