head img  
NTU Taiwan University img


Wind Rises with the Moon Halo, Rain Comes along with the Wet Base Stone of the Column: Unveiling the Myth of Analyzing Statistics and Detecting Causality, Professor Chih-hao Hsieh of NTU Institute of Oceanography Had His Recent Research Result Published on Science

Professor Chih-hao Hsieh of NTU Institute of Oceanography and Institute of Ecology and Revolutionary Biology and his international research team have just had the recent research result published on Science on October 26th: Detecting Causality in Complex Ecosystems. Their new approach of analyzing the real cause in complex ecosystems has successfully made it possible to detect the cause and effect, and furthermore they solved the myth of the related statistics in analysis of the causality, which have been baffling scientists for quite a long time.

It goes well with this old Chinese saying, “wind rises with the moon halo. Rain comes along with the wet base stone of the column.” However, the moon halo is not what causes the rising wind, and the wet base stone is not the cause of the rain. It is not exactly as it seems to be that two related objects could result in the cause and effect. Take a simple daily case for example, we tend to feel flighty and impetuous in the summer. And ice cream usually has the better sale in the summer, so it’s easily and very likely that we relate our being flighty and impetuous to the ice-cream sale in the summer, when it comes to the statistics. However, we should bear this in mind that it is irrelevant to conclude the cause and effect in this case. Instead of that, it is only because of the influence by the weather condition in the summer. This hilarious example shows not only a ridiculous joke but also a typical case while doing the researches in the ecosystems, making a certain case the unsolved, baffling task. To be more specific, the species’ close connection in the natural world might be resulted from the species’ interaction between these two species as well as the influence from the same environmental factors on the two species at the same time. The related statistics between them does not suggest the cause and effect in it; instead, it does not mean that there is no causality between two objects, which appear to be not related. The fact is, in the real ecosystems, it can get very complicated with the changing factors, and finding the real key factor is also as difficult as searching a needle in the vast ocean.

It was not until the new method from Professor Clive Granger and Professor Robert Engle that the myth in the natural world could be unveiled. The two professors gave up the conventional scientific approach by using the relativity; instead, they based on prediction to analyze the cause and effect. What they have predicted has already been successfully proved true in analysis of the data of finance and economics. This scientific breakthrough won them the Nobel Prize of economics in 2003. Professor Clive Granger and Professor Robert Engle designed their method according to the linear stochastic systems. Though this method has successfully solved parts of the causality and the correlation, it does not apply well in the nonlinear dynamic systems, which has been organized by rules in ecosystems.

In order to detect the causality and the correlation in complex ecosystems, Professor Chih-hao Hsieh has successfully developed convergent cross mapping with his NTU research team for the detection. In simple words, Professor Hsieh’s method is based on this Chinese saying, “It leaves the trace after every footstep.” His research team used data from convergent cross mapping to detect whether an object did leave the trace in another object’s history. If so, they can prove it true that this object is actually the causing effect. Their scientific approach is illustrated both by simple models and by application to real ecological systems.

In the publication on Science magazine, Professor Chih-hao Hsieh and his NTU research team used convergent cross mapping to analyze the history data of anchovy and sardine, the two major species in the fishery industry in the Californian current ecosystem. Their publication, Detecting Causality in Complex Ecosystems, presents how Professor Hsieh and his research team have successfully figured out the long-debating factors of causing the Californian sardines and anchovies have developed changes in their history. Over the past 100 years, there have been several times of anti-synchronistic happened in the two schools of sardine and anchovy, and this phenomenon shows the minus correlation obviously. Professor Hsieh presented a method, based on nonlinear state space reconstruction, which can distinguish causality from correlation (and which applies even when the dynamical system is chaotic.) It has been long debating that whether the species’ competition or the environment (the change of the sea temperature) resulted in the anti-synchronistic changes. In the research article on Science, Professor Hsieh presented that the anti-synchronistic between the two schools of major fish in Californian current ecosystem is caused by the different responses to the change of the sea temperature instead of the species’ competition in the ecological system based on their new scientific approach. This method can also be applied even when the dynamical is chaotic in other nonlinear state space reconstruction for the detection of the updated systems and the causality. It is highly expected that t can contribute a lot to the management of ecological systems and reservation in the near future.

Extended reading:

1. “Detecting causality in complex ecosystems” Science, Oct. 26, 2012.

2. “Causality test could help preserve the natural world” New Scientist, Sep. 28, 2012.

3. The front page explains how to detect the causality and the correlation by using the scientific method of State Space Reconstruction. (Click the Youtube video clips.)

(a) Time Series and Dynamic Systems http://youtu.be/7ucgQE3SO0o

(b) Takens Theory and Shadow Manifolds http://youtu.be/rs3gYeZeJcw

(c) Convergent Cross Mapping http://youtu.be/NrFdIz-D2yM


Chinese version