Anomalous variations of lithosphere magnetic field before several earthquakes

Based on the geomagnetic vector data measured each year since 2011 at more than 500 sites with a mean spatial interval of ~70km.we observed anomalous variations of lithospheric magnetic field before and after over 15 earthquakes having magnitude > 5. We find that the field in near proximity (about 50km) to the epicenter of large earthquakes shows high spatial and temporal gradients before the earthquake. Due to the low frequency of repeat measurements it is unclear when these variations occurred and how do them evolve. We point out anomalous magnetic filed using some circles with radius of 50km usually in June of each year, and then we would check whether quake will locat in our circles during one year after that time (June to next June). Now we caught 10 earthquakes of 15 main shocks having magnitude > 5, most of them located at less than10km away from our circles and some of them were in our circles. Most results show that the variations of lithosphere magnetic filed at the epicenter are different with surrending backgroud usually. When we figure out horizontal variations (vector) of lithosphere magnetic field and epicenter during one year after each June, we found half of them show that the earthquakes will locat at "the inlands in a flowing river", that means earthquakes may occur at "quiet"regions while the backgroud show character as"flow" as liquid. When we compared with GPS results, it appears that these variations of lithospere magnetic field may also correlate with displacement of earth's surface. However we do not compared with GPS results for each earthquake, we are not clear whether these anomalous variations of lithospere magnetic field may also correlate with anomalous displacement of earth's surface. Future work will include developing an automated method for identifying this type of anomalous field behavior and trying to short repeat measurement period to 6 month to try to find when these variations occur.