There is growing concern about seismicity triggered by human activities, whereby small increases in stress bring tectonically loaded faults to failure. Examples of such activities include mining, impoundment of water, stimulation of geothermal fields, extraction of hydrocarbons and water, and the injection of water, CO₂ and methane into subsurface reservoirs¹. In the absence of sufficient information to understand and control the processes that trigger earthquakes, authorities have set up empirical regulatory monitoring-based frameworks with varying degrees of success^2,3. Field experiments in the early 1970s at the Rangely, Colorado (USA) oil field⁴ suggested that seismicity might be turned on or off by cycling subsurface fluid pressure above or below a threshold. Here we report the development, testing and implementation of a multidisciplinary methodology for managing triggered seismicity using comprehensive and detailed information about the subsurface to calibrate geomechanical and earthquake source physics models. We then validate these models by comparing their predictions to subsequent observations made after calibration. We use our approach in the Val d’Agri oil field in seismically active southern Italy, demonstrating the successful management of triggered seismicity using a process-based method applied to a producing hydrocarbon field. Applying our approach elsewhere could help to manage and mitigate triggered seismicity.

人们越来越关注人类活动引发的地震活动，即应力的小幅增加会导致构造承载的断层破裂。此类活动的示例包括采矿、蓄水、地热田增产、碳氢化合物和水的开采以及将水、CO ₂和甲烷注入地下储层¹。在缺乏足够的信息来理解和控制引发地震的过程的情况下，当局建立了基于经验的监管监测框架，并取得了不同程度的成功^2,3。20 世纪 70 年代初在美国科罗拉多州 Rangely 油田的现场试验⁴建议可以通过使地下流体压力循环高于或低于阈值来打开或关闭地震活动。在这里，我们报告了一种多学科方法的开发、测试和实施，该方法使用有关地下的全面和详细信息来校准地质力学和震源物理模型来管理触发的地震活动。然后，我们通过将它们的预测与校准后的后续观察结果进行比较来验证这些模型。我们在意大利南部地震活跃的 Val d'Agri 油田使用我们的方法，展示了使用应用于生产油气田的基于过程的方法成功管理触发地震活动。在其他地方应用我们的方法可以帮助管理和减轻触发的地震活动。