Monitoring & conservation
Counting nature means estimating detection probability; the instrument matters less than the inference about what you failed to see.
The cluster rests on one quantity that rarely appears in reports: detection probability. Any method — DNA from water, sound in a night forest, a colony under a paper microscope — yields not a census but a sample with an unknown miss rate, and all of the cluster’s honest statistics grew out of fighting “not detected” as a synonym for “absent”.
The second recurring idea is measurement infrastructure: data becomes measurement only against a standard. Certified reference materials and characterised strains, reference genomes, pangenomes — without them, inter-laboratory results become disputes rather than calibratable offsets. The third idea is the limit of intervention: genetic rescue and even de-extinction run not into genome editing but into everything else a phenotype is made of — behaviour, ecology of return.
Start with eDNA invasive-species biomonitoring: it introduces the detection-probability vocabulary the rest of the cluster leans on.
Articles 8
| Environmental DNA for species monitoring Organisms shed DNA into water and sediment, PCR can find a single copy of it, and everything hard about eDNA monitoring lives in the gap between those two facts. |
| Passive acoustic monitoring Sound carries species identity in its frequency and time structure; propagation and noise decide how far it carries, and what the data deliver is detection probability, not a headcount. |
| De-extinction: what the biology allows Ancient DNA is fragmentary and damaged, editing rewrites specified loci rather than a species, and gestation plus learned behaviour hold the last mile — what de-extinction can honestly reconstruct is a proxy. |
| The optics of a paper microscope A ball lens gives magnification almost for free; resolution is where the physics collects its debt — and yet cheapness still buys real coverage of the observable world. |
| Why reference materials make measurements comparable A certificate is not a quality label but a chain: a stated value, a stated uncertainty and a documented path to an agreed scale — without which laboratories measure the same thing and cannot agree on the number. |
| Reference genomes as measurement infrastructure A genome database is not an archive but the ruler: reads are interpreted only against it, and a database that lacks your species or your population distorts every measurement made with it. |
| The built-environment microbiome Indoor microbial communities are an ecology with sources, sinks and dispersal routes; sterilization ambition collides with that ecology, and the collision usually selects for the organisms you want least. |
| Altitude training: the mechanism and its limits Thin air stabilises a transcription factor, the factor switches on erythropoietin, and more red cells carry more oxygen — the chain is real, and so are the reasons controlled trials keep returning modest and uneven results. |