Flood & Fire Early Warning System
FINCAPES FFEWS
A Full-Stack IoT Sensing & Decision-Support Platform for Forestry Conservation and Peatland Hydrology Monitoring in Indonesia.
Peatland Conservation & Karhutla Vulnerabilities
Indonesian tropical peatlands face dual hydrological extremes: intense dry-season fire risks (Karhutla) and wet-season localized flooding.
Peat Fire (Karhutla) Vulnerability
When Groundwater Level (TMAT) drops and surface/deep peat moisture dries out, smoldering underground peat fires ignite easily and produce toxic haze.
- Requires dual-depth soil moisture monitoring (Surface vs Deep).
- Needs real-time smoke & ambient temperature anomaly detection.
Localized Peatland Flooding
Sudden monsoonal rain events saturate degraded peat catchments, resulting in flash floods and infrastructure damage in low-lying river basins.
- High-frequency water level & hydrostatic pressure tracking.
- Threshold exposure calculation to monitor inundation duration.
Operational Challenges in Forestry Field Sites
Remote observation stations face frequent power outages, intermittent cellular connectivity, physical sensor installation variations, and multi-agency coordination barriers across central, provincial, and local forest stations.
End-to-End System Architecture
Decoupled, high-concurrency microservices combining Python FastAPI for telemetry ingestion and Next.js 14 for GIS situational awareness.
Raspberry Pi / Gateways (Soil, Water, Weather, Power)
Port 8000 (Telemetry Ingest, Shadow Sync, ACK)
Port 5432 (TimeSeries Data & Device Metadata)
Port 3000 (Leaflet GIS, RBAC, i18n, Analytics)
FastAPI Ingestion
Handles batch telemetry JSON payloads with client timestamps, supporting offline data buffering & battery health metrics.
Prisma & PostgreSQL
Optimized relational schemas segregating high-frequency time-series telemetry from static device installation metadata.
Next.js Dashboard
React 18 + Tailwind CSS + Leaflet map. Interactive GIS color coding, real-time alert triggers, and CSV report export.
Monitored Telemetry & Physical Depth Metadata
Differentiating transient environmental readings from permanent sensor physical installation depths for accurate volumetric calibration.
1. Dynamic Telemetry Metrics (Payload)
| Parameter | Unit | Scientific Relevance |
|---|---|---|
| Surface Moisture | % | Topsoil desiccation & surface ignition risk |
| Deep Moisture | % | Deep peat water reserve retention |
| Water Level / Pressure | m / bar | Groundwater table (TMAT) & flood head |
| Smoke / Temp / Flame | % / °C / bool | Direct fire & smoldering detection |
| Wind Speed / Direction | m/s / Deg | Fire spread direction situational awareness |
2. Static Sensor Installation Depths
Each physical device maintains 3 static installation depth metadata records (in cm). Maintained exclusively by assigned Station Operators:
Scientific Value: Prevents duplicating depth data in every telemetry row while providing IPB researchers exact depth baselines for volumetric water content (VWC) calibration.
Peatland Dual Early Warning Logic
Automated multi-factor threshold evaluations for simultaneous fire (Karhutla) and flood risk flagging.
Fire Warning Criteria (Karhutla)
Triggered automatically when any of the following compound environmental conditions are met:
- High Smoke Level (exceeding custom station threshold)
- High Ambient Temperature or Detected Flame
- Humidity dropped below safety threshold (%)
- Surface or Deep Peat Moisture dropped below dryness threshold (%)
Flood Warning Criteria
Triggered automatically when hydrological indicators breach safety limits:
- High River / Water Table Level (metres)
- High Cumulative Rainfall (mm)
- High Surface or Deep Peat Saturation
- Hydrostatic Pressure breaching configured lower limit
City-Wide Default Threshold Templates
Admins set regional baseline thresholds (DeviceDefaultSetting), broadcasting rules across all city stations automatically via Device Shadow.
