Assessing the impact of different sources of topographic data on 1-D hydraulic modelling of floods

Abstract

Abstract. Topographic data, such as digital elevation models (DEMs), are essential input in flood inundation modelling. DEMs can be derived from several sources either through remote sensing techniques (spaceborne or airborne imagery) or from traditional methods (ground survey). The Advanced Spaceborne Thermal Emission and Reflection Radiometer (ASTER), the Shuttle Radar Topography Mission (SRTM), the light detection and ranging (lidar), and topographic contour maps are some of the most commonly used sources of data for DEMs. These DEMs are characterized by different precision and accuracy. On the one hand, the spatial resolution of low-cost DEMs from satellite imagery, such as ASTER and SRTM, is rather coarse (around 30 to 90 m). On the other hand, the lidar technique is able to produce high-resolution DEMs (at around 1 m), but at a much higher cost. Lastly, contour mapping based on ground survey is time consuming, particularly for higher scales, and may not be possible for some remote areas. The use of these different sources of DEM obviously affects the results of flood inundation models. This paper shows and compares a number of 1-D hydraulic models developed using HEC-RAS as model code and the aforementioned sources of DEM as geometric input. To test model selection, the outcomes of the 1-D models were also compared, in terms of flood water levels, to the results of 2-D models (LISFLOOD-FP). The study was carried out on a reach of the Johor River, in Malaysia. The effect of the different sources of DEMs (and different resolutions) was investigated by considering the performance of the hydraulic models in simulating flood water levels as well as inundation maps. The outcomes of our study show that the use of different DEMs has serious implications to the results of hydraulic models. The outcomes also indicate that the loss of model accuracy due to re-sampling the highest resolution DEM (i.e. lidar 1 m) to lower resolution is much less than the loss of model accuracy due to the use of lowcost DEM that have not only a lower resolution, but also a lower quality. Lastly, to better explore the sensitivity of the 1-D hydraulic models to different DEMs, we performed an uncertainty analysis based on the GLUE methodology.

Application of CryoSat-2 altimetry data for river analysis and modelling

“…Even when correcting the SRTM for vegetation and submerged parts, a relevant error remained. Similar work was done by Md Ali et al (2015), who compared water levels in a hydrodynamic model based on the SRTM DEM with those from a model based on more accurate lidar data. The resulting simulated water levels showed relevant differences.…”

Sensitivity of habitat hydraulic model outputs to DTM and computational mesh resolution

“…Several studies that examined the effect of the mesh resolution in flood inundation modelling, and mapping support the fact that the use of smaller mesh elements reduces the terrain truncation errors and flow truncation errors (Begnudelli & Sanders, 2007;Begnudelli, Sanders, & Bradford, 2008;Horritt, Bates, & Mattinson, 2006;Schubert, Sanders, Smith, & Wright, 2008). Moreover, numerous studies that examined the effect of the DTM accuracy and resolution in flood inundation modelling and mapping support the fact that detailed and accurate representation of the river and riverine area has significant impact on the hydraulic-hydrodynamic modelling results, not only concerning the flood extent but the water depth as well (Courty, Soriano-Monzalvo, & Pedrozo-Acuña, 2019;Lim & Brandt, 2019;Md Ali et al, 2015;Papaioannou, 2017;Papaioannou et al, 2016;Papaioannou, Loukas, & Georgiadis, 2013;Vozinaki, Morianou, Alexakis, & Tsanis, 2017). Nevertheless, a limited number of studies in fish habitat hydraulic modelling have examined the DTM and/or mesh resolution (Boavida et al, 2013;Crowder & Diplas, 2000;Grantham, 2013;Kolden, Fox, Bledsoe, & Kondratieff, 2016;Lin, Lin, & Wu, 2015).…”

“…Despite the fact that these techniques are well established, when they are applied for hydraulic-hydrodynamic modelling applications, they are subject to certain limitations such as the time required for the measurements, the coverage of the study area, the point or pixel density, the accuracy of the derived data sets and the interpolation techniques that can create erroneous areas in the DTM (Md Ali, Solomatine, & Di Baldassarre, 2015;Teng, Vaze, Dutta, & Marvanek, 2015). These restrictive factors can be exceeded with the use of new spatial tools that produce high-resolution digital elevation models leading to better hydraulic-hydrodynamic model configurations and accurate fish habitat hydraulic modelling.…”

National-Scale Built-Environment Exposure to 100-Year Extreme Sea Levels and Sea-Level Rise

“…Coastal flood maps are reliant on available high-resolution digital elevation models [13]. LiDAR DEMs with centimetre-scale vertical resolution can be geographically limited at national-scales.…”

“…However, satellite measurements provide metre or several decimetre-scale vertical errors that constrain their meaningful application in coastal flood exposure assessments to only the highest SLR projections. Recent sensitivity analyses of satellite DEMs in flood exposure assessments, including the Shuttle Radar Topography Mission (SRTM) DEM, demonstrate substantial underestimates of flood exposure compared to DEMs derived from airborne light detection and ranging (LiDAR) [13][14][15]. LiDAR measurements can provide centimetre-scale accuracy that facilitate flood exposure assessments for smaller SLR increments resolvable over decades; however, DEM coverage is often limited to sub-national levels [16].…”