Why it matters
Without AI, inaccuracies in readings, delayed responses, or missed trends can cause welfare issues, unsafe water, and failed inspections, risking animal health and accreditation.
From design to safety
Design
AI-guided water system design
Design uses historical readings and target ranges to simulate turnover, dosing, and filtration, creating alert rules and data schemas for ongoing monitoring, with documented data lineage and accessibility.
- Ammonia sensor array
- pH and ORP probe
- Salinity refractometer/probe
- Central water quality monitor panel
Build
Commissioning and threshold setup
During build, sensors calibrate, thresholds set, integration with SCADA and ZIMS tested, and alarm workflows documented for operations handoff, with calibration certificates retained.
- Ammonia sensor array
- pH and ORP probe
- Central water quality monitor panel
- Lab sampling intake port
Operation
Real-time monitoring and auto-adjustments
Live AI reads ammonia, nitrite, nitrate, pH, ORP, and salinity, triggers dosing or water change orders, surfaces trends for daily rounds, and records decisions in the digital logbook.
- Ammonia sensor array
- Nitrite/Nitrate sensor
- pH and ORP probe
- Central water quality monitor panel
Maintenance
Calibration, drift checks, and records
Routine calibration, drift flags, and maintenance logs ensure data integrity, enabling audit trails and sustained accuracy in parameter readings, with quarterly verification of reference standards.
- pH and ORP probe
- Salinity refractometer/probe
- Central water quality monitor panel
- Lab sampling intake port
Safety
Alarms, licensing, and welfare safeguards
Automated alarms are routed to staff, with escalation to veterinarians and licensing bodies when thresholds breach safety limits, and emergency protocols trigger isolation and containment steps.
- Central water quality monitor panel
- Lab sampling intake port
- Ammonia sensor array
- pH and ORP probe
Part by part
- 1
Ammonia sensor array
- Watch for
- Drift or fouling reducing sensitivity; sudden spikes in readings at filter outflow; sensor saturation near biofilter.
- AI action
- Detect persistent drift; trigger recalibration and filter inspection; escalate if thresholds breached twice.
- 2
Nitrite/Nitrate sensor
- Watch for
- Gradual baseline shift or sensor saturation in denitrification loop; unexpected nitrate spike.
- AI action
- Flag anomaly; initiate confirmatory lab test and adjust dosing or water changes.
- 3
pH and ORP probe
- Watch for
- Electrode polarization or corrosion causing slow response; calibration drift noted.
- AI action
- Issue calibration reminder; alert maintenance about electrode replacement; log calibration event.
- 4
Salinity refractometer/probe
- Watch for
- Salinity drift affecting halogen dosing and species tolerance; temperature compensation error.
- AI action
- Flag drift; prompt recalibration and verify water source consistency; record temperature reading.
- 5
Lab sampling intake port
- Watch for
- Delayed sample flow or contamination affecting lab test accuracy; syringe contamination risk.
- AI action
- Trigger sample routing check and lab assay queue reprovisioning; flag repeat sample needed.
- 6
Central water quality monitor panel
- Watch for
- Communication fault or data backlog across sensors; clock drift in system.
- AI action
- Switch to failover channel; notify operators and begin data integrity check; archive old data.
Worked example
Preventing a pool toxin spike before a dolphin show
Illustrative scenario with sample data.
During mid-day operations, AI detects rising ammonia coupled with nitrate uptick and informs staff to verify biofilter performance and prepare a staged partial water change before dolphin activity resumes.
Signal in
sensor.ammonia_ppm=0.08, sensor.nitrite_ppm=0.002, sensor.nitrate_ppm=25, sensor.ph=7.8, sensor.orp_mV=310, sensor.salinity_psu=35, lab.last_test=ammonia=0.03; date=2026-11-08
System out
alert to waterquality@zoo.org; workOrder: ammonia_high_nitrate_trend; recommended actions: partial water change 12%, check biofilter flow, recheck ammonia and nitrate in 15 minutes; log: event_id=AQ-7062A
Human sign-off: Life-support engineer and aquatic veterinarian review within 30 minutes and approve action plan, with veterinary oversight for potential species sensitivity.
How to roll it out
Phase 1
Weeks 1 to 2
Baseline sensor calibration, thresholds finalized, operator training completed, and documentation stored in the ZIMS water module.
Phase 2
Weeks 3 to 6
Pilot across three habitat systems; refine alarms, dashboards, sampling cadence, and reporting; establish maintenance windows.
Phase 3
Weeks 7 onward
Full zoo-wide deployment with routine audits, calibration schedule, and annual review of lab test protocols.
Where people stay in charge
- Only qualified life-support engineers or veterinarians may approve changes to alarm thresholds or dosing actions, and all changes must be documented in the change log for regulatory review.
- All lab data must be corroborated by a certified test method before action is executed, to prevent erroneous decisions impacting welfare.
- Senior facilities management signs off on any modifications that affect licensing, accreditation, or public safety standards.
Questions owners and crews ask
What parameters should AI monitor in a zoo aquatic system?
AI should monitor ammonia, nitrite, nitrate, pH, ORP, salinity, and support lab test data to guide dosing, water changes, and system turnover decisions.
Who reviews AI-suggested actions for water quality?
The life-support engineer and the aquatic veterinarian review and approve actions, with escalation to licensing bodies if required to meet safety standards.
How does AI handle sensor drift or calibration gaps?
AI flags drift, schedules recalibration, routes maintenance tickets to technicians, and requires confirmatory lab tests before executing large changes.
What records are created by AI water monitoring?
Automated logs include sensor readings, calibration events, action decisions, work orders, and compliance reports for audits and licensing inspections.