(June 10, 2008) The chief engineer of the Regional Geology Investigation Team of the Sichuan Geology and Mineral Bureau talks about the possibility that the Zipingpu dam induced China’s deadly May 12 earthquake.
Interview by Wang Yongchen
On May 27, 2008, one of China’s most celebrated radio journalists and founder of the environmental organization, Green Earth Volunteers, Wang Yongchen interviewed Fan Xiao, chief engineer of the Regional Geology Investigation Team of the Sichuan Geology and Mineral Bureau about the possibility that the Zipingpu dam induced China’s deadly May 12 earthquake.
Wang Yongchen: Now, two weeks after the powerful Wenchuan earthquake, what are your thoughts?
Fan Xiao: This earthquake was a Richter scale magnitude 8 and, strictly speaking, it is very unusual for this area. The historic records show that the highest recorded earthquake in this area was magnitude 6.5, and no seismic activities of more than magnitude 7 occurred along the Longmenshan seismic belt before. (see map here)
These records, which begin in 1900, record seismic activities in general and big earthquakes in particular with earthquake instruments and so represent accurate recordings. But for seismic activities before 1900, we know about them only from history books. For that reason, it is difficult to know how big they were exactly. And it‚Äôs impossible for us to know about the seismic activity that occurred in remote regions, where almost no people lived.
Wang Yongchen: How about the Diexi earthquake1?
Fan Xiao: In the case of Diexi, there are two seismic belts rather than one (Editor’s note: i.e. the Longmenshan belt), with different structures and geological conditions though. In the case of the Longmenshan seismic belt, the earthquake authority conducted studies and predicted that an earthquake could occur, but not with a magnitude above 7.
Wang Yongchen: But Li Youcai’s2 forecast was magnitude 7.5.
Fan Xiao: Li Youcai wasn’t referring specifically to the Longmenshan seismic belt but to the China south-north earthquake belt, which covers a vast area and consists of a dozen seismic belts, as a matter of fact. As far as the China south-north earthquake belt is concerned, earthquakes with magnitude of not only 7, but of magnitude 8 have occurred in history. Thus, forecasting earthquakes in this area isn’t hard because there have been so many. What worries me most is that more and more hydro dams have been or will be built on the earthquake belts with high risk and high magnitude of earthquakes, such as the middle and lower Jinsha River (Note: the main channel of the upper Yangtze) and the Dadu and Yalong rivers (Note: both tributaries of the Yangtze). The cascade of dams such as Xiluodu (Note: under construction) and Wudongde (Note: approved) on the Jinsha River are on the Dongchuansonglin earthquake belt on which earthquakes with magnitude 8 have occurred before. For earthquakes, generally speaking, there is a cycle of occurrence every 100 or 200 years, for instance. But the phenomena of RIS (reservoir-induced seismicity) is likely to change the timing, the location of the epicentre, and the seismic intensity of an earthquake. The reservoir exerts an external influence, and is likely to interrupt the cycle.
Wang Yongchen: Do you think this unusual quake was triggered by the reservoir?
Fan Xiao: Very likely, given its geological conditions, and the epicentre being so close to the reservoir.
Wang Yongchen: Where?
Fan Xiao: Zipingpu. The reservoir of Zipingpu is only five kilometres away from the epicentre. Typically, RIS occurs within a dozen kilometres from a reservoir, and the distance between the epicentre and Zipingpu is only five kilometres.
Wang Yongchen: So, you are saying that the reservoir is likely to have induced the earthquake, but how? What are the conditions that trigger an earthquake?
Fan Xiao: There are many conditions, but from past experience and cases, both inside and outside China, at least three conditions are really important. First, the dam should be high enough, at least higher than one hundred metres. Second, the dam should be big enough, generally with a storage capacity of 1 billion cubic metres or more, because a small reservoir wouldn’t create significant threats. Third, and very important, the reservoir should be located in a fault zone. The case of the Zipingpu meets all three conditions: it has a height of 156 metres, a storage capacity of 1.126 billion cubic metres, and the reservoir is built in a fault zone. It wouldn’t be crucial if the reservoir was located in an inactive fault zone, but it does matter if a reservoir is located on an active fault belt. It’s unfortunate that Zipingpu is exactly this case: it is located on an active fault belt, with the tail of the reservoir really close to Yingxiu, where the quake occurred.
