How can statistical models predict volcano eruption?


  • Author: Carlos Alberto Gómez Grajales
  • Date: 02 August 2013
  • Copyright: Image appears courtesy of iStockPhoto

The National Institute of Geography and Statistics in Mexico recently presented a digital model developed to study and predict the threats associated with an eventual eruption of the most important active volcano in the country: The Popocatépetl. With a height of over 5,000 meters, it is one of the biggest volcanoes in North America. It is located in the centre of the country, about 72 km south of Mexico City. The Popocatépetl has been unusually active lately, with cataloged intermediate volcanic activity, releasing smoke and accumulating volcanic material, which could eventually produce explosions and lead to some potential lava eruption. Although no major cities would be affected by a hypothetical eruption (besides some annoying and lightly unhealthy ashes caused by the smoke expelled), minor towns are in the proximities of the mountain, so a good amount of effort has been devoted to prepare contingency plans, therefore soliciting some useful aid from statistics.

thumbnail image: How can statistical models predict volcano eruption?

Volcanic activity has indeed been a fascinating topic for statisticians, at least for a while. Although predicting specific events within volcanic activities is still far from us, security measures and policies have been greatly enhanced when statistics are used to understand the systems that take place after a volcanic explosion. As it occurs with most natural phenomenons, a deterministic model isn't able to capture the complexities that rest behind volcanic systems. As a consequence, researchers have used many forecasting statistical tools to understand the responses that a specific volcano would produce if eruption occurs. Some of the first analysis used simple Poisson models, though more sophisticated forecasting techniques have been developed recently, which have allowed to model the expected behavior of possible mudslides, as well as the flow that the lava expelled may take. This information is particularly important to design safety policies and contingency plans.

Even when some recent developments have increased the reliability of these type of models, many challenges still exist when studying volcanic activities. When researchers try to model the activities of a single volcano and model it as Time Series data, they often find that volcanic behavior can be practically nonexistent for long periods of time. Then, suddenly and without notice, a significant spike in activity occurs for a relatively short period of time. This is very common, since most volcanoes are inactive for long periods until they "wake up". The fact that only a minor amount of data is available for the moments with significant volcanic activity has proved to be a meritorious obstacle.

However, it is very appropriate to remember the words that the late statistician Professor George E.P. Box immortalised :"All models are wrong but some models are useful". Based on some of the latest advances in the field, the National Statistics Center of Mexico has just presented a digital version of a volcanic model for the Popocatépetl, in an effort to predict the effect that a volcanic reaction could produce. This digitized model, completed with some really cool three-dimensional maps that accurately represent the area of impact for the volcano, has been made available to the official center for disaster prevention in Mexico. This model takes into account the geographic conditions, the geology and the environment surrounding the volcano, as well as the demographic characteristics of the towns that may be affected by any unfortunate situation. With this tool, authorities may be able to forecast the impact that different volcanic reactions would have, not only in the habitants within that region but also the impact that a possible volcanic reaction would yield on the infrastructure or even within the agricultural activities in the region. This information will be key to draft emergency plans and to prepare for some eventual disastrous scenarios. Even though no one wishes to think about tragedies, reliable information and forecasts are irreplaceable tools to correctly manage them.


Related Topics

Related Publications

Related Content

Site Footer


This website is provided by John Wiley & Sons Limited, The Atrium, Southern Gate, Chichester, West Sussex PO19 8SQ (Company No: 00641132, VAT No: 376766987)

Published features on are checked for statistical accuracy by a panel from the European Network for Business and Industrial Statistics (ENBIS)   to whom Wiley and express their gratitude. This panel are: Ron Kenett, David Steinberg, Shirley Coleman, Irena Ograjenšek, Fabrizio Ruggeri, Rainer Göb, Philippe Castagliola, Xavier Tort-Martorell, Bart De Ketelaere, Antonio Pievatolo, Martina Vandebroek, Lance Mitchell, Gilbert Saporta, Helmut Waldl and Stelios Psarakis.