A probabilistic methodology to generate “risk maps” for determining the spatial extent and associated probability of groundwater leakage risk at Geologic Carbon Storage (GCS) sites was recently developed and is tested for well permeability distribution sensitivity here. Although the risk map methodology is general and may be used to assess leakage of either CO2 or brine, the degree to which risk may occur via plugged and abandoned (P&A) brine leakage is arguably the most uncertain. This uncertainty stems from the possibility of a high number of undocumented wells at a given injection site which, even when properly plugged, may yield a number of different leakage pathways. Furthermore, determining the spatial extent and probability of brine leakage through P&A wells is not straightforward. The aforementioned risk maps are probabilistic and are based on the premise that two of the greatest sources of uncertainty in brine leakage through P&A wells are the location of the unknown well with respect to the injection well and the permeability of the leaky well (which can span over several orders of magnitude). The methodology utilizes either analytical solutions or numerical simulations in conjunction with probability theory to generate spatial distributions of risk. Risk is defined with either no-impact levels (i.e. zero brine concentration above background in the drinking water aquifer) or other user defined thresholds such as the secondary maximum contaminant level (SMCL) for total dissolved solids, a non-health related standard related to water hardness, color, and taste. These probabilistic maps can be used to provide risk-based descriptions of the injection area to inform site selection prior to injection and monitoring during or after injection has ceased, similar to the United States Environmental Protection Agency's definition of the area of review (AoR). However, unlike the AoR definition, we differentiate risks according to a tiered approach, where CO2 leakage risk and brine leakage risk through P&A wells and open wellbores are assessed separately given that not only do the risks differ considerably, the spatial extents of each hazard do as well. A strength of this methodology is its demonstrated efficiency, which unlike computationally expensive Monte Carlo techniques that could require several thousands of simulations, requires only on the order of tens of simulations. Here, we assess the sensitivity of risk maps of a hypothetical, yet realistic GCS leakage scenario. In this sensitivity analysis, risk maps are generated with varying leaky well permeability distributions including differences in distribution modality and distribution variance. This work is motivated by the unknown quality to which many wells are sealed and/or the degree to which they degrade over time, in addition to the sheer unknown quantity of which P&A wells may be present at a single Geologic Carbon Storage site. Quantitative probabilities of exceeding contamination thresholds in the USDW are used for discussion, a practical metric for decision-makers and stakeholders. Results show that the modality of the leaky P&A distribution plays an important role in the risk profile and that larger variances in leaky P&A well permeability distribution may result in an overall lower profile of risk, depending on the upper-threshold of the distribution.