Author(s): Maria Fernanda Riano Neira; Diana Catalina Beltran Huertas; Nestor Alonso Mancipe Munoz

Linked Author(s): Maria Fernanda Riano Neira, Diana Catalina Beltran Huertas

Keywords: DEM (Digital Elevation Models); Resampling; Watershed; Regional hydrological modeling

Abstract: Colombia has five hydrographic areas with high water production potential that is constantly threatened by the accelerated development of urban areas and some anthropological activities. During the last decade, the government has boosted a rational management of water resources in different economic and productive sectors of the country. However, lack of reliable monitoring data and terrain data obstruct many efforts to efficiently use the water resources in Colombia. Regional hydrological modeling has proven to be a useful water resource management tool by estimating regional water balances when there is lack of monitoring data. However, developing these regional hydrological models may generate challenges given the pre- and post-processing of remote sensing products (e.g. digital elevation models -DEM) needed to achieve a desired spatial and temporal resolution. This research analyzes the challenges in generating input information associated with the DEMs used in regional hydrological models at different spatial resolutions. A study area located in the hydrographic area of the Magdalena - Cauca with approximately 30,000 km2 is used in this study. The research is developed in four phases: [1] A DEM is obtained from the HydroSHEDS platform and projected to the Colombian geographical reference system at a 90m resolution through four resampling techniques (i.e., nearest neighbor, majority, bilinear and cubic). Then, the watershed delineation process is performed in ArcGIS using the ArcHydro tool; [2] For each 90m projected DEM a resampling is executed to 2700m resolution using the same four resampling methods and the watershed delineation of these resampled DEMs is performed; [3] Contour lines and hydrographic subzones are obtained from the official Colombian platforms as a reference data to compare the results of the watershed delineation performed in phases 1 and 2; and [4] The results of phase 3 are assessed and compared using four metrics associated with morphometric parameters (i.e. area, length of the main channel, mean slope of the watershed and mean elevation of the watershed). Preliminary results suggest that the four resampling techniques do not drastically affect the morphometric parameters assessed if the spatial resolution is kept at 90m when projecting the original DEM. However, all the resampling methods exert negative impacts on the morphometric parameters analyzed when the DEMs are resampled to 2700m, especially in the mean slope of the watershed due to the leveling of the original elevations. Resampling techniques such as nearest neighbor and cubic may provide better results when projecting a DEM in Colombia is required. But combining bilinear projection with other resampling techniques tend to reduce differences in the morphometric parameters when compared with official Colombian data. Finally, the DEM reconditioning process is found to be a critical step for improving drainage line and flow direction representation in the 2700m resampled DEMs.

DOI: https://doi.org/10.3850/IAHR-39WC252171192022956

Year: 2022