SOIL GAS VELOCITY AND FLUX MODELLING USING LONG TERM RN MONITORING: A NEW TOOL FOR DEEP CO_2 ESCAPE DETECTION

摘要

Carbon dioxide has been injected into the oil reservoir at the Weyburn oilfield, Saskatchewan, Canada since September 2000 as part of a commercial enhanced oil recovery operation. As part of a European/North American research programme, soil gas data were obtained at sites selected from systematic measurements of many soil gas species and surface CO_2 flux. This work is being done to monitor the oilfield for possible leaks of injected CO_2 as part of the IEA Weyburn CO_2 Monitoring and Storage Project, which is being run in conjunction with the EnCana Weyburn CO_2 flood. If leakages were to occur, the one of the most likely sources of these would be via fault systems, which are also pathways for the Earth's natural degassing processes. Such structural features could potentially penetrate the reservoir caprock and may allow some CO_2 seepage to the surface. Potential natural gas leaks were searched for by soil gas surveys with the in situ measurement of CO_2, helium and radon at 60-80 cm depth along traverses at 25 m intervals. The traverses were sited to cover anomalies seen on soil gas data from a wider grid of measurements at 200m spacing. Helium (a deep gas tracer) and radon (a gas velocity tracer) are the best indicators of suspected natural gas leaks. Six radon detectors ('Barasols') were inserted into 1.5 m long PVC pipes, buried to 40 cm and covered by a waterproof (GoreTex) membrane and finally covered with soil at sites where there was a significant excess of helium relative to the atmosphere (Figure 1). Hourly records of radon concentrations since the autumn of 2001 show changes in the gas flow regime and seasonal soil permeability (Figure 2). Signal processing of soil-air-radon variations against atmospheric pressure indicates coherent maximum gas velocities of 5 to 15 cm/hour, in the range of usual values obtained at faults. At the opposite end of the gas velocity spectrum, diffusion processes control background values. The sensitivity of the monitoring equipment means that any precursor of pressure regime variation from depth would be recorded. Calculation of the CO_2 flux at 1.9 m depth gives values 10 to 20 times lower than ground surface measurements suggesting that flux monitoring would be better carried out at this depth, where biogenic CO_2 production is much reduced. Soil radon monitoring at Weyburn, over 2 1/2 years indicates that soil degassing appears to be affected by surficial parameters such as soil temperature, moisture, etc. and that there is presently no indication of deep CO_2 leakage.