Flooded with Error: Handling Uncertainty in SRTM for the Assessment of Sea Level Rise in the Mississippi River Delta

• Published in: The Professional geographer Vol. 73; no. 3; pp. 404 - 412
• Main Authors: Kadhim, Ameen A., Shortridge, Ashton M.
• Format: Journal Article
• Language: English
• Published: Washington Routledge 03-07-2021; Taylor & Francis Ltd
• Subjects: Bias; Canopies; elevación del nivel del mar; Elevation; Error analysis; error propagation modeling; Flooding; Floods; Hydrologic data; inundation model (bathtub); Least squares method; Marshes; Mathematical models; Mississippi River Delta region; modelado de la propagación del error; modelo de inundación de tina; región del delta del Mississippi; Regression models; Rivers; Sea level rise; Statistical analysis; Systematic errors; Topography; Uncertainty; Vegetation; Vegetation cover; 密西西比河三角洲地区; 海平面上升; 淹没模型（bathtub）; 误差传播模型; Mississippi River Delta
• ISSN: 0033-0124; 1467-9272
• DOI: 10.1080/00330124.2021.1898992

Digital elevation data are essential to estimate coastal vulnerability to flooding due to sea-level rise. Shuttle Radar Topography Mission (SRTM) 1 arc-second global is considered the best free global digital elevation data set available. Inundation estimates from SRTM, however, are subject to uncertainty due to inaccuracies in the elevation data. Small systematic errors in low, flat areas can generate large errors in inundation models, and SRTM is subject to positive bias in the presence of vegetation canopy, such as along channels and within marshes. In this study, we conducted an error assessment and developed a statistical error model for SRTM to improve the quality of elevation data in the Mississippi River Delta (MRD) region. Vegetation cover, SRTM elevation, and slope were found to be closely associated with SRTM error for a random sample of 10,000 small sites across the MRD region, with an ordinary least squares regression model using these variables explaining over 80 percent of the variation in error. Residuals from this model were spatially autocorrelated, and a variogram model was readily fit to them. We conclude by speculating on the utility of application of this model, developed for the MRD region, to similar near-coastal riverine regions around the world.