What have we learned about CO2 leakage from field injection tests?

Legislation and guidelines developed for Carbon Capture and Storage (CCS) have set performance requirements to minimize leakage risk, and to quantify and remediate any leaks that arise. For compliance it is necessary to have a comprehensive understanding of the possible spread, fate and impacts of any leaked CO2, and also the ability to detect and quantify any leakage. Over the past decade, a number of field scale CO2 release experiments have been conducted around the world to address many of the uncertainties regarding the characteristics of near-surface expression of CO2 in terms of the impact and quantitation of CO2 leaks. In these experiments, either free phase or dissolved CO2 is injected and released into the shallow subsurface so as to artificially simulate a CO2 leak into the near-surface environment. The experiments differ in a number of ways, from the geological conditions, surface environments, injection rates and experimental set-up-including the injection and monitoring strategy. These experiments have provided abundant information to aid in the development of our scientific understanding of environmental impacts of CO2 while assessing state of the art monitoring techniques. We have collated a global dataset of field-scale shallow controlled release experiments that have released CO2 at depths shallower than 25 m. The dataset includes 14 different field experiment locations, of which nine intended to release CO2 to surface, and the remaining sites intended for CO2 to remain in the shallow subsurface. Several release experiments have been conducted at half of these sites, and so in total, 42 different CO2 release tests have taken place at the 14 sites we examine. These experiments and their results are scrutinised to establish: (i) the range of experimental approaches and environments explored to date (such as the environment, subsurface conditions, injection strategy and whether gaseous or dissolved CO2 were injected and in what quantities); (ii) the range of CO2 injection and surface release rates at these experiments; (iii) the collective learnings about the surface and subsurface manifestation of the CO2 release, the spread and fate of the CO2, rates of CO2 flux to surface, and methods of measuring these; (iv) how successfully current approaches can detect and quantify CO2. This allows us to highlight where uncertainties remain and identify knowledge gaps that future experiments should seek to address. We also draw on the collective experiences to identify common issues or complications, and so recommend best practice' guidelines for experiment design and reporting at future CO2 release experiments.(C) 2017 The Authors. Published by Elsevier Ltd.