The goal of this effort was to assess if we could use spatial statistical tools provided within ArcGIS 10.1 to evaluate geospatial relationships to differentiate between natural subsurface leakage pathways and anthropogenic leakage pathways. Evidence for gas migration within the subsurface is well documented through petroleum system and CO2 sequestration related studies. In addition, there are a number of examples of paleo- and modern natural gas seepage from deep subsurface systems to the near-surface in both onshore and offshore environments. Gas/fluid migration are key to the documentation of fluid flux relative to both resource/storage potential estimates and improving constraints on potential environmental impacts associated with gas migration from subsurface to groundwater systems and the atmosphere. Migration of gas from subsurface systems is particularly visible in the marine environment where gas plumes into the water column have been observed and studied by a number of researchers worldwide.
Migration of gas from subsurface systems to near surface systems has been the target of relatively few investigations, although a better understanding of these processes is of increasing interest due to the focus on potential storage of CO2 in geologic media and on methane recovery using hydrofracking techniques. To date, many of the studies evaluating the flux and migration of gas from subsurface to near surface systems have focused on anthropogenic concerns such as potential leakage related to CO2 storage, landfills, and natural gas storage facilities. Given the variability of the natural geologic system and the limited amount of existing field data and studies examining these systems, questions about the timing and nature of gas migration in many regions persist. Evaluating available data to assess whether interactions between subsurface and near-surface systems are related to naturally occurring mechanisms or are the result of anthropogenic activities offers additional challenges. However, improving understanding of mid to large-scale, field- and regional-scale controls on gas flux will help further constrain the relationships between geologic media, and the nature and timing of gas flux in a variety of subsurface systems. Through the evaluation of field datasets utilizing laboratory, geospatial and geostatistical analytical methods patterns related to gas occurrences and concentrations in subsurface systems can be identified.

To support these evaluations in a test scenario for CO2 geologic storage potential in the Appalachian Basin’s Oriskany formation, we utilized geostatistical tools to assess a variety of parameters associated with the reservoir characterization of the Oriskany formation including thickness, porosity, subsurface depth to top of formation, pressure, temperature, and a variety of pore-filling media relationships such as brine density, salinity, etc. The data leveraged for this analysis was largely based upon borehole geophysical log interpretations from West Virginia, Ohio, and Pennsylvania. The distribution of these data were irregular, and not every data point had values for all the reservoir parameters being evaluated. The variability of the wellbore data points is highlighted in the figure below.

To fill in the gaps for areas of low data density or even no data density to develop a complete dataset with values for all parameters to assist with CO2 storage estimations, each parameters needs to have a value (at least a minimum, maximum, and average value) for all parameters. As a result, the following tools and approaches were evaluated within ArcGIS for this dataset:
1. Use the min, max, and average values for the entire dataset and assign those values to all empty grids
2. Inverse Distance Weighting (IDW) evaluated to fill in the empty grids
3. Kriging also evaluated as a tool to fill in the empty grids
Example of these results for porosity data:

The graph above shows the relative distribution and compares average % porosity values for the raw data points against interpolated values using ArcGIS’ kriging versus inverse distance weighting (IDW) approaches. While offering generally similar results, the IDW approach for this dataset appears to interpolate and fill in data gaps in a manner that more closely follows the trends of the actual average porosity data versus that of the kriging interpolation algorithm.
Summary and preliminary conclusions:
1. Recommend using IDW interpolations to fill gaps
a. Best fit to original well data for parameters
b. Represent natural fluctuations in data range
c. Correlates parameters across spatial distances
2. Intent to represent “field-” to basin-scale features and variability not reservoir/borehole scale features.
3. Help inform technology development, risk evaluation, knowledge gaps
4. For extrapolation to more complex systems and problems dealing with 3D, multi-system components:
a. Basic geostatistical tools are sufficient for preliminary analyses and filling in gaps
b. But more complex analyses will require other software/approaches