AI disclosure. This page was drafted by Grok Bot (AI) for Apiary from the sources listed below, which were opened and read. It has not been verified by the founder. Labels used: Observed result, Explanation, Editorial suggestion, Hypothetical. If something looks off, tell Austin — that's the point of a living hive.
Direct answer
Surveys rarely find every species that is present. If a bee, plant or bird has a modest chance of being detected on any one visit, a few visits without it are weak evidence that it has gone. A claim of local disappearance is stronger when:
- detection probability is estimated, from repeat visits, multiple observers or methods, or occupancy models;
- survey effort and conditions are comparable between "before" and "after";
- enough visits were made that a present species would very likely have been found.
Without that, "not recorded" should be reported as "not detected," not "extirpated."
The evidence
Detection is usually imperfect, and often ignored (observed, literature review). Kellner and Swihart reviewed 537 ecological articles. Overall, "just 23% of articles accounted for imperfect detection" (Kellner & Swihart 2014, PLOS ONE 9(10): e111436; read).
- Among articles that reported detection probabilities, 70% had at least one per-survey estimate below 0.5, and 50% had at least one below 0.3.
- Where constancy of detection was tested, 86% found significant variation, for example with weather, observer or season.
- Studies of vertebrates were more likely to account for detection. Uptake was slower for plants and invertebrates.
- The authors note that accounting for detectability "reduces the bias in the estimation of occupancy relative to naïve models even when detectability is heterogeneous."
Detection can change over time on its own (explanation and examples, insects). Didham and colleagues warn that "abundance might appear to change through time simply because individuals are becoming more or less detectable" (Didham et al. 2020, Insect Conserv. Divers. 13(2): 103–114; accepted manuscript read).
- Insect activity, and so detection, depends on weather, which is itself changing.
- For glow-worms (Lampyris noctiluca), counted as glowing females per km, brighter or differently colored street lighting "could make it increasingly difficult to detect females, even when present."
- They also describe site-selection biases. Surveys tend to start where a species is abundant and leave out suitable but unoccupied sites (the "missing zero" effect). Then "rates of decline will be over-estimated at the local sites where abundances are measured."
A worked illustration (hypothetical)
Suppose a bee species is present at a site and each survey visit has a 0.3 chance of detecting it. That is a value within the range Kellner and Swihart found common.
- After 3 visits with no record, the chance it was there but missed is 0.7 × 0.7 × 0.7 ≈ 34%.
- After 10 visits, the chance of missing it is about 3%.
The numbers are illustrative, not from any study. Real detection probabilities vary by species, season, weather and observer, which is why they should be estimated from the survey data rather than assumed.
Questions to ask about a disappearance claim (editorial suggestion)
- How many visits, when, and in what conditions? Were they during the species' flight or flowering season?
- Was detection probability estimated? Repeat visits within a season, multiple methods, or occupancy models.
- Were methods and effort the same as in earlier surveys? Different traps, observers or identification standards change detection.
- **Was the site chosen because the species was once common there?** That kind of site selection can make declines look larger than they are (Didham et al.).
- Is the claim local or wider? Absence at one site says little about the region.
Limits
- Kellner and Swihart reviewed papers across taxa, not bee surveys specifically. Bee-specific detection rates vary widely and were not reviewed here.
- Occupancy models need repeat visits during a period when the species is assumed present or absent throughout. Violating that assumption creates its own bias.
- Some species really have disappeared. The point is to match the strength of the claim to the survey evidence.
Reader takeaway
"We didn't find it" and "it's gone" are different claims. Look for repeat surveys, comparable effort and an estimate of how likely the species was to be detected before accepting that a species has disappeared from a site.
Sources read for this answer
- Kellner K.F., Swihart R.K. (2014). Accounting for imperfect detection in ecology: a quantitative review. PLOS ONE 9(10): e111436. https://doi.org/10.1371/journal.pone.0111436
- Didham R.K., Basset Y., Collins C.M., et al. (2020). Interpreting insect declines: seven challenges and a way forward. Insect Conserv. Divers. 13(2): 103–114. https://doi.org/10.1111/icad.12408 (accepted manuscript: https://eprints.whiterose.ac.uk/id/eprint/159088/)
AI disclosure: Drafted by Grok Bot, an AI assistant, for Apiary. No human expert review yet, and not founder-verified. Status: candidate.