Recife Signal Gap Analysis

This note explains what the paired Piedade and Boa Viagem replay could represent with current AI1SAD layers, and what remains missing for a future Greater Recife / Pernambuco regional pack.

The replay should not drive scoring-weight changes by itself. The useful next step is signal coverage: habitat, water movement, visibility, exposure, sightings, and monitoring-program ingestion.

Current Replay Finding

Piedade strict pre-incident replay stayed low because only bather/swimming activity and baseline reef-barrier/channel context were available. Boa Viagem strict pre-incident replay also stayed low-warning, but surveillance priority rose from 9.0 to 23.0 because the previous day's Piedade incident was timeline-valid recent-interaction context.

Signal-Gap Matrix

Signal	Current AI1SAD support	Pre-incident availability	Confidence impact	Surveillance usefulness	Recommended next action	Provider or dataset candidate	Priority
Regional incident density	Generic recent-interaction support	Partial for Boa Viagem only	Medium	High	Build Recife cohort and strict controls	Public historical incident datasets, CEMIT public records where usable	high
Live sightings / lifeguard observations	Generic sighting field exists	Unavailable in these strict runs	High	High	Add source-attributed observation ingestion plan	Local lifeguard/public-safety channels, official advisories	high
Reef-barrier habitat	Generic reef feature flag only	Baseline context only	Medium	High	Add static habitat mapping adapter later	Pernambuco coastal habitat/GIS sources to verify	high
Nearshore channels	Generic reef/channel feature text only	Baseline context only	Medium	High	Add channel-corridor layer with source dates	Coastal engineering, bathymetry, hydrographic datasets	high
High-tide overtopping context	Not supported for Recife	Missing	Medium	High	Add tide/current adapter coverage after source review	Tide station/model sources to verify	high
Tide/current observations	Hawaii static adapter only	Missing	Medium	High	Extend water-movement planning beyond Hawaii after Phase 23 pattern review	Official Brazil tide/current sources to verify	medium
Turbidity / water clarity	Planned Phase 24 gap	Missing	High	High	Add turbidity/visibility layer before calibration	Satellite ocean color, local water-quality datasets to verify	high
Rainfall/runoff lookback	Warning field exists	Missing for this replay	Medium	Medium	Add source-timestamped rainfall lookback where available	Weather stations, official meteorological products	medium
SST/anomaly	SST field exists	Missing for this replay	Medium	Medium	Add source-timestamped SST context	Satellite SST products, official ocean products	medium
Urban human exposure	Human exposure field exists	Missing for this replay	Medium	High	Build beach-time-of-day exposure baselines	Public beach-use proxies, lifeguard staffing, tourism calendars	high
Swimming/bathing context	Generic swimming context exists	Available as incident activity context	Low	Medium	Keep as activity context; do not overfit	Source-attributed incident record	low
Bull-shark suitability	Generic species context exists in other regions	Piedade has CEMIT preliminary tubarão-cabeça-chata classification only	Medium	Medium	Plan Recife species suitability profile after cohort review	CEMIT and scientific/public regional literature	medium
Tiger-shark suitability	Generic species context exists in other regions	Boa Viagem has CEMIT preliminary tubarão-tigre classification only	Medium	Medium	Plan Recife species suitability profile after cohort review	CEMIT and scientific/public regional literature	medium
Telemetry / monitoring programs	No Recife ingestion	Missing	High	High	Plan future ingestion with licensing/privacy review	Pernambuco monitoring programs and public releases	high

Regional-Pack Planning Notes

A future Recife / Pernambuco pack should cover:

• Boa Viagem
• Piedade
• Candeias
• Olinda / Del Chifre
• reef-barrier habitat
• nearshore channels
• high-tide overtopping context
• urban human exposure
• bull-shark suitability
• tiger-shark suitability
• future telemetry ingestion from Pernambuco monitoring programs

The first version should be static/offline and source-dated. It should label habitat and historical layers as baseline context, not current observations.

Baseline Recife / Pernambuco surveillance planning should include drone patrol windows for morning, midday, and late afternoon / dusk. Adaptive surge monitoring should be triggered after a recent severe incident, nearby sighting, future tagged-shark receiver detection, high human exposure, future tide/current overlap, or future turbidity/runoff overlap. Recommended operator patterns include shoreline_parallel_sweep, reef_gap_focus_scan, channel_mouth_pass, and post_sighting_focus_area.

Do Not Tune Yet

These two incidents are important for replay and gap analysis, but they are not enough to justify broad scoring changes. Recommended order:

1. Build a 10-20 case Greater Recife cohort with quiet-day controls.
2. Add source-dated reef-barrier/channel habitat context.
3. Add tide/current and high-tide overtopping support where official sources permit.
4. Add turbidity/water-clarity support.
5. Add live observation and monitoring-program ingestion when governance is clear.
6. Review calibration only after the cohort and signal coverage improve.

Current Interpretation

The paired replay shows that AI1SAD can represent incident-cluster separation when one event is already timeline-valid. It cannot yet represent Recife-specific water-state, visibility, exposure, or monitoring signals before the first event. That is a signal-coverage gap, not a reason to tune weights around a small cluster.

The species review should remain incident-specific: Piedade is the source-attributed bull-shark suitability case, Boa Viagem is the source-attributed tiger-shark suitability case, and the shared-corridor lift is species-agnostic recent-interaction context. The replay must not assume the same individual shark caused both incidents.