Resolution Effects on Hydrologic Modeling Parameters and Peak Discharge

The writers examine the effects of digital elevation model (DEM) grid size on hydrologic modeling. The Natural Resources Conservation Service TR-55 model is used to estimate peak discharges from 44 watersheds in central Maryland ranging in size from 0.33 to 39 km2. A geographic information system based hydrologic analysis package is used to extract hydrologic model parameters from DEMs with cell sizes of 12, 36, 60, and 96 ft, and 30 and 90 m. The writers investigate the effects of grid size, data source, and aggregation on parameters such as drainage area, flow length, relief, slope, curve number, time of concentration, and peak discharge. The value of the model parameter derived from the 12-ft data is used as the reference value, with estimates of model parameters for DEMs of coarser resolutions compared against it. Simple regression models are calibrated to express the relationship that exists between derived hydrologic parameter values as a function of resolution. The writers report systematic errors or biases in parameter estimates for flow length, relief, slope, time of concentration, and ultimately peak discharge. Estimates of flow length, relief, and time of concentration all decreased consistently as the resolution grew more coarse. These biases lead to a tendency to overestimate peak discharge when coarser data are used. The writers also present confidence envelopes for the relative errors in each hydrologic parameter with respect to resolution and the parameter value itself.