As climate disasters grow more intense and frequent, nations across the globe have ramped up investments in weather forecasting and modeling. But a stark question remains: what use is a forecast if it’s not trusted, understood, or acted upon?
Recent catastrophes have laid bare the limits of conventional early warning systems. In 2023, as wildfires devastated Maui in Hawai’i, sirens remained silent. Residents, unfamiliar with the alerts or how to respond, were left vulnerable. Two years earlier, during the 2021 eruption of Mount Semeru in Indonesia, no warnings were issued at all. The failure stemmed from applying a U.S.-based warning framework that hadn’t been adapted to local contexts despite heavy rains triggering the explosion.
These are not isolated failures. Even the most advanced forecasting tools struggle when real-world complexity enters the equation aging infrastructure, densely populated cities, low public trust in institutions, and misaligned communication methods. When crucial information arrives too late or worse, is mistrusted or irrelevant the cost is measured in lives.
But amid these gaps, communities often don’t wait. They act. Neighbors share information, organize evacuations, and check in using social media and messaging apps. These community-driven responses, rooted in local trust and proximity, are powerful—but they are often unsupported, invisible to official systems, and easily drowned out by misinformation and panic.
What if these same platforms people already use WhatsApp, Twitter, Facebook could be transformed into structured, reliable disaster communication networks?
A Grassroots Revolution: Technology Built by and for Communities
Enter Yayasan Peta Bencana (Disaster Map Foundation), a Jakarta-based nonprofit that has reimagined disaster communication from the ground up. Since 2017, their open-source software has enabled residents to report real-time disaster conditions through social media. Using AI and crowdsourcing, the system turns chaotic online chatter into structured, verified, actionable information.
Here’s how it works: When someone posts about a disaster using words like “flood” or “earthquake,” an AI chatbot replies, asking them to confirm their situation. A quick, user-friendly survey collects location, severity, optional photos, and a short description. Verified posts are then displayed on a live, interactive map PetaBencana.id in Indonesia and MapaKalamidad.ph in the Philippines.
No new app is needed. It works within platforms people already know, removing barriers and increasing participation. In essence the system doesn’t ask communities to adapt to technology it adapts to them.
The impact is extraordinary. In 2024 alone, the platform supported over 200 million residents and more than 900 humanitarian agencies. It’s fully integrated into local and national emergency systems and is recognized by many agencies including NASA and Indonesia’s Meteorological Department as the fastest source of verified disaster information available.
One emergency manager in Sumatra summed it up:
•Collective Intelligence, Shared Decisions
In the middle of a crisis, people make millions of micro-decisions:
•Is this road flooded? Where can I find shelter? Can I reach my family?
These decisions are deeply local and time-sensitive. Yet most disaster management systems rely on centralized authority and one-way alerts. PetaBencana turns that on its head democratizing access to real-time, ground-level data and allowing people to act together based on shared knowledge.
A flood report might inform a hydrologist’s model, a taxi driver’s route, a disaster official’s plan, and a family’s evacuation. The value lies not just in the data, but in the coordination it enables across sectors and communities.
And crucially, the system elevates local knowledge as a critical asset what researchers call high-context data. This includes hyperlocal insights satellites can’t capture: a blocked road, a damaged bridge, a stranded family. Such observations now validate satellite predictions and shape national planning tools.
Redefining Who Counts as an Expert
Beyond just collecting information, PetaBencana’s model challenges long-held assumptions:
•That only institutions hold expertise.
•That “top-down” is the only way to inform.
•That data must flow in one direction from agencies to people.
Instead, it builds dynamic, networked infrastructures where information flows both ways, and communities are not just passive recipients but active participants in shaping, governing, and responding to emergencies.
The system has expanded through South-South cooperation, with pilots underway in India, Pakistan, Panama, and Viet Nam. Each version is co-designed with local partners to reflect cultural and institutional realities. Experiences in Jakarta influence Manila, which in turn informs Hanoi and Karachi. This loop of mutual learning is a powerful counter to one-size-fits-all disaster tech.
Listening to the Margins
As climate disasters grow, there’s a larger question:
•Whose knowledge is being amplified in global systems? Whose is being excluded?
Without scrutiny, algorithms and datasets can replicate systemic blind spots. PetaBencana insists on inclusion—not only as a matter of ethics but of effectiveness. “If no district is to be left behind,” the team says, “then every district must be empowered to speak, to be heard, and to act.”
As the world races to improve resilience in a warming world, this model offers a powerful truth:
Forecasts are only useful if they reach people if they are trusted, relevant, and shared.
And sometimes, the most powerful infrastructure is not what’s built in labs or launched into orbit, but what’s already in the hands of communities waiting to be recognized.