Abstract
Floods are one of the most wide-spread, frequent, and devastating natural disasters that continue to increase in frequency and intensity. Remote sensing, specifically synthetic aperture radar (SAR), has been widely used to detect surface water inundation to provide retrospective and near-real time (NRT) information due to its high-spatial resolution, self-illumination, and low atmospheric attenuation. However, the efficacy of flood inundation mapping with SAR is susceptible to reflections and scattering from a variety of factors including dense vegetation and urban areas. In this study, the topographic dataset Height Above Nearest Drainage (HAND) was investigated as a potential supplement to Sentinel-1A C-Band SAR along with supervised machine learning to improve the detection of inundation in heterogeneous areas. Three machine learning classifiers were trained on two sets of features dual-polarized SAR only and dual-polarized SAR along with HAND to map inundated areas. Three study sites along the Neuse River in North Carolina, USA during the record flood of Hurricane Matthew in October 2016 were selected. The binary classification analysis (inundated as positive vs. non-inundated as negative) revealed significant improvements when incorporating HAND in several metrics including classification accuracy (ACC) (+36.0%), critical success index (CSI) (+39.95%), true positive rate (TPR) (+42.02%), and negative predictive value (NPV) (+17.26%). A marginal change of +0.15% was seen for positive predictive value (PPV), but true negative rate (TNR) fell −14.4%. By incorporating HAND, a significant number of areas with high SAR backscatter but low HAND values were detected as inundated which increased true positives. This in turn also increased the false positives detected but to a lesser extent as evident in the metrics. This study demonstrates that HAND could be considered a valuable feature to enhance SAR flood inundation mapping especially in areas with heterogeneous land covers with dense vegetation that interfere with SAR.
Highlights
According to the World Disasters Report, flooding has affected over 830 million people, caused over 342 billion US dollars in damages, and led to over 57 thousand fatalities globally from 2006 to 2015 [1]
This study has demonstrated how incorporating an auxiliary dataset, Height Above Nearest Drainage (HAND), to Sentinel-1 C-Band synthetic aperture radar (SAR) dual-polarized flood inundation predictions could significantly improve upon SAR-only flood inundation mapping in heterogeneous areas with dense vegetation and anthropogenic development
Three study areas in eastern North Carolina during the record floods of Hurricane Matthew were selected and United States Geological Survey (USGS) generated inundation maps were used as training and validation labels for three supervised machine learning classifiers, quadratic discriminant analysis (QDA), support vector machines (SVM), and k-nearest neighbor (KNN)
Summary
According to the World Disasters Report, flooding has affected over 830 million people, caused over 342 billion US dollars in damages, and led to over 57 thousand fatalities globally from 2006 to 2015 [1]. Flooding is expected to play a growing role in the future as the number of flood prone people, impacted cropland, and overall flood risk is expected to increase by the year 2050 according to a study that used 21 different climate models to simulate conditions [2]. Having more accurate and reliable flood inundation information on both retrospective and near real-time (NRT) (within 24 h) flood events could be a major asset to flood forecasters, public officials, first responders, and the general public. These stakeholders could leverage this information for a variety of applications including model validation, insurance underwriting, risk management, NRT evacuation mapping, and emergency response routing. Having more information on the spatial extents of inundation would be of great value to a breadth of stakeholders
Sign-in/Register to view highlights & summary Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a callDisclaimer: All third-party content on this website/platform is and will remain the property of their respective owners and is provided on "as is" basis without any warranties, express or implied. Use of third-party content does not indicate any affiliation, sponsorship with or endorsement by them. Any references to third-party content is to identify the corresponding services and shall be considered fair use under The CopyrightLaw.
