Will We Ever Be Able To Forecast Volcanic Eruptions Like Weather?

Episode Summary

Executive Summary: The episode explores why volcanic eruption forecasting remains difficult and whether it can one day become as routine as weather prediction. Volcanologist-turned-writer Robin George Andrews explains that volcanoes are deeply idiosyncratic, mostly hidden underground, and only intermittently monitored, but argues that better sensors, AI pattern recognition, and more universal physics could eventually enable probabilistic forecasts and safer evacuations.

Main Topics: Why volcano forecasting is so hard (Priority: 5/5): Volcanoes are difficult to predict because magma is hidden underground, each volcano behaves differently, and eruptions can switch styles without clear warning signs. What drives eruptions (Priority: 5/5): Eruptions are fundamentally about pressure: magma rises when heat, gas, earthquakes, or fractured rock allow built-up pressure to escape, with explosivity depending on magma viscosity and gas trapping. Current forecasting methods and signs (Priority: 4/5): Scientists use seismicity, ground deformation, and gas emissions to estimate eruption likelihood, especially at frequently erupting volcanoes where patterns can be learned over time. Limitations of monitoring and data collection (Priority: 4/5): Many volcanoes are poorly instrumented, and the crucial physics remains buried underground; remote sensing helps, but it cannot fully replace direct subsurface measurements. AI and dense instrumentation as a path forward (Priority: 5/5): Researchers are increasingly saturating volcanoes with sensors and using machine learning to identify tiny signals and causal relationships that humans might miss. Toward a universal physics of volcanoes (Priority: 4/5): Andrews argues a shared underlying physics likely exists, analogous to weather forecasting, and that archetypes of volcano behavior may eventually support generalized prediction models. High-stakes volcanoes and public safety (Priority: 5/5): The discussion highlights especially dangerous, densely populated volcanoes such as those in Indonesia, the Philippines, and Campi Flegrei near Naples, where improved forecasting could save many lives.

Key Arguments: Volcanic forecasting is closer to probabilistic weather prediction than to exact prediction; useful forecasts will likely be odds-based rather than certain. The major barrier is not just complexity but inaccessibility: the relevant processes occur underground, beyond direct observation. Volcanoes share physical principles, especially pressure-driven behavior, even if individual volcanoes vary widely. Repeated observation of active volcanoes can reveal recurring precursors such as earthquakes, deformation, and gas changes. AI is useful not as a magical solution but as a pattern-recognition tool that can process vast, subtle datasets and support hypothesis testing. A universal forecasting framework may emerge by combining archetypal volcano classes, physics equations, and continuous sensor data. The biggest obstacle may be human inertia: societies often act only after disasters, even when risks are known.

Data Points: People living near volcanoes: about 800 million - Hannah Waters cites the number of people living within 60 miles of a volcano worldwide. Distance threshold: 60 miles - Defines the proximity used to estimate the population at risk from volcanoes. Episode release schedule: every other Thursday starting June 11th - Promotional mention for season five of The Joy of Why at the start of the transcript. Depth of magma chambers: as shallow as 1–2 miles or as deep as 10–20 miles - Andrews explains how far beneath the surface magma reservoirs may lie. Frequency of eruption at Mount Yasur: several times an hour - Used to describe the near-continuous eruptive activity of the volcano in Vanuatu. Duration of Mount Yasur activity: hundreds of years - Mount Yasur has been erupting nearly continuously for a very long time. Distance of Campi Flegrei from Naples population: 2 or 3 million people - Andrews notes the population living essentially on the edge of the supervolcano system.

Pivotal Quotes: "forecast volcanic eruptions in the way we forecast the weather" — Robin George Andrews: Summarizing the story’s central ambition: probabilistic forecasting rather than exact eruption prediction. "they're a bit like cats. They're all idiosyncratic in their own ways" — Robin George Andrews: Explaining why lessons from one volcano often cannot be directly applied to another. "The dream basically is to study so many volcanoes over such a large scale that you come up with like archetypes." — Robin George Andrews: Describing the long-term vision for building a generalized prediction framework.

Implications: Volcano forecasting may become a probabilistic, AI-assisted science that improves evacuation timing and hazard planning. Progress depends on sustained monitoring, shared physics models, and political will to invest before catastrophes occur.