Consider a hypothetical acoustic survey. A recorder operates for an hour, its recording is usable, and the agreed review method detects no calls of the target type. Zero detections is a defensible entry. If the battery is flat, there is no comparable hour. If noise masks the relevant frequencies, the recording exists but its ability to support a detection claim has changed. None of these states proves that animals were absent.

Keep observation effort beside the count

Suppose, purely for illustration, one week contains ten detections in ten usable recording hours and another contains ten in one hundred. The totals match; the rates do not. Even a rate needs the definition of usable effort, including recording schedule, review coverage, and the detection procedure. Dividing by all calendar hours would quietly treat downtime as successful listening. Dividing by all recorded hours could hide masking or corrupted files.

Our account of listening for narwhals explains why a hydrophone records a soundscape rather than a population census. For a comparable time series, retain both the count and its exposure: the duration examined under the stated conditions. If a classifier changes, document the change and assess comparability before interpreting a jump as a change in animal behaviour.

Name the reason a value is absent

Not measured, not yet processed, rejected by quality control, below a detection limit, and withheld under a governance agreement are different meanings. A single public-facing gap may stand for any of them, but the underlying record should preserve the distinction at the appropriate access level. Otherwise a later user cannot know whether to wait, exclude a period, change a method, or request permission.

The CF conventions on missing data explain how attributes such as _FillValue identify undefined or missing values in netCDF. Their flag conventions support separate descriptions of instrument status and quality assessments. The important principle travels to an ordinary spreadsheet: a code needs an explicit definition, and it must not be mistaken for a measured number.

A result below a laboratory reporting limit deserves particular care. It is not automatically zero, nor is it the reporting limit itself. Preserve the limit, unit, and qualifier. The analysis may need a method designed for such observations. Replacing every qualified result with a convenient number can create a spurious average and conceal the uncertainty that prompted the qualifier.

Ask whether the gaps follow the conditions

Missingness is especially troublesome when it depends on the thing being studied. Rough weather might prevent servicing or sampling. Cloud can obstruct an optical view. A route may be inaccessible during the very conditions a project hopes to understand. The remaining observations then describe the accessible conditions more faithfully than the full season.

The NOAA observing-network assessment describes uneven spatial and seasonal coverage, satellite limitations, and gaps in field validation. Those limits cannot be repaired by colouring an empty map cell to match its neighbours. An interpolated estimate may be useful, but its assumptions, source observations, and uncertainty need to remain visible.

Before calculating a monthly mean, inspect coverage across days, times, and conditions. State the completeness rule and the number of accepted observations. There is no universal percentage that makes every monthly average representative: a missing storm peak and a scattered set of missing minutes have different consequences. A rule should reflect the quantity and decision, not simply produce a complete-looking chart.

Let a chart admit what it does not know

Leave a visible break where observations are unavailable, or distinguish estimates from measurements with a clear explanation. Include effort in a companion table or panel when changing coverage affects interpretation. Do not use the same pale colour for a genuine zero and no data. A reader should not have to open the processing code to discover that the quietest month was the month the recorder failed.

These are small editorial decisions with substantial consequences. The Open observations guide puts missing-value meanings alongside methods and calibration because all three govern what a number can support. A blank cell is not an embarrassment to conceal. Properly labelled, it tells the next observer where knowledge ends and the next useful question begins.

This is a question about what an instrument could have found, not about an illustrative encounter with a particular species. NOAA Fisheries’ acoustics guide explains that hydrophones have different listening ranges, human-made noise can limit detection, and converting calls to population density requires information about call rates. Each qualification changes how a blank interval should be read.

Separate the links in the detection chain

First, an animal must be within a relevant area. Next, it must produce the kind of sound being monitored. That sound must reach the instrument within its usable frequency range and above the effective detection threshold. Finally, an analyst or detector must recognise it. Failure at any link can produce the same recorded outcome: no confirmed call.

Listening range is not necessarily a permanent circle around the hydrophone. Propagation changes with water structure, depth, bottom and ice conditions, and frequency. Background sound changes too. A nearby faint call and a distant strong call need not have the same probability of detection. A map should not imply a sharp, fixed surveillance boundary unless the method can justify one.

What changing recording effort can do

Take an invented comparison. Recorder A listens continuously for 24 hours. Recorder B listens for ten minutes in each hour, giving four hours of recording across the same day. If calls occur in short bouts, B may miss a bout entirely. Dividing its detections by four does not necessarily recreate the continuous record, because the timing of the gaps matters.

Now suppose a later deployment uses a detector that recognises weaker calls. More detections could reflect better sensitivity rather than more animals. Conversely, increased vessel noise could reduce detections while the animals remain present. The first comparison should concern recording schedules, usable hours, settings, noise conditions, and detector versions, before interpreting the ecological change.

