Abstract
Digital elevation models (DEMs) are commonly used to illustrate and model topographic information related to river or stream channels. DEMs are frequently generated using interpolation to approximate elevation values in between points that are directly surveyed. Data collection using surveying is time-consuming; therefore, minimizing the number of surveyed locations while accurately capturing the topographic features is important. Survey data of geomorphological features and five cross sections were collected along a 100m length of the Arikaree River, a short grass prairie stream in eastern Colorado. Inverse distance weighted (IDW) and kriging interpolations were performed with 100 iterations of a random selection of 50, 60, 70, 80, and 90 percent of the points. The cross-section measurements were excluded both as a whole and for selected stream morphology features. For each sample density, the root mean square error (RMSE) of the five cross sections was found by comparing the measured elevation values to the interpolated values. The RMSE values were compared with the cross section of origin and the sample density as predictor variables. Sample density did not have a statistically significant effect on accuracy. The effect of interpolation on sample density varied between cross sections. There was a statistically significant difference in accuracy when the sample density of different morphology types was reduced while the rest of the data was not. Future surveys should focus on topography, the quality of sampling, and should reduce the amount of area where no points are being collected.