Closed-Loop Incident Resolution Workflow
Operators inspect, acknowledge, and mark alerts as Resolved in FlaggedAlerts, maintaining audit logs of responder actions.
Field Reliability & Device Control Features
Designed specifically for remote, harsh forest environments with intermittent power and connectivity.
Battery Level Monitoring
Real-time reporting of telemetry battery percentage (`batteryLevel`) in both batch uploads and lightweight heartbeats (`/sensors/heartbeat`) to prevent sudden field outages.
Offline Buffer & Batch Ingest
Edge nodes buffer telemetry locally during cell towers dropouts. Uploads contain client-side timestamps, allowing FastAPI to archive continuous historical trends seamlessly.
Active Alert Status Lock
Prevents operators from changing a device's operational status (e.g. to Maintenance/Offline) while it has an unresolved active alert—enforced by frontend validation and backend API PATCH rules.
Device Shadow & Remote Command Dispatch
Operators can dispatch remote actions (such as reboot) directly from the Web Dashboard. Edge devices poll commands via heartbeats and acknowledge execution via the Command ACK endpoint (/sensors/commands/ack).
Multi-Tier RBAC & Bilingual i18n Support
Supporting multi-agency collaboration between central administrators, regional forest authorities, and local station officers.
1. Role-Based Access Control (RBAC)
Global system access, user management, and system-wide threshold configurations.
Regional coordination scoped strictly to assigned cities and stations (e.g. Pontianak office).
Station-level monitoring and exclusive management of static sensor installation depth records.
2. Bilingual Support (i18n)
Powered by next-intl for seamless real-time language switching without page reloads:
Analytics & Scientific CSV Data Export
Empowering IPB faculty and graduate students with structured hydrological and meteorological datasets for secondary modeling.
TimeSeries Trends
Interactive multi-parameter trend charting across selectable date ranges, highlighting threshold-crossing events visually.
Threshold Exposure
Calculates cumulative days of low peat moisture or high water table head to assess ecosystem vulnerability duration.
Trend Prediction
Calculates rate-of-change projections to forecast imminent water level rises or peat desiccation before thresholds breach.
Scientific CSV Export
Exports clean telemetry logs bundled with the current station installation depth metadata for instant import into R, Python, or ArcGIS.
Sample Export Data Schema
5-Minute Live Demo Walkthrough for IPB
Step-by-step presentation script designed to showcase user experience and scientific relevance smoothly.
GIS Situation Map & Bahasa Indonesia Toggle
Switch language to Bahasa Indonesia, select Pontianak city, and inspect station risk markers.
Peatland Telemetry & Dual Fire/Flood Alert Flagging
Open station readings, highlight surface vs deep peat moisture, and show active fire alert triggers in FlaggedAlerts Page.
Device Status Lock & Remote Reboot Command
Demonstrate status lock blocking accidental device disablement during active alerts, then dispatch a reboot command.
Station Operator Depth Metadata Maintenance & CSV Export
Show sensor installation depth editing under Station Details, followed by one-click CSV research report export.
Conclusion & Future Academic Collaboration
Building an integrated, science-backed digital twin for Indonesian peatland and forest risk management.
Summary of System Impact
- Rapid Response: Reduces disaster warning response times across pilot regions.
- Scientific Integrity: Integrates physical depth metadata with real-time moisture & groundwater readings.
- Field Resilience: Battery monitoring, offline buffering, and safety status lock for rugged field deployment.
Future Research Collaboration with IPB
- FDRS / FWI Integration: Embedding Indonesian Fire Danger Rating System algorithms into the backend.
- Peat Carbon Emissions: Correlating TMAT drawdown rates with peat oxidation and carbon loss models.
- Multi-Source Fusion: Cross-validating ground IoT telemetry with satellite hotspot data (Sentinel/MODIS).
Terima Kasih! • Thank You!
Questions & Academic Discussion • Faculty of Forestry & Environment (IPB)