The example contains no measured call rates or assumed animal abundance. It simply shows why ‘days deployed’ is not always the relevant denominator. Days powered but not recording, files lost, and intervals rejected for quality should remain separate from periods that were analysed and contained no detected calls. Our open-observations guide explains the value of keeping those missing-data meanings distinct.

A detector threshold is a choice with consequences

Automated detection can make large archives manageable, but every classifier has errors. A permissive threshold can capture more genuine calls while also admitting more false positives. A stricter one may miss faint or unusual calls. Reporting only the chosen threshold hides whether the result is robust to that choice.

A useful evaluation checks detections and non-detections against manually reviewed material that represents the conditions being analysed. Quiet examples alone cannot show how the detector performs during ice noise or heavy vessel traffic. Review should also preserve an uncertain category where species identity is unresolved, rather than requiring every sound to become a confident label.

Combining visual and acoustic observations can help establish which sounds belong to which species, as NOAA describes. It does not remove all ambiguity: animals can be out of sight, several species can overlap, and the visual survey has its own detection limits. Independent evidence is most helpful when its limitations differ, not when both methods repeat the same blind spot.

What the result can responsibly say

‘No confirmed calls were detected in the usable recordings under this method’ is a real finding. ‘No whales were present’ requires a stronger detection model and supporting assumptions. Similarly, counting calls is not counting individuals: one animal may call repeatedly, while another remains silent. Behavioural context and call-rate knowledge are essential if the question concerns density rather than acoustic occurrence.

NOAA’s passive-acoustics portal provides tools for exploring detection patterns across instruments and projects. When reading such a display, look for effort, methods, and coverage alongside coloured detections. The wildlife and food-web overview places sound within the wider habitat. A well-described silence can be valuable evidence. It becomes misleading only when we make it say more than the listening system could establish.

Lowering an ear into the water

A hydrophone converts pressure changes into an electrical signal. Before it enters the water, a useful record already needs instrument identity, calibration, clock, sampling settings, cable configuration, position, depth plan, and the condition of connectors. Once deployed, sea state, vessel activity, ice, weather, and handling noise become part of the context.

The boat itself is loud. Engines, pumps, hull movement, cable vibration, and voices can enter a recording. A deployment protocol may include quiet periods and notes marking unavoidable noise. Analysts need those notes because a clean-looking signal can come from equipment, and a biological call can disappear beneath overlapping sound.

Methods recorded with every deployment

1. Confirm instrument identity, calibration, clock, and recording settings.
2. Record position at a suitable public scale, water depth, hydrophone depth, and deployment geometry.
3. Log sea state, ice conditions, weather, vessel state, and nearby vessel activity.
4. Mark handling, engine, cable, and other known noise in the field log.
5. Review data quality before leaving if conditions and equipment allow.
6. Store the recording with access, sensitivity, and location-governance terms.

A small context table

From a click to a classification

Narwhals use clicks and other sounds for orientation and communication. Classification may compare frequency, duration, shape, sequence, and context with reviewed examples. Confidence should be recorded. A signal can remain unknown, and forcing it into a label does not improve the evidence. Human review and automated methods each need a documented threshold and error process.

Vessel noise can mask calls, especially when frequencies and timing overlap. A period with no detected call is therefore not proof that no narwhal was present. Detection depends on distance, orientation, propagation, background noise, instrument state, and the classifier. Effort and conditions make an absence interpretable.

A changing ocean context

The NOAA Arctic Report Card chapter on Atlantification describes changes in Arctic Ocean structure and ecological context. An acoustic record can contribute detail, but it cannot attribute a regional change on its own. Sound, species observations, vessel information, ice, and ocean measurements need to be read together.

A quieter-ocean action can be operational and policy-based: reduce avoidable noise, plan routes and speeds with current evidence and authority, maintain equipment, and monitor whether a measure changes exposure. The specific choice depends on safety, law, vessel type, place, and season. Visit Take action for a general route from evidence to an accountable request.

Listen without pretending certainty

The strongest outcome of one deployment may be a well-documented file that another analyst can understand, including the moments too noisy to classify. That is not a disappointing result. It is the basis for comparing seasons and methods without hiding the conditions that shape what the hydrophone could hear.

Location can be sensitive

A precise acoustic location may reveal habitat use or activity that should not be made openly searchable. Data agreements can preserve the scientific record while limiting public coordinates, raw audio, or timing. Those terms should be attached to the file before upload and retained in derived products, not remembered only by the original team.

Human voices create another privacy issue. A hydrophone near a boat may record conversation unintentionally. A review process should identify and restrict or remove personal sound according to agreed policy. Open ocean data do not erase the rights of people whose voices happen to enter the channel.

When results are published, representative clips should retain their processing history. Filtering can make a call easier to hear, but the unprocessed source, settings, and reason for the edit need preservation. A compelling audio player should not conceal how much interpretation sits between pressure in the water and sound from a speaker.