Author(s): Rotchild Louis; Gethro Dauphin; Yves Zech; Adermus Joseph; Nyankona Gonomy; Sandra Soares-Frazao
Linked Author(s): Yves Zech
Keywords: Bathymetric data; UAV-based photogrammetry; Haiti
Abstract: In Haiti, it is difficult to monitor the morphological evolution of rivers because of a lack of regularly updated topographic data. Due to a lack of economic means, the latest topographic data available for the entire country dates back to 2016, based on data acquired in 2014 thanks to a financial support from the World Bank. However, considering the extreme climatic events occurring in the country, erosion of the bed and banks as well as excessive sediment deposition are often observed in some rivers with the consequence that official topographic data rapidly become out of date. In order to propose a fast and easily deployable solution for updating bathymetric data and monitoring morphological changes, the UAV-based photogrammetry (Unmanned Aerial Vehicle) was used on a section of the Rouyonne River (Leogane) and on a complex site of the L’Acul River (Dubreuil). A bathymetric survey was carried out manually with a differential GPS (DGPS), which was used as reference data, for the comparison with (1) the results obtained from the UAV-based photogrammetry technique using the MetaShape software, and (2) the official data. These comparisons aim at assessing the quality of the photogrammetric technique and to highlight the significant modification of the Rouyonne river between 2014 and 2020. The results obtained from the MetaShape software are very promising. In terms of elevation accuracy, the digital terrain model (DTM) obtained from the UAV-based photogrammetry over all sampled cross sections shows a mean square error (MSE) between 5 to 20 cm when compared to the manual survey. In terms of resolution, at Leogane (Rouyonne River) the pixel size on the obtained orthophoto is 4.5 cm × 4.5 cm, which gives a better visualization than the official orthophoto (0.5 m × 0.5 m) of the National Center for GeoSpatial Information (CNIGS) in Haiti. At the Dubreuil site (L'Acul River), the RMSE for the cross-section of the dam is 0.062 m for the MetaShape DTM, and 0.260 m between the CNIGS and DGPS data. For the longitudinal profile, drone photogrammetry better represents this situation with an RMSE of 0.058 m against 1.74 meters for the CNIGS DTM that does not include recent morphological changes. Faster to obtain, easier to update, less expensive and with a better resolution than the official data currently available in Haiti, UAV-based photogrammetry is a promising tool to be promoted for monitoring the morphological evolution of rivers.