But what if the premises of the present model employed by the US Geological Survey and research centres outside the US are seriously flawed because they are based on an unnecessarily limited range of precursors? What if these models can be refined by a broader and richer portfolio of indicators?

The International Earthquake and Volcano Prediction Centre in Orlando, Florida, has assembled a panel of distinguished experts from Russia, Japan, India, Italy, Australia and the US who have developed an alternative model called the "Catastrophic Geophysical Event Monitoring and Warning System". It provides a potentially powerful tool in the arsenal of mainstream seismologists.

Here is the essential idea.

The system monitors earthquake precursor signals, beginning with those of the longest time frame from the tentatively predicted date. It then marches through what amounts to an earthquake prediction countdown from two to three years out and ends with precursor signals that can predict a major quake (with a magnitude of more than 6.5) within one hour or less from the predicted time.

As the time-phased process proceeds, each successive precursor typically becomes more accurate. The longest-term precursors have the lowest probability for predictive accuracy, and the shortest time-frame precursors have the highest accuracy. However, it is the integration of all these signals within a coherent framework that yields the most effective results.

During the last six months of 2012, the model correctly identified the time frame, magnitude and epicentre of these earthquakes:

• Off the Kamchatka Peninsula in the Russian Far East - a deep sea quake resulting in a swarm of magnitude 4.6 to magnitude 5.8 quakes within an eight-day period;

• A magnitude 6.0 deep-sea quake in the Celebes Sea where no major fault was involved;

• A highly accurate prediction of a magnitude 6.8 quake in Myanmar, an inland quake involving a known fault.

This year, there was a successful prediction of the northern Chilean Tarapaca earthquake (magnitude 6.7) on March 16, which was soon followed by a gigantic magnitude 8.2 main shock on April 1.

These findings suggest it may be possible to construct an earthquake model applicable to a diverse range of seismic conditions occurring anywhere in the world.

The warning system suggests a high probability of a large earthquake within the next eight months with its epicentre in the deep sea, 400-500 kilometres south of the Philippine coast of the Molucca Sea.

The prediction is based on an analysis of several prominent precursors, including the strongest deep precursors that occurred in the Celebes Sea in July 2010, geological structure, changes in sea surface temperature and anomalies in outgoing long wave radiation. As the earthquake approaches, other short-term signals, such as total electron content, a very low frequency electromagnetic wave propagation anomaly, and animal behaviour reported by local residents will be observed, analysed and assessed.

If the alert is accurate, 2.4 million people living in the area of the Molucca Sea, including the Philippines, are in jeopardy. The coming earthquake cannot be prevented; but it can be anticipated, and a non-hysterical, reasonable emergency response can be prepared.

US President Barack Obama has just announced a new security pact between the US and the Philippines, anticipating even greater friction with China in the South China Sea. By refining their outdated seismic models, both countries can save many lives, while simultaneously extending a protective shield for all of humanity, including China.