predicting earthquakes

• Not an exact science. 
• Data goes back 100 years. 
• No single method is reliable for accurate prediction. 
• Precursor events vary between earthquakes. 
• Successful predictions are minimal. 
• • Return periods 
  • Seismic gap concept 
  • Patterns of foreshocks 
  • Groundwater monitoring 
  • Ground deformation 
  • Radon gas monitoring 
  • Electrical resistivity 

   
• Return period in Parkfield, California. 

  Lies on the San Andreas fault. Within a 100 years 6 earthquakes with a magnitude of 6.0 occurred, one every 22 years. A predicted earthquake arrived outside the 10 year window of return. 

   
• Seismic gaps are segments of active faults that have not moved for a long time compared to other segments along the same structure. Over a long period of time the displacement of each segment should be equal. Any areas with a longstanding gap are more likely to suffer an earthquake. 

   
• Seismic velocity changes (p waves). Velocity has a background rate, before an earthquake the p waves drop and rise - when returns to normal the earthquake occurs. Length of the drop directly relates to the magnitude. Limited evidence from a graph can be drawn to predict the magnitude of the quake. 

   
• Seismic velocity changes:

  • Water saturated rocks under stress increase in volume and fracture. 
  • Opening of micro-cracks that increase porosity - initially no water fills the spaces.
  • Lack of water (empty space) reduces p wave velocity. Water fills the cracks and increases the p wave velocity. 
  • The rock contains more water and is weaker. 
  • Length of time taken to restore water pressure indicates magnitude. 
• Patterns of foreshocks:

  Foreshocks are tremors with magnitude of 2.5 on the Richter scale - taking place before a mainshock of 5.4 magnitude. 
• Foreshocks:

  50 % of major earthquakes have foreshocks, 5 - 10 % of small earthquakes are foreshocks - false warnings. The east pacific rise transform faults show foreshock activity. 
• Foreshock:

  Foreshock activity combined with groundwater level and animal behaviour in China meant that millions of people were evacuated before an 7.3 magnitude earthquake. 
• Groundwater levels:

  Gradual lowering of water levels over time before an earthquake. Faster lowering of in the immediate time before an earthquake. In the last few hours before a quake the water table rebounds to original level. 
• Groundwater levels:

  Fluctuations of water due to tide, season and abstraction. China and Japan have dug about 100 wells each to monitor groundwater levels. 

   
• Groundwater temperature:

  Water temperature is monitored with a high accuracy. The day before the Kanto quake in Japan geysers became active in some hot springs. Monitoring showed a 0.3 dc rise in temperature and a small decrease before the event. 

   
• Radon gas analysis:

  Increase of radon in wells can hint at an earthquake. Radon is soluble in water so is easily monitored in wells and springs. 
• Radon:

  Stress is released and micro-fractures in rocks at depth. The radon escapes into the water that flows into the fractures. This pattern was seen before the Kobe quake in Japan. 
• Ground deformation-uplift and tilting:

  Measurements taken close to active faults show changes in height and gradient due to swelling of rocks caused by stress build up. 

   
• Ground deformation:

  • Tiltmeters - changes in gradient angle 
  • Creepmeter - two fixed piers over a fault. Measures displacement over time 
  • Strainmeter - placed down a borehole to measure change in shape
• Resistivity:

  Rocks are poor conductors, water is better at conducting electricity. When cracks develop, water flows into them - resistivity decreases and conductivity increases. Just after an earthquake the resistivity returns to normal. 5 - 10 % drop in resistivity has been observed before an earthquake.