Field sampling, tracking, remote sensing and environmental DNA

Observation coverage is part of the inference

A field sample records selected places, times or organisms within a stated wider system. Counts, occurrence records and environmental measurements can each be informative, but they refer to different quantities. Absence from a record can reflect absence in the observed area, imperfect detection or a part of the system that was never observed. The appropriate spatial boundary and inclusion rule therefore belong with the result.

The plot account makes detection complete inside four included areas, while leaving everything outside them unknown. Their counts sum to sixteen individuals across eighty square metres. Dividing by that observed area gives density 0.2 individuals per square metre. Applying it to the entire region would require a justified relation between selected areas and the wider system; the arithmetic does not establish representativeness.

Accessibility, habitat coverage, seasonal timing and observer conditions can affect which records exist. Nearby observations can also share conditions, so the number of plots does not automatically settle independence. A presence-only collection shows where observations were made, while unrecorded places are not automatically confirmed absences. Clear coverage and detection evidence help separate a biological pattern from a pattern of observation.

A position record is not a continuously observed route

Tracking associates observations with the same identified organism over time. Positions can reveal displacement, visited areas and relationships with recorded environmental conditions. They do not by themselves establish every behaviour between observations. Missing locations, positional error, observation intervals and effects of an observation device can all matter for the question being asked.

The exact point case separates two geometric accounts. Start-to-end displacement is the straight vector from the first position to the last. The recorded polyline adds straight distances between every successive pair of positions. If the intervening movement was not observed, that polyline is a lower bound on the travelled path, because a curved or backtracking journey could be longer.

The same individual supplies all three positions in this case. More positions can improve knowledge of that individual’s history without becoming more independently observed animals. A selected group of tracked individuals also does not automatically represent an entire population. Evidence about the observation scale and which individuals are included is necessary before a movement record explains a species-wide mechanism or population pattern.

Remote sensing records physical signals over a footprint

Remote sensing observes signals from a distance, often over a sensor’s spatial footprint and selected spectral bands. A pixel can summarise a mixture of surfaces and biological states. The recorded quantity is a physical observation; relating it to vegetation, habitat or organisms requires compatible conditions and a biological interpretation rather than a direct change of units.

The normalised difference vegetation index (NDVI) compares near-infrared and red reflectance as their difference divided by their sum. In the two hypothetical records, one pixel has values 0.6 and 0.2 and the other 0.3 and 0.1. Both give index 0.5 even though the band values differ. The ratio therefore does not retain every part of the original optical record.

Nor does this index identify the species, native status or ecological interactions beneath a pixel. Illumination, atmosphere, viewing geometry, surface composition and observation timing can affect interpretation. Field evidence can help connect a physical proxy to a particular biological question, but an equal greenness index is not proof of equal biodiversity, and a changing index need not identify which species changed.

Environmental DNA needs an origin and detection account

Environmental DNA is genetic material observed in an environmental sample, potentially originating from organisms, their cells or released material. An analytical detection provides evidence about the included DNA under the assay’s specificity and observation conditions. Current local living presence is a further claim: DNA can be transported from elsewhere or retained from an earlier source, and detection does not alone show a viable population.

A negative record can occur even with a local living source because production, transport, degradation and observation affect detectability. The hypothetical probability case fixes local presence and assigns a 0.6 positive-record chance at each of three conditionally independent opportunities. The chance of no positive record is 0.4 cubed, or 0.064. This is a record probability given presence, not the probability of absence after three negative records.

A separate case fixes local living absence while allowing positive records with probability 0.05 at each opportunity. The model can include transported or retained DNA and analytical errors in that ecological-record probability; it does not call every such record an analytical false positive. No prior occurrence probability is supplied, so a posterior presence estimate cannot be calculated from either case alone. Abundance, identity, origin, persistence and detectability need appropriately different evidence.

Worked example

Check your understanding

1. Does observed density 0.2 in the four complete plots establish eighty individuals in the whole region?

   Show reasoning

   No. That regional estimate requires the included density to represent the unobserved area under justified selection and observation conditions. The plots themselves establish sixteen individuals in eighty square metres. The other 320 square metres were not counted, so the regional total is still unknown.

2. Why are the seven-metre recorded polyline and five-metre displacement not competing measurements of the same path quantity?

   Show reasoning

   Displacement compares only the starting and ending positions. The polyline includes the intermediate position and sums straight links, giving a lower bound on actual travel. The intervening route is unobserved and can be longer. Both calculations concern one same individual and answer different geometric questions.

3. Do three negative DNA records give absence probability 0.064, and does a positive prove current local living presence?

   Show reasoning

   No. The value 0.064 is the probability of three negatives conditional on the supplied presence case, not absence conditional on negatives. Positive DNA can also have transported or historical origins. Inferring current local state needs additional origin, detection and prior evidence; the model supplies no posterior occurrence probability.

What to carry forward

Keep coverage, identity, time, physical proxy and conditional detection explicit before extending a field record to a wider biological conclusion.