Runoff estimation in small rural watersheds using DEMS in North West of Argentina

Basic information from analytical-descriptive methodologies on the relief of watersheds provides the necessary physical parameters for studying natural resources for, e.g., integrated management of watersheds, environmental impact, soil degradation, deforestation, conservation of water resources, and so forth. Geographic information systems can be used for all of these processes, which are linked to a strong spatial component. Digital elevation models (DEM) and their derivatives are a relevant component of these data sources. The parameters found from these models, such as the slope, are used directly or indirectly (as a component of these factors) in many surface runoff estimation equations. In Latin America, the Rational Method has been and continues to be one of the most widely used for the study of microdrainage in small watersheds. The experiment was conducted in the Yatasto district of the Province of Salta, North West Argentina and has an approximate surface of 270 hectares. This zone is located on the piedmont plain of the Metán sierra, where crops are grown, with relatively high slopes and high potential of soil erosion. In this area were studied five small rural watersheds all of them within the total DEM study area and none of them over 81 hectares. This work studied whether potential surface runoff found from the Rational Method (RAMSER) in small rural watersheds shows significant differences depending on whether the mean slopes are found from the DEM derived model or by the usual field methodology. We found that the small-sized grids increase the average slope (between 045% and 0.79%) found over those found from field data. The differences observed in volume of the peak runoff (between 0.003 m3 s−1 and 0.062 m3 s−1) were not significant. The results assure that there are no differences in the parameter evaluated (average slope) under the studied methodology and conditions.

Research highlights▶ The model was generated with the 10 m grid size; it is the highest accuracy rate. ▶ This result shows that ideal grid size which joins the best possible RMSE. ▶ In the paper we found that the three grid sizes studied are around 0.73–0.74 m. ▶ This RMSE is compatible with the minor “informatics’ load” or computational load.