METHODS AND DATA

The product is not the chart. It is the chain that makes the chart inspectable.

ACAE methods are organized around source provenance, versioned protocols, validation, uncertainty, and publication boundaries.

DATA ARCHITECTURE

Four layers separate collection, validation, analysis, and public release.

Assign stable identifiers to field records, images, acoustic clips, instruments, sites, protocols, and external datasets.

Validate required fields, units, categorical values, timestamp formats, coordinate precision, and null reasons.

Attach flags for missing context, uncertain identification, observer disagreement, device issues, or protocol deviation.

Preserve raw data, derive transformed tables separately, and retain code versions and environment manifests.

Bundle methods, data dictionary, provenance, limitations, and a human-readable interpretation with each public output.

MULTIMODAL EVIDENCE

No single data stream should be treated as the whole ecological condition.

VisualPresence, count, state, morphology, habitat

AcousticActivity timing, call structure, soundscape

EnvironmentalTemperature, humidity, rainfall, light, water

SpatialEdge distance, corridor position, canopy, surface

Human activityTraffic, visitors, maintenance, noise, lighting

EXAMPLE DATA DICTIONARY

Field	Type	Purpose	Validation
`observation_id`	string	Stable record identifier	Unique and immutable
`site_class`	category	Habitat and transition context	Controlled vocabulary
`protocol_version`	string	Collection procedure reference	Must resolve to archived protocol
`observed_state`	category	Behavior or ecological condition	Allowed state set
`confidence`	integer	Observer confidence from 1 to 5	Required with uncertain records
`quality_flags`	array	Missingness and deviation status	Enumerated values
`source_media`	reference	Linked image, video, or audio	Hash and metadata check