The tragic flooding that hit Spain’s Valencia region in late October, killing more than 200 people, was a reminder that many so-called natural disasters are, in fact, the result of extreme events enhanced by human factors.

Climate change likely drove the intense rainfall behind the floods, which was caused by a weather phenomenon called an isolated high-level depression, also known as a “cold drop” or “cut-off low.” This is where an area of low pressure gets detached from the jet stream and remains static for a long period, potentially drawing in moist air and subjecting the region beneath to prolonged adverse weather.

The Intergovernmental Panel on Climate Change warns that climate change has already intensified extreme precipitation in some regions, and that a warmer planet will experience more severe weather events. In addition, urban growth in areas with a history of flooding aggravates the situation—putting more people in harm’s way and lessening the ability of the land to handle heavy rainfall. Today, as a result of how we live, about 23 percent of the world’s population will be exposed to a flood every 100 years.

How then can we best cope with these scenarios? Much of the answer lies in accurately understanding severe weather phenomena, reducing the prediction time for them, effectively communicating the risk of each event to the population, and prioritizing nature-based solutions to lessen their impact.

Humanity has made the most progress on that first point: Many meteorological phenomena are well documented and understood—including cold drops. José María Bodoque, professor of geology and resources at Castilla La Mancha University in Spain, explains that a cold drop forms in Europe when a cold polar front of air migrates southward. If this happens just when the Mediterranean Sea is very warm, the contact between the air masses with different temperatures causes a cold drop, with its characteristic torrential rains. Knowing this helps meteorologists make predictions.

And the quality of weather forecasts for rainfall, even in times of climate change, has improved substantially in European countries in recent years, including in Spain. Models that simulate atmospheric processes based on physical laws are constantly improving, Juan Ballesteros, a researcher at the Spanish National Research Council, explained in a virtual meeting organized by the Science Media Center of Spain after the Valencia floods. Forecasts are quite good at reducing (but not eliminating) uncertainty, and there is ongoing research that is assimilating weather data with artificial intelligence for improved predictions. Nevertheless, it is still difficult to indicate the intensity and magnitude of extreme weather events.

In the case of the cold drop in Valencia, its identification was adequate, despite its being a particularly intense case, though the extent to which it can be attributed to climate change is complex and still unclear. Ballesteros points out that although we have historical records that we can compare the recent event to, from a statistical point of view, the size of our sample is not sufficient to confidently attribute the intensity of the Valencia weather to climate change.

But there are indications that climate change influences the occurrence and magnitude of these phenomena. One is the weakening of the jet stream, which is being induced by climate change, facilitating the descent of cold air from polar regions. Bodoque also indicates that a warmer sea brings more humidity, which “injects gasoline into the phenomena.” If the increase in sea surface temperatures seen in recent years is sustained, he points out, “we could have more intense, more frequent, and more extended cold drops—that is, outside their usual seasons, which are between the end of summer and the beginning of autumn.”

With other phenomena that cause torrential rains, such as hurricanes, the situation is similar: We know the ingredients, but we do not fully understand how they interact with our changing climate to have the impacts that we’re seeing. From a technical point of view, the modeling of torrential rainfall isn’t the hardest part of the equation. It is still difficult to anticipate the epicenter of a storm, for example. And Ballesteros points out that in terms of floods, we must consider that climate change affects soils, generating drier environments that are less able to absorb water, or which increase the velocity of water that flows off the land.

Regarding structural measures, Bodoque suggests mitigating risks by working to restore the natural functions of the soil—such as making sure water filters into it properly—as well as restoring the natural water cycle in ecosystems such as rivers and watersheds, rather than relying on gray infrastructure such as reservoirs and dams.

Ballesteros adds that progress must be made on policies to reduce the consumption of fossil fuels, the main cause of climate change. In terms of adaptation, he points out the importance of improving watersheds through restoring forests and mitigating nondestructive flows—for example, by improving connectivity with the floodplain, giving rivers space to flood without affecting people.

It’s Time to Get Real About Vulnerability

But while climate change was likely partly to blame for the loss of life following Spain’s recent cold drop, so too was a lack of preparation. An expert in flood risk assessment, Bodoque says the regional administration in Valencia did not warn the population sufficiently far in advance for them to take any action. In fact, its warning came 12 hours after the Spanish Meteorological Agency reported the high danger of the event.

In the case of the cold drop in Spain, the timely translation of the red alert into concrete actions on the ground failed. It is of little use to identify the danger in advance if communication is not effective or if the population does not know how to act. “One of the main lessons to be learned from these events is the importance of improving the early warning infrastructure to ensure that all people have access to accurate information on climate risks,” explains Natalia Alonso Cano, head of the office for Europe of the United Nations Office for Disaster Risk Reduction.

According to Bodoque, we need to improve flood-risk mapping too. There is a need to characterize vulnerability holistically, which implies considering the social, economic, physical, institutional, and cultural dimensions of what makes a community vulnerable to the weather. It is necessary to understand all components of what heightens people’s risk: not just their exposure to extreme weather, but how sensitive they are to it, and how resilient. Bodoque’s own research has found that most of the literature on vulnerability to natural disasters usually considers only two dimensions—the social and economic—with institutional and cultural qualities of regions being neglected.

As for the challenges of integrating flood-prone area mapping into regional decisionmaking, Bodoque points out that in the European Union there is a regulatory framework that includes a preliminary flood risk assessment, as well as hazard maps in which risk must be calculated according to the population and exposed assets. “There is quite a lot of room for improvement; the flood hazard maps present quite a lot of uncertainty.” In part, he explains, this is because flooding is a random process. It is very likely that where an intense flood has already occurred, another one will occur later, but it is not possible to know if it will happen in five or 300 years.

In addition to this, Bodoque explains, there is another issue. The parameters that feed the risk maps are not fixed values, but ranges—you can feed in upper, middle, or lower values, as desired. Yet the maps used in Spain and many other countries are deterministic; that is, they indicate only floodable and nonfloodable areas. In other words, they only see black and white. “I am providing a single cartographic output, when for each of the parameters and for range I have infinite outputs,” Bodoque says. Uncertainty is flattened into a deterministic map that can then generate a false sense of security.

It is necessary, Bodoque says, to change this method of generating maps that represent the probabilities of risk in flood-prone areas. This approach would better reflect the uncertainty inherent in flood events. However, this probabilistic model entails a high computational cost.

To better address the risks associated with torrential rains, Bodoque stresses the importance of making the population aware of the danger they face. In Spain, he and his colleagues have found that people exposed to natural weather processes do not perceive that they are at risk, partly because extreme weather events do not occur every year.

This low perception of risk has deadly consequences, as it encourages imprudent decisions in risky situations. Against this, Bodoque suggests developing communication plans for different audiences. In an article published in the Journal of Hydrology, of which he is a coauthor, Bodoque indicates that while “risk management based on a technocratic approach can give people a false sense of security,” the implementation of a good risk-communication strategy would facilitate a better response to emergency alerts.

This story originally appeared on WIRED en Español and has been translated from Spanish.