Effect of Groundwater on Noise-Based Monitoring of Crustal Velocity Changes Near a Produced Water Injection Well in Val d'Agri (Italy)

We study the crustal velocity changes occurred at the restart of produced water injection at a well in the Val d'Agri oil field in January-February 2015 using seismic noise cross-correlation analysis. We observe that the relative velocity variations fit well with the hydrometric level of the nearby Agri river, which may be interpreted as a proxy of the total water storage in the shallow aquifers of the Val d'Agri valley. We then remove from the relative velocity trend the contribution of hydrological variations and observe a decrease in relative velocity of approximate to 0.08% starting seven days after the injection restart. In order to investigate if this decreasing could be due to the water injection restart, we compute the medium diffusivity from its delay time and average station-well distance. We found diffusivity values in the range 1-5 m(2)/s, compatible with the observed delay time of the small-magnitude (M-L <= 1.8) induced seismicity occurrences, triggered by the first injection tests in June 2006 and with the hydraulic properties of the hydrocarbon reservoir. Our results show that water storage variations can not be neglected in noise-based monitoring, and they can hide the smaller effects due to produced water injection.

Automated summary

Through seismic noise analysis, we found significant effects of water storage changes on velocity variations after water reinjection at the Val d'Agri oil field. After removing the effect of hydrological variations, a decrease in relative velocity was observed. The diffusivity values and delay time of small seismic events were found to be compatible.