The entire Longmenshan fault belt consists of three parallel fault lines (see map): with the Wenchuan-Maoxian Fault on the west, the Dujiangyan-Jiangyou Fault on the east, and the Yingxiu-Beichuan Fault in the middle. It was on the Yingxiu-Beichuan Fault, which is called the main central fault, where the quake with magnitude 8 occurred. So this is why the Zipingpu reservoir is not far from the epicentre.
It’s interesting to note that the quake is called the “Wenchuan earthquake,” so everybody thought that the epicentre was in the county seat of the Wenchuan area. But actually the epicentre is much nearer to Dujiangyan than to the county seat of Wenchuan, despite the fact that the epicentre area is within the county of Wenchuan. This is why the losses in the county seat of Wenchuan were not as severe as in the county seat of Beichuan.
The main central fault, or the Yingxiu-Beichuan Fault, is several hundred kilometres long, and the Zipingpu reservoir has been filled for more than three years, which is in the period of time in which the reservoir is most likely to trigger earthquakes.
The Zipingpu reservoir started filling in December of 2004 from a water level of 700 metres above sea level, and was filled to its NPL (Normal Pool Level) of 877 metres above sea level by October 2006. According to the operating mode, the reservoir is raised to its highest water level in late September and early October, and lowered to its lowest water level in late April and early May, before the flood season arrives.
Therefore, Zipingpu has experienced the procedure of raising and lowering the reservoir level at least twice since it was put into operation – with the highest water level in the fall of 2006 and the lowest level in late April 2007, and again the highest water level in the fall of 2007 and the lowest level this late April and early May. I believe this significant change, and the repeated raising and lowering of the reservoir water level, might be a factor triggering the earthquake. Therefore, it’s understandable that the powerful quake occurred on May 12, when the water level was going down.
Wang Yongchen: Cheng Genwei, senior researcher at the Chengdu Institute of Mountain Hazards and Environment of the Chinese Academy of Sciences, said the quake should be named the “Sichuan earthquake,” but someone else argued it should be called the “May 12 earthquake.”
Fan Xiao: In my opinion, it should be called “Sichuan Longmenshan earthquake,” because whether it was the quake or aftershocks, they all occurred along the Longmenshan fault belt. Moreover, the main destruction is also concentrated in the Longmenshan fault belt.
Wang Yongchen: So Gansu (Note: north of Sichuan; see map) is also affected by the Longmenshan fault belt?
Fan Xiao: Right.
Wang Yongchen: So the affected regions are all near the mountains of the Longmenshan area?
Fan Xiao: True. Even Dujiangyan is on the edge of the Dujiangyan-Jiangyou fault.
Wang Yongchen: I have heard that residents in Dujiangyan don’t dare live in their houses for now, due to their great fears. Is that true?
Fan Xiao: True. If you take a closer look, it becomes clear that the Zipingpu reservoir is actually sandwiched between two fault lines. The dam site is less than two kilometres away from the Dujiangyan-Jiangyou Fault in east, while the nearest point of the reservoir is only 500 metres away from the Yingxiu-Beichuan Fault in the west. Moreover, there are many smaller fault lines parallel to these two main fault lines just below the reservoir itself. Those smaller fault lines are already seriously crushed, so after building a big dam on it, the reservoir exerts hydraulic pressure, crushing the fault below it and causing a disturbance in the fault belt, as a result. Furthermore, with rock stratums in the fault belt below the reservoir crushed and broken, more and more cracks would appear and develop, helping the water from the reservoir seep in deeply, which would significantly influence activities in the fault belt.
In general, there are two types of RIS. The first is called “rapid reaction type” which refers to seismic activities that occur shortly after filling the reservoir; the second is called “delayed reaction type” in which earthquakes take place several years after the reservoir is filled, such as two years, three years, five years, and even ten years, and so on. Based on global studies, the second one is much more severe than the first one because greater hydraulic pressure is exerted by the reservoir and has a greater influence on the fault lines.
