Fire Crisis Dashboard
Real-time wildfire monitoring and threat assessment for the Maramures region, Romania. Integrates satellite imagery, weather forecasts, air quality sensors, and emergency protocols into 6 live bilingual dashboards.
The Problem
Romania's Maramures region experiences recurring wildfire threats, amplified by traditional agricultural burning practices and increasingly dry summers. Existing response infrastructure lacked:
- Integrated monitoring combining satellite hotspot detection with ground-level air quality
- Bilingual (Romanian/English) public-facing threat communication
- Automated threat level assessment using multiple independent data sources
- Historical context for distinguishing agricultural burns from genuine wildfires
- Decision support linking weather forecasts to fire spread probability
The challenge was not data scarcity but data fragmentation: 8 separate sources, each with different update cadences, formats, and reliability profiles. A human operator manually correlating these sources could not maintain real-time awareness.
FDRP Application
Eight Integrated Data Sources
Six Hats Convergence
De Bono's Six Thinking Hats method was applied to reach 10 binding decisions on threat assessment methodology, dashboard design, and communication protocols.
Pure data: satellite detection latency, sensor coverage gaps, historical burn patterns for Maramures agricultural cycle.
Emotional urgency: public fear management, avoiding both panic and complacency in threat communications.
Risk assessment: agricultural burning base rate means most hotspots are not emergencies. False positive cost vs. missed threat cost.
Value: early detection enables pre-positioning of resources. Bilingual dashboards serve both local and EU audiences.
Creative: URL-based language switching (?lang=en), shared navigation bar across all pages, dynamic threat level with auto-refresh.
Process: structured convergence from 6 independent analyses to 10 binding decisions. Disagreements forced to evidence-backed resolution.
Key Outcomes
- Source Health Monitoring
- Automated staleness detection for each data source. If FIRMS data is more than 6 hours old, the dashboard flags degraded confidence and identifies which threat assessments are affected.
- Agricultural Burn Discrimination
- Base rate analysis showed that agricultural burning in Maramures is extremely common in spring and autumn. The threat model incorporates seasonal burn calendars to reduce false positive rates without weakening genuine threat detection.
- Bilingual Public Communication
- 6 dashboards deployed at e.loftrek.ro with URL-based language switching. All pages share navigation, EFFIS WMS map layers, and dynamic threat level indicators.