Technology stack powers PH disaster readiness
THE technology was tested on June 8, when the magnitude 7.8 earthquake struck off Maasim, Sarangani. Within minutes, the system was compiling earthquake data, tsunami warnings, landslide risks, damaged infrastructure, and population information into a common operating picture used by local government disaster responders.
Rather than waiting for fragmented reports from the field, emergency managers were able to prioritize search-and-rescue operations and allocate resources based on continuously updated analytics. The response demonstrated that the value of the platform lay not in any single technology but in its ability to integrate multiple data streams into actionable intelligence.
That is the future envisioned by the Pacific Disaster Center (PDC), whose disaster intelligence platform is quietly powering parts of the Philippines’ emergency management through a system called PhilAWARE.
PhilAWARE is the government’s disaster intelligence platform, originally developed by the PDC with funding from the US Agency for International Development (USAid) before being transferred to the Office of Civil Defense (OCD). The system remains in operation even after USAid was largely dismantled during the Trump administration’s government restructuring led by the now defunct Department of Government Efficiency (DOGE).
During an interview with The Manila Times, PDC Director of Information Technology Cassie Stelow described a multiple-technology ecosystem that combines artificial intelligence, geospatial analytics, authoritative scientific data, and automated communications to transform raw hazard bulletins into life-saving information.
Unlike conventional warning systems that simply forward weather advisories, DisasterAWARE functions as a continuously operating intelligence platform. Every day it gathers information from meteorological agencies, volcano observatories, seismic networks, humanitarian organizations, and government partners around the world. In the Philippines, that includes agencies such as Pagasa and the Philippine Institute of Volcanology and Seismology (Phivolcs).
The incoming information does not simply appear on a dashboard. Instead, it enters a processing pipeline where the system immediately overlays hazard data with extensive baseline information that PDC has already collected. These databases include hospitals, schools, roads, power infrastructure, demographics, poverty indicators, language distribution, and previous disaster impacts.
“If the bulletin drops in, that’s when I can extract the points on the map,” Stelow explained. “Where is the typhoon heading? What is the wind radius? Then we intersect it with the critical infrastructure that we already have mapped.”
The result is a far richer understanding than a standard weather forecast. Rather than simply identifying where a typhoon will make landfall, DisasterAWARE estimates how many hospitals fall within the danger zone, how many elderly residents live there, whether communities are already recovering from previous disasters, and which roads or evacuation routes may become inaccessible.
The platform currently monitors roughly 30 different hazard categories ranging from typhoons, floods, earthquakes, volcanoes, and tsunamis to geopolitical crises. Secondary hazards are also tracked. A cyclone, for example, automatically triggers additional flood and storm surge models, while major earthquakes generate aftershock monitoring and damage assessments.
Artificial intelligence has become increasingly important, but not in the way many people imagine.
Rather than relying on one massive language model, PDC uses multiple specialized AI models for different tasks. One model scans authoritative news sources worldwide for hazards in countries lacking sophisticated warning systems. Another simplifies highly technical weather bulletins into language ordinary citizens can understand. Other models analyze infrastructure and demographic data or summarize lengthy situation reports from organizations such as the World Health Organization and disaster management agencies.
Crucially, humans remain part of every important decision.
“We always have a human in the loop,” Stelow said, noting that analysts validate AI-generated information before it becomes operational. This safeguard is particularly important as misinformation becomes increasingly common during emergencies.
Localization is another area where AI is reshaping disaster management.
PDC is working with partners to move beyond simple translation by teaching AI to recognize regional terminology and dialects. Lessons from projects in Colombia are expected to influence future deployments in multilingual countries such as the Philippines, where local expressions often differ significantly even within the same language family. The objective is to ensure emergency instructions are delivered using terminology people naturally understand during moments of crisis.
The technology extends beyond forecasting disasters to supporting decisions after they occur.
As reports arrive from emergency responders, hospitals, humanitarian organizations, and local governments, AI helps organize the information into constantly updated maps. Emergency managers can quickly identify damaged hospitals, operational evacuation centers, disrupted transportation corridors, and areas requiring immediate assistance.
PDC ultimately envisions a future in which citizens interact directly with the system.
Instead of searching social media for unreliable information, residents could ask questions such as where the nearest operational health center is located, whether floodwaters are expected to rise in their neighborhood, or which evacuation center still has available capacity. Because DisasterAWARE already integrates hazard information with infrastructure and demographic databases, those personalized answers become technically possible.
As climate change increases both the frequency and severity of natural disasters across Southeast Asia, platforms like DisasterAWARE illustrate how disaster management is evolving from reactive response into predictive intelligence. The goal is no longer simply to warn people that danger is approaching, but to tell every person exactly what is happening, how it affects them, and what action they should take before it is too late. with RGBT
