Streamflow, biogeochemical cycling, and flux transport models rely on digital representations of river networks. At local to regional extents, such representations can be very detailed and account for individual hydrologic features such as dams and river diversions. However, at continental to global extents, these hydrologic features are often far less resolved. This lack of detail can lead to a mismatch between the resolution of hydrologic features relative to the resolution of the network itself. One solution to such mismatches is to impose a “global” standard hydrologic feature resolution. However, this approach may fail to provide critical information that is essential for accurate modeling because it removes hydrologic feature data (such as lakes) that could otherwise be passed to calibration and fitting routines. In this research, we test how variations in river network resolution may introduce such resolution mismatches. Using Viti Levu, Fiji as a case study, we show that even small “coarsening” of network resolution has a significant effect on lake representation and has carry-on effects on overall network transport. Because these effects were less pronounced for networks with larger lakes, this indicates that including lakes even in coarse models may be very informative given the extent to which available chemical and hydrologic data is skewed towards larger lakes.