Huang Xiangning is among a small legion of diehard mainland seismologists who favour the study of earthquake forecasting, a field he says could die out in 10 years. The 74-year-old retired academic recounts his work in a career spanning more than 40 years and how he was marginalised for his beliefs. How did you get involved in earthquake forecasting research in the first place? I had worked for a unit of the now defunct Ministry of Petroleum since graduating from Beijing College of Geology, now known as China University of Geosciences, before I was recruited in 1967 by Li Siguang, a renowned Chinese geologist, for a special unit tasked with earthquake forecast research. Li, who was later to become minister of geology, was commissioned by late premier Zhou Enlai to set up the China Seismological and Geological Team, the predecessor of the China Earthquake Administration, along with another scientist, Weng Wenbo, after a devastating 7.2-magnitude quake in Xingtai, Hebei province, in March 1966 that killed more than 8,000 people. Could you briefly describe how you have approached earthquake forecasting research? The school of thought behind our research is the application of Li Siguang's geological mechanics, in the belief that any crustal movement inside the Earth will lead to changes in the amount of geo-stress or crustal stress, and that analysis of such changes collected at monitoring stations across the nation would help us determine where and when an earthquake would happen as well as its intensity. Li, a mentor of mine, likened earthquake forecasting studies to the way a doctor of traditional Chinese medicine diagnoses someone by taking his or her pulse. What are the highlights of your research? The first highlight was a forecast we made in May 1971, when our studies led us to believe that there would be an earthquake of magnitude 5 in the area straddling Changzhi, Pingyao and Linfen in Shanxi province between June 4 and 12. A cluster of quakes ranging between magnitude 4.8 and 5.2 struck Heshun county in Shanxi on June 5. That forecast was documented in the country's first report card for earthquakes of destructive magnitude (above magnitude 4.75). On July 14, 1976, our research also pointed to a possible magnitude-5 earthquake in Laoting county in Hebei province between July 20 and August 5. A 7.8-magnitude quake hit Tangshan in Hebei on July 28, but our predicted magnitude was lower because we believed we did not factor in the impact of a magnitude-6.4 jolt in Helinggeer county in Inner Mongolia the previous month. However, that forecast was still impressive. Your forecasting wasn't taken seriously, was it? Unfortunately not, as the forecast was snubbed by the 'revolutionaries' who controlled the China Seismological Bureau in the heat of the Cultural Revolution and overlooked by the nation's leader at the time, Hua Guofeng , who was torn between opposing voices among seismologists over the validity of earthquake forecasting studies. Hua could, to some extent, be held to account for ensuing earthquakes like the killer quake in Sichuan on May 12, 2008, because he failed to help put in place a concrete strategy for earthquake studies. How would you define whether an earthquake forecast is successful? A successful near-term forecast should give a location within a radius of 150 to 200 kilometres for the earthquake within 10 days from when it actually occurs and its magnitude should remain within a tolerance of half a magnitude. How accurate has your research been so far? I would say an accuracy of 70 per cent at best and no worse than 30 per cent. What's your take on the magnitude-8 quake in Sichuan in 2008? I began to notice an increasingly higher possibility of a major quake in some areas in Sichuan from 2005 and in a long-term report back in 2007 I raised the alarm over a magnitude-8.3 quake in an area along a large area covering part of the Yarlung Zangbo River to the Yajiang River in Sichuan, Jianchuan in Yunnan province and all the way to the northeastern part of India between May 2007 and April in 2008. As a long-term assessment, it was good because a magnitude-8 quake did strike Sichuan on May 12. However, it was impossible to follow it up with any near-term forecast research because of the abolition of more than 100 geo-stress monitoring stations around the country, including one in the town of Yingxiu in Wenchuan county, the epicentre of the May 12 quake, as a result of a policy shift in quake research since the early '80s. What was behind the policymakers' shift in attitude towards earthquake forecasting? There have been two opposing voices in seismological research circles over earthquake forecast research and the scientists who do not believe quakes can be foretold gradually gained the upper hand after the death of Li Siguang in 1971. Most of the 2,000 people Li helped assemble for earthquake forecast studies have left the field and those who continued to pursue such studies, like myself, have been sidelined and largely denied access to funding. Do you think that what justified the policy shift is the belief of the mainstream seismologists in the world, particularly the United States and Japan? I think that's why a lot of Chinese seismologists oppose earthquake forecast studies in China. I'm still a strong believer in what Li Siguang left us - a treasury of studies of earthquake forecasting - and if we had been allowed to continue doing what we were doing, a breakthrough would have been very likely in the next 10 years. How would you respond to concerns over the massive panic that an earthquake alert could cause, particularly if it is a false alarm? Those who say they are worried about panic or instability from a false alert are actually more worried about financial losses than the loss of lives. So I would propose a law to guarantee the public the right to information over a potential threat to their living environment. People can choose to leave or take precautions over their investments in volatile regions.