When we talk about early warning systems, much of the discussion focuses on the dissemination of warning information: a hazard is detected, a warning is issued, people receive it and they evacuate. This is an important part of early warning, but it is not the whole story.
A warning does not arrive in an empty space. People receive information with previous experiences, expectations and assumptions about what a hazard means. They have to make sense of what they are hearing and decide whether the situation unfolding in front of them is familiar or whether something is different this time. This is particularly important for rare, rapidly evolving or unfamiliar events — what scientists describe as “uncertain events”.
The experience of Rajendra Dawadi during the recent flood in Rasuwa, Nepal, illustrates this human side of warning. Dawadi, the principal of Tribhuvan Trishuli Secondary School, managed to evacuate around 1,200 students to higher ground before the school itself was washed away by the flood. His account shows that, during a rare and uncertain event, receiving a warning is only the first step. The life-saving difference comes from how multiple uncertain signals are interpreted, discussed, judged and transformed into action.
It is therefore important to distinguish between two sides of warning: warning dissemination and warning interpretation for action. That distinction matters for how we understand risk communication and how we design future early warning systems.
The warning did not arrive as one warning
Dawadi’s account describes information arriving incrementally. At around 9:25 on Wednesday morning, he was teaching in a classroom. It was a normal morning and there was no rain. Hundreds of students were inside the school, studying in the morning shift, when one of his staff members came running into the school and shouted, “The flood is coming! The flood is coming! Everything has been swept away from upstream! Everything has been swept away!”
Dawadi came down to the ground floor to see what had happened outside. At first, he thought nothing would happen because everything appeared normal. Then another staff member, a social studies teacher, reported that there was flooding.
He came shouting, “Sir, I have left the classroom. There is a huge flood upstream and everything has been swept away.” He also stressed that his elder brother had told him not to stay there: “The Trishuli is not safe.”
The staff member then got on his motorcycle and left. After he had gone, Dawadi began thinking about the hundreds of students still inside the school.
Just as the teacher was leaving, another colleague called and said, “Sir, the flood has come from upstream.”
A colleague standing with Dawadi began asking, “What should we do now?” Then a parent entered through a nearby gate on a scooter. He told Dawadi, “Sir, the flood is very large. I have come to collect my child. I don’t know whether the flood will reach the school or not, but apparently everything upstream has already been swept away.”
By then, Dawadi had heard several different reports. His phone was ringing continuously and he was becoming increasingly worried. The colleague with him also said, “Sir, this might actually be serious.” Dawadi realised that he had to make a decision. As he explained:
“I thought: if someone had exaggerated the situation and the flood was not actually severe, then we would simply lose one day of teaching. But if the flood really had arrived, then we had to act.”
None of these pieces of information, considered individually, necessarily provided a complete picture of what was about to happen. The warning did not arrive as a single, formal message. Instead, information emerged through staff reports, telephone calls, a parent’s account and discussion among colleagues. Each signal was uncertain on its own, but together they changed Dawadi’s understanding of the situation and prompted him to weigh the consequences of acting against those of waiting.
His decision to evacuate was therefore not based on certainty. It was based on interpreting multiple signals and recognising that the potential cost of waiting was far greater than the cost of a false alarm. The warning did not simply turn into action. It emerged through interpretation, judgement and collective sensemaking.
Dawadi and his colleagues therefore did something fundamental to risk communication: they interpreted information in context. They considered whether the information was credible, discussed it with colleagues and weighed the consequences of acting on information that might turn out to be wrong. The challenge was not simply to communicate what was happening. It was to decide what the information meant and what should be done about it.
Risk communication is not information communication
This is where Dawadi’s experience challenges a conventional understanding of risk communication. A warning does not move directly from a scientific or monitoring system into human action. Between information and action sits interpretation.
People ask: What does this information mean? How credible is it? Is it relevant to us? How serious could the consequences be? What happens if we act and the warning is wrong? What happens if we do not act and the warning is right? What should we do now?
These questions become particularly important when hazards are rare, rapidly evolving or difficult to predict. In such situations, waiting for certainty can itself become a dangerous decision. Dawadi did not need certainty that the flood would reach the school. He and his colleagues needed to determine whether the available information was sufficiently credible and consequential to justify action. That is a fundamentally different way of understanding risk communication.
From information to meaning
Dawadi’s account shows that a warning is not simply transmitted from a technical system to a person. It develops through a chain of interpretation and action.
Signal: A signal is an individual piece of information suggesting that something may be happening, but it does not yet provide a complete or certain picture.