As mentioned above, the Zipingpu reservoir has been filled for longer than three years since it began filling in December 2004, so it is definitely the case of the second type. According to state regulations, a seismic risk assessment should be done before building a hydro dam. Actually, the China Institute of Water Resources and Hydropower Research did such a risk assessment, and predicted that the highest magnitude earthquake that would occur there was a 5.5 magnitude. The state regulations also require that a seismic monitoring network be established before filling a big reservoir. In the Zipingpu’s case, a seismic monitoring network was set up. The problem was not that the government or dam authority failed to do this, nor that the experts involved didn’t take the seismic issue into account, but that nobody thought such a powerful earthquake of 8 on the Richter scale would occur. The experts who participated in the feasibility study for Zipingpu predicted or estimated that an earthquake only as big as magnitude 6 could occur.
Recently, I looked at a scholarly article published by the Bureau of Seismology but it contained data from the seismic monitoring network only for 2004-2005. Complete data should cover a period from the very beginning of filling the reservoir (Note: December 2004) to May 12, 2008, when the earthquake occurred. But even that one year of data showed that seismic activity was, not only more frequent, but highly concentrated in the two faults: the Dujiangyan-Jiangyou and Yingxiu-Beichuan Faults in general, and in the Yingxiu-Beichuan Fault in particular.
As the above scholarly article revealed, as many as 730 seismic events were recorded from December 2004 to December 2005, with the highest at magnitude 3 or a little bit higher. The recorded earthquake waves also indicated that the seismic activity was concentrated in a line, which was consistent with the earthquake surface of the quake that occurred on May 12, 2008. In other words, the direction and locations of this earthquake and aftershocks tallied with the records obtained from the seismic monitoring stations set up around Zipingpu, though the recorded magnitude of seismic events in the December 2004-2005 period were not that high.
But the data from the end of 2005 to May 12, 2008 has not been published yet. It would be much easier to do a good analysis if the data was available.
Many experts are doing studies on RIS in China. As far as I know, at least two institutes in China have been involved in this discipline: the Chengdu-based Institute of RIS at the Sichuan Bureau of Seismology, and the Beijing-based Institute of RIS at the China Institute of Water Resources and Hydropower Research.
Wang Yongchen: So they definitely have the data.
Fan Xiao: Definitely. This data would really help to determine whether or not Zipingpu induced the quake, though the data would not be sufficient to explain the earthquake on the Richter scale of magnitude 8.
Wang Yongchen: Forecasting earthquakes is still a difficult thing to do. And what we really want to know is not whether earthquakes can be predicted but whether the Zipingpu dam reservoir triggered the quake. One of the experts from Beijing commented that there have been cases of reservoir-induced earthquakes, but he has also argued that no reservoir has ever triggered such a large earthquake. Thus a better explanation, he said, is that the May 12 Sichuan earthquake was a result of diastrophism3 rather than being induced by a reservoir. Currently, most people have accepted that.
Fan Xiao: Right. Even if the Zipingpu reservoir hadn’t been built, the quake would have occurred in this area, but not necessarily with a magnitude of 8, because the Longmenshan belt had never before recorded a quake bigger than magnitude 7. Secondly, if Zipingpu hadn’t been built, the epicentre would not have been in the current position, so close to the reservoir, but possibly somewhere else. Thirdly, if Zipingpu hadn’t been built, the quake would likely have occurred 100 or even 200 years later, but not now.
Before the earthquake, this fault had actually been moving, not with fast breakage but slow deformation, which is hardly perceptible to humans. In a region like the Longmenshan belt, the earth’s crust extrusion movement has been very strong, as if it was looking for the weakest link to break through. At this moment, the reservoir introduced an external influence, which not only allowed the seismic stress to be released at an earlier date, but in a truly unexpected and devastating manner. This is why the epicentre was really close to the reservoir. Because, in general, RIS occurs within a dozen kilometres from the reservoir.
