Regulating Subsidence and its uncertainty in the Dutch Wadden Sea

Abstract

Abstract. At the start of gas production its effects on land subsidence are not certain. There are uncertainties in mechanisms, models and parameters. Examples are non-linear deformation of reservoir rock, fault transmissibility, behaviour of overlaying salt and aquifer activity. Looking back at historical cases in the Netherlands, a factor two or three difference between initial prediction and final outcome is quite common. As the Dutch regulator, SSM is tasked with assuring proper management by operators of the risks associated with land subsidence from natural gas production in The Netherlands. Large initial uncertainties can only be tolerated if operators can demonstrate that timely actions can still be taken when predefined subsidence limits are at risk of being exceeded now or in the future. The applied regulatory approach is illustrated by the case history of gas production induced subsidence in the Dutch Wadden Sea area. This environmentally highly sensitive UNESCO World Heritage Site is a natural gas province. Extensive legal, technical and organisational frameworks are in place to prevent damage to its natural values. Initial uncertainties in the predicted subsidence (rate) were later exacerbated by the detection of strong non-linear effects in the observed subsidence behaviour, leading to new concerns. It was realised that – depending on the underlying physical cause(s) – there will be a different impact on future subsidence. To assure proper management of the additional uncertainty by the operator, several improvements in the regulatory approach have been implemented. Possible underlying mechanisms had to be studied in depth and improved data analysis techniques were requested to narrow down uncertainties as time progresses. The approach involves intensified field monitoring, scenario's covering the full range of uncertainties and a particle filter approach to handle uncertainties in predictions and measurements. Spatial-temporal double differences, production data and the full covariance matrix are used to confront scenario predictions against measurements and to assess their relative probability. The regulator is actively involved in assuring this integrated control loop of predictions, monitoring, updating, mitigation measures and the closing of knowledge gaps. The regulator involvement is supported in the Mining law and by appropriate conditions in the production plan assent. With the approach it can be confidently assured that subsidence (rate) will remain within the allowed range.