雅思阅读 131: Reading the Mountain（读懂大山）

改编从 Quanta Magazine / Piton de la Fournaise research / WMO（2025-2026年）。雅思阅读 Section 3 难度，约 1050 词。 素材来源：https://www.quantamagazine.org/will-we-ever-be-able-to-forecast-volcanic-eruptions-like-weather-20260508/

Reading Passage

A. When Mount St Helens erupted in 1980, the lateral blast caught even the scientists who had been watching the mountain for weeks. Today, the science of volcanic forecasting is more sophisticated, but it remains a humbling discipline: unlike meteorologists, who have centuries of daily observations and a planet-spanning sensor network, volcanologists must wait years or decades between eruptions of any given volcano, and many of the world's most dangerous peaks are barely instrumented. The hope, driving a wave of new monitoring programmes and machine-learning projects, is to narrow the warning window from weeks to days — and eventually to hours — so that the next large city threatened by an ash column has time to evacuate. The ambition is not to predict eruptions with certainty, which most geophysicists consider impossible, but to issue probabilistic warnings accurate enough to act on. The tools now being deployed range from old-fashioned seismometers to fibre-optic cables buried in volcanic soil and artificial-intelligence models trained on decades of global eruption data. The difference between a successful evacuation and a catastrophe is often measured in hours, and every incremental reduction in that uncertainty window directly translates into lives saved.

B. A signal identified only in recent years illustrates how much remains to be discovered. Researchers monitoring Piton de la Fournaise on Réunion Island noticed subtle ground movements — very-low-frequency transients in both horizontal motion and tilt — that preceded eruptions by hours. These so-called "Jerk" signals appear to be generated as fractures open in the rock above an ascending magma body, and they can be detected with a single broadband seismometer, a simpler and cheaper instrument than the multi-sensor networks traditionally required. Fully automated Jerk detection is now running at Réunion and has correctly flagged pre-eruption unrest on multiple occasions. Separately, researchers studying the seismic b-value — the relative frequency of small versus large earthquakes on a volcano — have found that a falling b-value typically precedes rock failure, because increasing stress in the edifice produces more medium-sized shocks relative to tiny ones. Analyses of eight well-monitored volcanoes, from Iceland to Japan, confirmed that b-value shifts provide a useful probabilistic warning when fed into traffic-light alert systems.

C. Machine learning has accelerated this work in two distinct ways. First, it can process the enormous volume of seismic, geodetic and gas-emission data that dense modern networks now produce — far faster than human analysts — and identify patterns that no individual operator would notice. A 2025 study introduced a universal machine-learning approach that used just four seismic features across different volcanic settings, showing that the same algorithm can be trained on one volcano and applied to another. Second, Japanese researchers reported in late 2025 that an AI model trained on decades of data from Mount Aso in Kumamoto Prefecture could forecast eruptions several months in advance at roughly 70 per cent accuracy. Neither result is perfect — false alarms are as dangerous as missed predictions — but they suggest that data-driven methods are reaching practical utility. The World Meteorological Organization, in a 2025 workshop report, called for expanding monitoring capacity in high-risk regions, adopting a global Common Alerting Protocol, and combining satellite Earth-observation with ground networks. The same report noted that many of the world's most dangerous volcanoes — those near dense populations in the developing world — still lack even basic seismic monitoring, meaning that any AI trained on well-instrumented volcanoes may fail when applied to a poorly instrumented one.

D. The 2025 unrest on Santorini shows how complex the picture still is. The Greek archipelago, long quiet beneath its tourist hotels, began showing seismic swarms and ground deformation in early 2025. Scientists initially feared a magmatic intrusion like the one that produced the 2011–2012 unrest, which ended without eruption. Detailed analysis, published in Science later that year, revealed instead a rebounding magmatic dike — a body of magma that had intruded and then partially withdrawn, triggering seismicity as it moved. The episode illustrates a core difficulty: not all unrest ends in eruption, and issuing a false evacuation can be as socially damaging as missing a real event. Tourists fled Santorini in April 2025 despite expert consensus that the risk of imminent eruption was low, precisely because warning language is hard to calibrate for a public that understands only "yes" or "no". Probabilistic forecasts, the WMO emphasises, must be phrased in plain language and supported by community exercises if they are to be trusted when a real threat arrives.

