Real-time multi-level CO2 concentration monitoring in vadose zone wells and the implication for detecting leakage events

Abstract Multi-level wells screened at different depths in the vadose zone were installed and used for CO 2 and carbon isotope monitoring. Well CO 2 time series data were collected along with subsurface and atmospheric parameters such as air pressure, temperature, wind speed, and moisture content. Our aim was to determine the natural factors affecting the variation of CO 2 concentration and how the influence of these factors varies with time of day and seasons of the year. We were motivated to understand the cause and extent of CO 2 natural fluctuations in vadose zone wells in order to separate natural variation from signals due to anthropogenic CO 2 leaks anticipating future monitoring using these wells. Variations of seasonal mean and variance of CO 2 concentrations at different depths seem to follow the diurnal trend of subsurface temperature changes that reflect the atmospheric temperature but with time delay and amplitude damping due to heat transport considerations. The temperature in the ground lags behind the change in the atmospheric temperature, thus, the deeper the depth, the longer the time delay and the smaller the amplitude of the change. Monitored seasonal variation as shown in Appendix A shows the temperature-dependent depth-dependent CO 2 production in the soil zone indicating higher CO 2 concentrations in the summer and fall seasons with high concentrations ranging between 10,990 and 51,600 ppm from spring to summer, and 40,100 and 17,760 ppm from fall to winter. As the temperature in the organic-rich topsoil layer changes from daytime to nighttime, the concentration of CO 2 in the soils also changes dynamically in response to chemical and biological reactions. When a screened well is installed in the vadose zone the dynamic temporal and depth difference in CO 2 production is further complicated by upward (out of the subsurface) or downward (into the subsurface) gas flow, which will amplify or attenuate the temporal and vertical biochemically produced differences. Nested wells screened at different depths in the vadose zone and wells fully screened through the vadose zone were used for comparison. In addition, experiments changing the well from open to surface air to sealed at the top were conducted. The flow rates of inhaled (downward) and exhaled (upward) gas were estimated based on multi-level monitoring data. Based on time-series monitoring data, we proposed a time-dependent conceptual model to explain the changes of CO 2 concentration in wells. The conceptual model was tested through analytical model computations. This conceptual model of natural variation of CO 2 will be helpful in utilizing the vadose zone well as a method for monitoring CO 2 leakage from subsurface storage or anthropogenic CO 2 -producing activities.