Days ago, at a press conference by the Ministry of Water Resources, Liu Ning, chief engineer of the MWR, said that, after this earthquake, there is a need to re-assess the water resources development plan, especially the geological implications. Despite what he said, I don’t think there will be significant changes made in the major plans.
Wang Yongchen: Do you think there is a possibility of having more earthquakes?
Fan Xiao: It’s hard to say. It depends on whether the reservoir continues to store water. If the reservoir is filled again, it will definitely have an impact on the recurrence of earthquakes if a link between the reservoir and quakes exists. However, it would take time, given that the energy built up was released already. In general, there are intervals between one quake and the next.
Wang Yongchen: I wonder if anybody refutes your point of view?
Fan Xiao: No problem. Anybody can challenge me. Actually I talked to the South Urban Daily about that before. It is possible to determine whether Zipingpu triggered the earthquake through careful analysis and study with the already established seismic monitoring network and with complete data from the time the reservoir was filled to when the earthquake occurred. The experts at the Bureau of Seismology can do that for sure.
Wang Yongchen: I have got a story to tell you: a senior engineer in dam design in Yingxiu, now in his eighties, was invited to be in charge of dam projects many times, but he firmly refused, saying it’s too dangerous to do so because the mountains were too intensively used.
Fan Xiao: Human activity can be seen everywhere in the upper Min valley, such as building hydro dams in a massive way, developing industrial development zones, and expanding roads and so forth. More and more man-made projects and large-scale excavation have done so much harm to the environment, disturbing the stability of mountains in this area. As a matter of fact, landslides had already become a serious problem in many places, even before the May 12 quake occurred.
Therefore, it is no wonder that so much infrastructure collapsed as a result of the rock and landslides as soon as the big quake came. It’s sad to see, when you look from an airplane, that the Min valley has been afflicted with all kinds of man-made ills. It would be hard for the valley to recover completely in a short period of time. I believe it would take several decades and even up to one hundred years to recover.
Wang Yongchen: One of the experts says this is a new orogenic4 movement.
Fan Xiao: An orogenic movement would take a very long time, but we are truly shocked by such widespread destruction and change in such a short period of time. I believe that inappropriate human activities played a significant role in the earthquake, making the degree of destruction greater and the losses bigger. We have to say that conditions in this area make earthquakes inevitable, but the dam as the factor triggering the earthquake is too significant to ignore.
Translated by Three Gorges Probe
 On August 25, 1933 a 7.5 earthquake occurred north of Maowen, Sichuan province. The city of Diexi and about 60 villages in the area were completely destroyed. Damage and casualties also occurred at Chengdu and it was felt in Chongqing and Xi’an. Landslides created four lakes on the Min Jiang River. Over 2,500 of the 9000 casualties occurred 45 days after the earthquake, when the lakes broke through the slides and inundated the valley.
 Li Youcai is a 69-year-old senior engineer at the Sichuan Bureau of Seismology and is now retired. He successfully predicted seismic activity in Sichuan on several previous occasions and warned that earthquakes could occur in the Longmenshan area. He also insisted that the Zipingpu dam should be designed to have a higher capability to withstand earthquakes (at least 9 (IX) rather than 7 (VII) of seismic intensity) but the decision makers did not listen to him.
 Also called tectonism.The large-scale deformation of the Earth’s crust by natural processes, which leads to the formation of continents and ocean basins, mountain systems, plateaus, rift valleys, and other features by mechanisms such as lithospheric plate movement (that is, plate tectonics), volcanic loading, or folding.
Source: Britannica Online http://www.britannica.com/eb/article-9030295/diastrophism
 A mountain-building event, generally one that occurs in a geosyncline. Orogeny tends to occur during a relatively short geologic time frame. It is usually accompanied by the folding and faulting of strata and by the deposition of sediments in areas adjacent to the orogenic belt. Orogenies may result from continental collisions, the underthrusting of continents by oceanic plates, the overriding of oceanic ridges by continents, and other causes. See also Acadian orogeny, Alleghenian orogeny, Alpine orogeny, Laramide orogeny, Taconic orogeny.