Like other sedimentary plains, the Po Plain in Northern Italy has largely subsided due to natural processes and human activities. Displacements of the Earth surface of hydrological origin are caused by groundwater changes, which in turn, are expected to be related to rainfall changes. In the Bologna metropolitan area (located in the Southeastern border of Po Plain), the 2010 politic decision of dismissing civil water supply from the groundwater withdrawal has provided us the opportunity to test a methodology for the retrieval of an anthropic effect in two different data sets: vertical displacements measured by continuous GNSS sites and piezometric water table fluctuations. The data sets have been analyzed by means of the Principal Component Analysis (PCA) and compared to rainfall time series from the Po Plain rain gauges. Several piezometers undergo a clear increase in the water level following the withdrawal decrease. Differently, the anthropic induced surface displacements are significantly smaller than the ones induced by rainfall. Accordingly, without a multivariate analysis such an effect on vertical displacements would have remained hidden in the raw time series. Only looking at the spatial distribution of the principal components we have highlighted that anthropic effects are local and present even in GNSS data, entailing for the 2010 case a decrease of about 4 mm/y of vertical velocity in some sites closest the withdrawal wells. Moreover, the multivariate analysis allowed us to assess that, on time scales larger than months, the rainfall-related hydrological response of vertical displacement depends on the geological setting. In the Apennines chain a water level increase causes subsidence, in agreement with the predictions of elastic models, whereas in the Po Plain it causes uplift, suggesting a dominant poro-elastic response, in agreement with the guess that the subsidence of the Po Plain is related to soil compaction. Our results suggest that in cases of the aquifers over-exploitation, a PCA analyses and the combined use of different observables such as GNSS, piezometers time series, rainfall data, geological setting allow getting a correct identification of the anthropic and natural signals.

像其他沉积平原一样，意大利北部的蒲平原在很大程度上由于自然过程和人类活动而消退。水文起源的地球表面位移是由地下水变化引起的，而地下水变化又与降雨变化有关。在博洛尼亚都会区（位于Po Plain的东南边界），2010年从地下水取水中取消民用供水的政治决定为我们提供了机会，以两个不同的数据集来测试一种人为效应的检索方法：由连续的GNSS站点和测压地下水位波动测得的垂直位移。数据集已通过主成分分析（PCA）进行了分析，并与Po Plain雨量计的降雨时间序列进行了比较。随着抽水量的减少，数个压力计的水位明显增加。不同的是，人为引起的表面位移明显小于降雨引起的位移。因此，如果没有多元分析，这种对垂直位移的影响将一直隐藏在原始时间序列中。仅查看主要成分的空间分布，我们就强调了人为影响是局部的，甚至在GNSS数据中也存在，对于2010年的情况，在最靠近抽水井的某些位置，垂直速度降低了约4 mm / y。此外，多变量分析使我们能够评估，在大于数月的时间尺度上，与降雨有关的垂直位移水文响应取决于地质环境。在亚平宁山脉链中，水位的增加引起沉降，这与弹性模型的预测相符；而在波普平原，它引起隆升，表明是主要的孔隙弹性反应，这与波普平原的沉降为与土壤压实有关。我们的结果表明，在含水层被过度开采的情况下，PCA分析并结合使用不同的可观测数据（如GNSS，测压仪时间序列，降雨数据，地质环境）可以正确识别人为和自然信号。