Study of seasonal variation of groundwater and atmospheric loads : Possible influence on seismicity and astrometry

我们定量地评估了日本全国GPS网(GeoNet)位置时间序列中影响季节成分的各种因素。我们发现，日本群岛表面负荷的季节变化是造成季节变化的原因。其中，雪荷载是其他荷载的几倍。大气负荷、地下水负荷(土壤水分和水库蓄水)、非潮汐海洋负荷被认为是次要的。我们利用气象资料和模型估计了这些载荷的季节变化，计算了预期的地壳季节性形变，并与实际观测的形变进行了比较。我们发现，除了这些负荷变化外，与未知来源的季节尺度变化和大气延迟梯度相关的明显变化也是重要的贡献。我们建立了考虑所有这些因素的经验季节变化模型，这些模型预测的季节变化与实际GPS数据一致。季节分量被一种自适应方法去除，即用最小二乘法估计一年和两年一次的分量，以便分离出瞬时地壳形变信号。我们的调查表明，使用我们的基于先验知识的模型(非自适应校正)去除季节性分量可以产生与自适应方法一样好的结果。我们的研究成果发表在《地球状况》一书(AGU地球物理专著系列第150卷)上。我们还计算了这种季节性载荷引起的库仑破坏应力，发现它们对日本内陆的地震活动有一定的影响。众所周知，在积雪覆盖的地区，春季/夏季的地震多于秋季/冬季的地震，这与我们的预测是一致的。这篇论文发表在地球科学期刊上，并被《自然》、《新科学家》等主要科学期刊S介绍。

We quantitatively evaluated various factors governing the seasonal components in the position time series of the Japanese nationwide GPS network (GEONET). We found that seasonal changes of surface loads in the Japanese Islands are responsible for seasonal changes. Among them, the snow loads are several times as large as other kinds of loads. Atmospheric loads, groundwater loads (soil moisture and empoundment in reservoirs), non-tidal ocean loads were found to be of secondary importance. We estimated the seasonal changes of these loads using meteorological data and models, and calculated expected seasonal crustal deformation, compared them with actually observed deformations. We found that, in addition to these load changes, significant contributions come from the apparent changes associatedd with seasonal scale changes of unknown origin and atmospheric delay gradients. We established empirical seasonal change models considering all these factors, and the seasonal changes predicted by these models were consistent with actual GPS data. Seasonal components have been removed by an adaptive method, i.e. by estimating annual and biannual components using least-squares method in order to isolate transient crustal deformation signals. Our investigation showed that removal of seasonal components using our model based on a priori knowledge (non-adaptive correction) gives rise to as good results as the adaptive method. Our results were published in the book "State of the Planet" (AGU Geophysical Monograph Series Vol.150). We also calculated the Coulomb Failure Stress caused by such seasonal loads and found they have some influence on the inland seismicity in Japan. It has been known that there are more earthquakes in spring/summer than in fall/winter in snow covered regions, which is consistent with our prediction. This was reported as a paper in Earth Planet.Sci.Lett., and was introduced by major scientific journal s such as Nature and New Scientist.