Eddies of varying size, which characterize the convective boundary layer, are typically sub-grid scale in numerical weather prediction simulations, and thus are normally parameterized. The most simplistic parameterization, K-theory, can instantaneously transfer quantities between adjacent grid boxes, along the local gradient of the quantity. The reliance of K-theory on the local gradient, results in K-theory accounting for only the smallest of the convective eddies; neglecting the behavior and properties of the larger eddies. This is especially evident in the boundary layer observed during lake effect snow events. Attempts at modeling the convective boundary layer of lake effect snow events have shown that standard K-theory requires a deep superadiabatic layer to develop in order for sub-grid scale turbulence to transfer entropy and moisture through the depth of the convective boundary layer. In reality, large eddies transfer entropy and moisture through locally stable areas. A nonlocal correction term was added to the K-theory formulation in the UW-NMS (University of Wisconsin - Nonhydrostatic Modeling System) in order to account for the transport done by the larger convective eddies. Simulations with and without nonlocal vertical transport are presented for cases where the eddies are completely parameterized and for cases where the eddies are partially resolved. The nonlocal term was quite successful in improving the simulations with unresolved clouds. Simulations where cloud structures were partially resolved showed a double accounting problem, similar to those encountered with cumulus parameterizations.