Environmental monitoring underpins conservation, infrastructure planning, climate adaptation, and risk assessment. Without reliable data on biodiversity, we are making billion-dollar decisions in the dark. (And we lose out on the joy of understanding the natural world around us.)
The biodiversity databases that exist today are invaluable. We wouldn’t try to assign them a value — but we can begin to estimate what it would cost to generate a similar volume of observations.
Let’s start with insects. In GBIF, there are ~284 million geo-referenced insect occurrence records
Across workflows, species-level insect identification commonly costs ~$3 per species observation, depending on the method and accounting assumptions.
Therefore, GBIF insect records represent ≈ $852 million
There are many ways to collect insects — Malaise traps, light traps, pitfall traps, etc. At Jonah, we’ve been exploring the potential of airDNA and doing some back-of-the-envelope calculations.
Our work to date suggests that a typical airDNA sampler, using our standard workflow, detects about 30 insect species in a 24-hour period.
If so, how many samples would it take to match existing observation volumes?
To reach GBIF’s 284M records → 284,000,000 ÷ 30 ≈ 9.5 million samples
Now assume we deploy 100,000 airDNA samplers globally — a large but technically feasible network.
At one sample per day per unit: GBIF equivalent → ~95 days
With sufficient infrastructure, decades of accumulated observational volume could theoretically be generated in months.
To analyze the samples, we would need to scale lab infrastructure — but 10M samples is on the order of the number of records in BOLD and far below the ~1B qPCR SARS CoV-2 analyses performed in the US during the pandemic.
What would it cost?
Basic capital cost per unit:
• $10 fan
• $30 brackets or tripod
• $50 battery + charger
= $90 per sampler
Deploying 100,000 units would require roughly $9 million in hardware
The larger expense is lab processing.
Assume $100 per sample (a round, potentially conservative estimate):
• 9.5M samples → $950M
So, roughly $1B to generate GBIF-level insect observation volume using this approach (ignoring deployment, administrative, or data-management costs).
And this is assuming we don’t learn to identify more insects in a single airDNA sample. If we double the number of species we detect in a sample, costs halve.
Of course, there are always caveats:
• AirDNA samples insect assemblages differently than traditional methods
• Spatial placement of 100k samplers would be constrained
• Not all observation types are substitutable — a pinned specimen may hold different long-term value than a DNA detection
Still, this exercise helps bracket the scale of the problem.
Regardless of method, biodiversity monitoring at global scale–whether aquatic or terrestrial–is an infrastructure, processing, and systems design challenge.
But more importantly, in summary, it’s doable.