In Dawadi’s case, the signals included:
- the accountant running in and saying, “The flood is coming!”;
- a teacher reporting a huge flood upstream;
- the teacher’s brother warning that the Trishuli was not safe;
- telephone calls from colleagues saying the flood had come from upstream; and
- a parent arriving on a scooter and reporting that everything upstream had been swept away.
Interpretation: Interpretation is the process of asking: What does this signal mean?
Dawadi initially looked outside and thought nothing would happen because everything appeared normal. This shows that the same signal can be interpreted differently depending on what people can see and what they already expect.
As further information arrived, his interpretation began to change. The question shifted from, “Is there really a flood?” to, “Could this be serious enough to threaten the students?”
Sensemaking: Sensemaking goes beyond interpreting one piece of information. It is the process of putting multiple pieces together to develop an understanding of what is happening.
This is particularly visible in Dawadi’s account. He had heard several reports from different people, his phone was continuously ringing and he was discussing the situation with colleagues. The individual reports gradually formed a larger picture: destruction upstream, a parent’s confirmation and continuing calls pointed to something potentially much more serious than a normal flood. Sensemaking was therefore collective and cumulative rather than simply Dawadi receiving one authoritative message.
Decision on evacuation: Once Dawadi had developed an understanding of the situation, he had to decide what to do despite the remaining uncertainty. This is where his reasoning becomes particularly important.
If the information was wrong, they might lose one day of teaching. If it was right, people could lose their lives. He did not know exactly how high the flood would be or when it would reach the school. But he could compare the consequences of two choices: waiting for greater certainty or acting despite uncertainty. He decided that the potential consequences of waiting were too serious.
Trust is part of the warning system
Dawadi’s story also raises an important question about trust. The information became actionable partly because it was embedded in relationships. Staff members communicated with one another, colleagues consulted each other and a parent arrived with information from outside the school. Telephone calls provided further reports as the situation developed.
Dawadi was therefore not responding to an anonymous message from a distant warning system. He was making sense of information within a network of people he knew, trusted and could consult. Those relationships helped him assess whether different signals were credible and whether, taken together, they pointed to a situation serious enough to require action.
Trust and relationships, then, are not peripheral to risk communication infrastructure. They are part of it. A technically sophisticated warning system may detect a hazard and disseminate information quickly, but detection alone does not guarantee that people will recognise its significance, trust the information or know when to act. In uncertain situations, trust can allow fragmented information to become credible enough to support a decision.
The problem is not uncertainty or rare events
Dawadi’s decision also illustrates something central to my thinking about risk communication: disaster decisions do not necessarily fail because uncertainty exists. They can fail when the meaning of uncertainty for action is poorly communicated.
The information available to Dawadi was incomplete. He did not know precisely what would happen. But uncertainty did not prevent action. Instead, uncertainty became part of the decision. The question was essentially one of consequences:
What is the consequence of acting when the information is wrong, compared with the consequence of not acting when it is right?
This is particularly important for low-frequency and rapidly evolving hazards, where prediction may never provide the level of certainty that decision-makers would ideally like. Risk communication should therefore not be designed around the assumption that uncertainty must first be eliminated. It should help people interpret uncertainty in relation to decisions and consequences.
A warning system is more than sensors and messages
This has implications for how we invest in early warning systems. It is tempting to define an effective warning system through its technical components: sensors, monitoring networks, forecasting models, communication channels and alert technologies. These are essential. But Dawadi’s experience suggests that another form of infrastructure is equally important: the capacity to interpret information and turn it into action.
The lesson from Dawadi’s experience is therefore not simply that warnings need to be faster. It is that we need to think more carefully about what happens between receiving information and taking action. An early warning system should not be judged only by whether it successfully detects a hazard or issues an alert. We should also ask whether it creates the conditions in which uncertain information can be recognised and interpreted for action.
Alongside warning delivery, we need to focus on action capacity. This changes how we understand risk communication.
Risk communication is not simply about getting the right information to the right people at the right time. It is about enabling people and institutions to determine what information means, what it requires them to do and when that judgement needs to change.
Dawadi’s story is powerful precisely because he does not describe himself as a hero. He describes himself and his colleagues as doing their duty. That may be the most important lesson of all.
Effective risk communication does not necessarily depend on extraordinary individuals making extraordinary decisions. It depends on ordinary people having the relationships, information, judgement and institutional capacity to make consequential decisions before certainty arrives.
The warning is not the end of communication. It is the point at which interpretation has to become action.
