Extreme Waters. A Collaboration between Mexico and Illinois on Hurricanes and Rainfall

The collaboration between the Tropical Climate Modelling group at UNAM’s National School of Earth Sciences (ENCiT) and Professor Zhuo Wang and her team at the Department of Climate, Meteorology, and Atmospheric Sciences of the University of Illinois, grew quite naturally from a shared research interest: how far can we take hurricane prediction with the models available today? Both teams had been working, from different angles, on tropical cyclones (which both in Mexico and the United States are known as hurricanes), as well as on sub-seasonal prediction, an emerging area that focuses on forecasts beyond the 10 days scope.

 Monserrat García Silva

Through our conversations, we realized that although we can currently predict hurricane tracks with reasonable accuracy, we still do not have a clear understanding of how far in advance tropical cyclone occurrence is actually predictable, or why forecasts become much less reliable beyond two weeks. As a result, we lack a complete picture of how much we can really trust actual forecasts when we try to anticipate their impact beyond about one week, particularly in terms of rainfall over Mexico and the U.S. 

It soon became clear that the key question was not only if we can predict a hurricane, but also what we can say, two or more weeks in advance, about the rain it will bring. In that sense, water became a central focus of our collaboration: the rainfall associated with tropical cyclones, its distribution in time and space, and the consequences for highly vulnerable regions to both drought and flooding in both countries.

Satellite image of hurricane Otis hitting Guerrero’s coast in October, 2023.

 Colorado State University (CSU/CIRA)

Hurricanes are often associated with adverse impacts, but many studies show that they can also play a positive role in the hydrological balance of some regions, contributing more than 30 percent of total rainfall in places such as Baja California Sur and Florida. In other regions, like Nuevo León, the rainfall from a single hurricane can put an end to prolonged droughts. At the same time, this rainfall is modulated by other climate variability modes, such as El Niño–Southern Oscillation (ENSO) [see box]. Understanding how these factors interact, and how much of that variability is actually predictable, is a major scientific challenge. Water management, from local decisions to national policies, depends to a large extent on this “available rainfall”: when it arrives, how much it rains, and what the chances are that it will be intense or extreme. 

THE PROJECT
Our project focuses precisely on this: understanding the causes behind the presence, absence, and excess of hurricane-related rainfall on weekly to sub-seasonal time scales. This framework comes with several challenges. The first is conceptual and theoretical: we need a better understanding of the processes that control rainfall from week to week and from year to year, and of how the climate “organizes” itself to enhance or suppress tropical cyclone activity in a given region. The second challenge is more practical and numerical: we need a clear and methodologically rigorous assessment of how good current models actually are when we try to use them beyond the usual seven to 10-day forecast range. In forecasting, this “goodness” is often summarized in what we call forecast skill, which essentially measures how much better a prediction is than simply relying on past averages.

Events such as hurricane Otis remind us of how catastrophic they can be. That is why characterizing these aspects of prediction is at the core of our collaboration. On the one hand, we study the climatic factors that modulate cyclone activity and rainfall distribution over Mexico and the U.S. On the other hand, we evaluate the forecast skill of sub-seasonal prediction models in representing that variability and turning it into information that can be useful for decision-making. 

CHARACTERIZING THESE ASPECTS OF PREDICTION IS AT THE CORE OF OUR COLLABORATION

COLLABORATING
There is also a collaborative, intercultural dimension within this scientific agenda, and in a very practical sense. Partnership between Professor Wang’s team in Illinois and our team at UNAM has been built gradually, based on shared interests and realistic constraints on both sides. We are making progress step by step, with good ideas and ongoing discussions, while managing a project that started in the middle of many other responsibilities and commitments. Recognizing both our strengths and limitations helps us split tasks in a sensible way, emphasizing what each group does best, whether it is running experiments, handling large datasets, or performing detailed analysis. 

We have also benefited from the involvement of doctor Chia-Ying Lee, from Columbia University, who contributes as a collaborator and helps bridge different perspectives and expertise. Overall, our approach is to be supportive and realistic: we try to organize ourselves, be patient with the pace of progress, and help each other deal with institutional and logistical limitations in both countries. 

BEHIND THIS SCIENTIFIC AGENDA THERE IS A STORY OF TRUST AND INTERCULTURAL UNDERSTANDING

From a practical standpoint, the potential benefits of our work are mostly related to prevention. As natural scientists, our main focus is to better understand and anticipate risks. If we can improve, even incrementally, our ability to anticipate rainfall patterns with forecasts two to three weeks in advance, we will be able to provide more robust inputs for planning the operation of reservoirs and irrigation systems, as well as tools for regions where dependence on tropical-cyclone rainfall makes water supply especially vulnerable to climate variability and change.

Looking ahead, we hope this collaboration will leave a legacy at several levels. Scientifically, we aim for our results to improve understanding of the links between tropical cyclones, climate variability, and water availability in Mesoamerica and Southern U.S. Our hope is to make a modest contribution as one of the early binational efforts addressing this topic from a sub-seasonal prediction perspective. Institutionally, we hope the project will open doors to further academic exchange, co-supervision of students, and joint projects between UNAM and the University of Illinois, so that the collaboration can continue beyond the individuals who initiated it.

---