E. The longer-term goal is weather-like forecasting — not a binary alert but a daily updated probability, computed from continuously streaming data. Projects now underway in the Canary Islands, Iceland and Japan will deploy hundreds of seismometers and fibre-optic cables that turn kilometres of ordinary telecommunication cable into distributed strain sensors, recording the tiniest earthquakes during periods of apparent calm. Machine-learning models will be trained on these recordings to distinguish "background" seismicity from the faint rumble of moving magma. Affordable remote microwave radiometers, adapted from radio-astronomy components, offer all-weather detection of volcanic plumes at a fraction of the cost of conventional thermal satellites. Volcanologists are under no illusion that they will soon predict eruptions as precisely as meteorologists predict rain: each volcano has its own plumbing, its own magma chemistry and its own behavioural history, and the feedback loop between a moving dike and the rock around it remains imperfectly understood. What they can do — and are beginning to do — is shorten the uncertainty interval from months to weeks, and weeks to days, for the population living on the slopes. The ultimate goal, though still distant, is a volcanic forecast that updates daily the way a hurricane track does: a probabilistic cone, with confidence intervals, that emergency planners can use to decide which roads to close and which neighbourhoods to evacuate.

Questions 1-4

Choose the correct heading for paragraphs B, C, D and E from the list of headings below.

List of Headings i. New physical signals — the Jerk and the b-value ii. How machine learning changes forecasting iii. The Santorini episode and the problem of false alarms iv. The long-term vision: weather-like probabilistic warnings v. The history of volcanic monitoring vi. How satellite cameras work vii. Why tourists dislike Santorini

1. Paragraph B: ____
2. Paragraph C: ____
3. Paragraph D: ____
4. Paragraph E: ____

Questions 5-8

Choose the correct letter, A, B, C or D.

5. What is the "Jerk" signal? A. A loud noise produced by erupting lava. B. A very-low-frequency ground motion from opening fractures before eruptions. C. A type of volcanic ash. D. A satellite measurement of heat.

6. What did the Japanese AI model achieve for Mount Aso? A. 100 per cent accuracy in predicting eruptions. B. Roughly 70 per cent accuracy forecasting eruptions months in advance. C. It predicted eruptions only hours in advance. D. It failed completely.

7. What did the 2025 Santorini unrest actually turn out to be? A. A full catastrophic eruption. B. A rebounding magmatic dike that partially withdrew. C. An earthquake unrelated to magma. D. A hoax by tourists.

8. What is the WMO calling for? A. Closing all volcano observatories. B. Expanding monitoring capacity and adopting a global Common Alerting Protocol. C. Stopping volcanic research. D. Relying only on satellite data.

Questions 9-13

Do the following statements agree with the claims of the writer?

Write:

• TRUE if the statement agrees with the information
• FALSE if the statement contradicts the information
• NOT GIVEN if there is no information on this

9. Volcanologists expect to predict eruptions with absolute certainty.
10. A falling b-value typically precedes volcanic rock failure.
11. The Santorini event of 2025 produced a major explosive eruption.
12. Fibre-optic cables can be used as distributed strain sensors.
13. Volcanic eruptions are more predictable than hurricanes.

Questions 14-15

Complete the summary below using NO MORE THAN TWO WORDS from the passage.

New monitoring tools include the Jerk signal, detectable with a single (14) __________ seismometer, and fibre-optic cables that record tiny earthquakes. Machine learning models now forecast eruptions with probabilistic accuracy, aiming to shorten the warning window from weeks to (15) __________.

答案与解析

← 上一篇 | 返回雅思焦点 | 下一篇 →